New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

QUERY: What is Grbl being used for? #332

Open
chamnit opened this Issue Jan 16, 2014 · 74 comments

Comments

Projects
None yet
@chamnit
Member

chamnit commented Jan 16, 2014

What are you using Grbl for? I'd love to see how Grbl is being used by everyone, as I have absolutely no clue. This would be tremendously useful in determining the future for Grbl and what we should prioritize for new features and developments!

Please post descriptions, pictures, and/or links to your projects!

@kmaxo

This comment has been minimized.

Show comment
Hide comment
@kmaxo

kmaxo Jan 16, 2014

well I am using it to control my laser cutter, using a tba6550.
It is working, with M03/M05 inverted to turn on turn off the laser , as for now I could not find the high trigger for laser. (the old controler use the Low to trigger the laser). AS tba6550 has the optical isolation , the motors are too noise, but I will jump it as next steps too.
So for now I Have the plans

  1. replace the optical couples by jumpers
  2. Find out where is the High trigger for laser
  3. put the home to work, as I do not have Z , I am jumping the limit gnd and try to reverting the logical to other axes
  4. Replace the communication with computer for a pendrive, with the file I want to workout!
    But GRbl ... is fantastic
    congratulations !!!

kmaxo commented Jan 16, 2014

well I am using it to control my laser cutter, using a tba6550.
It is working, with M03/M05 inverted to turn on turn off the laser , as for now I could not find the high trigger for laser. (the old controler use the Low to trigger the laser). AS tba6550 has the optical isolation , the motors are too noise, but I will jump it as next steps too.
So for now I Have the plans

  1. replace the optical couples by jumpers
  2. Find out where is the High trigger for laser
  3. put the home to work, as I do not have Z , I am jumping the limit gnd and try to reverting the logical to other axes
  4. Replace the communication with computer for a pendrive, with the file I want to workout!
    But GRbl ... is fantastic
    congratulations !!!
@twforeman

This comment has been minimized.

Show comment
Hide comment
@twforeman

twforeman Jan 16, 2014

I run GRBL on my ShapeOko mill. GRBL works great for me and 0.9b is even better. :)

You can see my projects on my blog: http://timf.anansi-web.com/wp/category/shapeoko-mill/

My current project is a crossbow pistol that shoots mini-marshmallows. It's not on my blog yet, but here are a couple of photos:

http://imgur.com/RnmSXXt

http://imgur.com/aZPThtL

All the wooden parts are milled using GRBL on my ShapeOko.

twforeman commented Jan 16, 2014

I run GRBL on my ShapeOko mill. GRBL works great for me and 0.9b is even better. :)

You can see my projects on my blog: http://timf.anansi-web.com/wp/category/shapeoko-mill/

My current project is a crossbow pistol that shoots mini-marshmallows. It's not on my blog yet, but here are a couple of photos:

http://imgur.com/RnmSXXt

http://imgur.com/aZPThtL

All the wooden parts are milled using GRBL on my ShapeOko.

@henols

This comment has been minimized.

Show comment
Hide comment
@henols

henols Jan 16, 2014

Contributor

I'm usin GRBL with my home made CNC, Atmega328 and Pololu drivers on a stripboard.
I mainly use it for milling pcb's and different engravings.

Thanks for an exellent pice of software! Hope I can contribute a bit more in the future.

Contributor

henols commented Jan 16, 2014

I'm usin GRBL with my home made CNC, Atmega328 and Pololu drivers on a stripboard.
I mainly use it for milling pcb's and different engravings.

Thanks for an exellent pice of software! Hope I can contribute a bit more in the future.

@tmpvar

This comment has been minimized.

Show comment
Hide comment
@tmpvar

tmpvar Jan 16, 2014

Contributor

I've been using grbl for a couple years on 3 homebuilt cnc's

mdf based router

key components: nema 23 @ ~200oz/in, 1/2-10 acme rod, mdf, rollerblade bearings, and aluminum L bracket

build process - https://plus.google.com/photos/103607741810677023482/albums/5698695539719986129
milling some wax - http://youtu.be/gVgpxa3yEAE?t=5m24s

This thing worked exceptionally well for my first attempt at a full 3 axis cnc, grbl was pivotal to getting this working as most of my time was spent building the hardware. Using this machine I was able to mill the bits for a midi-drum pad, which was the original intent of building a cnc! I gave a talk on that process here: http://2012.nodeconf.com/theatre/elijah_insua.html

aluminum cnc

First attempt was way to big and not very fast or rigid; cutting aluminum was one of the scariest things I've done. so I built a better machine!

key components: bosch extrusion, linear rails, T5 timing belt, nema 34 @ ~900oz/in

There are relatively no build pictures with this one, because it was mostly like a lego project.

videos: http://www.youtube.com/watch?v=WvyWAcof9q0, http://www.youtube.com/watch?v=NhNrayFjvXk

carryon cnc

So, I've given a few talks about using grbl and what I'm trying to accomplish with the tpad (see: nodeconf 2012 video), but I felt that people we're not getting the point of a cnc. Primarily because the rigs I took to demo were single axis and performed no real operation. Being a software developer by trade, there is something magical about taking an idea and implementing it in a physical object that you can hold. Extremely gratifying!

Anyhow, back on target..

key components: 9mm GT2 belting, steel rod, linear bearings, maker beam, helicopter RC motor, and a straight shank er-11 collet holder

pics: http://imgur.com/a/4vsQV

The solid aluminum parts were milled on the aluminum cnc. Due to deadlines I had to mill around the clock for ~5 days. I gave a (not great) talk about taking apart old electronics and exploring the mechanics of printers at nodeconf.eu. Spent more time building the rig than the presentation! Oh well, I did end up milling some wood on stage which was fun and I think inspired a few people to at least take a peek at this type of thing.

I've carried this onto an airplane a bunch of times, and TSA/Customs has never given me any grief about it!

I guess we can say I use grbl to build more machines :) - @chamnit I can't thank you enough for your hard work on this project. You've always been very helpful and extremely patient, I owe you at least a couple cases of beer.

Contributor

tmpvar commented Jan 16, 2014

I've been using grbl for a couple years on 3 homebuilt cnc's

mdf based router

key components: nema 23 @ ~200oz/in, 1/2-10 acme rod, mdf, rollerblade bearings, and aluminum L bracket

build process - https://plus.google.com/photos/103607741810677023482/albums/5698695539719986129
milling some wax - http://youtu.be/gVgpxa3yEAE?t=5m24s

This thing worked exceptionally well for my first attempt at a full 3 axis cnc, grbl was pivotal to getting this working as most of my time was spent building the hardware. Using this machine I was able to mill the bits for a midi-drum pad, which was the original intent of building a cnc! I gave a talk on that process here: http://2012.nodeconf.com/theatre/elijah_insua.html

aluminum cnc

First attempt was way to big and not very fast or rigid; cutting aluminum was one of the scariest things I've done. so I built a better machine!

key components: bosch extrusion, linear rails, T5 timing belt, nema 34 @ ~900oz/in

There are relatively no build pictures with this one, because it was mostly like a lego project.

videos: http://www.youtube.com/watch?v=WvyWAcof9q0, http://www.youtube.com/watch?v=NhNrayFjvXk

carryon cnc

So, I've given a few talks about using grbl and what I'm trying to accomplish with the tpad (see: nodeconf 2012 video), but I felt that people we're not getting the point of a cnc. Primarily because the rigs I took to demo were single axis and performed no real operation. Being a software developer by trade, there is something magical about taking an idea and implementing it in a physical object that you can hold. Extremely gratifying!

Anyhow, back on target..

key components: 9mm GT2 belting, steel rod, linear bearings, maker beam, helicopter RC motor, and a straight shank er-11 collet holder

pics: http://imgur.com/a/4vsQV

The solid aluminum parts were milled on the aluminum cnc. Due to deadlines I had to mill around the clock for ~5 days. I gave a (not great) talk about taking apart old electronics and exploring the mechanics of printers at nodeconf.eu. Spent more time building the rig than the presentation! Oh well, I did end up milling some wood on stage which was fun and I think inspired a few people to at least take a peek at this type of thing.

I've carried this onto an airplane a bunch of times, and TSA/Customs has never given me any grief about it!

I guess we can say I use grbl to build more machines :) - @chamnit I can't thank you enough for your hard work on this project. You've always been very helpful and extremely patient, I owe you at least a couple cases of beer.

@scottrcarlson

This comment has been minimized.

Show comment
Hide comment
@scottrcarlson

scottrcarlson Jan 16, 2014

Home built 3axis cnc here. approx. 12" cubed work volume. I just couldn't stand the thought of running mach3 with a windows box on my baby!

Future work (before end of year)., FDM Printing and 4th axis/Lathe.

Grbl Rocks!

scottrcarlson commented Jan 16, 2014

Home built 3axis cnc here. approx. 12" cubed work volume. I just couldn't stand the thought of running mach3 with a windows box on my baby!

Future work (before end of year)., FDM Printing and 4th axis/Lathe.

Grbl Rocks!

@matthewSorensen

This comment has been minimized.

Show comment
Hide comment
@matthewSorensen

matthewSorensen Jan 17, 2014

10'x4'x4' plasma cutter converted to a rather large FDM printer. We just got the Grbl-based system up and running and we're still evaluating it, but the hardware has printed some large objects with its former control system.

matthewSorensen commented Jan 17, 2014

10'x4'x4' plasma cutter converted to a rather large FDM printer. We just got the Grbl-based system up and running and we're still evaluating it, but the hardware has printed some large objects with its former control system.

@lakeside53

This comment has been minimized.

Show comment
Hide comment
@lakeside53

lakeside53 Jan 17, 2014

My use is simple, but I have far greater plans!

Currently, it uses an UNO and drives a geared stepper motor simulating an A axis. A gemstone is placed on the platter and a fixed macro camera takes 360 degree pictures- typically 90 shots in 30 seconds. The G-Code uses the G4 (dwell) function for camera shutter and other timing, and M codes for firing the shutter etc. I use the programmable acceleration/decel to not move the stones off center when moving fast. The z-axis will be connected to the camera rack for staged focus shots.

CNC plasma cutter coming up…

-Andy

From: Sonny Jeon [mailto:notifications@github.com]
Sent: Thursday, January 16, 2014 5:56 AM
To: grbl/grbl
Subject: [grbl] QUERY: What is Grbl being used for? (#332)

What are you using Grbl for? I'd love to see how Grbl is being used by everyone, as I have absolutely no clue. This would be tremendously useful in determining the future for Grbl and what we should prioritize for new features and developments!
Please post descriptions, pictures, and/or links to your projects!

Reply to this email directly or view it on #332 GitHub. https://github.com/notifications/beacon/4221088__eyJzY29wZSI6Ik5ld3NpZXM6QmVhY29uIiwiZXhwaXJlcyI6MTcwNTQxMzM4OCwiZGF0YSI6eyJpZCI6MjM2NTgyODJ9fQ==--7f488570487aeb4ea4318b4f80653bd44dbd62e7.gif

lakeside53 commented Jan 17, 2014

My use is simple, but I have far greater plans!

Currently, it uses an UNO and drives a geared stepper motor simulating an A axis. A gemstone is placed on the platter and a fixed macro camera takes 360 degree pictures- typically 90 shots in 30 seconds. The G-Code uses the G4 (dwell) function for camera shutter and other timing, and M codes for firing the shutter etc. I use the programmable acceleration/decel to not move the stones off center when moving fast. The z-axis will be connected to the camera rack for staged focus shots.

CNC plasma cutter coming up…

-Andy

From: Sonny Jeon [mailto:notifications@github.com]
Sent: Thursday, January 16, 2014 5:56 AM
To: grbl/grbl
Subject: [grbl] QUERY: What is Grbl being used for? (#332)

What are you using Grbl for? I'd love to see how Grbl is being used by everyone, as I have absolutely no clue. This would be tremendously useful in determining the future for Grbl and what we should prioritize for new features and developments!
Please post descriptions, pictures, and/or links to your projects!

Reply to this email directly or view it on #332 GitHub. https://github.com/notifications/beacon/4221088__eyJzY29wZSI6Ik5ld3NpZXM6QmVhY29uIiwiZXhwaXJlcyI6MTcwNTQxMzM4OCwiZGF0YSI6eyJpZCI6MjM2NTgyODJ9fQ==--7f488570487aeb4ea4318b4f80653bd44dbd62e7.gif

@chamnit

This comment has been minimized.

Show comment
Hide comment
@chamnit

chamnit Jan 17, 2014

Member

Wow. This is some awesome stuff! Keep it the posts coming! @tmpvar I especially love that suitcase CNC deployable awesomeness.

Member

chamnit commented Jan 17, 2014

Wow. This is some awesome stuff! Keep it the posts coming! @tmpvar I especially love that suitcase CNC deployable awesomeness.

