Skip to content
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.

Sign up for GitHub

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

Jump to bottom

Improving heat distribution #210

Open
bgrigoriu opened this issue Jun 23, 2021 · 71 comments
Open

Improving heat distribution #210

bgrigoriu opened this issue Jun 23, 2021 · 71 comments

Comments

@bgrigoriu
Copy link

Hi everyone,
Just received my T962 and doing some initial testing.
After reading a lot of post the thing that was still apparent for me is that the heat distribution is uneven. Using a fan improves it but power is insufficient to heat-up enough fast and high.
Did anyone test adding one/two lamps (apart from the issues regarding power dissipation over the triac) ? Does it improves anything ?
I am not a fan of radical changes but mechanically it is do-able.
I would like to spare time if anyone has already done it (Just ADDING the same type of lamp).

Second question: it seems evident that there is left to right gradient of temp . Did somebody has an sound explanation for this ?
@GitLang
Copy link

GitLang commented Jun 23, 2021

I'd need to look at the schematic, but on my earlier 220V model, the lamps are switched by a solid state relay which is well overrated.

@bgrigoriu
Copy link
Author

Hi,
The actuals lamps seems to have between 300 and 400W. So we would need one or two more that means 1200W to max 1800 W.
The actual triac is a 16A BT139- 600E. Did not find easily a replacement for it in the same size and the same (10mA) gate sensitivity. Bigger (20/25 or 40A) may work but I would need to test and thus to have another T962 electronic board which I lack.
However with a change of the heat sink can be done easily and improve max current.
I will test and post.

In the meantime I have disassembled mine and I think I have an explanation for the huge right to left difference in temperature that some of the users report.

The lateral part of the oven are very poorly isolated with chucks of Si/Mg plaques that leave a lot of free space and some metal exposed.
I have replaced them with a continuous plaque and the difference between the right and the left part average now only 3.3°C during the heating phase and 6.3 °C during the cooling phase when the fan is on (which is expected).
I will retest with the new firmware but this is a thermodynamic issue and no depreciation/improvement can be expected.
If anyone is interested I have the photos and a video with instructions but I do not ( yet?) know how to post them on GitHub (newbie).

@GitLang
Copy link

GitLang commented Jun 24, 2021

So they have done away with the 25A solid state relay, and replaced it with a lower rating triac?. How many lamps of what power are fitted as standard?. Is this the small A version or the larger B version (or don't they have separate versions now?). 3C to 6C difference is much better than a lot of us get. What is a silicon manganese plaque? (sorry for so many questions)

@bgrigoriu
Copy link
Author

On my machine T962 (small version produced 04/ 2021 if I read correctly) there is a 16A triac on the PCB.
If I read ok the images on various web pages only the larger T962A machine has a SSR fitted on the back. I do not think that there is enough space in the back to place a SSR on the smaller T962 machine.

I think it is possible to use the actual triac (16A) at 8A (i.e. 1600W). As for now the temp of the triac seems to not be very high but I still need to make precise measures..
At 8A it will dissipate around 12W ( compared to around 6W now). A lager heat sink can compensate this. I will try to see how to fit one.
For the difference of temp yes I got a left to right difference 3 °C during heating phase and a bit over 6 when cooling ( which is unimportant in my opinion). The silica/Mg plaque is a "stupid" thermal shield your plumber is using for gas welding. I am buying it from my local DIY store (https://www.bricodepot.fr/leers/ecran-thermique-souple/prod5967/# - do not take as a advertisement) but you can find similar things on ebay
(https://www.amazon.fr/WELDTEAM-Ecran-pareflamme/dp/B00BBU60Z2/ref=sr_1_9?__mk_fr_FR=%C3%85M%C3%85%C5%BD%C3%95%C3%91&crid=1MXURE4SRLAN1&dchild=1&keywords=ecran+thermique+soudure&qid=1624536152&sprefix=ecran+thermique+%2Caps%2C150&sr=8-9 for ex).

Thank you for asking, I am also asking a lot of questions as I am an (old) newbie and starting to reflow .

By the way I think that a clear procedure on how to calibrate the TC ( using the provided OA) would be useful. Some direction are given but they are not quite clear and refer to the use of mathlab (which I do not have being alone in my kitchen).