@tmpvar

This comment has been minimized.

Show comment
Hide comment
@tmpvar

tmpvar Jan 17, 2014

Contributor

@chamnit thanks!

Contributor

tmpvar commented Jan 17, 2014

@chamnit thanks!

@rverhag

This comment has been minimized.

Show comment
Hide comment
@rverhag

rverhag Jan 17, 2014

Just a plain 3 axes CNC (29.5" x 19.6" x 4.9" / 750 x 500 x 125 mm).

Wood/MDF/aluminum/plastics

Used Grbl 0.9a for a couple of months now. Runs perfect for me! Now testing with 0.9c.

My blog about the development of the machine http://hobbycnc.vlindervanwalcheren.nl/ (only dutch, sorry for that)

img_1977 250x167
img_1979 167x250

rverhag commented Jan 17, 2014

Just a plain 3 axes CNC (29.5" x 19.6" x 4.9" / 750 x 500 x 125 mm).

Wood/MDF/aluminum/plastics

Used Grbl 0.9a for a couple of months now. Runs perfect for me! Now testing with 0.9c.

My blog about the development of the machine http://hobbycnc.vlindervanwalcheren.nl/ (only dutch, sorry for that)

img_1977 250x167
img_1979 167x250

@metropt

This comment has been minimized.

Show comment
Hide comment
@metropt

metropt Jan 20, 2014

I'm using in on a live paint board system :)

metropt commented Jan 20, 2014

I'm using in on a live paint board system :)

@r01walker

This comment has been minimized.

Show comment
Hide comment
@r01walker

r01walker Jan 25, 2014

the usual a 3d printer. jig boring style low pressure plastic extruder. i am surprised their is not more support for a x,y axis, x being linear and y being rotary.
i am also surprised that the capacity of the stepper boards is not being utilised that being the step control commands. full half 1/4 1/8 step. some really fantastic feed rates could be achieved with that capacity. all that would have to happen is when coming from 1/8 step step to the 1/4 half or nearest full step position and full step until the last full step of the distance commanded and then micro step to position. for me i would be able to run a much larger extruder and create parts much faster. of course i don’t know the code so i don’t know the problems modding the code. yet. the loss is 4 pins off the ardunio instead of two. per axis. on a mega the pin outs would not be a issue.
of course idiot lights. jog control, sd card or hard drive support and generally more documentation.
I am pleased with the software it seems to work well, my complements and sincere thanks, and as i can i will contribute.
great job

r01walker commented Jan 25, 2014

the usual a 3d printer. jig boring style low pressure plastic extruder. i am surprised their is not more support for a x,y axis, x being linear and y being rotary.
i am also surprised that the capacity of the stepper boards is not being utilised that being the step control commands. full half 1/4 1/8 step. some really fantastic feed rates could be achieved with that capacity. all that would have to happen is when coming from 1/8 step step to the 1/4 half or nearest full step position and full step until the last full step of the distance commanded and then micro step to position. for me i would be able to run a much larger extruder and create parts much faster. of course i don’t know the code so i don’t know the problems modding the code. yet. the loss is 4 pins off the ardunio instead of two. per axis. on a mega the pin outs would not be a issue.
of course idiot lights. jog control, sd card or hard drive support and generally more documentation.
I am pleased with the software it seems to work well, my complements and sincere thanks, and as i can i will contribute.
great job

@chamnit

This comment has been minimized.

Show comment
Hide comment
@chamnit

chamnit Jan 25, 2014

Member

@r01walker : Thanks for your machine query post and your suggestions and comments.

In response to some of your comments, Grbl is designed to be a simple 3-axis motion control system and doesn't support rotary axes, sd cards, and other things on purpose. The primary reason is that this allows us to concentrate on the core functionality and improve it quicker without having to worry about the other stuff breaking. Another reason some other things are not available is from the lack of resources, from I/O pins to available flash memory and RAM. Grbl tries to stick with the limited Uno processor (328p) because of its ubiquity and popularity. This makes entry into the CNC world very cheap, highly accessible, and virtually risk-free. Keep in mind that Grbl is open-source, and that anyone, including yourself, can fork and add whatever they want to the firmware. Variants of Grbl, like Marlin (requires a Mega), have already added all these bells-and-whistles that you suggested and is designed for 3D printing. I suggest that you take at look at their project, although they broke off of Grbl quite some time ago and may not have as fine and capable motion control as Grbl does, as we have since made lots of progress on this front.

This isn't to say Grbl won't officially upgrade to a more capable processor. Grbl will and likely soon. The development milestones of the motion control core upgrades are almost done. Once finished, we'll move on.

In regards to using I/O pins as a way to control the micro step size in realtime, this has already been discussed offline as a possibility, but you've already hit the nail on the head eluding to I/O pins are in short supply. The most difficult thing with this is making sure steps are not lost when switching between the micro stepping states in a robust manner. Plus, this also can be sensitive to stepper drivers and how they operate, so numerous configurable conditions may be needed to be able support a large array of drivers. The only way we will do this is, if we have more I/O pins and memory to work with, but as I said, this isn't now.

Member

chamnit commented Jan 25, 2014

@r01walker : Thanks for your machine query post and your suggestions and comments.

In response to some of your comments, Grbl is designed to be a simple 3-axis motion control system and doesn't support rotary axes, sd cards, and other things on purpose. The primary reason is that this allows us to concentrate on the core functionality and improve it quicker without having to worry about the other stuff breaking. Another reason some other things are not available is from the lack of resources, from I/O pins to available flash memory and RAM. Grbl tries to stick with the limited Uno processor (328p) because of its ubiquity and popularity. This makes entry into the CNC world very cheap, highly accessible, and virtually risk-free. Keep in mind that Grbl is open-source, and that anyone, including yourself, can fork and add whatever they want to the firmware. Variants of Grbl, like Marlin (requires a Mega), have already added all these bells-and-whistles that you suggested and is designed for 3D printing. I suggest that you take at look at their project, although they broke off of Grbl quite some time ago and may not have as fine and capable motion control as Grbl does, as we have since made lots of progress on this front.

This isn't to say Grbl won't officially upgrade to a more capable processor. Grbl will and likely soon. The development milestones of the motion control core upgrades are almost done. Once finished, we'll move on.

In regards to using I/O pins as a way to control the micro step size in realtime, this has already been discussed offline as a possibility, but you've already hit the nail on the head eluding to I/O pins are in short supply. The most difficult thing with this is making sure steps are not lost when switching between the micro stepping states in a robust manner. Plus, this also can be sensitive to stepper drivers and how they operate, so numerous configurable conditions may be needed to be able support a large array of drivers. The only way we will do this is, if we have more I/O pins and memory to work with, but as I said, this isn't now.

@r01walker

This comment has been minimized.

Show comment
Hide comment
@r01walker

r01walker Jan 25, 2014

thank you for your comments kind sir. 

in regards t the rotary axis. since you already control the axis with a stepper motor. it would be nothing but mounting and wiring. the issue is in software. i don't know of any professional cam system that supports that configuration. and if it could be made to work it would double the axis capacity. but the functions of that axis would be x and y being a arc to get any kind of strait line. 
I appreciate that their are limitation issues that as i am mega i do not have. at least to the severity of a uno biased system. 
I do understand the market niche that is being exploited. again great respect. I hesitated to mention anything as i am used to professional cnc machines. 
for a question is the spindle coded for a stepper or a dremel tool?. i understand that the voltage wont let me connect the dremel directly but through a h-bridge circuit. or relay. 
to put this out their. have you considered using a daughter board plunging into some of the pins on the inside of the uno and sending command info down line in a binary manner. this would let the main board have the target market you want and allow even mega additional and expandable capacity.
my thanks again i will look at the  Marlin admittedly i have the grbl 8.0 installed on the mega and functioning. except for the extruder. but i will look and see. 

On Saturday, January 25, 2014 1:10 AM, Sonny Jeon notifications@github.com wrote:

@r01walker : Thanks for your machine query post and your suggestions and comments.
In response to some of your comments, Grbl is designed to be a simple 3-axis motion control system and doesn't support rotary axes, sd cards, and other things on purpose. The primary reason is that this allows us to concentrate on the core functionality and improve it quicker without having to worry about the other stuff breaking. Another reason some other things are not available is from the lack of resources, from I/O pins to available flash memory and RAM. Grbl tries to stick with the limited Uno processor (328p) because of its ubiquity and popularity. This makes entry into the CNC world very cheap, highly accessible, and virtually risk-free. Keep in mind that Grbl is open-source, and that anyone, including yourself, can fork and add whatever they want to the firmware. Variants of Grbl, like Marlin (requires a Mega), have already added all these bells-and-whistles that you suggested and is designed for 3D printing. I suggest that you take at look at
their project, although they broke off of Grbl quite some time ago and may not have as fine and capable motion control as Grbl does, as we have since made lots of progress on this front.
This isn't to say Grbl won't officially upgrade to a more capable processor. Grbl will and likely soon. The development milestones of the motion control core upgrades are almost done. Once finished, we'll move on.
In regards to using I/O pins as a way to control the micro step size in realtime, this has already been discussed offline as a possibility, but you've already hit the nail on the head eluding to I/O pins are in short supply. The most difficult thing with this is making sure steps are not lost when switching between the micro stepping states in a robust manner. Plus, this also can be sensitive to stepper drivers and how they operate, so numerous configurable conditions may be needed to be able support a large array of drivers. The only way we will do this is, if we have more I/O pins and memory to work with, but as I said, this isn't now.

Reply to this email directly or view it on GitHub.

r01walker commented Jan 25, 2014

thank you for your comments kind sir. 

in regards t the rotary axis. since you already control the axis with a stepper motor. it would be nothing but mounting and wiring. the issue is in software. i don't know of any professional cam system that supports that configuration. and if it could be made to work it would double the axis capacity. but the functions of that axis would be x and y being a arc to get any kind of strait line. 
I appreciate that their are limitation issues that as i am mega i do not have. at least to the severity of a uno biased system. 
I do understand the market niche that is being exploited. again great respect. I hesitated to mention anything as i am used to professional cnc machines. 
for a question is the spindle coded for a stepper or a dremel tool?. i understand that the voltage wont let me connect the dremel directly but through a h-bridge circuit. or relay. 
to put this out their. have you considered using a daughter board plunging into some of the pins on the inside of the uno and sending command info down line in a binary manner. this would let the main board have the target market you want and allow even mega additional and expandable capacity.
my thanks again i will look at the  Marlin admittedly i have the grbl 8.0 installed on the mega and functioning. except for the extruder. but i will look and see. 

On Saturday, January 25, 2014 1:10 AM, Sonny Jeon notifications@github.com wrote:

@r01walker : Thanks for your machine query post and your suggestions and comments.
In response to some of your comments, Grbl is designed to be a simple 3-axis motion control system and doesn't support rotary axes, sd cards, and other things on purpose. The primary reason is that this allows us to concentrate on the core functionality and improve it quicker without having to worry about the other stuff breaking. Another reason some other things are not available is from the lack of resources, from I/O pins to available flash memory and RAM. Grbl tries to stick with the limited Uno processor (328p) because of its ubiquity and popularity. This makes entry into the CNC world very cheap, highly accessible, and virtually risk-free. Keep in mind that Grbl is open-source, and that anyone, including yourself, can fork and add whatever they want to the firmware. Variants of Grbl, like Marlin (requires a Mega), have already added all these bells-and-whistles that you suggested and is designed for 3D printing. I suggest that you take at look at
their project, although they broke off of Grbl quite some time ago and may not have as fine and capable motion control as Grbl does, as we have since made lots of progress on this front.
This isn't to say Grbl won't officially upgrade to a more capable processor. Grbl will and likely soon. The development milestones of the motion control core upgrades are almost done. Once finished, we'll move on.
In regards to using I/O pins as a way to control the micro step size in realtime, this has already been discussed offline as a possibility, but you've already hit the nail on the head eluding to I/O pins are in short supply. The most difficult thing with this is making sure steps are not lost when switching between the micro stepping states in a robust manner. Plus, this also can be sensitive to stepper drivers and how they operate, so numerous configurable conditions may be needed to be able support a large array of drivers. The only way we will do this is, if we have more I/O pins and memory to work with, but as I said, this isn't now.

Reply to this email directly or view it on GitHub.

@Blaisun

This comment has been minimized.

Show comment
Hide comment
@Blaisun

Blaisun Feb 6, 2014

I have built a fairly generic MDF 3 axis router... and am having good luck with GRBL !!!

Blaisun commented Feb 6, 2014

I have built a fairly generic MDF 3 axis router... and am having good luck with GRBL !!!

@Mgilbride

This comment has been minimized.