Currently I have a difference of about 8 °C between the two TC (tested at lower temps which is probably not relevant) BUT questions remaining are:

  1. Do you need (is it of any utility) to calibrate first using the pot's on the PCB and then by software or all can be done by the software
  2. What is the best way to obtain the two temp steps for calibration (a separate reflow program included in the firmware would be great).

@xnk
Copy link
Member

xnk commented Jun 24, 2021 via email

@bgrigoriu
Copy link
Author

Thanks !
I was going to do this today and then test
I will try find some time next week to solder a 31850/31855 PCB and then compare.

@bgrigoriu
Copy link
Author

I have used the bake mode to set the two temperatures needed to calculate the gain and offset adjustments. I’m trying to remember whether the pots affected the gain or offset, but you shouldn’t have to touch them and instead compensate in firmware (as long as you won’t hit the ADC limit). The offset adjustment simply adjust the readout up or down, the gain adjustment will adjust the gain +/-10% so by calculating the “real” temperature difference and the setpoint difference between the two temperatures you can calculate what the compensation needs to be.

Can you tell how to isolate left and right TC channels ? if one is absent how is temp is calculated

@bgrigoriu
Copy link
Author

Sorry, I hit enter to fast.
How is temp calculated with only one channel ?

@xnk
Copy link
Member

xnk commented Jun 24, 2021 via email

@bgrigoriu
Copy link
Author

Thank you to all of you for the help.
I just finished to calibrate the TC. Not a very fast task.
I confirm that during the heating phase, with the change of the lateral insulation. the maximum difference between the right and the left side is no more than 3 degrees C in average. I have only <10 seconds a 4°C difference (highest one).
I looked with a thermal camera and the part that a losing most heat are in order : the front part of the drawer and the bottom lateral part both exceeding 40°C during heating phase.
However, during the cooling phase, the left side is cooling hyper fast and the differences are HUGE; I even had more than 30 °C during almost one minute. I think that air circulation is a mess and I do not have any modality to know where the issue is. I am not sure that makes a great difference anyway (and I do not know who to ask).
Thank you again for the great work done by all the contributors to the project. You deserve a medal.
I have just another short question. is there a documentation on what the firmware transmits over the serial and how to use it ?

@bgrigoriu
Copy link
Author

Hi again,
I also looked at the taliesin post about introducing an air straightener with the claim that it reduced strongly the temp differences during cooling phase. Did anyone also tested it ?
And if yes do you know if slimmer 120mm fans exists?

@xnk
Copy link
Member

xnk commented Jun 29, 2021 via email

@DimaSM338
Copy link

Did anyone test adding one/two lamps (apart from the issues regarding power dissipation over the triac) ? Does it improves anything ?

I installed two additional lamps of the same since my board is 200x140 mm. The result is positive. I was able to complete the task.

Without additional lamps, the edges of the board did not warm up. With additional lamps everything is OK

@bgrigoriu
Copy link
Author

I installed two additional lamps of the same since my board is 200x140 mm. The result is positive. I was able to complete the task.

Without additional lamps, the edges of the board did not warm up. With additional lamps everything is OK

Hi,
Did you do any changes ? which lamp did you use ? 110V or 220V ? which power
The triac was "happy" ? did you measure a surface temp on the triac ?
On my machine with the original configuration the triac heats up to 50°C during a ramp test to 245°C.
Based on specs it should dissipate 8W in this situation. with two more lamps in wil go up to 16W.
Can you share some photos/ details ?

@bgrigoriu
Copy link
Author

Hi does anyone has the exact dimmension of the PCB ( and the position of the fixing screws ?) I would like to build a thermocouple board which will include an Wifi interface and a USB and I will need for designing it.
I tried to measure on the board but it it not enough precise

@GitLang
Copy link

GitLang commented Jul 4, 2021 •
edited
Loading

Buy some digital calipers, they are cheap now. Close calipers over a stud/fixing hole, and zero the caliper. Now measure over a pair of studs/holes and the reading is the centre distance between them, Do this for the two holes in 'X' and two holes in 'Y' You should get better than 0.1mm accuracy total. Leave 0.5mm clearance on your PCB holes and you should be fine. More accurate than measure the fixing points on the machine, would be to remove the board and measure the holes centres on the board using the same method. Measuring the board outline itself is trivial with calipers, but don't forget tor reset the zero on the calipers :).

@bgrigoriu
Copy link
Author

bgrigoriu commented Jul 4, 2021 via email

@DimaSM338
Copy link

I installed two additional lamps of the same since my board is 200x140 mm. The result is positive. I was able to complete the task.
Without additional lamps, the edges of the board did not warm up. With additional lamps everything is OK

Hi,
Did you do any changes ? which lamp did you use ? 110V or 220V ? which power
The triac was "happy" ? did you measure a surface temp on the triac ?

Hi,
In the form in which the oven is sold, it can hardly be called a soldering oven, therefore, yes, there were many changes. Below I will try to tell you more.

  1. First, eliminated all uncontrolled heat loss. I bought a rubber oven door seal and stuck it around the edge of the PCB tray. In the photo, all 4 re lamps are just visible.

image

Glued with special high-temperature silicone.
image

The profile of the rubber seal is like this. The bottom of the card tray is also papered.
image

Also inside the oven, I taped two of the four lower air inlet grooves towards the fan.
image

In addition, I added a lot of mineral insulation to the case (when working with it, use a respirator and gloves). All slots were closed with Kapton high-temperature tape.
This improved the situation, but still prevented me from melting large boards - which I am working with around 185 x 130mm. The center was burnt, the edges of the board were not melted.

  1. After that I decided to install two additional lamps. I developed a modernization plan and blueprint and bought lamps.
    image

I bought lamps in China on Alliexpress (links inside) - exactly the same as they are in the oven.
image

image

They cost about $ 50. The lamps themselves are designed for 110 volts, but since I have a voltage of 220 volts, naturally they are connected in series with me. Series connection gives less current through the triac. If you have 110 volts in the network, then you will not be able to do this trick. The lamps will have to be switched on in parallel. With triac, you have to do something. I did not photograph the process of installing the lamps - I did not think that I would have to talk about it.

I can only give a drawing and say that I cut the holes for the new places of the lamps with such a tool with a diameter of 14mm.
image

In the lid where the sensors are located, I had to work like this below
image

The horizontal grooves were cut with a conventional grinding machine with a circle of 125 mm. The lower table did not have to be dismantled from the body. We managed to do everything on the assembled case. Angles at 45 degrees in the grooves were cut with ordinary metal scissors.
I sealed the old places where the lamps were installed with aluminum self-adhesive tape and Kapton tape.

Here you can see the places of the old lamps and how I organized the sensors.
image

Note - When installing the lamps, use only special silicone wires, since the temperature there is about 350 degrees.
With 4 re-lamps on board - the ramp test results turned out like this.
image

Cooling in my opinion is not enough - this is with half-taped slots to the fan. I'll try to open them later.

I will answer your question on triac.
triac was replaced by another - instead of BT139-600E.127 (16A) I installed BTA140-800 (26A). They have the same package and pinout. So I soldered it to the same place. I have not changed any other details in the triac harness. Additionally, I added the largest possible heatsink and used quality Arctic thermal grease. ... The radiator profile is like this:
image

Radiator size 37x30x15 mm. The hole for the triac attachment is drilled slightly offset.
image

And since I am not a supporter of noise, I immediately turned off the small fan that was in the oven. I made an additional window in the case on top - in the oven lid, with a grill for a thin 120x120 mm fan and planned to install it.
image
I haven’t set it yet, as everything works well without it.

All these activities allowed to leave the triac in the same place. During test switching on twice the number of lamps, the triac did not heat up above 60 degrees. (I could hold on to the radiator with my hand !!! (With the power off  )
I recently made a series of 10 boards, nonstop for 1 hour. BTA140-800 stayed alive. But I still made the two external lamps switchable, so as not to burn off electricity when they are not needed - when I solder a narrow board, I work on internal lamps.All these activities allowed to leave the triac in the same place. During test switching on with twice the number of lamps, the triac did not heat up above 60 degrees. (I could hold on to the heatsink with my hand !!! With the power off
3. Also, all firmware updates have been made for the furnace - interfaces for K-type MAX31855 sensors have been added - 4 pieces and a USB connector. Added 2 additional sensors and arranged sensor locations according to the blueprint:
image

I do not use contact sensors. I use the same sensors that I used, but I lowered them as low as possible to the board. See drawing for details. Above there is a photo where you can see how I installed the sensors in sleeves with a diameter of 5 mm.
4. In the PCB tray, I made nuts and swivel sliding mounts and raised the PCB over the tray by 8 mm - and somewhere the same distance I lifted the lamps when installing new ones.
image

This allowed us to always install the boards the same for the series of boards - and get the same results on the series of boards. It is important to always center the boards in the center of the oven - note that it does not coincide with the center of the tray. Therefore, the center of the oven on the tray is marked with a marker.
image

  1. Now I'm trying to solve an issue with a program for Windows - Reflow Controller.exe in an adjacent branch - Executable file for Reflow Controller software #201 (comment) - I will be grateful for any help in fixing code.