Show comment
Hide comment
@Mgilbride

Mgilbride Feb 8, 2014

Plain 3 Axis CNC. Still rough around the edges. Moderate size 4x5 work bed. Mills ABS, PVC just fine. Have not tried aluminum / soft metals yet. Primarily meant for soft to moderate density foam carving for iterative design model making. So far, it's been mostly CNC making!! Note, wiring in pictures does not reflect updates to schematic or other posts. Belt drives need a bit of adjustment but work great. Do regret no going rack and pinion now that I found a cheap source of racks. So greedy!

schematic custom 2

img_20140203_172402 1 custom

img_20140203_172416 1 custom

img_20140203_172424 1 custom

img_20140203_172346 custom

img_20140203_172355 custom

img_20140210_120707 large

img_20140210_120714 large

Running a foam cutting job. Still having signal problems.

Mgilbride commented Feb 8, 2014

Plain 3 Axis CNC. Still rough around the edges. Moderate size 4x5 work bed. Mills ABS, PVC just fine. Have not tried aluminum / soft metals yet. Primarily meant for soft to moderate density foam carving for iterative design model making. So far, it's been mostly CNC making!! Note, wiring in pictures does not reflect updates to schematic or other posts. Belt drives need a bit of adjustment but work great. Do regret no going rack and pinion now that I found a cheap source of racks. So greedy!

schematic custom 2

img_20140203_172402 1 custom

img_20140203_172416 1 custom

img_20140203_172424 1 custom

img_20140203_172346 custom

img_20140203_172355 custom

img_20140210_120707 large

img_20140210_120714 large

Running a foam cutting job. Still having signal problems.

@tmpvar

This comment has been minimized.

Show comment
Hide comment
@tmpvar

tmpvar Feb 8, 2014

Contributor

@Mgilbride looking good! Few questions:

  • how's the makerslide extrusion working out for you?
  • it looks like you are using a sort of servobelt design for your x/y - looks pretty slick. does it work as advertised?

I tried doing something similar (servobelt), but the belt I was using liked to snake out of it's "zippered" position. Might have been that the belt was to narrow? I'd love to hear your thoughts/experiences

Contributor

tmpvar commented Feb 8, 2014

@Mgilbride looking good! Few questions:

  • how's the makerslide extrusion working out for you?
  • it looks like you are using a sort of servobelt design for your x/y - looks pretty slick. does it work as advertised?

I tried doing something similar (servobelt), but the belt I was using liked to snake out of it's "zippered" position. Might have been that the belt was to narrow? I'd love to hear your thoughts/experiences

@Mgilbride

This comment has been minimized.

Show comment
Hide comment
@Mgilbride

Mgilbride Feb 8, 2014

Thanks Elijah.

Makerslide is great stuff. Really a marvel in it's economy of providing
linear motion, structural attachment, cable routing slots. Other extrusion
based solutions looked great too. But I like this one best. Looking for
next level of integration with rack gear integrated into extrusion. think
that would need to be a machined post process. Long span working well wo
central support. But planning to add one in future. Triple extrusion beam
on x-axis really is nice and rigid.

Belts work great. could benefit from a spring loaded bolt tensioner for
easy maintenance and adjustment. Should be easy to rig one up. I tried
bonding my lower belt with a rubber cement. didn't work as well as I had
hoped. Really, I am starting to doubt the utility of the bottom belts.
Guess they help but most systems work fine without them. I have occasional
belt alignment problems. Mostly because I need to adjust the gear on the
stepper.

UPDATE 2.23.2014: I have made some changes over the last couple of weeks that pertain to this discussion. I removed the underlying belt that I was utilizing as a sort of rack gear for the drive belt. So far the system runs well. I have done a lot of additional research and reading about belts during that time and am thinking that a couple of additional things could be considered when implementing a belt drive. Belts make a very efficient drive system. Especially under low loads like 3D printers and laser cutters. But they can and do stretch and benefit from structures that can implement them under higher tension. The underlying belt I removed from the system may well help counter stretching and slippage in the drive system. Especially in one that will see decent loads while CNC router cutting. Additonally, I think I failed to do my homework when scaling up and adapting the shapeoko model of this system. I am using MXL belts which have, to date, worked just fine. However, I think the system would benefit a good deal from wider belts that can both resist stretching forces and have greater contact area with the drive gear. I may just implement that change. At least on the dual Y axis. Have to see what parts and belt widths / pitches are available.

On Friday, February 7, 2014, Elijah Insua notifications@github.com wrote:

@Mgilbride https://github.com/Mgilbride looking good! Few questions:

  • how's the makerslide extrusion working out for you?
  • it looks like you are using a sort of servobelt design for your x/y
  • looks pretty slick. does it work as advertised?

I tried doing something similar (servobelt), but the belt I was using
liked to snake out of it's "zippered" position. Might have been that the
belt was to narrow? I'd love to hear your thoughts/experiences

Reply to this email directly or view it on GitHubhttps://github.com//issues/332#issuecomment-34531733
.

Mgilbride commented Feb 8, 2014

Thanks Elijah.

Makerslide is great stuff. Really a marvel in it's economy of providing
linear motion, structural attachment, cable routing slots. Other extrusion
based solutions looked great too. But I like this one best. Looking for
next level of integration with rack gear integrated into extrusion. think
that would need to be a machined post process. Long span working well wo
central support. But planning to add one in future. Triple extrusion beam
on x-axis really is nice and rigid.

Belts work great. could benefit from a spring loaded bolt tensioner for
easy maintenance and adjustment. Should be easy to rig one up. I tried
bonding my lower belt with a rubber cement. didn't work as well as I had
hoped. Really, I am starting to doubt the utility of the bottom belts.
Guess they help but most systems work fine without them. I have occasional
belt alignment problems. Mostly because I need to adjust the gear on the
stepper.

UPDATE 2.23.2014: I have made some changes over the last couple of weeks that pertain to this discussion. I removed the underlying belt that I was utilizing as a sort of rack gear for the drive belt. So far the system runs well. I have done a lot of additional research and reading about belts during that time and am thinking that a couple of additional things could be considered when implementing a belt drive. Belts make a very efficient drive system. Especially under low loads like 3D printers and laser cutters. But they can and do stretch and benefit from structures that can implement them under higher tension. The underlying belt I removed from the system may well help counter stretching and slippage in the drive system. Especially in one that will see decent loads while CNC router cutting. Additonally, I think I failed to do my homework when scaling up and adapting the shapeoko model of this system. I am using MXL belts which have, to date, worked just fine. However, I think the system would benefit a good deal from wider belts that can both resist stretching forces and have greater contact area with the drive gear. I may just implement that change. At least on the dual Y axis. Have to see what parts and belt widths / pitches are available.

On Friday, February 7, 2014, Elijah Insua notifications@github.com wrote:

@Mgilbride https://github.com/Mgilbride looking good! Few questions:

  • how's the makerslide extrusion working out for you?
  • it looks like you are using a sort of servobelt design for your x/y
  • looks pretty slick. does it work as advertised?

I tried doing something similar (servobelt), but the belt I was using
liked to snake out of it's "zippered" position. Might have been that the
belt was to narrow? I'd love to hear your thoughts/experiences

Reply to this email directly or view it on GitHubhttps://github.com//issues/332#issuecomment-34531733
.

@chamnit

This comment has been minimized.

Show comment
Hide comment
@chamnit

chamnit Feb 9, 2014

Member

@Mgilbride : Awesome machine! Looks to be perfect for foam cutting. I'd be a little leery on the X-axis stiffness for cutting metals though. When the y-axis is midspan, there will probably some flexing there when the x-axis moves and there are some cutting forces, but this isn't hard to fix.

I'm pretty dang stoked to see big machines like this working. For my next big project, I've been planning to build a 5'x5' (or 5'x8') vertical CNC router that can be also be cantilevered down to be horizontal. I was thinking about using the Makerslides, but was concerned about how stiff and durable they were. Have you had any issues with the Makerslide rails or delrin bushing wearing out? How accurate/straight are the Makerslide and how does it affect the accuracy of your machine?

Member

chamnit commented Feb 9, 2014

@Mgilbride : Awesome machine! Looks to be perfect for foam cutting. I'd be a little leery on the X-axis stiffness for cutting metals though. When the y-axis is midspan, there will probably some flexing there when the x-axis moves and there are some cutting forces, but this isn't hard to fix.

I'm pretty dang stoked to see big machines like this working. For my next big project, I've been planning to build a 5'x5' (or 5'x8') vertical CNC router that can be also be cantilevered down to be horizontal. I was thinking about using the Makerslides, but was concerned about how stiff and durable they were. Have you had any issues with the Makerslide rails or delrin bushing wearing out? How accurate/straight are the Makerslide and how does it affect the accuracy of your machine?

@Mgilbride

This comment has been minimized.

Show comment
Hide comment
@Mgilbride

Mgilbride Feb 9, 2014

UPDATE 2.23.2014: Sonny, I have been thinking about your comments about
machine stiffness while cutting heavier materials. When you smack or poke my
rig when the x axis is in the middle of the Y span you can see the whole structure
a good deal. Poking around for some ideas I noticed that Improbable Construct
from the Shapeoko community has compared open rail to makerslide in a couple
of videos you may want to check out. Something to consider when thinking about
what to implement where. Think I will double the rails for stiffness.

Open Rail: http://www.youtube.com/watch?v=UTIf5EOQYrw
Makerslide: http://www.youtube.com/watch?v=-x-35kdj0Ns

Thanks Sonny. Agreed on Y axis stiffness. Am thinking through some frame
revisions to add stiffening. That said, most metal cutting will probably be
small, near home, and use very different feeds and speeds. Also want to
revise z axis to be more robust, rigid, and compact. Also want better quick
tool swapping capabilities.

Love your plan. Have mulled the vertical mill myself. I think that locks
you into lead screws. Belts & R&P would drop tools without some sort of
clutch or brakes wouldn't they. And you need a good structure that is rigid
in both orientations. Gonna motorize the vertical to horizontal change
over? ;)

Have not used machine enough to wear anything out. I think makerslide will
last well if well supported, kept free of abrasive materials, and used with
delrin rather than steel wheels. Delrin is used a lot in industrial
material handling applications. You will need to replace it but it is quite
durable and long lasting. Especially in a rolling rather than sliding
bearing or wear pad application.

Sorry. forgot about accuracy question. I am not currently looking for crazy accuracy. I have done some quick checks with calipers on a number of jobs and everything is close enough to make me happy at the moment. Accuracy and repeatability are two metrics I thankfully do not need to chase right now given intended application of the mill. That said, there are some folks on the Shapeoko forums that are accuracy nuts and have done an impressive job of both testing and optimizing their machines. If you want to delve into that subject I would head over there and do some reading. I remember at least one guy who did some FEA analysis to see how much deflection rails would experience under load. SO there is a wealth of info out there.

This remarkably inexpensive genre of machines seem to be capable of amazing accuracy for the price. At least over short periods of a few jobs. But if you are looking for something rock solid over long periods of heavy use I think you will want a much more robust build. I also think smaller machines lend themselves to some heroic efforts in producing accurate, repeatable results. In short, While I did not scrutinize my rails I did not encounter any obvious flaws that I found problematic or disappointing. The makerslide drawings posted at http://www.buildlog.net/documents/b17022_rev_3.pdf have some tolerance info you can use for design purposes.

On Feb 9, 2014 9:57 AM, "Sonny Jeon" notifications@github.com wrote:

@Mgilbride https://github.com/Mgilbride : Awesome machine! Looks to be
perfect for foam cutting. I'd be a little leery on the X-axis stiffness for
cutting metals though. When the y-axis is midspan, there will probably some
flexing there when the x-axis moves and there are some cutting forces, but
this isn't hard to fix.

I'm pretty dang stoked to see big machines like this working. For my next
big project, I've been planning to build a 5'x5' (or 5'x8') vertical CNC
router that can be also be cantilevered down to be horizontal. I was
thinking about using the Makerslides, but was concerned about how stiff and
durable they were. Have you had any issues with the Makerslide rails or
delrin bushing wearing out? How accurate/straight are the Makerslide and
how does it affect the accuracy of your machine?

Reply to this email directly or view it on GitHubhttps://github.com//issues/332#issuecomment-34575926
.

Mgilbride commented Feb 9, 2014

UPDATE 2.23.2014: Sonny, I have been thinking about your comments about
machine stiffness while cutting heavier materials. When you smack or poke my
rig when the x axis is in the middle of the Y span you can see the whole structure
a good deal. Poking around for some ideas I noticed that Improbable Construct
from the Shapeoko community has compared open rail to makerslide in a couple
of videos you may want to check out. Something to consider when thinking about
what to implement where. Think I will double the rails for stiffness.