What else I am planning is to make an additional screen on the bottom of the oven - since 4 re-lamps will significantly heat the surface under the oven, especially when a small board is melted.

And yes .. I plan to replace the LEDs on the heating and cooling indicator, otherwise it's not very good with the white insulating tape :)

@DimaSM338 DimaSM338 mentioned this issue Jul 4, 2021
Open
@bgrigoriu
Copy link
Author

Thanks a lot for these detailled explanations !

@GitLang
Copy link

GitLang commented Jul 5, 2021 •
edited
Loading

The rubber seal round the PCB tray will greatly reduce the cooling airflow if it is anything like my next larger model, as the air is exhausted via that route. You can see on your temperature profile that the final temperature drops slowly and exponentially rather than the fast linear slope I get without a seal. Try removing all or removing sections of the seal. That final cool down is important on the time-temperature product of a lot of components.

Permanent standoffs is a good idea as long as you can position your board to not get soldered to them :). I use MF 6mm M3 brass standoffs fitted one in each board mounting hole with a nut. If you use a dome nut or long studded standoff you may have clearance enough to turn the board over to do the other side.

@bgrigoriu
Copy link
Author

bgrigoriu commented Jul 5, 2021 via email

@zian31
Copy link

zian31 commented Jul 14, 2021

Wow DimaSM338, I'm impressed with your upgrades !
I would prefer buy a new T962A with 4 IR lamps instead of my T962 than add 2 IR lamps.
I my case, I don't use MAX31855 sensors, but MAX31850KATB+ (I've added the PNOXI board).
I disconnect the 2 original TC of the T962 and put the 4 TC in the 4 angles on the PCB board to solder (with Kapton), but the firmware is only using 2 TC values for it's algo : I will change it in the C code to use the 4 TC values next.
Now I'm always working on a new windows software with curves dialoging with UART Port Com, it's almost working.

I'm wondering why you don't prefer to add a SSR-25A-DA (as existing in T962A) for the heating part instead of this new TRIAC+dissipator, it seems more simple to do : just a little place to find in the T962. And also possible to add another one for the cooling fan.

I think it's very important to put the TC sensors on the PCB because it's the source of the PID control.

Anyway, my problem is not the same as yours with your 200x140mm PCB : in my case with this kind of boards and the custom#2 profile, the time of cooling is too long, the heating time is more acceptable. And I have the T962Z new model (hard to say what are the exact differences with the original T962).

@bgrigoriu
Copy link
Author

Now I'm always working on a new windows software with curves dialoging with UART Port Com, it's almost working.

And I have the T962Z new model (hard to say what are the exact differences with the original T962).

Hi,
I did not know that a Z version exists. I looked on the net and the images as identical with the “classical” T962 with the 2020 upgrade on the exhausting tube connection ( which on my occasional use I am not sure is very useful).

Do you have some Photos of the Interior design and the PCB ?

It is wonderful that you could develop a windows software for communicating /controlling the Owen. Can you give mode details?

I am very interested by this. I put on the ISP connector an ESP-01 with ESP-link inside and it work well in terms of connection but it is not very practical ( apart getting rid of a wire) because the oven firmware does dot expects any other character after each command and is not easy to use a terminal connected with it. Your software could be a solution.

I wonder also if your code could be easily re-writed as a web page as the ESPlink can host a web page and thus the oven could be controlled in a platform-independent manner by a web browser.

@zian31
Copy link

zian31 commented Jul 14, 2021

I ever used two ESP-01 to speak together with UART on each side, but I didn't know this ESP-link with Wifi, it's interesting.
Unfortunatly, my windows soft will not be compatible with a web browser, but it will work also on Linux.
I use the PNOXI board with a FTDI chip (or equivalent) to create a virtual COM port with a USB cable.