Open Rail: http://www.youtube.com/watch?v=UTIf5EOQYrw
Makerslide: http://www.youtube.com/watch?v=-x-35kdj0Ns

Thanks Sonny. Agreed on Y axis stiffness. Am thinking through some frame
revisions to add stiffening. That said, most metal cutting will probably be
small, near home, and use very different feeds and speeds. Also want to
revise z axis to be more robust, rigid, and compact. Also want better quick
tool swapping capabilities.

Love your plan. Have mulled the vertical mill myself. I think that locks
you into lead screws. Belts & R&P would drop tools without some sort of
clutch or brakes wouldn't they. And you need a good structure that is rigid
in both orientations. Gonna motorize the vertical to horizontal change
over? ;)

Have not used machine enough to wear anything out. I think makerslide will
last well if well supported, kept free of abrasive materials, and used with
delrin rather than steel wheels. Delrin is used a lot in industrial
material handling applications. You will need to replace it but it is quite
durable and long lasting. Especially in a rolling rather than sliding
bearing or wear pad application.

Sorry. forgot about accuracy question. I am not currently looking for crazy accuracy. I have done some quick checks with calipers on a number of jobs and everything is close enough to make me happy at the moment. Accuracy and repeatability are two metrics I thankfully do not need to chase right now given intended application of the mill. That said, there are some folks on the Shapeoko forums that are accuracy nuts and have done an impressive job of both testing and optimizing their machines. If you want to delve into that subject I would head over there and do some reading. I remember at least one guy who did some FEA analysis to see how much deflection rails would experience under load. SO there is a wealth of info out there.

This remarkably inexpensive genre of machines seem to be capable of amazing accuracy for the price. At least over short periods of a few jobs. But if you are looking for something rock solid over long periods of heavy use I think you will want a much more robust build. I also think smaller machines lend themselves to some heroic efforts in producing accurate, repeatable results. In short, While I did not scrutinize my rails I did not encounter any obvious flaws that I found problematic or disappointing. The makerslide drawings posted at http://www.buildlog.net/documents/b17022_rev_3.pdf have some tolerance info you can use for design purposes.

On Feb 9, 2014 9:57 AM, "Sonny Jeon" notifications@github.com wrote:

@Mgilbride https://github.com/Mgilbride : Awesome machine! Looks to be
perfect for foam cutting. I'd be a little leery on the X-axis stiffness for
cutting metals though. When the y-axis is midspan, there will probably some
flexing there when the x-axis moves and there are some cutting forces, but
this isn't hard to fix.

I'm pretty dang stoked to see big machines like this working. For my next
big project, I've been planning to build a 5'x5' (or 5'x8') vertical CNC
router that can be also be cantilevered down to be horizontal. I was
thinking about using the Makerslides, but was concerned about how stiff and
durable they were. Have you had any issues with the Makerslide rails or
delrin bushing wearing out? How accurate/straight are the Makerslide and
how does it affect the accuracy of your machine?

Reply to this email directly or view it on GitHubhttps://github.com//issues/332#issuecomment-34575926
.

@chamnit

This comment has been minimized.

Show comment
Hide comment
@chamnit

chamnit Feb 12, 2014

Member

@Mgilbride : Thanks for the Makerslide info! It's interesting to know how accurate they are (0.0125"/foot straightness and 0.5deg/foot twist). Not terrible and acceptable for large router type machines that would be used for wood.

Lol. FEA analysis is a bit overkill for looking at deflection of rails. :) Simple beam analysis would have been plenty, but if you got the tools available, why not?

So, the plan for the vertical mill is to base the entire structure on a "torsion-box", rather than building a rigid frame. These are popular among woodworkers to create a very flat(~0.001"-0.005") and stable tabletop, but they are simply hollow-core boxes or honeycomb structures. The woodworker versions are WAY over-built and heavy. I'm thinking of something along the lines of a super-rigid aircraft wing-like design: thin spars, thin skins, and thicker/meaty attachment points. The idea is to fabricate everything with a ShapeOko2 (when I get my hands on one) so I can make all of the parts quickly, identical, and accurate.

I've been looking at mounting schemes and will probably use a gantry-rail type system like the ShapeOko. It's hard to beat the cost efficiency. But instead, I'll probably fabricate my own MakerSlide type rail system with some bigger and stiffer 80/20 and attach steel v-rails on them so I can use steel bearings. These rails would be mounted to the sides of the "torsion-box". I don't think I will necessarily need leadscrews, because I will be off-loading the weight of gantry and heads with constant force springs when in the vertical position. The only thing I'm concerned about is the stretching of the belts when they get really long. A chain, or something else, might make more sense, but I haven't gotten that far into the design yet.

As for moving from vertical to horizontal, I'm going to try to keep it simple and attach slides/rollers features somewhere on the "torsion box" where I would make some curved rails (kinda looks like elephant tusks) to let gravity rotate it from the wall and into the horizontal position. And maybe a cheap HB wench to help out. This thing will probably weigh a hundred pounds or more.

Member

chamnit commented Feb 12, 2014

@Mgilbride : Thanks for the Makerslide info! It's interesting to know how accurate they are (0.0125"/foot straightness and 0.5deg/foot twist). Not terrible and acceptable for large router type machines that would be used for wood.

Lol. FEA analysis is a bit overkill for looking at deflection of rails. :) Simple beam analysis would have been plenty, but if you got the tools available, why not?

So, the plan for the vertical mill is to base the entire structure on a "torsion-box", rather than building a rigid frame. These are popular among woodworkers to create a very flat(~0.001"-0.005") and stable tabletop, but they are simply hollow-core boxes or honeycomb structures. The woodworker versions are WAY over-built and heavy. I'm thinking of something along the lines of a super-rigid aircraft wing-like design: thin spars, thin skins, and thicker/meaty attachment points. The idea is to fabricate everything with a ShapeOko2 (when I get my hands on one) so I can make all of the parts quickly, identical, and accurate.

I've been looking at mounting schemes and will probably use a gantry-rail type system like the ShapeOko. It's hard to beat the cost efficiency. But instead, I'll probably fabricate my own MakerSlide type rail system with some bigger and stiffer 80/20 and attach steel v-rails on them so I can use steel bearings. These rails would be mounted to the sides of the "torsion-box". I don't think I will necessarily need leadscrews, because I will be off-loading the weight of gantry and heads with constant force springs when in the vertical position. The only thing I'm concerned about is the stretching of the belts when they get really long. A chain, or something else, might make more sense, but I haven't gotten that far into the design yet.

As for moving from vertical to horizontal, I'm going to try to keep it simple and attach slides/rollers features somewhere on the "torsion box" where I would make some curved rails (kinda looks like elephant tusks) to let gravity rotate it from the wall and into the horizontal position. And maybe a cheap HB wench to help out. This thing will probably weigh a hundred pounds or more.

@r01walker

This comment has been minimized.

Show comment
Hide comment
@r01walker

r01walker Feb 12, 2014

i am of the mindset of making the tables themselves out of a torsion box arrangement. had a setback on the first machine. voltage spike fried my stepper boards. have new ones on order. first proves the system and makes the parts for the second. It seems that you are in engineering phase. so having some practical experience. and with no disrespect i will share some thoughts. I don't know if this is within the scope of this blog. i had the ardunio plugged into the computer and the stepper boards plugged into a old computer power supply. computer supply blinked and boards gone. i had a capacitor on the power lines. so next time i will put a cap and a battery on the stepper power lines along with diodes. 40 dollar lesson is preventative design is better than mail order and flying power supply. i went caveman with this assembly. i used a piece of foam board, in a old dvd case, put a fan in the hole on the top. and wired the ardunio and the stepper control's
to it. i used the pins of old floppy drives. i used 40 pin ide cable to join the components. use a butane torch to unsolder the components you want. i soldered the wire to the stepper boards and pins on the ardunio end. next time i will put the pins on the stepper and breadboard them mounting the breadboard beside the ardunio. (cleaner setup). if you have the cash buy a 4 stepper controller. three for x, y, z, then if a extruder machine feeder #4 stepper. their is a small possibility of using a android IOIO board to expand the uno. i think that a mega has additional port capacity and would accept the IOIO device easier. again i don't know if it would be connectable. their is a capacity to use the three volt line on the ardunio to allow for encoders (dc motors)  something that can be robbed out of dvd drives. power the motor and read the output 3V. baseline. rotate ccw 1.5v. cw 4.5v as a example. put a resistor across the poles to get a crisp volt out
and read step and dir. as a thought. 

i am considering finding some long keystock 12" or 18" pin the key stock and captivate it to hold it together or drill and tap and run all thread. possibly ground square stock use 3 bearing per corner blocks on fixed table to mount the slide. the working ares needs to be flat you can have a table that is 100 inches long and if the actual operating area is 7" square all you have to precisely support is the 7" square area. directly under the spindle. yes the machine needs to be rigid enough to hold itself and the possible components (belt, tensioned lead screw, anti backlash nuts, fixtures, etc. but the table not in the "precise" area does not need the "precise" support of the working areas. ideal for a torsion box type of table construction. repeatable yes. if you are going with a hollow core then i would look into some kind of vibration dampening system (foam). also pay attention to the bearings. dust and dirt pushes the bearing/slide out of alignment.
 i have seen metal chips get under box ways. (please don't ask, boss man was a caveman/blacksmith with no respect, and a big pry bar), and it caused some real fun until it was found. if you are making a extruder vibration, inertial forces, and rigidity is the big issue. i have been considering using a toothed belt drive instead of a lead screw. bicycle chain would be rigid enough and when tensioned vibration dampening to a point, definitely zero slip, i would think a bit jerky in small motions. might want to rob some old deraliers for parts if you go that way.

At worst case in a low side pressure application keep the steppers on and let them hold the machine axis. bigger power supply. higher side pressure machines use a lead screw/worm gear/stepper setup. tension the lead screw to keep it strait and running true. the tighter the better without distorting the threads. a lead screw doing a wobbly dance at high speed throwing everything off is hard to diagnose. 1.5 in dia bar hanging out of a lathe 5ft. at 300 rpm will bend. or put a 1/4-20 12" long piece of all thread in a hand drill  spin it up and do the oh shit dance. toughing up a machine to hold a lead screw under tension is not profitable for light duty work.

their are three type of machines gantry style, knee mill style, and jig borer style. gantry is the largest capacity. jig borer is the most precision and knee mill is old machine shop a good all around general purpose machine. i am considering a knee mill for the second machine. hold z axis guides as far apart as possible and i get the stability that i want. the extruder is the heaviest component. i am going with a extruder that has a worm gear feeder to enable the use of pellets, and the plastic i run through a shredder. the extruder is going to have to be rigid mounted. that point is the 0,0,0 point. 

On Wednesday, February 12, 2014 10:17 AM, Sonny Jeon notifications@github.com wrote:

@Mgilbride : Thanks for the Makerslide info! It's interesting to know how accurate they are (0.0125"/foot straightness and 0.5deg/foot twist). Not terrible and acceptable for large router type machines that would be used for wood.
Lol. FEA analysis is a bit overkill for looking at deflection of rails. :) Simple beam analysis would have been plenty, but if you got the tools available, why not?
So, the plan for the vertical mill is to base the entire structure on a "torsion-box", rather than building a rigid frame. These are popular among woodworkers to create a very flat(~0.001"-0.005") and stable tabletop, but they are simply hollow-core boxes or honeycomb structures. The woodworker versions are WAY over-built and heavy. I'm thinking of something along the lines of a super-rigid aircraft wing-like design: thin spars, thin skins, and thicker/meaty attachment points. The idea is to fabricate everything with a ShapeOko2 (when I get my hands on one) so I can make all of the parts quickly, identical, and accurate.
I've been looking at mounting schemes and will probably use a gantry-rail type system like the ShapeOko. It's hard to beat the cost efficiency. But instead, I'll probably fabricate my own MakerSlide type rail system with some bigger and stiffer 80/20 and attach steel v-rails on them so I can use steel bearings. These rails would be mounted to the sides of the "torsion-box". I don't think I will necessarily need leadscrews, because I will be off-loading the weight of gantry and heads with constant force springs when in the vertical position. The only thing I'm concerned about is the stretching of the belts when they get really long. A chain, or something else, might make more sense, but I haven't gotten that far into the design yet.
As for moving from vertical to horizontal, I'm going to try to keep it simple and attach slides/rollers features somewhere on the "torsion box" where I would make some curved rails (kinda looks like elephant tusks) to let gravity rotate it from the wall and into the horizontal position. And maybe a cheap HB wench to help out. This thing will probably weigh a hundred pounds or more.