The main interest of a software is the see curves of differents TC in live (without it's possible to do it on Excel after), and I want to be able to upgrade the oven EEPROM from this soft (with also a firmware modification needed) to change the custom#1 profile as needed.

Concerning the T962Z version, it's the name of this 2020 upgrade with the exhausting tube, here are the differences I've noticed with the T962 :

  • Power Earth connexion already done with the rack
  • Circular output fan on behind
  • With the T962 air comes from behind and get out with the bottom ? With the T962Z, the air seems to come from the center of the circular fan and get out by the borders of the circular fan output, and there is almost no circulation with the bottom of the rack : it's strange

@bgrigoriu
Copy link
Author

Thanks for the response.
I tested the ESP 01 and virtual TCP to serial port (either comtocom or HW VSP3). They work very well even with the existing reflow controler software or everything else.
If you finish you software and are eager to share the code, one can try to copy the algorithm and implement it on the webserver of the ESP link. I think it will be widely useful.
I also designed a PCB wich include en ESP plus 4 TC controllers ( versions with 31855 and 31850 available) but I was not yet able to test them as I am waiting for the PCB from China.

@George-2005
Copy link

In the meantime I have disassembled mine and I think I have an explanation for the huge right to left difference in temperature that some of the users report.

The lateral part of the oven are very poorly isolated with chucks of Si/Mg plaques that leave a lot of free space and some metal exposed.
I have replaced them with a continuous plaque and the difference between the right and the left part average now only 3.3°C during the heating phase and 6.3 °C during the cooling phase when the fan is on (which is expected).

If anyone is interested I have the photos and a video with instructions but I do not ( yet?) know how to post them on GitHub (newbie).

It sounds great, actually! I've just bought T-962 so I'm preparing now for oven improvements. Please let me ask: Have you replaced original Si/Mg insulation completely (even at the top where thermocouples are situated)? Or have you replaced it at the sides (left & right only)? And is the just one piece of "Écran thermique souple - CASTOLIN" sufficient or it has to be used more pieces?

I'm very interested in watching your photos and video of replacing the insulation. Could it be possible to provide them? Thank you

@bgrigoriu
Copy link
Author

Please let me ask: Have you replaced original Si/Mg insulation completely (even at the top where thermocouples are situated)? Or have you replaced it at the sides (left & right only)? And is the just one piece of "Écran thermique souple - CASTOLIN" sufficient or it has to be used more pieces?

Hi,
I kept the original insulation " as is". I just extracted the lateral parts then inserted the new continous insulation (the one I had in my garrage was 1 cm thick and had to put two layers of it) and the reinserted the old fragmented insulation that was on the owen.
I wil try to retreive he photos and put then on the next post. The Top insulation is uniform, continuos and should not be touched I think. The heat loss on the sides are probably due to the uneven insulation.

@George-2005
Copy link

Hi,
I kept the original insulation " as is". I just extracted the lateral parts then inserted the new continous insulation (the one I had in my garrage was 1 cm thick and had to put two layers of it) and the reinserted the old fragmented insulation that was on the owen.
I wil try to retreive he photos and put then on the next post. The Top insulation is uniform, continuos and should not be touched I think. The heat loss on the sides are probably due to the uneven insulation.

Hi, thank you very much for your quick reply and all valuable information! If you could be able to post the photos (or even e.g. link to video etc.) I would appreciate it very much, actually! Thanks again for your help.

@George-2005
Copy link

Just another issue regarding the oven insulation - some people say that the insulation material inside the oven (including inside the drawer's door) could be carcinogenic asbestos (e.g. please see #215 below). Is that really true? If yes, what to do with it to avoid the cancer risk?

@affetus
Copy link

affetus commented Feb 6, 2022

During cool-down the exhaust air gets cooler very rapidly so I don't think any metal fan will have problems. But in any case the air is definitely warmer than what it would be when sucked in thru the rear.

What I tried to say is, my oven was already equipped with an all-metal Papst fan from factory which is quite surprising given the total price I paid.

@xnk
Copy link
Member

xnk commented Feb 8, 2022 via email

@zian31
Copy link

zian31 commented Feb 9, 2022 •
edited
Loading

Pulling in even small amounts of cool air directly below the PCB seems to be messing things up. The original firmware really isn’t impacted as the fan isn’t running during reflow.