Reply to this email directly or view it on GitHub.

r01walker commented Feb 12, 2014

i am of the mindset of making the tables themselves out of a torsion box arrangement. had a setback on the first machine. voltage spike fried my stepper boards. have new ones on order. first proves the system and makes the parts for the second. It seems that you are in engineering phase. so having some practical experience. and with no disrespect i will share some thoughts. I don't know if this is within the scope of this blog. i had the ardunio plugged into the computer and the stepper boards plugged into a old computer power supply. computer supply blinked and boards gone. i had a capacitor on the power lines. so next time i will put a cap and a battery on the stepper power lines along with diodes. 40 dollar lesson is preventative design is better than mail order and flying power supply. i went caveman with this assembly. i used a piece of foam board, in a old dvd case, put a fan in the hole on the top. and wired the ardunio and the stepper control's
to it. i used the pins of old floppy drives. i used 40 pin ide cable to join the components. use a butane torch to unsolder the components you want. i soldered the wire to the stepper boards and pins on the ardunio end. next time i will put the pins on the stepper and breadboard them mounting the breadboard beside the ardunio. (cleaner setup). if you have the cash buy a 4 stepper controller. three for x, y, z, then if a extruder machine feeder #4 stepper. their is a small possibility of using a android IOIO board to expand the uno. i think that a mega has additional port capacity and would accept the IOIO device easier. again i don't know if it would be connectable. their is a capacity to use the three volt line on the ardunio to allow for encoders (dc motors)  something that can be robbed out of dvd drives. power the motor and read the output 3V. baseline. rotate ccw 1.5v. cw 4.5v as a example. put a resistor across the poles to get a crisp volt out
and read step and dir. as a thought. 

i am considering finding some long keystock 12" or 18" pin the key stock and captivate it to hold it together or drill and tap and run all thread. possibly ground square stock use 3 bearing per corner blocks on fixed table to mount the slide. the working ares needs to be flat you can have a table that is 100 inches long and if the actual operating area is 7" square all you have to precisely support is the 7" square area. directly under the spindle. yes the machine needs to be rigid enough to hold itself and the possible components (belt, tensioned lead screw, anti backlash nuts, fixtures, etc. but the table not in the "precise" area does not need the "precise" support of the working areas. ideal for a torsion box type of table construction. repeatable yes. if you are going with a hollow core then i would look into some kind of vibration dampening system (foam). also pay attention to the bearings. dust and dirt pushes the bearing/slide out of alignment.
 i have seen metal chips get under box ways. (please don't ask, boss man was a caveman/blacksmith with no respect, and a big pry bar), and it caused some real fun until it was found. if you are making a extruder vibration, inertial forces, and rigidity is the big issue. i have been considering using a toothed belt drive instead of a lead screw. bicycle chain would be rigid enough and when tensioned vibration dampening to a point, definitely zero slip, i would think a bit jerky in small motions. might want to rob some old deraliers for parts if you go that way.

At worst case in a low side pressure application keep the steppers on and let them hold the machine axis. bigger power supply. higher side pressure machines use a lead screw/worm gear/stepper setup. tension the lead screw to keep it strait and running true. the tighter the better without distorting the threads. a lead screw doing a wobbly dance at high speed throwing everything off is hard to diagnose. 1.5 in dia bar hanging out of a lathe 5ft. at 300 rpm will bend. or put a 1/4-20 12" long piece of all thread in a hand drill  spin it up and do the oh shit dance. toughing up a machine to hold a lead screw under tension is not profitable for light duty work.

their are three type of machines gantry style, knee mill style, and jig borer style. gantry is the largest capacity. jig borer is the most precision and knee mill is old machine shop a good all around general purpose machine. i am considering a knee mill for the second machine. hold z axis guides as far apart as possible and i get the stability that i want. the extruder is the heaviest component. i am going with a extruder that has a worm gear feeder to enable the use of pellets, and the plastic i run through a shredder. the extruder is going to have to be rigid mounted. that point is the 0,0,0 point. 

On Wednesday, February 12, 2014 10:17 AM, Sonny Jeon notifications@github.com wrote:

@Mgilbride : Thanks for the Makerslide info! It's interesting to know how accurate they are (0.0125"/foot straightness and 0.5deg/foot twist). Not terrible and acceptable for large router type machines that would be used for wood.
Lol. FEA analysis is a bit overkill for looking at deflection of rails. :) Simple beam analysis would have been plenty, but if you got the tools available, why not?
So, the plan for the vertical mill is to base the entire structure on a "torsion-box", rather than building a rigid frame. These are popular among woodworkers to create a very flat(~0.001"-0.005") and stable tabletop, but they are simply hollow-core boxes or honeycomb structures. The woodworker versions are WAY over-built and heavy. I'm thinking of something along the lines of a super-rigid aircraft wing-like design: thin spars, thin skins, and thicker/meaty attachment points. The idea is to fabricate everything with a ShapeOko2 (when I get my hands on one) so I can make all of the parts quickly, identical, and accurate.
I've been looking at mounting schemes and will probably use a gantry-rail type system like the ShapeOko. It's hard to beat the cost efficiency. But instead, I'll probably fabricate my own MakerSlide type rail system with some bigger and stiffer 80/20 and attach steel v-rails on them so I can use steel bearings. These rails would be mounted to the sides of the "torsion-box". I don't think I will necessarily need leadscrews, because I will be off-loading the weight of gantry and heads with constant force springs when in the vertical position. The only thing I'm concerned about is the stretching of the belts when they get really long. A chain, or something else, might make more sense, but I haven't gotten that far into the design yet.
As for moving from vertical to horizontal, I'm going to try to keep it simple and attach slides/rollers features somewhere on the "torsion box" where I would make some curved rails (kinda looks like elephant tusks) to let gravity rotate it from the wall and into the horizontal position. And maybe a cheap HB wench to help out. This thing will probably weigh a hundred pounds or more.

Reply to this email directly or view it on GitHub.

@chamnit

This comment has been minimized.

Show comment
Hide comment
@chamnit

chamnit Feb 12, 2014

Member

@r01walker : Thanks for the insights. I've built a few CNC machines but never this large, so any prior lessons learned are always helpful.

Anyhow, the torsion box is planned to be around 5' wide, 5' to 8' long, and 6"-10" thick. I'll have to make sure it is constructed carefully and accurately since its the basis for the whole machine. I'm a structural engineer by trade, so I'll make certain any vibration problems are taken care of by adjusting the torsion-box spar dimensions and densities or, if I have to, foam the cavities as you suggested.

I'm opting for a gantry style mill because it doesn't require the machine move the workpiece. The gantry mass and dynamics always stays the same, so you don't have to over-engineer the motors and power requirements of having to move something very large, like a thick, 70lb MDF panel that this machine may be cutting. With a rail-bearing system, the bearings getting clogged up is definitely a real problem, but I think it'll be ok since this will be somewhat a light duty machine and bearings are very cheap and easy to replace.

Member

chamnit commented Feb 12, 2014

@r01walker : Thanks for the insights. I've built a few CNC machines but never this large, so any prior lessons learned are always helpful.

Anyhow, the torsion box is planned to be around 5' wide, 5' to 8' long, and 6"-10" thick. I'll have to make sure it is constructed carefully and accurately since its the basis for the whole machine. I'm a structural engineer by trade, so I'll make certain any vibration problems are taken care of by adjusting the torsion-box spar dimensions and densities or, if I have to, foam the cavities as you suggested.

I'm opting for a gantry style mill because it doesn't require the machine move the workpiece. The gantry mass and dynamics always stays the same, so you don't have to over-engineer the motors and power requirements of having to move something very large, like a thick, 70lb MDF panel that this machine may be cutting. With a rail-bearing system, the bearings getting clogged up is definitely a real problem, but I think it'll be ok since this will be somewhat a light duty machine and bearings are very cheap and easy to replace.

@r01walker

This comment has been minimized.

Show comment
Hide comment
@r01walker

r01walker Feb 12, 2014

i agree. as a suggestion. get some .031 aluminum sheet and fabricate bearing guards. superglue or jb weld or gorilla glue leather around the edges and or slides, possibly depending on design use the leather as a wipe for the bearing surface to where the dirt goes anywhere but the bearing. 

as a complement i had only considered only gantry style where the table moved. so i will have to think a bit. but i can see where a fixed table is useful. i am looking at multiple extruders so i can build the same part multiple times on one run. that would be a plus. 

one extreme thought along those lines i had was to make a rotary transfer style and extrude parts of a product per station. like different colors. not much different than a multiple spindle machine but i think more robust. if the center rotator is located with dowel pins or linear bearings like a punch press die pins. then i don't see any issue but the control of multiple axis per cycle. out of my league right now. but to put it out their. 

as a further thought i made a dust collector for my dads work shop look to instructables.com and search for dust collection systems. the documentation was really long and lots more involved as he was also interested in noise reduction. i used a old vacuum a 5 gal bucket and a old safety cone and made a really good dust extractor. it run off a 110v line so it could possibly be run of a relay on a ardunio. lungs work better when not filled with dust, don't understand why.

good luck and make lots of play toys. 

On Wednesday, February 12, 2014 4:24 PM, Sonny Jeon notifications@github.com wrote:

@r01walker : Thanks for the insights. I've built a few CNC machines but never this large, so any prior lessons learned are always helpful.
Anyhow, the torsion box is planned to be around 5' wide, 5' to 8' long, and 6"-10" thick. I'll have to make sure it is constructed carefully and accurately since its the basis for the whole machine. I'm a structural engineer by trade, so I'll make certain any vibration problems are taken care of by adjusting the torsion-box spar dimensions and densities or, if I have to, foam the cavities as you suggested.
I'm opting for a gantry style mill because it doesn't require the machine move the workpiece. The gantry mass and dynamics always stays the same, so you don't have to over-engineer the motors and power requirements of having to move something very large, like a thick, 70lb MDF panel that this machine may be cutting. With a rail-bearing system, the bearings getting clogged up is definitely a real problem, but I think it'll be ok since this will be somewhat a light duty machine and bearings are very cheap and easy to replace.

Reply to this email directly or view it on GitHub.

r01walker commented Feb 12, 2014

i agree. as a suggestion. get some .031 aluminum sheet and fabricate bearing guards. superglue or jb weld or gorilla glue leather around the edges and or slides, possibly depending on design use the leather as a wipe for the bearing surface to where the dirt goes anywhere but the bearing. 

as a complement i had only considered only gantry style where the table moved. so i will have to think a bit. but i can see where a fixed table is useful. i am looking at multiple extruders so i can build the same part multiple times on one run. that would be a plus. 

one extreme thought along those lines i had was to make a rotary transfer style and extrude parts of a product per station. like different colors. not much different than a multiple spindle machine but i think more robust. if the center rotator is located with dowel pins or linear bearings like a punch press die pins. then i don't see any issue but the control of multiple axis per cycle. out of my league right now. but to put it out their. 

as a further thought i made a dust collector for my dads work shop look to instructables.com and search for dust collection systems. the documentation was really long and lots more involved as he was also interested in noise reduction. i used a old vacuum a 5 gal bucket and a old safety cone and made a really good dust extractor. it run off a 110v line so it could possibly be run of a relay on a ardunio. lungs work better when not filled with dust, don't understand why.

good luck and make lots of play toys. 

On Wednesday, February 12, 2014 4:24 PM, Sonny Jeon notifications@github.com wrote:

@r01walker : Thanks for the insights. I've built a few CNC machines but never this large, so any prior lessons learned are always helpful.
Anyhow, the torsion box is planned to be around 5' wide, 5' to 8' long, and 6"-10" thick. I'll have to make sure it is constructed carefully and accurately since its the basis for the whole machine. I'm a structural engineer by trade, so I'll make certain any vibration problems are taken care of by adjusting the torsion-box spar dimensions and densities or, if I have to, foam the cavities as you suggested.
I'm opting for a gantry style mill because it doesn't require the machine move the workpiece. The gantry mass and dynamics always stays the same, so you don't have to over-engineer the motors and power requirements of having to move something very large, like a thick, 70lb MDF panel that this machine may be cutting. With a rail-bearing system, the bearings getting clogged up is definitely a real problem, but I think it'll be ok since this will be somewhat a light duty machine and bearings are very cheap and easy to replace.

Reply to this email directly or view it on GitHub.

@ZabojnikM

This comment has been minimized.

Show comment
Hide comment
@ZabojnikM

ZabojnikM Mar 28, 2014

I would use GRBL for poject cyclone cnc.
This project is designed primarily for milling PCB.
It's really small CNC. My version PortalCyclone also be used for cutting plastic and plexiglass.
Now we use electronics from RepRap 3D printer.
It would be great to use GRBL. But for our project it is necessary to use probing.
My simple pages about project http://www.cyclonecnc.4fan.cz/

ZabojnikM commented Mar 28, 2014

I would use GRBL for poject cyclone cnc.
This project is designed primarily for milling PCB.
It's really small CNC. My version PortalCyclone also be used for cutting plastic and plexiglass.
Now we use electronics from RepRap 3D printer.
It would be great to use GRBL. But for our project it is necessary to use probing.
My simple pages about project http://www.cyclonecnc.4fan.cz/

@chamnit

This comment has been minimized.

Show comment
Hide comment
@chamnit

chamnit Mar 29, 2014

Member

@ZabojnikM : Thanks for sharing! FYI, the newest version of Grbl supports probing, along with a lot of other stuff. It should be released shortly, but it's available as beta in the dev branch, so there might be a few minor bugs still there.

Member

chamnit commented Mar 29, 2014

@ZabojnikM : Thanks for sharing! FYI, the newest version of Grbl supports probing, along with a lot of other stuff. It should be released shortly, but it's available as beta in the dev branch, so there might be a few minor bugs still there.

@xP373Rx

This comment has been minimized.

Show comment
Hide comment
@xP373Rx

xP373Rx Mar 31, 2014

Im using grbl to drive my cnc with a bluetooth arduino and a custom ,made shield..
http://gsgrid.net/grblshield.jpg
The shield is capable of driving the steppers with pololu drivers drv8825 or a4988. i like the drv8825 better since they can stop the feed on fault of a driver.( I just dont know if the FAULT signal is fast enough to stop the feed before loosing steps ^^) It also can be connected to a external dsub25 driver board connector.. i just need to solve the pin assign issue , right now i do it with jumper cables if external driver board is connected.

I also tried serval cam tools but they all seem to be to complicated to do stuff fast .. I wanned something portable and easy to use to cut parts for my LED projects . So i started to look for ways i could generate some basic shape/curve/text code or import svg or generate patterns of shapes (like 20 holes in a row with distance of 10mm for example). I have found some javascript vector api called paper.js and the gcodephplib that i converted to java script so it all works local in the browser.
Here some preview.
http://gsgrid.net/gcode/test.php

It is pretty alpha right now but it generates already some messy but usable gcode per shape.

shift+ shift- zoom
shift* reset matrix
shift/ zoom to positive axes

xP373Rx commented Mar 31, 2014

Im using grbl to drive my cnc with a bluetooth arduino and a custom ,made shield..
http://gsgrid.net/grblshield.jpg
The shield is capable of driving the steppers with pololu drivers drv8825 or a4988. i like the drv8825 better since they can stop the feed on fault of a driver.( I just dont know if the FAULT signal is fast enough to stop the feed before loosing steps ^^) It also can be connected to a external dsub25 driver board connector.. i just need to solve the pin assign issue , right now i do it with jumper cables if external driver board is connected.

I also tried serval cam tools but they all seem to be to complicated to do stuff fast .. I wanned something portable and easy to use to cut parts for my LED projects . So i started to look for ways i could generate some basic shape/curve/text code or import svg or generate patterns of shapes (like 20 holes in a row with distance of 10mm for example). I have found some javascript vector api called paper.js and the gcodephplib that i converted to java script so it all works local in the browser.
Here some preview.
http://gsgrid.net/gcode/test.php

It is pretty alpha right now but it generates already some messy but usable gcode per shape.

shift+ shift- zoom
shift* reset matrix
shift/ zoom to positive axes

@zanderbier

This comment has been minimized.

Show comment
Hide comment
@zanderbier

zanderbier Apr 4, 2014

I am using GRBL in a personal project, constructing a paper cut with plotter blade.

The frame is almost done, just have to deal with some sort of solenoid to the blade support.

Design and patches are here: https://github.com/zanderbier/papercut

Thanks for this great code that GRBL is!

zanderbier commented Apr 4, 2014

I am using GRBL in a personal project, constructing a paper cut with plotter blade.

The frame is almost done, just have to deal with some sort of solenoid to the blade support.

Design and patches are here: https://github.com/zanderbier/papercut

Thanks for this great code that GRBL is!

@pratipo

This comment has been minimized.

Show comment
Hide comment
@scottrcarlson

This comment has been minimized.

Show comment
Hide comment
@scottrcarlson

scottrcarlson Feb 3, 2015

I think you're looking for a cam package that understands edm tool path
strategies more than the gcode, drive system having special edm features..

I.e. backing out of a cut is all part of the toolpath/gcode generator... or
if used manually is the job of the upper level app+ui

s.

On Sun, Jan 11, 2015, 06:11 Jaakko Fagerlund notifications@github.com
wrote:

Wire EDM

This I have done previously (check http://www.youtube.com/TheWireEDM)
with Arduino MEGA 2560 R3 board and my own code where I used just the G
code interpreter from the grbl project and implemented everything else. It
didn't offer any acceleations, just very slow (50 mm/min maximum) feed
rate, so stopping/starting suddenly didn't lose any steps.

I'm now actively trying if I can use grbl as such to control my machine
(2-axis wire EDM), just need to flash an Arduino and make the necessary
connections and do a simple system to toggle the feed hold/resume buttons
based on the spark gap voltage.

But the problem will be that grbl doesn't understand how to back-out from
the cut without external means. I'm gonna write a separate discussion of
this, as Google found only one discussion reagrding EDM and it was closed.

If the grbl dev team is interested in collaborating with EDM use, I'm all
for it and ready to test things out, as I have the machine and means :) I
can be contacted through email most easily through my profile.


Reply to this email directly or view it on GitHub
#332 (comment).

scottrcarlson commented Feb 3, 2015

I think you're looking for a cam package that understands edm tool path
strategies more than the gcode, drive system having special edm features..

I.e. backing out of a cut is all part of the toolpath/gcode generator... or
if used manually is the job of the upper level app+ui

s.

On Sun, Jan 11, 2015, 06:11 Jaakko Fagerlund notifications@github.com
wrote:

Wire EDM

This I have done previously (check http://www.youtube.com/TheWireEDM)
with Arduino MEGA 2560 R3 board and my own code where I used just the G
code interpreter from the grbl project and implemented everything else. It
didn't offer any acceleations, just very slow (50 mm/min maximum) feed
rate, so stopping/starting suddenly didn't lose any steps.

I'm now actively trying if I can use grbl as such to control my machine
(2-axis wire EDM), just need to flash an Arduino and make the necessary
connections and do a simple system to toggle the feed hold/resume buttons
based on the spark gap voltage.

But the problem will be that grbl doesn't understand how to back-out from
the cut without external means. I'm gonna write a separate discussion of
this, as Google found only one discussion reagrding EDM and it was closed.

If the grbl dev team is interested in collaborating with EDM use, I'm all
for it and ready to test things out, as I have the machine and means :) I
can be contacted through email most easily through my profile.


Reply to this email directly or view it on GitHub
#332 (comment).

@JaakkoFagerlund

This comment has been minimized.

Show comment
Hide comment
@JaakkoFagerlund

JaakkoFagerlund Feb 4, 2015

@scottcarlson : Uh no, CAM is not the place for machine control. CAM is used to make the tool path, which is then run through a post processor to get the machine language you need (Heidenhain, G code etc.), which is then sent to the machine. In the machine the motion controller or more properly the machine controller decides how the machine should operate under different conditions.

CAM is out of the cycle once the tool path is done. The only thing that has a sense of what the machine is doing and how the cut is going is the machine controller unit and it is the only thing that can be used.

Thus, grbl is the place to modify for such a need.

JaakkoFagerlund commented Feb 4, 2015

@scottcarlson : Uh no, CAM is not the place for machine control. CAM is used to make the tool path, which is then run through a post processor to get the machine language you need (Heidenhain, G code etc.), which is then sent to the machine. In the machine the motion controller or more properly the machine controller decides how the machine should operate under different conditions.

CAM is out of the cycle once the tool path is done. The only thing that has a sense of what the machine is doing and how the cut is going is the machine controller unit and it is the only thing that can be used.

Thus, grbl is the place to modify for such a need.

@blinkenlight

This comment has been minimized.

Show comment
Hide comment
@blinkenlight

blinkenlight Feb 4, 2015

@JaakkoFagerlund : Strictly speaking, that's not exactly true. A motion controller has no business taking the machine anywhere the CAM path didn't instruct it to, so to be fair the CAM software and the motion controller are supposed to be just about equally aware of what's where when (we can discuss independent movement as part of toolchanges and linear manual jogging if you insist, but neither is particularly relevant to this). With that in mind, there's nothing preventing an EDM-aware CAM software to finish the tool path with an appropriate exit sequence. A suitably devised controller is just as able to do this at any moment of course, but it's by no means "the only thing that can be used".

That aside though, the main reason you're barking up the wrong tree is that grbl is simply not in a position to accommodate arbitrary path reversals even if it would want to, considering grbl is never in possession of any significant portion of the G-code it is executing. At any given moment it only stores as much as it fits into its (rather small) buffers, relying on some kind of streamer counterpart to keep if fed, and properly reversing would require random access to the entire G-code being executed. There is of course nothing preventing said streamer software to implement a "stop and reverse from here" feature, but that has nothing to do with grbl; the one who can help you is the author of the streaming GUI you use...

blinkenlight commented Feb 4, 2015

@JaakkoFagerlund : Strictly speaking, that's not exactly true. A motion controller has no business taking the machine anywhere the CAM path didn't instruct it to, so to be fair the CAM software and the motion controller are supposed to be just about equally aware of what's where when (we can discuss independent movement as part of toolchanges and linear manual jogging if you insist, but neither is particularly relevant to this). With that in mind, there's nothing preventing an EDM-aware CAM software to finish the tool path with an appropriate exit sequence. A suitably devised controller is just as able to do this at any moment of course, but it's by no means "the only thing that can be used".

That aside though, the main reason you're barking up the wrong tree is that grbl is simply not in a position to accommodate arbitrary path reversals even if it would want to, considering grbl is never in possession of any significant portion of the G-code it is executing. At any given moment it only stores as much as it fits into its (rather small) buffers, relying on some kind of streamer counterpart to keep if fed, and properly reversing would require random access to the entire G-code being executed. There is of course nothing preventing said streamer software to implement a "stop and reverse from here" feature, but that has nothing to do with grbl; the one who can help you is the author of the streaming GUI you use...

@JaakkoFagerlund

This comment has been minimized.

Show comment
Hide comment
@JaakkoFagerlund

JaakkoFagerlund Feb 4, 2015

@blinkenlight : As a background, I'm a tool & die maker and use the EDM's almost daily, so I kind of know what I'm talking about.

The EDM motion controllers will of course follow the path the code says, BUT: when there is a short-circuit (and there will be), it will halt the execution and reverse back out the same way it came in until either A) it runs out of memory to back-out, B) program start point is reached or C) the short doesn't go away in a distance dictated by machine parameters. IF the short goes away, it will then come back to the point where it reversed, apply all the cutting conditions as they were and get back to executing the code.

CAM software is usually a CAD/CAM package, but it doesn't sit on the actual machine but on a PC. Sure it stands for Computer-aided Manufacturing, but that is way more than machine controller and usually refers to software sitting on a PC. Actual machine has its own controller, usually made by the machine manufacturer or bought as a unit, such as Fanuc, Siemens, Heidenhain. Nobody says those are CAM programs; they are machine controllers.

grbl as such isn't capable of reversing, that's true, and raping the feed hold & cycle start inputs just doesn't work properly (though it works for a while), so to use grbl in an EDM will require throwing the current motion planner out of the window and building on that base then.

Reversing motions is not an issue, just needs to store XY(UV) points as much as can fit in the memory (UNO might be just a tad small for a 4 axis controller) to get enough distance (which usually is a few millimeters). Just halt executing the G code, pick up points from the memory and travel there and if the short clears, come back those steps and then resume executing the G code. Requires zero participation from the G code feeder.

The fact is that (especially) with a wire EDM, if it shorts out and doesn't recover in a few millimeters, something is just so wrong that no amount of reversing would rescue. Thus there is no need to have access to the whole G code, just the last visited points in a ring buffer.

But this is getting seriously side tracked from the topic, so I would advice posting anything relating to EDM use in the proper thread that I just a few weeks ago did.

JaakkoFagerlund commented Feb 4, 2015

@blinkenlight : As a background, I'm a tool & die maker and use the EDM's almost daily, so I kind of know what I'm talking about.

The EDM motion controllers will of course follow the path the code says, BUT: when there is a short-circuit (and there will be), it will halt the execution and reverse back out the same way it came in until either A) it runs out of memory to back-out, B) program start point is reached or C) the short doesn't go away in a distance dictated by machine parameters. IF the short goes away, it will then come back to the point where it reversed, apply all the cutting conditions as they were and get back to executing the code.