Thanks xnk for these important informations.
With your (excellent) firmware, the fan is always working during reflow (can you confim that it's only to mix air inside ?) : what is the really difference with the original firmware concerning the final soldering results (which as you said is not using the fan during reflow), is it really better ?
Maybe in your firmare it's interresting not to use the fan during the period which temp is over Liquidus solder paste ? (or for example already 30°C before this Liquidus temp) ? In order to avoid cold air to climb during this critical "Liquidus" period ?

@xnk
Copy link
Member

xnk commented Feb 9, 2022 via email

@zian31
Copy link

zian31 commented Feb 9, 2022 •
edited
Loading

Sorry xnk, I've edited my question before your answer.
Your firmware has two other nice differences : cold junction correction & 1-wire reading (with MAX31850KATB+ in my case).

So I re-ask you my new question : "With your (excellent) firmware, the fan is always working during reflow (can you confim that it's only to mix air inside ?) : what is the really difference with the original firmware concerning the final soldering results (which as you said is not using the fan during reflow), is it really better ?
Maybe in your firmare it's interresting not to use the fan during the period which temp is over Liquidus solder paste ? (or for example already 30°C before this Liquidus temp) ? In order to avoid cold air to climb during this critical "Liquidus" period ?"

Or maybe another solution could be :

  • Keep T962+ with air flow from back to rear
  • Keep the slats
  • Use of your firmware modified without rear fan working during reflow
  • Add a slow Convection Fan inside (working during all reflow) close to lamp IR as Jerry Walker : https://www.youtube.com/watch?v=9026Odx_jOw&t=960s

@affetus
Copy link

affetus commented Feb 9, 2022 •
edited
Loading

Or maybe another solution could be :

* Keep T962+ with air flow from back to rear

* Keep the slats

* Use of your firmware modified without rear fan working during reflow

* Add a slow Convection Fan inside (working during all reflow) close to lamp IR as Jerry Walker : https://www.youtube.com/watch?v=9026Odx_jOw&t=960s

This is exactly what I am planning to do for my T-962, as soon as I get the parts I need. Plus, assemble new thermal insulation and seal the airflow so that air is not sucked from outside the chamber during cooldown.

@xnk
Copy link
Member

xnk commented Feb 9, 2022 via email

@zian31
Copy link

zian31 commented Feb 10, 2022

Thanks xnk ! (so I'll modify your firmware, hope you are ok with that ;) )

This is exactly what I am planning to do for my T-962, as soon as I get the parts I need. Plus, assemble new thermal insulation and seal the airflow so that air is not sucked from outside the chamber during cooldown.

I don't understand what you mean with "seal the airflow so that air is not sucked from outside the chamber during cooldown" : the air will be sucked from where in your new design ?
One of my interrogation will be that during cooldown, air will come from down to up (and rear Fan is at 100% during cooldown), with old T-962 it was from up to down directly on PCB, here PCB could be a kind of "barrier" in air flow.
I forget to write another thing in my new design :

  • 2 Thermocouples used directly put on 2 angles of the PCB (orginal 2 Thermocouples close to IR lamp will be disconnected, because these one don't represent usable temperature)

@affetus
Copy link

affetus commented Feb 10, 2022

I don't understand what you mean with "seal the airflow so that air is not sucked from outside the chamber during cooldown" : the air will be sucked from where in your new design ?

Maybe there are variations in the manufacturing process, but in my unit the rear fan was not completely sealed to the plate which holds the slats, so there was air leakage from the inside of the main chassis, i.e. the space which also contains the electronics.

@xnk
Copy link
Member

xnk commented Feb 10, 2022 via email

@xnk
Copy link
Member

xnk commented Feb 10, 2022 via email

@zian31
Copy link

zian31 commented Feb 10, 2022

Now I understand what you mean : you want that during cooldown, air comes only from back of the PCB, and not from other place in the T-962 (moreover because slats are limiting the good air way) ? Another solution could be one or two servo-motor to open the slats just during cooldown, with kind of "wires" to pull slats ? I prefer to try that :)

Have you seen this slow convection Fan inside (working during all reflow) close to lamp IR (Jerry Walker) : https://www.youtube.com/watch?v=9026Odx_jOw&t=960s
Or maybe will you use another AC motor ? I think Jerry Walkeris true when saying slow rotation is important, if it's too fast it's not working well...

@affetus
Copy link

affetus commented Feb 10, 2022 •
edited
Loading

Yes I have seen the video from his T-962A and I will try to apply that exact method for T-962.

The heating power of the smaller model is a bit underrated (about 670W), so I am trying to prevent all cool air from entering during the reflow and the air mixer should help in that. The goal is to not pull air from under the PCB but just circulate the hot air around like in a household convection oven.

The slats will stay closed nicely unless the rear fan runs with a higher speed, so I don’t think any extra control is needed for them. This also means that using the rear fan for circulation does not work, as it requires a higher flow to open the slats (especially if there are air leaks which I mentioned earlier).

Problem is that the rear fan is so thick that it limits the slats from opening completely. It is possible to increase the slat opening angle a bit (about 1.5mm) by removing four fiber washers which are between the fan and the chassis rear plate, but it requires some machining of either the chassis or the fan frame because those washers were used to add a small gap to fit the fan better. I need to investigate this more, maybe there is a way to modify the slats a bit in a way that they open up more, still sealing the airway when in a closed position.

@zian31
Copy link

zian31 commented Feb 10, 2022

Ok, I believed that slats were no fully openned during 100% PWM power with rear fan, so no need of extra servo-motors.
I've now read that slats are 400mils openned, it should be enough compared to ES technical special new piece (with other air way direction) : https://www.estechnical.co.uk/images/pdfs/Addendum_-_T962A_A_Fan_and_air_path_modification.pdf

So as you, I will put some aluminium to seal the air flow.

And the main modification is this new inside little fan motor : do you already find one ? It should also be thermally isolated. I think I will make it work all the time (so not with a firmware command), but I hesitate to find one with DC 12V or AC 230V.
I also use a T-962+ I've added a "pnoxi" board.

@affetus
Copy link

affetus commented Feb 10, 2022

I will use Mabuchi RF-500TB-12560 motor mounted on 25mm long standoffs so there is room for the insulation wool + aluminum tape to fit between the motor and the reflow chamber upper plate. Then a 2mm to 3mm shaft adapter which goes through the wool and rotates the air circulator blade. This should keep the motor cool enough, I will not plan to use the oven for production so the duty cycle will be very very low. Just need to make sure there is a bit of clearance for the shaft so that it does not catch the wool fibers when it rotates.

@zian31
Copy link

zian31 commented Feb 10, 2022

Perfect ! I'd be interested with your 2mm to 3mm shaft adapter reference and also your helix or blade reference :)
I think I will also use a 25mm metallic tube arround the shaft to protect from wool shaft catching.

@affetus
Copy link

affetus commented Feb 10, 2022 •
edited
Loading

This is what I will use as the adapter, motor shaft is 2mm:

https://www.aliexpress.com/item/1005002911230673.html

On the blade side, I will probably just use a regular M3 screw (so I ordered 2mm to 3mm adapters) and the blade will mount with help of star washers and a nut to lock it in place.

Blade design will need some more thinking but I will use Jerry’s design (from the Youtube video) as the starting point, as he had good results with it in T-962A.

@GitLang
Copy link

GitLang commented Feb 10, 2022

Werner, how are you going to control the temperature if you have no forced cooling element?. Without it, you have fast heat in from the lamps and very slow heat out from thermal insulation losses. I've tried to do this in the past, and it just cannot work unless you are on an extremely slow control bandwidth. For the job we did controlling a standards room with +/- 0.2F. the cooling apparatus was as much power as the heating. If you put energy in you have got to take it out, and unless you want to get into very non-linear control and stability problems, keep the cooling!. In fact, some schemes will heat the target with more power than required then control temperature with heat extraction..

@xnk
Copy link
Member

xnk commented Feb 10, 2022 via email

@zian31
Copy link

zian31 commented Feb 11, 2022

the direction of airflow is important, and I can only agree that I had huge issues with the T-962A+ which also had the new “improved” airflow. It looked like the solder oxidised really bad despite reachine the correct temperature. Reversing the airflow using a similar modification as estechnical describes resolved that issue for me. Pulling in even small amounts of cool air directly below the PCB seems to be messing things up. The original firmware really isn’t impacted as the fan isn’t running during reflow.

Hoping that the problem of "solder oxidised really bad" is only due to "small amounts of cool air directly below the PCB" during first phase before hitting the peak temperature, and not due to something else : another reason can appears during the starting of cooldown => with old T962 cool air is arriving directly on PCB top, but with new T962+ cool air is arriving on PCB bottom (not directly on PCB top) and some bad air circles turbulences could be created ?

@zian31
Copy link

zian31 commented Feb 11, 2022

Blade design will need some more thinking but I will use Jerry’s design (from the Youtube video) as the starting point, as he had good results with it in T-962A.
What do you think about these 4 leaves blades : https://aliexpress.com/item/4000977277776.html

@affetus
Copy link

affetus commented Feb 11, 2022

Blade design will need some more thinking but I will use Jerry’s design (from the Youtube video) as the starting point, as he had good results with it in T-962A.
What do you think about these 4 leaves blades : https://aliexpress.com/item/4000977277776.html

Might be quick and easy, but looks like they may create too much draft (i.e. pull air from the bottom or push it towards bottom. It may be better to use a blade which doesn't act like a fan, but rather as an air mixer. The air circulation doesn't have to be strong, it is enough to just mix the air inside the chamber.

@jakorten
Copy link

jakorten commented Aug 30, 2023 •
edited
Loading

Wow, great work @DimaSM338 do you also happen to have a part number for the heat sink? I could not find it. Or [bgrigoriu] do you have a suggestion? I still could dive into my parts 'heap' and might find one that fits, but ordering one would be easier/more reliable.

@DimaSM338
Copy link

Wow, great work @DimaSM338 do you also happen to have a part number for the heat sink?

Wow great job @DimaSM338 do you also have the part number for the radiator?

Unfortunately, I do not remember the model number of the radiator. I wrote its dimensions above -
Radiator size 37x30x15 mm and provided a profile picture. Choose a similar one.

Unfortunately, due to the war in my country, I do not follow the topic. Can you tell me if you managed to connect the oven to a Windows PC?

@jakorten
Copy link

jakorten commented Sep 3, 2023

Thanks Dimas, It should work without any problems. Windows sees the USB FTDI converter at least (but I use a Mac, Parallels sees the FTDI converter and asks me if I want it in Windows or macOS).

I'll try, I already got the new Triac from Farnell.

Really hope the war will end sooner than later!

@DimaSM338
Copy link

DimaSM338 commented Sep 3, 2023 via email

@PerennialNovice
Copy link

PerennialNovice commented Nov 27, 2023 •
edited
Loading

Concerning the windows connection, someone recently posted a link of another heavily modified T962
https://en.gradient-sg.com/t962/
They also somehow control the oven via PC (but I think there is a bluetooth module used to do so...

maybe it helps?

@zian31
Copy link

zian31 commented Nov 27, 2023 •
edited
Loading

Concerning the windows connection, someone recently posted a link of another heavily modified T962 https://en.gradient-sg.com/t962/ They also somehow control the oven via PC (but I think there is a bluetooth module used to do so...

maybe it helps?

Very interesting ! What a job !

After reading all this page, i'm wondering if they use the older or newer T-962, or the modified newer T-962 (fan direction inverted with unmount and reverted remount).

But one thing is sure : they don't use the PWM fan during heating to mix the heat inside, the curves show PWM fan = 0% during heating, so they find THE solution to avoid mixing the heat inside with little PWM fan during heating (or using a little fan blade to mix heat) ?

EDIT : I've done this little drawn :
T962 vs T962+ Airflow

@zian31 zian31 mentioned this issue Dec 8, 2023
Open
@darkroomai
Copy link

The unit I received this month (Feb 2024) definitely blows the air out the back. From the EST modification PDF:

"Our testing has shown the direction of the air flow within the oven has a large effect on the overall performance
of the reflow oven."

"14. Find the marking on the fan which shows the direction of airflow and fan rotation. Refit the fan to the
rear panel of the oven - ensuring that the airflow arrow points away from the rear panel (airflow directed
into the oven chamber)."

This modification they recommend is only for the new version with the airflow OUT the back, but not for the earlier version with the airflow coming IN the back.

The airflow mod seems pretty invasive, so I'll hold off on that for now.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
10 participants
@xnk @GitLang @jakorten @zian31 @DimaSM338 @bgrigoriu @PerennialNovice @George-2005 @affetus @darkroomai