CAM software is usually a CAD/CAM package, but it doesn't sit on the actual machine but on a PC. Sure it stands for Computer-aided Manufacturing, but that is way more than machine controller and usually refers to software sitting on a PC. Actual machine has its own controller, usually made by the machine manufacturer or bought as a unit, such as Fanuc, Siemens, Heidenhain. Nobody says those are CAM programs; they are machine controllers.

grbl as such isn't capable of reversing, that's true, and raping the feed hold & cycle start inputs just doesn't work properly (though it works for a while), so to use grbl in an EDM will require throwing the current motion planner out of the window and building on that base then.

Reversing motions is not an issue, just needs to store XY(UV) points as much as can fit in the memory (UNO might be just a tad small for a 4 axis controller) to get enough distance (which usually is a few millimeters). Just halt executing the G code, pick up points from the memory and travel there and if the short clears, come back those steps and then resume executing the G code. Requires zero participation from the G code feeder.

The fact is that (especially) with a wire EDM, if it shorts out and doesn't recover in a few millimeters, something is just so wrong that no amount of reversing would rescue. Thus there is no need to have access to the whole G code, just the last visited points in a ring buffer.

But this is getting seriously side tracked from the topic, so I would advice posting anything relating to EDM use in the proper thread that I just a few weeks ago did.

@blinkenlight

This comment has been minimized.

Show comment
Hide comment
@blinkenlight

blinkenlight Feb 4, 2015

It was a misunderstanding then, sorry - for some reason I was under the impression you were talking about fully backing out from somewhere halfway in the stock when you were actually talking about dynamic micro-back motions as part of the cutting process; that is indeed a different problem.

blinkenlight commented Feb 4, 2015

It was a misunderstanding then, sorry - for some reason I was under the impression you were talking about fully backing out from somewhere halfway in the stock when you were actually talking about dynamic micro-back motions as part of the cutting process; that is indeed a different problem.

@chamnit

This comment has been minimized.

Show comment
Hide comment
@chamnit

chamnit Feb 4, 2015

Member

@JaakkoFagerlund : To stir the pot some more, I don't think you have to throw out the current motion planner completely. It need refactoring to account for situations like you state. It'll take some work, but I'll keep some of these points in mind the next time I"m working on the planner.

Member

chamnit commented Feb 4, 2015

@JaakkoFagerlund : To stir the pot some more, I don't think you have to throw out the current motion planner completely. It need refactoring to account for situations like you state. It'll take some work, but I'll keep some of these points in mind the next time I"m working on the planner.

@JaakkoFagerlund

This comment has been minimized.

Show comment
Hide comment
@JaakkoFagerlund

JaakkoFagerlund Feb 4, 2015

@chamnit : Yeah, the problem is that I'm not familiar enough with the actual working of the planner, as most of the stuff doesn't happen by just reading code linearly but instead is affected by interrupts etc., so it really would be helpful to have a basic flow chart of how the program functions interact.

My solution to this problem was just recoding my own version from mainly scratch. It will accelerate/decelarate only on G0 seek and homing, all the feeds (G1) are done with some low maximum (50 mm/min) so that on every step it checks the gap voltage. So no interrupts used, as the Arduino has all the time in the world to do stuff like serial connection, parsing the G code etc., as the only fast movements are just rapid seeks from point A to B, needing no motion planning or anything happening during that move other than watching for limit switches.

But I'm very interested to see and hear if there comes a version that would handle this :) And keep us posted if you happen to have time some day to draw that flow chart ;)

JaakkoFagerlund commented Feb 4, 2015

@chamnit : Yeah, the problem is that I'm not familiar enough with the actual working of the planner, as most of the stuff doesn't happen by just reading code linearly but instead is affected by interrupts etc., so it really would be helpful to have a basic flow chart of how the program functions interact.

My solution to this problem was just recoding my own version from mainly scratch. It will accelerate/decelarate only on G0 seek and homing, all the feeds (G1) are done with some low maximum (50 mm/min) so that on every step it checks the gap voltage. So no interrupts used, as the Arduino has all the time in the world to do stuff like serial connection, parsing the G code etc., as the only fast movements are just rapid seeks from point A to B, needing no motion planning or anything happening during that move other than watching for limit switches.

But I'm very interested to see and hear if there comes a version that would handle this :) And keep us posted if you happen to have time some day to draw that flow chart ;)

@electrokean

This comment has been minimized.

Show comment
Hide comment
@electrokean

electrokean Feb 8, 2015

Hi all. Great to read about some of the interesting applications of Grbl. I've got one as well which may be of interest to others.
I'm in the process of implementing Grbl to control a 5-axis Non Destructive Testing (NDT) machine for automated checking rather expensive titanium parts for faults. I've added full-6 axis control on a Mega2560 based largely on the work of @EliteEng and so far it it seems to work great.
I'm now just working a bit on how soft limits are handled on rotating/tilting axes, and adding conditionals around my rotary axis changes in case someone wanted to merge it.
I'm be using the newly supported non-erroring G38.3 probe command for most movements to monitor the probe that can detect faults. The movement paths and G-code will be calculated via a custom Windows sender app with a UI, calibration process, and report generation specifically aimed at the test engineers.
So far I've only tested it on the bench with a few standalone motors and an oscilloscope, but I hope to try it on my 6040 mill, and then the real NDT machine, this coming week.

electrokean commented Feb 8, 2015

Hi all. Great to read about some of the interesting applications of Grbl. I've got one as well which may be of interest to others.
I'm in the process of implementing Grbl to control a 5-axis Non Destructive Testing (NDT) machine for automated checking rather expensive titanium parts for faults. I've added full-6 axis control on a Mega2560 based largely on the work of @EliteEng and so far it it seems to work great.
I'm now just working a bit on how soft limits are handled on rotating/tilting axes, and adding conditionals around my rotary axis changes in case someone wanted to merge it.
I'm be using the newly supported non-erroring G38.3 probe command for most movements to monitor the probe that can detect faults. The movement paths and G-code will be calculated via a custom Windows sender app with a UI, calibration process, and report generation specifically aimed at the test engineers.
So far I've only tested it on the bench with a few standalone motors and an oscilloscope, but I hope to try it on my 6040 mill, and then the real NDT machine, this coming week.

@darinwhite

This comment has been minimized.

Show comment
Hide comment
@darinwhite

darinwhite Apr 28, 2015

@chamnit: I used Grbl (and Easel) to reanimate an old TORCAM mill and lathe at a local high school here in Waterloo, Ontario Canada. The original turnkey PC that came with the system had long since rotted, so I substituted Grbl and a fly-wired shield to hook a modern laptop to the original motor drivers. Video of the mill firing up here: http://makebright.com/2014/11/torcam-cnc-mill-comes-alive-with-easelgrbl/ and some background on the reverse engineering required to interface Grbl to this beast http://makebright.com/2014/10/reverse-engineering-a-torcam-cnc-lathe-controller/. Thanks a lot for all your work on Grbl. DW

darinwhite commented Apr 28, 2015

@chamnit: I used Grbl (and Easel) to reanimate an old TORCAM mill and lathe at a local high school here in Waterloo, Ontario Canada. The original turnkey PC that came with the system had long since rotted, so I substituted Grbl and a fly-wired shield to hook a modern laptop to the original motor drivers. Video of the mill firing up here: http://makebright.com/2014/11/torcam-cnc-mill-comes-alive-with-easelgrbl/ and some background on the reverse engineering required to interface Grbl to this beast http://makebright.com/2014/10/reverse-engineering-a-torcam-cnc-lathe-controller/. Thanks a lot for all your work on Grbl. DW

@chamnit

This comment has been minimized.

Show comment
Hide comment
@chamnit

chamnit Apr 28, 2015

Member

@darinwhite : That is so awesome. I'm glad you were able to get something that probably would have been thrown away and to get it working for kids to start learning how to make. This made my day. :)

Member

chamnit commented Apr 28, 2015

@darinwhite : That is so awesome. I'm glad you were able to get something that probably would have been thrown away and to get it working for kids to start learning how to make. This made my day. :)

@ashelly

This comment has been minimized.

Show comment
Hide comment
@ashelly

ashelly May 21, 2015

Hi all. I ported Grbl to an Arduino Mega clone, added a 4th axis, and am using it to run all the motion in the KeyMe Kiosk, our robotic "Locksmith in a Box". We selected Grbl because of its clean, well structured implementation. It has been a pleasure to hack on it. Features added:

  • A "C" axis for a "conveyor" which delivers parts to the workspace. Uses the probe feature for alignment checks.
  • Axis 'Z' used for a gripper. Encoder support added. Reports counts with $e. Client uses this to check
    for lost counts
  • Allow each axis to be homed individualy with $HX. Per-axis home seek rates. Reports lost counts when homed.
  • Homing no longer blocks serial communications.
  • Option to report position continuously at compile-time configurable rate, currently set at 3Hz
  • The reported line number is incremented only when that line's move is complete.
  • Compiler option to extend the Idle Lock for up to 60 seconds on any axis
  • Serial protocol mods: Ensure each command gives 1 and only 1 response. Add checksum.
  • Sequenced start command, guaranteed not to be processed until previous line is done parsing.

I've fallen behind the master branch - when time permits I'm planning on updating. Until then, I'm happy to cherry-pick useful features into mainline Grbl if there is any demand (and any memory space) for them.

ashelly commented May 21, 2015

Hi all. I ported Grbl to an Arduino Mega clone, added a 4th axis, and am using it to run all the motion in the KeyMe Kiosk, our robotic "Locksmith in a Box". We selected Grbl because of its clean, well structured implementation. It has been a pleasure to hack on it. Features added:

  • A "C" axis for a "conveyor" which delivers parts to the workspace. Uses the probe feature for alignment checks.
  • Axis 'Z' used for a gripper. Encoder support added. Reports counts with $e. Client uses this to check
    for lost counts
  • Allow each axis to be homed individualy with $HX. Per-axis home seek rates. Reports lost counts when homed.
  • Homing no longer blocks serial communications.
  • Option to report position continuously at compile-time configurable rate, currently set at 3Hz
  • The reported line number is incremented only when that line's move is complete.
  • Compiler option to extend the Idle Lock for up to 60 seconds on any axis
  • Serial protocol mods: Ensure each command gives 1 and only 1 response. Add checksum.
  • Sequenced start command, guaranteed not to be processed until previous line is done parsing.

I've fallen behind the master branch - when time permits I'm planning on updating. Until then, I'm happy to cherry-pick useful features into mainline Grbl if there is any demand (and any memory space) for them.

@chamnit

This comment has been minimized.

Show comment
Hide comment
@chamnit

chamnit May 21, 2015

Member

@ashelly : Nice! Many of those features you've added are on my to-do list. I'll take a look at what's there and pull them in. :)

Member

chamnit commented May 21, 2015

@ashelly : Nice! Many of those features you've added are on my to-do list. I'll take a look at what's there and pull them in. :)

@109JB

This comment has been minimized.

Show comment
Hide comment
@109JB

109JB May 21, 2015

@ashelly Any chance you would be willing to share the 4th axis implementation?

109JB commented May 21, 2015

@ashelly Any chance you would be willing to share the 4th axis implementation?

@ashelly

This comment has been minimized.

Show comment
Hide comment
@ashelly

ashelly May 21, 2015

@109JB, The code is in my fork. But that version comes with enough caveats that it's unlikely you can use it as is:
I use the 4th axis completely independently, I have never tried running it in coordination with the other 3.
It only runs on our custom board, I stopped caring about compatibility with stock Arduino layouts, and broke CNC features I wasn't using.
The codebase is not up to date with the current master, and has tons of stuff you probably don't need.

I'll work on (or help chamnit with) a pull request for just the 4-axis feature.

ashelly commented May 21, 2015

@109JB, The code is in my fork. But that version comes with enough caveats that it's unlikely you can use it as is:
I use the 4th axis completely independently, I have never tried running it in coordination with the other 3.
It only runs on our custom board, I stopped caring about compatibility with stock Arduino layouts, and broke CNC features I wasn't using.
The codebase is not up to date with the current master, and has tons of stuff you probably don't need.

I'll work on (or help chamnit with) a pull request for just the 4-axis feature.

@polygontwist

This comment has been minimized.

Show comment
Hide comment
@polygontwist

polygontwist Jul 27, 2015

Hello, my Projekt is Laserengraver K403 with grbl:
http://www.a-d-k.de/20150611_213938-LasercutterK403.htm (german)
-Arduino UNO
-with modification of the water cycle (not ideal for continuous use, but it eliminates the bucket)
-with modification to Inkscape (laserengraver)
-i use Grbl Controller 3.0
salü

polygontwist commented Jul 27, 2015

Hello, my Projekt is Laserengraver K403 with grbl:
http://www.a-d-k.de/20150611_213938-LasercutterK403.htm (german)
-Arduino UNO
-with modification of the water cycle (not ideal for continuous use, but it eliminates the bucket)
-with modification to Inkscape (laserengraver)
-i use Grbl Controller 3.0
salü

@ilaro-org

This comment has been minimized.

Show comment
Hide comment
@ilaro-org

ilaro-org Mar 18, 2016

hi!
we just forked grbl to drive our grafitti
we are testing! drop us a line if you need help or have suggestions :D
https://github.com/ilaro-org/grbl-polar

ilaro-org commented Mar 18, 2016

hi!
we just forked grbl to drive our grafitti
we are testing! drop us a line if you need help or have suggestions :D
https://github.com/ilaro-org/grbl-polar

@bovo-fabiano

This comment has been minimized.

Show comment
Hide comment
@bovo-fabiano

bovo-fabiano Mar 28, 2016

I'm working on a XY system with belt linear actuators. At first I'm using a 2W diode laser but I would add others heads like pen drawing, pick and place, spindle, etc. I'm from Brazil.

img_20160328_200515570_hdr
img_20160328_200525591_hdr
img_20160328_200533385_hdr
img_20160328_200552693_hdr

https://youtu.be/AKZt2V5Amj4

bovo-fabiano commented Mar 28, 2016

I'm working on a XY system with belt linear actuators. At first I'm using a 2W diode laser but I would add others heads like pen drawing, pick and place, spindle, etc. I'm from Brazil.

img_20160328_200515570_hdr
img_20160328_200525591_hdr
img_20160328_200533385_hdr
img_20160328_200552693_hdr

https://youtu.be/AKZt2V5Amj4

@mkeyno

This comment has been minimized.

Show comment
Hide comment
@mkeyno

mkeyno Apr 13, 2016

hi guys, I have question , actually I decide to use GRBL to engrave image or letter on the metal ring , I've just intend to aligned my penomatic hammer with central line of rotary axis and use only axis A & Y ,instead of plannar X&Y axis any help or links really appreciated

mkeyno commented Apr 13, 2016

hi guys, I have question , actually I decide to use GRBL to engrave image or letter on the metal ring , I've just intend to aligned my penomatic hammer with central line of rotary axis and use only axis A & Y ,instead of plannar X&Y axis any help or links really appreciated

@langwadt

This comment has been minimized.

Show comment
Hide comment
@langwadt

langwadt Apr 13, 2016

@mkeyno I've done that to mill a spiral groove on a rod. I just looked down one axis and moved the wires driving to motor to the rotary axis, adjusted the steps/mm to match the circumference and wrote the g-code as if everything was "unwrapped"

langwadt commented Apr 13, 2016

@mkeyno I've done that to mill a spiral groove on a rod. I just looked down one axis and moved the wires driving to motor to the rotary axis, adjusted the steps/mm to match the circumference and wrote the g-code as if everything was "unwrapped"

@mkeyno

This comment has been minimized.

Show comment
Hide comment
@mkeyno

mkeyno Apr 14, 2016

thanks @langwadt but do you have any link or post that demonstrated such steps as you mentioned specially for > adjusted the steps/mm to match the circumference and wrote the g-code as if everything was "unwrapped"

mkeyno commented Apr 14, 2016

thanks @langwadt but do you have any link or post that demonstrated such steps as you mentioned specially for > adjusted the steps/mm to match the circumference and wrote the g-code as if everything was "unwrapped"

@langwadt

This comment has been minimized.

Show comment
Hide comment
@langwadt

langwadt Apr 14, 2016

@mkeyno I don't have any links, but I'll try to explain.
lets say you have rotary axis that is 3200 steps per revolution, and a part that is 20mm in diameter. one revolution of the part will be 20mm*pi = 62.83mm, 3200step/62.83mm = ~51steps/mm.

by unwrapping I mean you should make you gcode so that if you printed your design it on a piece
of paper and wrapped around your part you would get what you want

langwadt commented Apr 14, 2016

@mkeyno I don't have any links, but I'll try to explain.
lets say you have rotary axis that is 3200 steps per revolution, and a part that is 20mm in diameter. one revolution of the part will be 20mm*pi = 62.83mm, 3200step/62.83mm = ~51steps/mm.

by unwrapping I mean you should make you gcode so that if you printed your design it on a piece
of paper and wrapped around your part you would get what you want

@mkeyno

This comment has been minimized.

Show comment
Hide comment
@mkeyno

mkeyno Apr 14, 2016

thanks @langwadt , so you say I just need to set the right rotary steps/mm and then X gcode will do as it always do in planar x axis

mkeyno commented Apr 14, 2016

thanks @langwadt , so you say I just need to set the right rotary steps/mm and then X gcode will do as it always do in planar x axis

@lytke

This comment has been minimized.

Show comment
Hide comment
@lytke

lytke Apr 28, 2016

I'm very impressed of what the GRBL can be used for. Controlling XYZ are very generic and can be used in many ways.

I have used GRBL in three projects:
To build a clone of Handibot (Open source project by Shopbot)
Laser engraver
Plotter using coreXY

The last project is currently ongoing and I'm working on using the spindle PWM to make a servo lift the pen. It is looking promising and I have made good use of tips from #869

Thanks
Per.

lytke commented Apr 28, 2016

I'm very impressed of what the GRBL can be used for. Controlling XYZ are very generic and can be used in many ways.

I have used GRBL in three projects:
To build a clone of Handibot (Open source project by Shopbot)
Laser engraver
Plotter using coreXY

The last project is currently ongoing and I'm working on using the spindle PWM to make a servo lift the pen. It is looking promising and I have made good use of tips from #869

Thanks
Per.

@tworiverstay

This comment has been minimized.

Show comment
Hide comment
@tworiverstay

tworiverstay May 9, 2016

Hi
We had some stepper motors from a 3D printer, Went on YouTube and viewed NYC CNC.
Also saw drawing machine video from resonance studio, so made a drawing machine:
https://www.youtube.com/watch?v=r8SeBaYpAps

Now looking at introducing a fourth stepper motor to rotate pen and make more complex patterns.

Novice. Not sure what's going on - but good fun and enjoying the ride

Dave Smith, Physics Glenalmond College

tworiverstay commented May 9, 2016

Hi
We had some stepper motors from a 3D printer, Went on YouTube and viewed NYC CNC.
Also saw drawing machine video from resonance studio, so made a drawing machine:
https://www.youtube.com/watch?v=r8SeBaYpAps

Now looking at introducing a fourth stepper motor to rotate pen and make more complex patterns.

Novice. Not sure what's going on - but good fun and enjoying the ride

Dave Smith, Physics Glenalmond College

@Mgilbride

This comment has been minimized.

Show comment
Hide comment
@Mgilbride

Mgilbride May 10, 2016

Absolutely beautiful. I love it.
On May 9, 2016 3:29 PM, "tworiverstay" notifications@github.com wrote:

Hi
We had some stepper motors from a 3D printer, Went on YouTube and viewed
NYC CNC.
Also saw drawing machine video from resonance studio, so made a drawing
machine:
https://www.youtube.com/watch?v=r8SeBaYpAps

Now looking at introducing a fourth stepper motor to rotate pen and make
more complex patterns.

Novice. Not sure what's going on - but good fun and enjoying the ride

Dave Smith, Physics Glenalmond College


You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#332 (comment)

Mgilbride commented May 10, 2016

Absolutely beautiful. I love it.
On May 9, 2016 3:29 PM, "tworiverstay" notifications@github.com wrote:

Hi
We had some stepper motors from a 3D printer, Went on YouTube and viewed
NYC CNC.
Also saw drawing machine video from resonance studio, so made a drawing
machine:
https://www.youtube.com/watch?v=r8SeBaYpAps

Now looking at introducing a fourth stepper motor to rotate pen and make
more complex patterns.

Novice. Not sure what's going on - but good fun and enjoying the ride

Dave Smith, Physics Glenalmond College


You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#332 (comment)

@frett27

This comment has been minimized.

Show comment
Hide comment
@frett27

frett27 May 15, 2016

We Use GRBL for punching old mecanical barrel organ books :
here is it in action, using a modified grbl version.

prototype version video :
https://www.youtube.com/watch?v=WrCj02v7GNU

website :
http://www.barrel-organ-discovery.org/site/index.html

frett27 commented May 15, 2016

We Use GRBL for punching old mecanical barrel organ books :
here is it in action, using a modified grbl version.

prototype version video :
https://www.youtube.com/watch?v=WrCj02v7GNU

website :
http://www.barrel-organ-discovery.org/site/index.html

@Art999

This comment has been minimized.

Show comment
Hide comment
@Art999

Art999 May 20, 2016

lytke,

I am trying to implement a corexy plotter. I have built the mechanical/electrical system BUT I am unable to make the appropriate changes to grbl v0.9j . What I have tried is to simply un-comment the statement " #define COREXY" in the config.h file. That did not work although GRBL compiled normally.

What does one have to do to use GRBL with a corexy system???

Art999

Art999 commented May 20, 2016

lytke,

I am trying to implement a corexy plotter. I have built the mechanical/electrical system BUT I am unable to make the appropriate changes to grbl v0.9j . What I have tried is to simply un-comment the statement " #define COREXY" in the config.h file. That did not work although GRBL compiled normally.

What does one have to do to use GRBL with a corexy system???

Art999

@lytke

This comment has been minimized.

Show comment
Hide comment
@lytke

lytke May 30, 2016

@Art999,

Take a look at this link
#957 (comment)

/lytke

lytke commented May 30, 2016

@Art999,

Take a look at this link
#957 (comment)

/lytke

@Art999

This comment has been minimized.

Show comment
Hide comment
@Art999

Art999 Jun 1, 2016

lytke,

Thanks! . I had finally got it working BUT I only did #s 3 and 4. I will
now go ahead and do your #s 1 & 2 as well. Will homing work then as well?

Thanks again.
Art

On Mon, May 30, 2016 at 4:20 PM, Per Lytkemeyer notifications@github.com
wrote:

@Art999 https://github.com/Art999,

Take a look at this link
#957 (comment)
#957 (comment)

/lytke


You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#332 (comment), or mute
the thread
https://github.com/notifications/unsubscribe/ASlsLXb19OdoDEel3T7QL5s5d-RcLjddks5qG0aNgaJpZM4BaP8q
.

Art999 commented Jun 1, 2016

lytke,

Thanks! . I had finally got it working BUT I only did #s 3 and 4. I will
now go ahead and do your #s 1 & 2 as well. Will homing work then as well?

Thanks again.
Art

On Mon, May 30, 2016 at 4:20 PM, Per Lytkemeyer notifications@github.com
wrote:

@Art999 https://github.com/Art999,

Take a look at this link
#957 (comment)
#957 (comment)

/lytke


You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#332 (comment), or mute
the thread
https://github.com/notifications/unsubscribe/ASlsLXb19OdoDEel3T7QL5s5d-RcLjddks5qG0aNgaJpZM4BaP8q
.

@m0n5t3r

This comment has been minimized.

Show comment
Hide comment
@m0n5t3r

m0n5t3r Apr 6, 2017

600x600mm MPCNC machine (about 300x300 usable area), connected to an orange pi zero running cncjs; yet to make anything significant with it (I engraved a logo for a friend, had it drill holes in its own base for threaded inserts, still fumbling around with work holding); I still need to make a box for the electronics and am waiting for a 48v spindle I'm going to rig up with PWM control, etc.

actually, I just got it to control a 2.5W banggood laser (after much cursing the mosfet I had at hand, which wouldn't open at the 4.3 volts the arduino outputs, so needed another mosfet and a npn transistor to double-invert the signal); I was writing here about how I'm confused about the power / spindle speed setting when I figured I should set the max speed to 100 for laser (since cncjs uses percents for speed); I think stood for about an hour looking at the oscilloscope and and scratching my head :)

I also have a tiny Proxxon MF70 I'm planning to convert to CNC sometime in the future, it will probably run the same stack

m0n5t3r commented Apr 6, 2017

600x600mm MPCNC machine (about 300x300 usable area), connected to an orange pi zero running cncjs; yet to make anything significant with it (I engraved a logo for a friend, had it drill holes in its own base for threaded inserts, still fumbling around with work holding); I still need to make a box for the electronics and am waiting for a 48v spindle I'm going to rig up with PWM control, etc.

actually, I just got it to control a 2.5W banggood laser (after much cursing the mosfet I had at hand, which wouldn't open at the 4.3 volts the arduino outputs, so needed another mosfet and a npn transistor to double-invert the signal); I was writing here about how I'm confused about the power / spindle speed setting when I figured I should set the max speed to 100 for laser (since cncjs uses percents for speed); I think stood for about an hour looking at the oscilloscope and and scratching my head :)

I also have a tiny Proxxon MF70 I'm planning to convert to CNC sometime in the future, it will probably run the same stack

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment