plant_conductance_primary_perception

This serves notice that this project has now been made private for processing of copyright and patent applications

26 Aug 2022

Today I worked with my brother on the family security system, since he would otherwise have worked all by himself on it.

25 Aug 2022

I built a re-configurable rig so that I can quickly switch back to the Arduino and the RPi Pico W. The Arduino allows operation at the previous level of achievement. The Pico does not show operation at an attempted next level, but at least I can now troubleshoot. Time to call it a day for a few hours first.

23 Aug 2022

Powering the project up with the RPi Pico W resulted in a non-functional response. So I went back to Arduino merely to know that the hardware is OK. It is. I'll make the RPI Pico W move in two steps instead of one this time - the first step is to get the circuit working at 3.3 volts in Arduino first, THEN wire it back to the RPi Pico W.

...and...the circuit does NOT work at 3.3v. Time to have a very close look at relevant design parameters now.

And now the circuit works at 3.3v. Now to hook it up to the RPi Pico W

19 Aug 2022

I've come back to this project after needing to work on the family suveillance project. My first step back was to realize I must immediately port my code over to the Raspberry Pi Pico W instead of "just one more small tweak" for the Arduino to run it. The memory size of the Arduino just wouldn't allow it. The port is a bit more than trivial < sigh >.

11 Aug 2022

Early this morning...and just about 5 hours after I applied improvements to the family surveillance system...a burglary was attempted at the building. The burglary AND untold damage was prevented due to my diligence with that side project, in spite of it taking much time away from this project.

05 Aug 2022

I've been working on the family surveillance system instead of this project to try to stay ahead of [burglars and vandals](https://drive.google.com/drive/folders/1b0pf9T0j0jGG7v6EXj3Siou54XZk-6vi?usp=sharing).

03 Aug 2022

A few days of research brings me to firm up better a plan to take this project up a notch. I've fired up the new microcontroller platform that will broadcast the signal of interest in a standard HTML5 stream over wifi. I've loaded an HTML 1.2 server (per https://www.raspberrypi.com/news/how-to-run-a-webserver-on-raspberry-pi-pico-w/) in hopes of it being a suitable starting point. Next steps are to incorporate FreeRTOS and adding HTML5 streaming. I'll bee-bop between this sub-project and the original sub-project of improving the sampling algorithm in the old Arduino platform for later porting to the Pi Pico W. Because the Pi Pico W is a fast device, I'm hoping I'll have the option of using either MicroPython or c++ per the librariies available.

30 Jul 2022

Meeting Paul has ramped up the urgency of getting this project done even higher than bringing Eric and Pieter on board last year. I needed that, I'm embarassed to admit.

20 Jul 2022

I've decided to move away from Arduino microcontroller as soon as I can after proof of concept phase in favor of the production model using the Raspberry Pi Pico W microcontroller. The advances in microcontroller technology make the Arduino terribly obsolete and relatively limiting. The new microcontroller will mean the conductance signal of interest can be displayed on any nearby (and possibly internet-connected) device with an HTML5 browser and wifi instead of being limited to display on a laptop only.

Do you have an idea for features? Now is a great time for all interested parties to ask me about including capabilities you think would make this project as valuable as it can be. Just submit an "Issue" (see the tab above) describing the device feature that would interest you. Thank you! BEING PLANNED: Migrating/porting the code to Raspberry Pi Pico W away from Arduino allow building the appropriate adapter. This will have the signal of interest available wirelessly via HTML5 stream to your cell phone browser.

10 Jul 2022

I've produced a circuit version that condenses three stages at 5VDC into two (at the expense of precision) to operate at 3.3 VDC. Initial readings indicate the precision is still adequate, but if not, I'll add a voltage boost to a future design and get the precision back.

20 Nov 2021

The device has an annoying intermittent fault in it, but otherwise the circuit appears to be working as designed. I met with Pieter and used his plant to try to characterize the system. From that I realized I would need to modify the system a bit in various ways, and the lateness of the hour dictated that we continue some other day after my modifications are done and the intermittent fault is fixed. The mods are now done as of yesterday, and I am developing a troubleshooting technique in order to locate the fault.

09 Nov 2021

As I my efforts return to coding from circuitry, I am a bit bogged down from having to re-study the code for a couple days. I think I have re-acquainted myself enough to renew code advancements later today. My first and second steps are to code the hardware potentiometers into momentary switched to toggle up and down; then scale their displayed values up to enable the entire dPots' settings to fit within the plotter linespace. After that I'll code something to auto-adjust the dPots. I hope I can finish all that by year-end when President Trump should be back into his rightful office.

03 Nov 2021

BREAKTHROUGH! THANK YOU, LORD! I developed a troubleshooting technique to point me to the very pins giving me all my troubles with U8-U33. Once I addressed all those pins, I now have a electronically working prototype, except the software (firmware, if you will, since we're talking Arduino). Now, I will switch back to software development to produce a fully working prototype.

26 Sep 2021

In order to ensure the instability is not caused by feedback through the 3rd Stage Guard Ring I will change the source of the 3rd Stage Guard Ring from U2 pin 1 over to U2 pin 7.

24 Sep 2021

Was delayed about 3 days by home maintenance. Coming back to troubleshoot the instability.

20 Sep 2021

Got through the soldering challenge. I am now tracking down the source of an unwanted oscillation.

18 Sep 2021

Just now discovered what I am up against - the microvias inside the MSOP-10 pads have solder mask on them. I must keep solder mask from being applied to those specific pads, but how? For now, I'll try to deal with it by removing the solder mask or something.

16 Sep 2021

If I had to guess right now, I would say the LM334 U6 goes into current starvation when A0 is allowed to swing above 3.5 VDC. IOW, the operating current minimums are getting violated. At that point its impedance understandably lowers suddenly. I really need to find a sink for 1 μA or throw in another LM334 set to 1 μA to ground...or go with the original thought of making A0 into a virtual ground (306/1024) and displaying the DPOT values it takes to keep it at 306/1024. That's what I need to do.

15 Sep 2021

U5 was bad. Replaced. Circuit now working, but seems to have latchup tendency that I had hoped the new op-amp would fix. I'll have to think about this. For the record, I do have Q2 shorted across D-S which allows a much wider voltage swing of A0. Whether that is related to the latchup behavior is unknown right now. Time to think things through.... A positive note is that the new op-amp certainly does eliminate the crossover distortion like I intended it would do.

14 Sep 2021

Hmmmm I'm trying to characterize the solvent and counter-solvent control circuits but not seeing that the solvent control circuitry has any effect. After I understand the boundaries of both, I will step back and entirely re-think the code rather than try to modify it.

12 Sep 2021

Took time off of this project to make a road trip with wife. After coming back to this project, I am getting thwarted by the circuitry of the LM334 programming currents. Have to figure out troubleshooting method.

8 Sep 2021

I enlarged a number of more pads to mitigate delamination of them during soldering, but I can still use the boards I have right now. I also better aligned 7_A7,A6 for the SainSmart UNO.

I've been struggling with the auxiliary board with all the intricate soldering needed for that. I think I might have achieved a soldering technique for the aux board seperately that will do well going forward - squeegee on the lightest solder cream layer I have and work on several soldering passes without placing any of the MSOP-10s until the pads appear perfectly tinned with a little less tinning thickness than a fab house would leave. Then, starting with U20, place and solder one MSOP-10 after dragging the lead ends through a very light layer of solder cream. Then inspect the soldering result and run diags after each MSOP is placed and soldered. Proceed with each MSOP-10 until done. That way if you make a mistake at any MSOP-10, you'll know where you made it.

The main board solders just fine as is.

26 Aug 2021

I enlarged a copper pad to better secure it from delamination during soldering. The new design is uploaded to OshPark for public dissemination. This means that until I update links here to match the new project there, things are not synchronized, so go direct to OshPark and search their site to retrieve my latest design.

21 Aug 2021

My order was shipped and arrived at the town post office yesterday. It remains there now, but the route to me is mostly straight interstate highway. I have hopes to see the package by Tuesday, 24 Aug.

20 Aug 2021

I should have realized that there would be a difference between pre-tinned boards and ENIG boards in how much solder cream would be appropriate for the smallest components. Thankfully, I do have appropriate solder stencils for both, though that was unintentional. What happened was I'm having to learn from trial and error just how exacting I need to be placing and soldering the DPOTS on the auxiliary board. On my last try however, I applied too much solder cream to a pre-tinned board and got way too many solder bridges to continue with that solder job. So I have to remove those DPOTS, which subjects them to yet another thermal cycle. I've already subjected them to so many that I'm suspicious whether they're surviving. So I placed an order for new ones enough for two more sets. Wait time is probably a week plus a day or so. Hurry up and wait, as the good saying goes.

18 Aug 2021

After placing most SMD components on the main board, I discovered a manufacturing tolerance violation that had to be corrected with a knife on each board. I'll fix the three boards I am building with a knife, but I uploaded new designs to OshPark for future PCB orders: Combo board having the main and outboards attached is https://oshpark.com/shared_projects/8y2wPOtG. Main board only is https://oshpark.com/shared_projects/18dBWrLE. Outboard only is https://oshpark.com/shared_projects/zuaqf85h Please note as of this moment, these links supercede any gerbers for this project that you'll find elsewhere on this site. I'm not in a hurry to track down and fix the others.

25 Jun 2021

I restarted working lightly on this project yesterday with my grandson. I'll put greater intensiity into it imminently.

18 Jun 2021

Apparently my idea of using sparing amount of paste solder on the MSOP-8 packages is not a good idea. I tried minimizing the amount of paste in order to reduce solder bridging. I'll need to abandon that idea and order a new stencil having normal apertures for the MSOP-8 packages.

16 Jul 2021

16 Jun 2021

The oxidized contamination combined with my crazy sleep schedule overwhelmed me and made me miss my sunrise goal...

15 Jun 2021

I am nearly finished soldering the SMD components to the most populated side of the board. I expect to start soldering components to the other side by sunrise tomorrow. I hope to be able to apply power before Saturday 19th at the very, very latest. NOTE: I expect I'll erase all these development stage updates when the board becomes operational. The updates will then become informational on discoveries from me and others using the training aid.

A challenge I face is to recognize solder paste oxidizes in short order, the small amounts and just need to be discarded rather than salvaged. That is against my life's learning about resource management. What is happening is that I set a small amount of paste solder out to dip from, and as it oxidizes it becomes a source of contamination to the joints I try to add it to.

03 Jun 2021

I've had to work on the [Surveillance project](http://github.com/kenneth558/KingdomComeIMeanItAndIMeanNowSurveillance) instead of this one, but I'm nearing smoother sailing with than one to where I'll be able to come back to this project. In fact,I'm actually at a point where the final work on the Surveillance project can only be done on site, which means that while I am home I can work on this project. Thank you Lord!!!

03 Apr 2021

Basement water problem turned my work area into chaos. I can't find a clean PCB to build, so I placed another order yesterday for 3 more boards. It will be a number of days before they arrive.

06 Feb 2021

My other project is still distracting me from this one......

20 Jan 2021

Trying to squeegee the paste solder on to the PCB, I find that the stencil retains the paste inside many of the apertures. I'll see if I can adjust my technique to overcome this latest challenge.

19 Jan 2021

After dropping an SMD or two, I spread a white drape over my work bench and scrap linoleum under the work table to better allow the finding of such components I'll most surely drop again. Cost of time and momentum, but I'm returning...

17 Jan 2021

I received my new solder stencil. I'll get back to work on this project today after I get a night's rest and attend church later this morning.

12 Jan 2021

I just discovered that the solder stencil I've had since April had no alignment holes, making it unreasonably difficult to place over the PCB accurately. I ordered a new one, and it is on its way to me.

03 Jan 2021

I'm trying to develop a paste solder application scheme. I have no known finish date, yet.

27 Dec 2020

I'm just not satisfied with any of the USB microscopes I find for sale. So instead, I will put together various lenses I have. The advantage of that is high resolution and accurate color rendering, so I hope it will all be for the better.

24 Dec 2020

Sigh... In the re-arranging, my USB microscope broke. I really need that thing! Unable to disassemble it (it's probably epoxied). The good side is that the next one I order should be better quality, if I can make that happen. Shopping for it next...

23 Dec 2020

I plan to return to this project today. FINALLY!

19 Dec 2020

No updates again, but the security system development is done enough that I'll start now to arrange this project for resumption...after a good night's sleep, maybe two. During this interim period, I've had to physically re-arrange everything related to this project which makes it not merely a simple matter of picking up where I left off.

19 Nov 2020

No update again. Still distracted from this project by having to develop a security system for my father's property.

08 Sep 2020

No update again. Still distracted from this project by having to develop a security system for my father's property.

23 Aug 2020

I'm now being pulled away from this important project by other urgencies, like thieves and vandals to my father's property which make me develop a security system. His property is where my web server-to-be is located.

28 Jul 2020

After some weeks of distraction, I am slowly returning to this project. I think my next step is to place and solder componponents.

14 Jul 2020

While awaiting components, I am configuring a server for the GWAAMC web site. It will include a tight firewall ruleset from my poofitee project, knowing how there are so many malicious and mischievous malevolents on the internet.

11 Jul 2020

My component order ended up being two orders. One is scheduled to arrive next Thursday, 16 Jul 2020. The other order, consisting of Q1 and Q2, is pending. These two P-channel J-FETs one the second order are to demanding specs in noise level and V_gs cutoff voltage.

10 Jul 2020

This latest jig looks like it will work. The quality of print is slightly under past prints I've received, but at least the design that I sent to the printer seems to be correct this time. Loaded as many components as I had. Discovered I need to order more components. Did it. Waiting for those components to arrive.

09 Jul 2020

Given clues by the postal service, I found that they delivered to the wrong address. Jig retrieved, looks like I got it right this time, assembling a board. Files for printing a jig of your own here -> https://drive.google.com/drive/folders/1_j0E5dFC1kLGvFu55QTtv5UvltjqBdjs?usp=sharing.

08 Jul 2020

The postal service says they delivered the latest jig to me, but there was no jig in the mailbox or even in the secondary delivery location. I'll meet the postal delivery person tomorrow and have a talk with them about this.

02 Jul 2020

Final jig is ordered. No word if it will go into printing before the July 4 weekend.

01 Jul 2020

The new jig has some too tight capacitor cavities. Another one has to be made. I also realized there are a few components I had yet to order, so I placed component orders.

30 Jun 2020

While populating the jig, I found a couple issues that really need to be addressed still before it will work. A new jig will arrive tomorrow which might work.

27 Jun 2020

Done designing new jig and ordered it. Now building a board with this jig that I took a utiiity knife to and made it workable.

26 Jun 2020

Jig received. Building a board while tweaking the jig design in small ways.

24 Jun 2020

I will be receiving the jig by Saturday June 27.

19 Jun 2020

Latest form design ordered. As a clarification, this is the bottom side form I've been working on. The top side was very easy and not even necessary but accomlished early on. Later this day, I received notice that my order has already gone into production. Based on previous dealings, I expect to receive the jig in 6 or 7 days.

16 Jun 2020

Received my latest form/holder, and the VSSOP-8 wells are still too small. Going for one last design now.

11 Jun 2020

My son offered the term of "form" instead of "cassette" for these components holders. I didn't care for calling them "cassettes". I'll change their term to something like "holder" or "form". They can also be called a "jig", but only as a last resort.

09 Jun 2020

That cassette was better, but still not right. Made corrections to design and ordered another. A return notice is telling me the new cassette is already in production.

05 Jun 2020

I received notice that my latest cassettes will arrive Monday, 08 Jun.

03 Jun 2020

Again, The components wells were too small. Ordered ones with larger wells and using three different materials. I was promptly notified that my order went into production, for which I am pleased, of course.

01 Jun 2020

The projected arrival date of my newest cassette has been changed to be today in about 4 hrs.

28 May 2020

I had to modify the cavity sizes a bit larger in the bottom cassette and order a new one. It is scheduled to arrive Tuesday June 2.

17 May 2020

The cassettes are still in "Production started" status according to craftcloud3d web site.

13 May 2020

I placed a new order for top and bottom resin cassette. No word as of 14 May 9am whether the order will get honored, but my last resin order did get made for me. Product expected in 9 days. Once the cassettes arrive I can start constructing boards. I will also need to locate carbon fiber "tape". Most of what I see for sale uses the term "graphite" in the sales lit. Graphite and diamond are the WRONG forms of carbon for me to use. I wish everyone would stay away from using the term "graphite" unless the molecular structure is actually graphite. Bad terminology makes me have to dig deeper and even travel hundreds of miles to measure electrical resistance.

12 May 2020

Both my cassette orders got canceled due to too fine of features for ABS. I've arranged for someone to see if they can print the bottom. No word back from them, yet....

09 May 2020

Placed the order for the bottom cassette.

08 May 2020

Just placed an order for a new top aspect cassette. Still working on the bottom to add one last item - some tin whiskers cages for the larger IC packages. The design for MSOP package cages are done. These cages will span across the IC packages to contain any tin whiskers that might form to only the pin they originate from.

07 May 2020

Now making final changes to the bottom aspect cassette.

05 May 2020

My cassette top will arrive tomorrow. The design for the bottom is complete, sans any changes I will learn it needs from examining the top when I get it. If all is well, I will order a bottom aspect shortly after I approve the top.

03 May 2020

Still working on the cassette bottom. Since it is substantially more detailed than the top side, it is taking longer.

01 May 2020

I ordered a revised top aspect cassette (top "side" sounds strange) yesterday, and it is already being printed. I'm close to being finished designing the bottom aspect. It really could be done in an hour by someone with more stamina than I have. I'll see what I can do...

29 Apr 2020

I received the top side cassette from the 3D printing service. The top is the simplest side, so I finished designing it first and will be able to apply whatever I learn to the design of the bottom side cassette. I've learned the bare PCB needs greater clearance to slip on and off without forcing. Applying said knowledge to the bottom side design now.

23 Apr 2020

I've received my paste stencil, but without a cassette printed, I am hindered to lay out the components rapidly enough to before the paste dries. I am working on designing and 3D printing cassettes to place an entire side of components simultaneously.

12 Apr 2020

A very blessed Resurrection Sunday to you. Are you aware we are entering into a season many years long in which many resurrections will take place? Even some people that have had their bodies cremated will be resurrected! As I observe our society, I can't help but realize we need some help getting to that point. People need to learn to put their faith to work, and to work hard. That is EXACTLY what this project is all about! In Jesus' name

11 Apr 2020

I ordered my first solder cream stencil from OSH Stencils. I am so thankful they priced themselves below China, considering shipping. Delivery time is also as fast or faster than any Chinese fabricator can do.

10 Apr 2020

I changed the solder cream layers for the MSOP-10 packages to minimize the component pins mashing the cream into solder bridging between pins. Added notes into gerbers as well b/c the fab house was confused about guard rings and copper zones joining traces. Next, I am considering a unique technique to place the components down into previously laid solder cream. It will consist of a stepped (multilayered) transparent stencil into which the components will be placed upside down. The freshly creamed board will be laid down on top of the undersides of the components. The assembly will be flipped over and the top stencil carefully removed. Viola, one side ready to bake.

07 Apr 2020

I'm hoping this latest technique works to place MSOP-10s on double-tinned pads works. After double-tinning while pressing down on the solder (to better displace the gold), I use the hot air tip to smooth the pad flat. One last solder layer, this time we convert to bismuth, then scoop the pad flat again. That doesn't leave much solder there, but at least the MSOP-10 package stays put during hot air soldering, plus we've converted to low temp solder _on ENIG_ :-). Note that you should not allow leaded solder to be sitting anywhere around lest you mistakenly use it and thus corrupt the bismuth.

The sheer tediousness of the job compels me now to order stencils. Soliciting quotes next...

05 Apr 2020

MSOP10s are absolute bears! Still not done.

31 Mar 2020

Double-tinning without ending in shorts is not so straightforward. And after that enough gold too often remains so the wetting is impaired. Not impossible, though. Continuing the work.

29 Mar 2020

I am double-tinning the first board, after which I will flatten the solder, then place components. All of this goes slow so I can update the instructions as I go along while letting wife sleep in the same small room.

28 Mar 2020

I received my ckt boards from OSH Park and inspected them. So now I'll change the gerbers to make smaller solder mask holes over the vias so they will be less likely to short over to nearby copper during soldering. Until that new design is published, you'll just have to give extra attention when soldering those vias. Some of you may wonder why you'd solder a via? It's to hermetically seal inside the shields because the REF200 requires it for long-term stability if used for measurements in a laboratory environment. I'm just trying to permit that use. Note that I avoid using the term "measure" in normal intended use of this device - its normal use is for "observation", not measurement. Yet I'll try to design not to preclude measurement use.

BTW, when soldering vias, bismuth solder is preferred b/c it uniquely expands on solidifying. Make sure you don't mix it with leaded solder, though. Research it if you don't know why.

25 Mar 2020

I received notice today that my lead-free HSAL boards from PCBCart will ship April 1. My ENIG boards have been fabricated through OSH Park and are being shipped today my direction. I am putting my time to use researching methods of assembly. I'm planning to use an acrylic "footprint position" stencil made KiCAD's .pos file if I can determine how to render it into a stencil.

19 Mar 2020

While awaiting the arrival of boards with the two finishes, I'll state that I have high hopes in my board designs. What I'm not sure about is which op-amp to specify for U2. Right now it is the OPA2388. Regardless, the footprint on the board is VSSOP-8_3.0x3.0mm_P0.65mm, which is a TI specific footprint but I chose it hoping it will also work for similar sized SO-8 dual op amps.

14 Mar 2020

I am going to try to get boards in lead-free HASL instead of fighting with this ENIG that really needs a soldering fountain to wash off the gold. I have no soldering fountain, so I'm trying to do it with a soldering pencil, and it is too frustrating.

11 Mar 2020

In addition to keeping the two-board design, I had to combine boards into a single board design in order to make lead-free HASL quotes more convenient and reliable. Those new files are available for you but only if you find them in this directory https://github.com/kenneth558/plant_resistance_primary_perception/tree/Free/Where%20to%20find%20everything%20for%20making%20your%20own%20GWAAMC%20device/Files%20you%20need%20relative%20to%20the%20printed%20circuit%20board/v1.0/combined%20main%20plus%20aux%20on%20single%20board.

04 Mar 2020

Received 1 oz boards and am losing pads during the double-tinning ENIG wash. The 2 oz boards were more robust against this problem, but I am not given the ordering option of expediting 2 oz boards. I do not advise 1 oz ENIG boards. I will investigate ordering lead-free HASL boards from China unless I find a decent price locally. HASL boards don't require the double-tinning wash. In the meantime, I will try to get these other ENIG boards built.

25 Feb 2020

I have completed yet another PCB design of both the main project board and the auxiliary board. The project files at OSH Park are: main board https://oshpark.com/shared_projects/y7WQJ0Jt, aux board https://oshpark.com/shared_projects/kbQQkkZ4, both boards attached and combined into a single file https://oshpark.com/shared_projects/0zQW3Jms. I am awaiting delivery next week, then I'll have to build them out.

19 Feb 2020

I've received the newest PCBs and found numerous shorted traces the result of my design having violated manufacturing limits. I will have to increase the size of annulars on the auxiliary board in a new design and re-order boards. Day 2 of doing this: about 15% into the new design. At that rate it will be the end of the month before the design is complete. That rate is unacceptably slow, so I'll try harder.

11 Feb 2020

Had to cancel the previous expansion board order for better hole alignment. Now completed and new orders of both boards are being processed. Expected delivery is by 18 Feb 2020

24 Jan 2020

Received notice that my shipment of 3 main boards will arrive Monday (today is Friday). I'll be able to begin placing those components then, but in the meantime I need to finish design of the flex circuit and get 3 of those on order.

18 Jan 2020

Just finished fitting all copper traces onto the main PCB. I did have to move the HX711 ADC off to the side onto an optionally cut-off tab due to lack of room. I am now widening selected traces (Vcc and GND for example) and laying down the non-copper layers.

I am also having to redesign the auxiliary board that holds 25 MCP4262-502E/UN dPots. This aux board now has to fit underneath a shield with different dimensions than before. The good part is that I've already designed the pattern for that aux board, so seeing the aux board through from this point will be trivial. UPDATE - I'm taking a rabbit trail to learn how to use a plug-in that replicates sub-circuit traces since I have 25+ subcircuits to replicate and KiCAD keeps producing rounding errors with its own replicating methods. What should be trivial will now involve an additional learning curve.

26 Dec 2019

I've been pulled away from this project by other assignments, but the larger issue that has me entangled is that I'm re-designing the main circuit board from ground zero to accommodate the fact that there was no drop-in replacement for the AD8244 that offered zero crossover distortion. I would say two more months before I can finish the design, then another 4-5 weeks for PCBs of that design to arrive, then a week to build one up. Fourteen weeks until the latest design is working. April-ish.

21 Nov 2019

My LMC662 came yesterday. Other op amp types that I can compare with will arrive today. The LMC662 has better noise and input bias current characteristics than the other ones I've ordered, but it doesn't work at 3.3 VDC. We'll see if the noise characteristics are acceptable. I figure by the time I'm done, the AD8692 will win out.

18 Nov 2019

To avoid the input bias current caused "crossover distortion", as it seems to be called, I'll have to re-design stage 3 without the rail-to-rail input (RRI) AD8244 and use a non-RRI op-amp instead. That takes the expected project completion date into December. Sorry. (Anyone know of a non-RRI quad unity gain buffer so I don't have to deal with more pins that go along with the op-amp solution?). Until I change my mind, I'll be changing the design to use an LMC6036 and remove the AD8244 altogether.

17 Nov 2019

Back in business. Q1 needed to get a higher cutoff version (which I had anticipated as a possibility) and with Q2 out and Q2 footprint shorted Drain to Source (might get reverted later), there are no oscillations any more. The downside is that there will be less possible gain without Q2. Since I was only guessing around with gain anyway, I'm not going to sweat it until that extra gain shows itself being needed. Onward sans Q2 to track down a bit of undesired hysteresis in the current-to-voltage stage (aka, 3rd stage) which is surely from AD8244 bias current "...rail-to-rail input stages have bias currents that change direction as the common-mode voltage passes through the transition region" (https://www.analog.com/media/en/training-seminars/tutorials/MT-038.pdf). Exactly descriptive of what I'm seeing. My solution for now will be to subtract the transition region from the output and see how it looks.

16 Nov 2019

The latest thing that has slowed me down is needing to clean the circuit board at a location where there was no isopropyl alcohol. Instead of alcohol, I found what I thought might work instead because its odor indicated an isopropyl main ingredient. It was for cleaning lenses or something, so I used it. That was a bad idea. It left a seriously detrimental conductive residue underneath components. I didn't understand for several days after why the circuit was acting strange after that. I'll have to pull components and clean under them until I get proper operation back.

14 Nov 2019

I'm seeing a low level of undesireable positive feedback so as to make the analog output at A0 a little unstable right around mid-point between rails. It might be the effect of AD8244 bias current, but whatever the source, I'm intent to identify it and compensate.

09 Nov 2019

Removing (shorting across pins 1 and 2) Q1 and Q2 seems to alleviate the oscillations. The flip side of doing that is twofold - less gain can be gotten in the third stage and the voltage on the guard ring trace for U5 is slightly imperfect. (Shorting across Q1 is probably overkill, and I may experiment with undoing that or using one of the other JFETs with higher pinch-off voltage for Q1 and even Q2.) Neither issue is important enough to get wrapped around the axle over. Full testing resumes now.

08 Nov 2019

A better look at the oscillations produces evidence that they are produced by the 3rd stage high impedance creating a type of relaxation oscillation in the presence of any solute/DUT current at all. Will work more on cleaning and inspecting the main board, especially since I've left a bit of solder flux on the board, and studying the design and datasheets.

04 Nov 2019

Due to oscillations with apparent origins in inadequate power supply decoupling, I have changed the values of C3 and C5 to 10 µF. Testing tomorrow....

01 Nov 2019

I have gotten the manual tuning aspect of the sketch to function. I envision a better method of tuning, but that path of programming eludes me for the time being. The type of tuning that I hope to one day program into the sketch would treat the LM334 pair in opposing adjustment as one function block (entitled "DC level" or "DUT impedance balance") while that same LM334 pair working in in tandem adjustment as a different function block (entitled "dilution ratio"). That scheme in contrast to the easy-to-program manual tuning I'm resigned to run with for now. The manual tuning is at level 4 of Diagnostics rather than being built into the main aspect of the sketch. That is what I'll use to prove concept during tours so I can just get going and make this project known by demonstration. Tested at 5 VDC only for right now.

Engaged the DUT and the graph won't keep the trace inside the boundaries for me. From the looks of it, the circuit gain is too high. At least, that is my first suspicion. If that is the case, I will have to modify the circuit. Current signal circuits are very unusual, and their troubleshooting methodologies are different than those of normal circuitry. You can't just put a scope or meter on a node to see if operation is within normal limits. I may have to develop circuit modifications purely for troubleshooting.

I need to take a break right now. I'll get back to things later today.

22 Oct 2019

Yesterday I found a large difference in conductivity between attaching to the DUT plant while located at a tech library with fluorescent lighting vs. at a residential house with only incandescent lighting. I'm suspicious that the difference had nothing to do with the lighting but may be attributable to either a circuit board intermittent or a second stage current balancing micro-amperage level. The amount of current operating in the second circuit stage may end up being a very important tuning consideration. Regardless of any aforementioned item, I have no reason to doubt anything about the circuit design except that more power supply decoupling may be in order. In the next few days I'll add some via air-suspension to see if there is any reason to enhance the circuit with it.

21 Oct 2019

Attached to a plant for the first time today - one electrode around the stalk and one enclosing a leaf. I see far less conductivity than I expected. That being the case, the interference from the fluorescent lighting is also annoyingly significant. I'll need to get the sketch working and work on controlling the board's electrical and temperature environment next.

17 Oct 2019

Updated [this flex board](https://oshpark.com/shared_projects/59YhWYRZ) to contain a U7 expansion section, since it doesn't cost any extra and it will help me use up a number of old valued U7's that I bought before I settled on using the 50K values that are in the final design. If you don't want this cut-away piece on your flex board, just order normally and avoid using the link in the previous sentence.

15 Oct 2019

Once I soldered J9 better and then kept the flex board clean every time I add a dPot on it, all works well. I CONSIDER MY CIRCUITRY DEVELOPMENT TO BE SUCCESSFULLY COMPLETED NOW. Finishing the sketch is the next on my list after some car repairs...

13 Oct 2019

I'm beginning to suspect the flex board problem is a combination of J9 not being solidly soldered onto the main board and conduction on the flex board from traces of contamination that need to be cleaned off better. I'll get back to work in a few hours.

12 Oct 2019

I have the main board working consistently, but the flex board is misbehaving much of the time. That is what I'll be working on next. I'm quite sure the problem I'll find is a soldering problem rather than a board design problem.

11 Oct 2019

I changed the flex board to accommodate a through-hole GNDD pad dedicated for shield grounding.

09 Oct 2019

I've now verified operation of Rev. 8 and re-versioned the perfected rendition to Version 1.0. You'll see lower noise in the output if you enclose the flex board in foil EMI protection, please. I used aluminum foil with appropriate insulating tape for my prototype. Note that I won't develop temperature compensation in the circuit any time soon, so I am avoiding the term "measure" in favor of "observe" when I refer to electrical conductance of the DUT.

07 Oct 2019

I just corrected a short between GNDD and 5VDC due to the boards not getting fully etched. I corrected the design at that location to prevent it from happening again. Power applied and I'm now running and tweaking the diags...

03 Oct 2019

The solder joint issues with the fine pich components seem to be never ending. I'm still not done with the visual inspection, but when that is finished I'll apply power...

29 Sep 2019

I just finished placing all components onto this board rendition. Before I apply power, I'll have to spend a few hours examining the entire thing. First, however, comes a night's sleep....

28 Sep 2019

Yesterday I finished building out one rigid main board and built out one flex board. I had much difficulty soldering down SMD components that I had the "wise" idea of laying out on the adhesive side of my Kapton flex prototype. I'll re-build the entire flex board today after my night's sleep and after I clean those SMD components with contaminated leads. Then I'll spend several hours examining the boards with loupe and microscope before applying power. Then I'll repair whatever build mistakes I made, after which I'll know whether my design (sans wireless and SD Card) is all I hoped it would be. Wireless and SD Card operation may or may not work, but checking their operation is low priority for my time at this stage.

GOT TO WORK FOR THE DAY AND MY SOLDER STATION BROKE. All soldering work is on hold while I await heaters for my solder station.

27 Sep 2019

I received 4 copies and am building out one of the flex board. I don't know why they sent 4 when they say 3, but I'm happy about it. I also had to order some components in the LDO circuit stage that I overlooked buying. I expect that order to arrive today.

23 Sep 2019

While building my latest design boards, I will fine-tune various files here to reflect refinements I deem necessary. Just updated iBOM and board designs at the KiCad file-type level, but not the gerbers yet. New board designs are in recognition that solder mask seems to be an adequate hermetic seal for inside SHLD1. Therefore, I removed the solder mask blocks I had placed on those micro-vias.

Also while building these boards now, I'm focusing on getting the boards to working condition the quickest I can so I'm not video recording. I'll video record for you once I get a few more boards under my belt.

22 Sep 2019

Although I didn't know it until today, I received my order of 3 main boards (not the flex boards) yesterday. I will start ASAP to build them out.

21 Sep 2019

Building my server complex has been curtailed by ISP blocking port 80 traffic into my modem. This even though they assured me back when that they would not do such a thing. Negotiating with other ISPs that can give me a data drop. So far no success with that. Anyone want to help me out here? Open an Issue to let me know. My specific need is acquire and install my SSL certificate for e-commerce, which requires my server to receive port 80 traffic. Thank you!!!

15 Sep 2019

I tweaked the J9 flex connector footprint in a PCB edit to strengthen it mechanically - I moved some nearby PTH vias into its GND attachments so that SMD pads weren't the only means of it staying attached. With this change, solidified solder through the depth of the board under the flex connector's GND attachments will assist the board's SMD pads holding that connector down. Not only that, but I made those vias large enough to place guy wire ends into so that connector could actually also be held down with a guy wire around the top side of it for even greater strength. The OSH Park project is updated with this enhancement.

13 Sep 2019

While I await the [hopefully] final boards, I'm building my server complex to host gwaamc.com and all my other domain names with pages, videos, and forums. This will be a child-friendly and adult-friendly venue. Any hostile, uncivil, and profane language posts will be removed and responded to with a permanent and uncompassionate ban from the site. An iptables ruleset that I've developed myself will be modified as needed to ensure the utmost security for my server.

09 Sep 2019

The final correction (of a backwards connector) has been incorporated into the latest PCB design. Additionally, a flex board expansion board design is finished. I would venture to bet this new main board and the flex expansion board will work just fine. I have them on order from OSH Park now. When they arrive in two or three weeks, I will expect to take about a week or so to prove their operation. Mid-October-ish.

06 Sep 2019

I'm changing schematic symbol numbers for U1 and up for the ICs and changing chip select pins from the MCU. That means I'll be publishing a new schematic and any references to U numbers and CS pins in previous postings and documents are now obsolete. Be sure when examining any documents referencing U numbers and CS pins that you realize those changes that occurred as of this date. I'm hoping to correct a lot of it on GitHub within minutes.

03 Sep 2019

I've adjusted the PCB design to accommodate an LDO for USB power decoupling and have been designing a flex board to hold gobs of MCP4262-502's in parallel. It's tedious but coming along. All those MCP4262's in parallel are going to reduce their resistance step size down from 5KΩ/256 (19.5 Ω) of a single Dpot to a possible 5KΩ/256/26 (0.751 Ω) with 26 Dpots. The result will be much, much better fine tuning of the signal, biasing it to the best voltage level (1.494 VDC) for the 24-bit ADC (HX711). If 26 Dpots ends up being too many, you can just place on the flex board the number of Dpots that works the best. I intend to design that flex board to be able to be cut to size conveniently.

26 Aug 2019

I've decided that the USB cables that power this board contribute too much Vcc noise when they're worn from a lot of use, so a zener diode off U2 [old U number] VREF is required. Learning from performing the diags, I also want to add a couple current limiting resistors, one each for U4 and U5 [the new U numbering has changed these U references to U5 and U6]. That means shifting board traces to make room where there seems to be no room. That will be PCB Rev 8. Rev 7 does work, though, if you can't wait. If you want to get PCB Rev 7 simply because Rev 8 isn't ready yet, you can air-bridge the zener diode on the Rev. 7 board if you want to. You would put the diode cathode end on pins 2 &/or 3 of U2, the anode end to GNDA, but you'll likely have to add a little bit of a wire to one end or the other or both of the diode. You'll have to do without the resistors, however.

Another change forthcoming in Rev 8, and even more vital, is stackable U6 [the new U numbering has changed these U6 references to U8, 9, 10, etc.] in order to obtain higher resolution by paralleling 5K U6's in a stack, less the SDO (MISO) pin 9's which aren't connected to anything.

24 Aug 2019

The board now works fully at 5VDC after cleaning specks and other contamination from the area around U6. It is not fully stable at 3.3VDC when a transient overload condition occurs. Investigating this...

24 Aug 2019

The board works at 5VDC only. I've posted new gerbers at OSH Park to change some pin assignments that tended to cause confusion while running diagnostics, but any differences between the newest board and the previous version are operationally inconsequential. You may now buy PCB boards from wherever you want to with assurance that the product works at 5VDC.

In fact, this now works so well at 5VDC that the HX711 may not even be necessary!

I'm investigating now into 3.3VDC operation. The difference seems to be that the U6 dPot with the buffering resistor isn't changing settings at 3.3VDC when being told to, yet it works just fine at 5VDC.

23 Aug 2019

I've fixed everything I broke an am now ready to proceed with placing U3 and SHLD1, which should be done as a single step without allowing any interruptions so that U3 is not left to absorb moisture from the humidity in the air. Let me say that this project is very, very difficult to build...nearly impossible for me.

22 Aug 2019

I'm hopeful I found the last of the circuit problems...about half dozen broken traces and lifted pads and about twice that many solder joints that didnt take.

21 Aug 2019

I am having fits with an apparent open circuit in the programming resistor circuit of U5. Yet another nightmare, but everyone building this will benefit from the improved diags that it makes me write and PCB test points I lay down.

19 Aug 2019

Today I found that I had placed two U7s instead of a U6 and U7. Would have made definite progess otherwise. Replaced. Starting again writing the diags for this step.

18 Aug 2019

I've made more progress on the diagnostics after having to abandon my first U1 [the new U numbering has changed these U1 references to U2] due to solder cross contamination. Early on I tinned the leads of that component with leaded solder and have been plagued ever since with it rejecting other solders. I handled it so much that its leads started falling off. $6. The new U1 took low-temp paste solder like a champ, but the whole experience sure tested my patience. These new bismuth solders don't mix well at all with the old leaded solder! A word to the wise.

When I resume development, I'll be writing diags for U6 and U7, the dPots. I'm hoping the ones I've installed have no issues for me to troubleshoot, but I can't expect a perfect world.

After that comes placing ADC1 and writing those diags. Then U3 w/last of diags, SHLD1, J1, and the potentiometers.

17 Aug 2019

I've been working on developing the build instructions which are found in the BOM named ibom.html in this directory https://github.com/kenneth558/plant_resistance_primary_perception/blob/Free/Where%20to%20find%20everything%20for%20making%20your%20own%20GWAAMC%20device/On%20buying%20the%20components/. On my own board, I'm up to the diags right before placing ADC1. What remains after any troubleshooting the diags make me do is to place ADC1, U3, SHLD1, J1, and the potentiometers. BTW, I'm writing the diags as I go, so that adds to the slowdown.

12 Aug 2019

I'm to the stage of fine position adjusting so this board best fits an MCU while I'm powering up for build diagnostics. The components I have left to place are ADC1, U3, the RF shield, J1, the potentiometers, and the wireless module. These have to wait until I develop diagnostic code for the state of the build as is. However, I'm taking a family road trip now and won't be making any more progress on this until late this week.

10 Aug 2019

While building, I've just learned one has to place the two-leaded components and U2 [the new U numbering has changed these U2 references to U3] after the multi-leaded ones other than U2. U2 is so compact that it is as susceptible to collateral de-soldering as the two-leaded components are. In other words, I've been slowed down building because U2 had to be soldered on twice, and it is a bear to do it even once. Past that, now...

08 Aug 2019

Building a new board that I expect will become the first working rendition of this project. I'm placing components, then performing diagnostic tests on what I've just placed. I've succeeded in the first diagnostics step of getting function from U2 and now in process of installing LEDs and their driving circuitry.

I'm trying to work the word "Entanglement" into this project's title. I think it works best being the type of gardening..."Entanglement Gardening".

06 Aug 2019

I just made one last improvement - I freed up pin D4 so it may be stacked on top of an SD card shield to store the plot during learning sessions. Code has not yet been written for that feature, though.

03 Aug 2019

I've been making scores of cosmetic changes to the PCB design and some copper changes to enable ESP-01 wireless module to work. J6 is now 9 pins long for wireless modules. I emphasize that I have no code to operate any wireless right now. Rx of the MCU runs to Tx of the module and vice versa. Some modules may have pinouts that will require solder jumpering of nearby pads. I've eliminated solder mask in those places to facilitate such jumpering. My hopefully functional boards of a previous edit level should arrive in a day or two and we'll assemble, troubleshoot, and publish instructions immediately.

26 July 2019

I'm seeing that if you choose to use solid wire solder (not that I'm recommending that, but if that is a limitation...), you'll need to use Kapton polyimide tape over both the components and the two tiny pieces of appropriately sized solder (obtained from being mashed down then cut with scissors &/or knife blade) at the component ends, and a hot air solderer is best. The Kapton tape stands up to the heat and holds everything in place while the solder melts. To allow the hot air to reach the PCB traces (a necessity or the traces won't get hot enough to adhere to the solder), you'll need to have the edge of the tape follow the center lines of the three pieces lined up end-to-end - solder, component, solder: one half of each of them are under the tape. Aim the hot air into them from beside the tape. They will tend to fly out of place unless you also add some flux or rosin on them before starting the air. When finished soldering, the same piece of tape can be re-used for the next component, as tough as Kapton is.

I've added guide traces (fiducials) for pick and place contracting and have updated the PCB. That way you can have a vision-capable pick and place machine place the components.

25 July 2019

Even though I have the PCB on order, I'm looking through the PCB design for anything that I might have missed because OSH Park has been very accommodating with me pulling the orders and replacing them when up until they send it out for etching if I find any little thing I decide I can improve in the PCB design. Those changes would be mostly in minor silkscreen edits rather than anything affecting operational specifications. There are a couple places where i've had to compromise with circuit guard rings in the two layer design. A four layer or flex design would perfect those issues, but at a cost that I can't justify right now.

Well looky there - I just got now notified that OSH Park sent the board in to get fabbed. Nice! While I wait on this final board design, I'll use all the prior designed boards I've accumulated to perfect the special soldering techniques I need on these ENIG boards. When soldering onto a new ENIG pad, the gold layer migrates into the solder and raises that melting point noticeably by making the solder less eutectic. I had not anticipated that in my first placement attempts so had my soldering temperatures set to just barely melt the solder. That didn't work, and I had to find out why. I now realize I have to raise the soldering temperature to overcome that new-ENIGed-pad effect. One option I'll consider is to remove the gold layer with double tinning that is required industry standard practice now, but I'll prefer a single-pass placement rather than a two-pass. Why is this such an ordeal? Because I'm trying to avoid leaded solder, which leaves me possibly using higher temperature solder even while one of the components, U3 [the new U numbering has changed these U3 references to U4], specifies a soldering temperature lower than is common for electronics components. Besides, lead and gold don't play solder together at all, yet ENIG gold plate keeps the pads nice and flat for fine pitch components. Adding the place-assist holes around those fine pitch components like I've done in these later PCB designs may allow everyone now to use lead-free but not flat HASL instead of ENIG and allow single-pass placement. Lead-free HASL tinning melts at about 450F, so that is getting close to the limit for U3 where temperature control and limiting will absolutely be essential.

24 July 2019

I put alignment assist holes around U1, U2, U6, U7 and the shield and put the new PCB design. I finally feel very good about the state of this PCB design but emphasize it is still UNTESTED.

20 July 2019

I am developing a different method of soldering the center pad of U2 - I am putting a via underneath it that I will give a slightly oversized hole. That much is in the latest PCB files here on GitHub, but since I am also adding line-up holes around the more difficult to place components in a design later today, I pulled the OSH Park order until I finish that enhancement.

19 July 2019

Today I discovered that disparate solder types/compositions cannot be mixed. That tells me why I have been having so much trouble building my first boards. I've been cavalier about mixing 138C Bi/Sn eutectic solder and leaded solders, for instance. That causes Bi solder to fail when thermal cycled above 95 degrees C. Other types of cross contamination simply won't adhere by increasing the melting point too much. I'll have to start building a board all over with a clean, new board and clean, new components. Lesson learned. I expect much easier building after this! I also updated the PCB to space components better and fix silk screen imperfections. The PCB now also contains a warning against mixing solder types so we never forget.

16 July 2019

Soldering clearances were less than ideal. Increased them on PCB and put the new one here and on OSH Park.

15 July 2019

When I powered up the board-under-construction, the prelim diags showed I had given Q3-7 a wrong pinout at the footprint. I pulled that PCB design off and put up a new corrected design.

13 July 2019

I made some very minor adjustments on the PCB to move some words in silk screen and solder mask to where I think they would look more professional. That PCB design won't appear at OSH Park until my next order, but you do find it here at GitHub, as always.

12 July 2019

Added a jumper near ICSP1 pin 2 for ease of disconnect in case the power there doesn't match a 3V3 MCU Vcc. I discovered that will sadly be the case with most 3V3 MCUs. PCB files updated, new OSH Park order placed for the project on OSH Park.

11 July 2019

I discovered the Arduino 101 is a 3V device with onboard BLE that could allow this HC-05 capability to then be used for wifi instead of BT/BLE. Since that would be a nice feature, but the Arduino 101 is 3V only, I modified the PCB a little to allow easy jumpering to 3V operation without needing to cut any trace. Just pull out pin 5 of J2 if using the Arduino 101 and use part of it or a similar size wire segment to solder down as a jumper instead of cutting a trace. If you do this, don't ever use the board in a 5V environment without reversing this mod!

I have not tested 3V operation idea, and I would not expect Q1 and Q2 to allow for it unless those two JFETs are shorted across S to D. However, it would be very nice to find they don't need to be. There are 3 different versions of that JFET (MMBF5460, MMBF5461 & MMBF5462) - one or even all of them might work fine in the 3V environment without shorting them out of operation. I still haven't built a board completely, so I don't even know which JFET version I'll end up with anyway.

08 July 2019

I added BlueTooth module capability, which means a new PCB design is out there now. I have not added the BlueTooth module to the parts list...its up to you to find one that suits you. Get one that follows the same pinout as HC-05, a very common pinout for wireless modules in applications like this.

07 July 2019

I made a few tweaks on the PCB design and couldn't think of any more to make, so I ordered that latest PCB design which updated the PCB design published by OSH Park to the very latest. Still, be aware that I have NOT assembled any PCB fully enough to operate. I expect I will do so in about a week, but I cannot claim any of this project works in its present design. All developers of hardware and software know enough that future changes in circuit and software should still be expected at this stage of the project.

06 July 2019

Discovered I neglected to order one last(?) component needed for board completion. Ordered it from USA, and I expect arrival in a week maybe. I also discovered a silkscreen imperfection and posted a new PCB layout on this site only (not on OSH Park site until my next PCB order happens) correcting it.

03 July 2019

Today I placed my electronics component order that I expect will complete my purchase of all remaining components I need to populate and energize the first rendition of this project. Enough of the order should arrive in two days to get it working.

29 June 2019

After spending another couple hours trying again to solder C1 down, I realize that the problem of there being too much space between C1 pads for both to attach to C1 simultaneously, if this board version had to be salvaged, would require some makeshift piece of solid copper to bridge to one or the other pad. I'll get back to that possible workaround once I get a more workable board designed and ordered.

In my judgment, the specific C1 I'm working with right now is actually too small for the 0201 pad. It should have been spec'd for the next smaller size pad, which is actually not feasable with my PCB software nor fabricating house.

27 June 2019

Discovered that I spec'd C1 far too small physically to handle and solder onto the PCB. I lost about 8 of them from their small size while simply trying to develop a means of soldering them to the PCB with solid solder. That step took about 4 or 5 hrs and I ended up pulling the C1 footprint off the board. I'll have to wait until my low temperature paste solder arrives later to resume my attempt to build my first working PCB. BTW, this small C1 will fit into a space half or less the size of a pinhead and just might be the smallest component to claim a 0201 pad. Every time I observed one through the magnifying glass, I was newly amazed all over again how microscopic it is. It is not even long enough to connect between both pads without adding some sort of bridging conductor.

I'll be making a new PCB design once again <sigh...>

15 June 2019

PCBs on their way to me for arrival Monday 17 June.

09 June 2019

It won't be too much longer before I get my PCB (bare) I have on order. I have enough components to build it out. Depending on how many build mistakes I make I could have it built by July week 1. After that my guess is another 3 weeks sketch development time.

06 June 2019

My ultimate design has been put into PCB design and three PCBs have been ordered. Once those are built up, I'll test this latest and final circuit design. Note that I haven't tested this design due to the breadboarding challenges I didn't want to deal with.

31 May 2019

With the recent sub-nano scale enhancement, I realized I'll have to add shielding and more extreme guard tracing to gain the full benefit. Back to work on the PCB design, and I've cancelled the PCB order until the new one is designed...

25 May 2019

The annotation numbering was changed today for the components, and the signal input 4K stabilizing resistor was changed to two resistors in series: a 3900 and a 100. AND silk screen adjustments for better presentations. Newest files uploaded for you.

24 May 2019

Neither the BOM nor the sketch are updated fully for the latest circuit. Until I actually build a board using the latest design, I won't know if it can even be done, considering soldering iron clearances, etc. so please know that. Rev. 6 PCB is being fabricated now with an expected fab finishing date of 10 June. I've uploaded the project file and PCB gerber files for you.

20 May 2019

Pondering done, I will utilize the ADL5315 (for pico-scale adaptation and plant level signal current creation and duplication) along with two LM334s (for programability with MCP4262s, tempco offsetting, and micro-scale adaptation needed for the REF200AU) and the REF200AU (for high impedance current-to-voltage signal transcreation). The high impedance that the REF200AU offers is needed to maintain viability of a pico-scale signal current amongst the micro-scale operating current. Although pico-scale signal viability may end up being wishful thinking, I'm confident from reading the datasheets that at least nano-scale viability will be within reach.

17 May 2019

I have decided to dramatically change the front end of the circuit in v.6 to do away with the LM334's and the REF200AU in favor of the ADL5315 if I can figure out how to convert a few nA sourced signal current to a signal voltage centering at 1.5VDC without lowering its impedance into oblivion (where oblivion happens to be the very range of the MCP4262s that I'm using :-(). However, I've already ordered boards to be made at the v.5 level, and those boards will be delivered in a week or two. I'll probably build them out as my reference point while I ponder...

10 May 2019

While we await this latest PCB to be made and sent to me, I'll explain that I do expect a certain amount of undesirable temperature effect on the circuit and I don't intend to use mathematics to predict its amount and design it away. Instead, I'll wait and take the empirical approach to deal with that. Suffice it to say at this point that you should try to maintain an unvarying temperature on the circuit board during operation until v.6 of this project.

09 May 2019

The still untested v.5 circuit board design is here: https://github.com/kenneth558/plant_resistance_primary_perception/blob/Free/embeddeds/v.5%20training%20aid%20board%20gerbers.zip. No further circuit changes after that should be expected until I decide how to incorporate temperature compensation based on experimentation with v.5. Alternately, I may decide to incorporate the ADL5315 in v.6 to replace the two LM334's and REF200AU. Regardless, v.5 PCB version contains helpful construction-time diagnostic capability for helping the person building the board to catch any soldering mistakes (shorts and opens) early enough to repair them without having to remove umpteen other components that would otherwise block access to the soldering mistakes that will be inevitable sometime, somewhere. This PCB version also contains an untested sensitivity enhancement which can be jumpered out of operation in case it ends up not being viable.

26 Apr 2019

As of 26 Apr 2019, I have not gotten any PCB rendition of this circuit working, plus I've decided to incorporate additional circuitry onto the board which will enhance sensitivity to a more reasonable capability. The board space that the additional circuitry uses forces me to use SMD much more, thus requiring extensive layout redesign. That board will be Rev 4. [UPDATE: V.4 design aborted due to the coming of a second enhancement that caused me to pause v.4 development] Though I personally have not gotten the circuit working with the PCB design I publish here, I do believe someone with electronics skills will be able to gain a great deal of headway with it. (I am talking about Rev 3.). I am proceeding at a leasurely pace with Rev 4 [UPDATE: v.5] PCB design.

17 Apr 2019

My first try with MCP4262 dual dPots and a real printed circuit board is not working. Am troubleshooting....

08 Apr 2019

I have corrected the known errors in the printed circuit board design. These errors were: Q1-3 body shown in silkscreen was backwards and VR1-2 adjustment directions were backwards. I also spread apart the leads for Q1-3 and for U1-2. The new board is designated v.3 in the file name.

07 Apr 2019

After trying unsuccessfully to place components on my first three boards, I offer this word of advice: use low-temp solder on the MSOP packages. I delaminated those traces on two boards while the third board is left with shorts who knows where. Though I know the boards were subjected to higher soldering temperatures than my previous soldering jobs, I feel just the same that by paying the highest price in the world for this batch, I should have received more durable boards...thank you for bearing with the complaint.

01 Apr 2019

Increased value of R2 from 1Mohm to 1.5Mohm for possibly a little better temperature tracking with the DUT leg. Improved the Bill of Materials to be an interactive web page (download/offload it because GitHub won't serve it as its own web page like needs to be done for it). May populate it later with vendors, but its main use is to track components as they are placed on the board. Components need to be soldered lowest profile ones first so the taller ones don't get in the way of the soldering iron and burnt.

29 Mar 2019

I added a green LED for confidence that the electrodes are connected properly. I uploaded my PCB gerber files to here: https://github.com/kenneth558/plant_resistance_primary_perception/blob/Free/embeddeds/v.5%20training%20aid%20board%20gerbers.zip (link updated for v.5 09 May 2019). The BOM is near the bottom of this page: https://github.com/kenneth558/plant_resistance_primary_perception/blob/Free/More%20construction%20details.md

14 Mar 2019

I completed the PCB design and will use tomorrow to find a PCB maker beginning with local sources (a community college, a state university). Have received most of the components to build the innards of several training aids. When we have a couple working training aids, I will develop the best form of the rest of the training station design to include the project box. Top view of circuit board: https://github.com/kenneth558/plant_resistance_primary_perception/blob/Free/embeddeds/top%20side%20training%20aid%20extension.png Bottom view of circuit board: https://github.com/kenneth558/plant_resistance_primary_perception/blob/Free/embeddeds/top%20side%20training%20aid%20extension.png

26 Feb 2019

Note that the latest sketch version contains an untested feature - gain adjust via potentiometer or a 3 position switch with voltage divider resistors. Also I have not tested any gain other than 128, and to remind you, a gain of 32 is going to require INA+/- of the HX711 to be paralleled with INB+/- of that same HX711. I am still working on the PCB design, and until I master that, I don't foresee any other project advancement happening. UPDATE - 14 March 2019 - still working on the PCB design in KiCad.

20 Feb 2019

Using KiCad for making a PCB from my first schematic is SO un-intuitive for me b/c I set some traps for myself in the way I created the schematic. Seems like I'll have to create a new schematic from scratch while I simultaneously develop the PCB with each change I make in the schematic. This sets my expectations aback. On another note, I'll be adding a visual and/or audible indicator to assist the operator in achieving enough conductance while placing the plant-side (upper) electrode. If you remember, I decided that the standard electrode placement will be with the reference electrode pushed into the soil or other root-circuit substrate. Electrically, the reference electrode can be either one during this development phase via the 4-way switch, but mechanically it will be fitted with a probe instead of the flat conductive sheets.

15 Feb 2019

In studying the HX711 datasheet better, I understand how INA and INB (channel A and channel B inputs) are selected to be read from. To read at a gain of 32 I simply parallel INA+/- with INB+/- and compile the sketch for 32 gain. That is nice because the plant I attached to needed a substantially lower gain than the 128 I used for it. Since I left channel B inputs at NC (no connection), I will have to so some soldering and schematic modifying. Where the AD8244 buffers where paralleled, I will change the wiring to have one pair of buffers drive the INA and the other pair drive INB. Simple as that.

One single challenge remains to me: configuring the electrodes so as to acquire high enough conductivity for the LM334 circuit to work by surface contact electrodes. Right now, I can only achieve high enough conductance for meaningful operation if I pierce the leaf membrane with the signal electrode.

31 Jan 2019

I've been considering what the best electrode design is to be, and I've discovered a company named ERG Aerospace Corp. that makes what they call Reticulated Vitreous Carbon (RVC) that I believe would be a fine active electrode surface. I've ordered a sample from their U.S.A. distributor, Duocel. I'm sure hoping to obtain all subassemblies and parts from Judeo-Christian friendly economy sources, so I'm happy that this source qualifies.

30 Jan 2019

With the higher than anticipated resistance, my plan is that one electrode will be put into the ground and the other will be placed onto leaf surfaces (multiple, if needed) so that a much greater contact area can result in a lower resistance.

21 Jan 2019

A big Thank You! to the maker bunch at Turbine Flats in Lincoln, NE for letting me demonstrate the GWAAM-Sea training aid. What went wrong, you would ask. In reality, the problem wasn't exactly the electrode contact resistance, but rather the resistance through the leaves. That plant I brought to the demo was not selected based on the necessary criteria, which would have been its electrical conductivity. Instead, it was selected based on availability and price. It was from a funeral of an uncle of mine last summer. Its conductance through the leaves was just far higher than the sketch as of today is able to handle. It remains to be seen whether its conductance is too low for the circuit with LM334 being used. If I find out it is too low, I just may modify the bridge to use only dpots - a bit more expensive of a scenario. My first course of action will be to invert the initial linearity calibration such that the bridge balancing begins with maximum dpot resistance rather than minimum.

15 Jan 2019

YOU NOW HAVE THE PUBLISHED VERSION v.Free, DATED 15 JAN 2019 THAT IS FUNCTIONAL TO THE EXTENT I RAN IT. To read about the only configuration I've run it under, see [Current Operational Status](https://github.com/kenneth558/plant_resistance_primary_perception/blob/Free/Current%20Operational%20Status.md) Testing now has to be done to locate all bugs. Many defunct and obsolete remnants and printed messages remain. I do intend to add wifi connectivity and SD Card local storage in the future. I do expect bugs to be found, but they should be very easy ones to fix. THANK YOU, LORD!

08 Jan 2019

I've been working hard at this, which is due to several structuring changes that needed to be made in order to allow for various enhancements. The new structure is now complete. Bug track-downs are next. None of the remaing bugs should necessitate the whole scale re-writes that have taken so much time. They should be simple fixes, but I'll take a few hours of leisure right now.

17 Dec 2018

Taking time from sketch development in order to beef up the prototype circuit from residing on breadboard to more residing on perfboard. I wish I had a fiberglass perfboard lying around, but all I find is phenolic. That will have to do for now...

11 Dec 2018

A person could wonder why I'm not providing updates more often. There seem to be several reasons for that, one of which is that when I am in the process of investigating hinderance with the project, it may take several days for me to accurately discover down the cause. Such is the case for the latest hinderance. I thought I had it characterized correctly, but am having already to correct my post of a minute ago and I know better now than to even say what my initial incorrect posting was. SIGH.

08 Dec 2018

Sketch development is now the farthest along and working, just needing some cleanup before I put this latest version out for you. Give me a day or so, please. Thank you! The sketch you will then have will be the skeletal basis for future enhancemants and expansion.

03 Dec 2018

A follow-up to the idea of wifi capability: I believe I mention somewhere that I am writing this sketch specifically for the Arduino IDE Serial Plotter, which I am. Connection to that Serial Plotter is unidirectional data flow, making an inherently bidirectional wifi connection a level beyond the immediate scope of development. Not only that but the data flow also becomes incapable of connecting to the USB-only IDE plotter. Wifi capability eventually will be part of my planned client-side development as evidenced by the fact that I've already started accommodating it by means of extensive data now being included in the outgoing data stream. The Arduino serial plotter conveniently ignores this additional data, but a future client renderer will use it. When I do get around to coding for the wifi, my current thinking is that it will utilize HTML5 SSE (preferred) or WebSockets (with the plant side being a web server), both of which I have not yet developed for. Client-sent data would be via HTTP put() methods that would allow some means of operational adjustment at the plant end of the wifi connection.

My current state of development is debugging the predictive auto-adjust. I tripped over an attempted WEMOS XI compatibility effort and have to go back over those changes I made during that effort. All for nought, it seems, b/c I discovered the WEMOS XI has been discontinued.

UPDATE: Sollae Systems has developed Web Serial Plotter which uses WebSockets and integrates so well into the Arduino IDE that I would not have to code my own client-side renderer. NICE! The disadvantage is that my super-inexpensive ESP8266 modules will only be useful at the computer end of the wifi connection while the plant end would require a $$$ part costing more than all the rest of the circuitry combined, and I only hope their use of WebSockets over SSE is for the better. I'm sure the PHP server is in their shield so it would be well nigh impossible to reverse engineer the PHPoc source code to emulate the expensive board at the plant end while still using the Sollae Systems Web Serial Plotter at the computer end.

UPDATE: A more cost-effective solution would be a peer-to-peer wifi connection that down-converts the data back to USB. Then I could use duffster's (duff2013's or duff's) more fancy "Arduino Serial Plotter 2.0", if it will be made available.

30 Nov 2018

Thanks to Brook for suggesting that I add wifi connectivity, SD Card storage and a means of plotting plant growth. On attempting to include to first two, I've found RAM limitations could be a serious challenge with boards only having 32K RAM or less. The plant growth idea may not even need to connect into the Arduino, if it could be a standalone video camera. Unidirectional growth upward might easily be included, though. I don't plan on offering any of these enhancements in the free sketch version.

16 Nov 2018

Scalability changes made to the sketch. DO NOT USE PREDATED VERSIONS TO ADVANCE DEVELOPMENT!!!!!!! Many changes have been made as of this date to accommodate future scalability. This specific upgrade effort has been what has kept me busy lately. Had I not done this, I couldn't have improved the sketch beyond the very unimpressive level it stalled at. There are still places that need attention, but I needed to get this scalable style out here for community involvement. Somewhere along in the process I lost indenting, in case anyone would like to take that on as a little project. Just fill out an Issues entry so we don't get multple people on that one project.

2 Nov 2018

I'm rebuilding my prototype circuitry to improve robustness, considering that I'm traveling hither and yon with it. I've made more progress with sketch transition, but there is still more to do with it. The sketch posted in the base sketch directory here on GitHub is still as far along in performance as any version I am working on locally.

22 Oct 2018

Using the 20 Sep version to regain functionality, I'm taking smaller, baby steps to expand the sketch. My current goals are

• to allow for use of less than 3 and more than 3 DPots/leg
• to allow for using DPots without LM334 in bridges
• to allow for using unbridged DPots
• to allow for multiple HX711s and, of course
• to allow for faster conductance range changes when DUT conductance goes out of range.

In the meantime, I have coded a new pre-processor macro PLOT_ONLY_THE_SIGNAL_LEG_OF_SUPERIMPOSED_DPOT_LEG_BRIDGE_INBOARDS to suppress the reference plotlines of superimposed traces. Using this macro, however, may alter the intuition behind specifying how many inboard lines are to be plotted. I'm not concerned at all about that probable disconnect from intuition, since it is only a matter of very little experimentation for you to determine how many inboards to plot if you use the new macro. Specifically, if you define this new macro in your compilation while superimposing bridged traces, fewer traces will be plotted than specified by the NUM_OF_INBOARDS_PLOTTED is defined for. If this becomes confusing to users, I can consider changing it in the future. It works great as is, so I'm moving on to the other goals in the meantime.

11 Oct 2018

There was something to be said for the 20 Sep 2018 version even though it auto-adjusted slow. At least it auto-adjusted whereas changes since then have never worked as well. Am reverting back to 20 Sep in the auto-adjust respect only and re-publishing with other modifications included that do work fine. Hope to finish this publish goal today....

06 Oct 2018

Sketch versions dated today and later will now accommodate the dual pot version of MCP42xxx. The way the second pot in the package is controlled to that its pin number will have the pin number of first pot in the package OR'd with a hex value of 40 (0x40) as being the pin number for the second digipot in the package in the defines section of the sketch right where the digipot pins are currently defined.

04 Oct 2018

Its been very slow progress in recent days as I realized the sketch needed an overhaul to its inter-function flow to remove some algorithm duplicates when I found it overly difficult to get the bracketing within the code correct. Better to deal with it now than later.

One function to debug, then sketch will be published in beta test mode. That function is labeled with an #error preprocessor directive for ease of locating.

20 Sep 2018

Breakthrough with the software today. I'll post the latest code and begin working on the electrodes aspect of the project. It has non-predictive auto-adjust on the digipots and thus an auto adjust takes quite a bit of time. I think the free version will stay that way.

19 Sep 2018

Still working on this. I'm suspicious of my breadboard connections being intermittent. I will resort to soldering things together on a perfboard for increased robustness.

16 Sep 2018

I'm able to dedicate more time to finishing this project now. I need to give it the ability to auto-adjust to the widest range possible of operational conductance levels. Up until now, I've only been getting the project to function with the DUT being a fixed 1Mohm resistor, which now works very well in itself but is not very practical working with conductances other than 1Mohm. Obviously, you can't expect a plant to be exactly centered on that value, so I'll be working on that now...

26 Aug 2018

In the sketch, I've laid out the direction for the new array concepts in self-documenting macros. If you compile the sketch, it will stop at an explicit preprocessor error line stating as much. The rest of the sketch next needs to be brought into compliance with the macros as they are aptly named. This new array concept should at last bring the sketch into a bug-free and useable state. After a good night's sleep...

26 Aug 2018

The sketch bugs I have been struggling against have been due to me wanting to develop in too straight of a course in that some arrays are made too multi-purpose and have indices that are too different, both from each other and in the various possible configurations. That approach lends itself to too complex of troubleshooting. I am breaking out those arrays into smaller, more specifically named dedicated arrays with more specifically named indices and macros/variables specifying their sizes simply to speed up sketch development.

24 Aug 2018

Bug found relative to displaying multiple traces while one or more are from inboard analog pins in combo with magnified and who knows what. I am forced by the reality of things to strongly suggest that until this sketch has been bug-free at the latest revision level for at least three or four weeks, all potential users just hold off unless you are prepared to fix my sketch bugs. Sorry. The current "bug" might instead be circuit malfunction, but until I finish troubleshooting, I don't know.

23 Aug 2018

Added ability to reduce the magnification via a pot connected to a spare analog input pin specified with a new preprocessor define POT_WIPER_TO_THIS_ANALOG_INPUT_PIN_TO_ADJUST_MAGNIFICATION_FACTOR. Fixed all known bugs, specifically with magnification trace plotting. One thing left to code is auto-adjust when readings go off-scale. Code for this is commented out right now. It will use preprocessor macro LOOP_COUNTER_LIMIT_THAT_TRACE_IS_ALLOWED_TO_BE_OFF_CENTER. I have decided that the 10K digipots are better off being 5K instead. Ordered new ones. A few lines of code will have to be changed to use the lower values by adding 10 to some rollover literals.

22 Aug 2018

I'm able to code periodically. When I changed the plotting algorithm to handle larger values it generated a bit of required coding time that I'm in the middle of. I also purchased a few 74LVC138 decoders to allow more digipot banks. They should arrive today thanks to Arrow's free overnight shipping.

10 Aug 2018

A number of non-project tasks have arisen that I need to place in higher priority than this. Expecting nearly a month delay...

02 Aug 2018

Not a memory management problem but is a casting problem: The print command won't print ( long long ), (long) is the closest it can get, and when casting down to a (long) is loses value.

31 July 2018

I've noticed one sketch bug in the printing of the magnified traces in that they don't stay within bounds any more. Older sketch versions never had that problem. I'm not sure where it began, but it seems to be a memory management problem from my inital troubleshooting rather than any type of algorithm weakness. I'll work on that when I'm able...

29 July 2018

I finally feel confident in the condition of the sketch. I have corrected all bugs that I've ever found in it and have it completed to a useful level. I recommend using it with two LM334s in the Wheatstone bridge input stage - one LM334 in the positive half of each bridge leg. Each LM334 in turn has its current controlled by three digipots in series - two MCP4162-104 and a MCP4162-103. Note that the two MCP4162-104s may be able to be replaced by a single MCP4262-104 instead, but I have not tested that idea. I would also recommend yet another resistance in series with those three to prevent too low of a set resistor value if the digipots associated with an LM334 should all get set to zeros (very likely considering I've not prevented endless loops). Which tells me that there is still some sketch work to do relative to handling specific failure or over-limit conditions that can arise. For now, though, I have other projects needing my attention.

I also recommend using an AD8244 if you can solder tiny things. Perhaps I'll eventually employ small toroid inductors in series with the LM334s, but the noise level right now looks low enough.

To take the sketch operationally (the code will never be elegant in appearance) to a professional level, I will need to handle the input voltage level tracking on the electrode from the plant. I say voltage but you do realize, I hope, that the voltage level is simply produced by the Wheatstone bridge and it correlates to a plant resistance.

Eventually, I'll get a schematic published for you.

18 July 2018

My two HX711s read differently from each other, so I'll have to order two more. Expect them no earlier than 30 July, costing me $3.03 due to added shipping cost to get them that early. Until they come I'll work on repairing the differential conversion algorithm for negative 24-bit values, at least for two's complement ADC libraries like the HX711.

17 July 2018

AD8244 arrived at 09:07. Nice! However, it's MSOP pin spacing is well less than SSOP. I had to spread the pins a little bit to solder it onto the SSOP adapter board I had on hand. Initial power up indicates I have a wire misplaced, but this is too late at night to continue today. Tomorrow I would do well to focus on the garage roof instead of this project.

16 July 2018

Since I conclude that buffers like AD8244 will be needed in any case, the Wemo XI Analog Inputs and ADS1x15 become just as viable as any other ADCs. I will change comments in source code to that effect.

Received notice that my AD8244 has been shipped with overnight expidite, no extra charge. ETA tomorrow, by 4:30pm. Thank you, Lord!

15 July 2018

Just discovered the HX711 input Z is way too low. The datasheet hadn't specified it in any way, nor would I have deduced the implications anyway. Regrouping...will need to use High-Z buffers (AD8244 or equivalent) or a different 24-bit ADC as the base device. Ordered a AD8244....

08 July 2018

My efforts continue to be sidetracked by the urgencies of property maintenance and repair. My mower has broken irreparably, and with zero income, I haven't the money to do anything about it other than mow my lawn with a hand-held string trimmer. I had to fire the crew that was repairing my garage due to their admitted inability to pass a drug test after I had been dismayed with their poor workmanship. City hall has verbally indicated that the city is eyeing my property as being on the verge of getting written notices, only three which are allowed before city action to de-blight. A realty company is having a hard time seeing a win-win path for me to sell due to the cost of fixing the house and garage.

In light of all this, It would be irresponsible of me to work on this project at the expense of my house. Progress on this telempathic gardening training aid, therefore, can be expected to be minimal for months to come.

28 June 2018

My efforts have been sidetracked for some days by the urgencies of property maintenance and repair.

20 June 2018

I had assumed the excitation voltage from the HX711 across the bridge circuit would be higher than it is in reality. I though it would be high enough to include headroom for LM334s to operate in the upper halves of the bridge. Not so. The positive excitation voltage is less than 2.5 vdc using the high impedance scenario with the plant. Not only that, but the negative bridge power is about a diode drop above ground. I will have to supply the bridge from +5 and GND and see what that does to the noise floor. Datasheet for HX711 says to supply less than within .1 vdc of Vdd, so I am risking a $1 part here...

19 June 2018

I received the correct digi-pots and should be able to work on this project after I get some yard and house work done. I added a .png screen shot file of a comparison between one inboard Analog Input vs. HX711. Note that LM334s are being used as the top component of both bridge legs. That will be a standard inclusion in the circuit hereafter.

I am still unsure of the 24-bit differential non-magnified display amplitude. It seems far too insensitive because it should max out with a 20mV tweak up or down, and I don't see it maxing out at all. I'll investigate. The good news is that this is nothing more than my suspicions of display amplitude of the lesser important trace, so I'm not declaring it a bug and especially not a real problem.

17 June 2018

[OBSOLETE:Since the Wemos XI/TTGO XI and maybe other boards have worthless Analog Input pins, yet] two full-scale input pins connected to the two bridge compare points can be very useful when a 24-bit high sensitivity load cell ADC is used, I am contemplating how to change the sketch to accommodate, not just one, but two add-on ADCs. This will also enable the sketch to be used with micro-controller board without any Analog Input pins, and ARM micro-computers (which have no Analog Input pins). The alternative of using an analog buffer for the Wemos/TTGO would not be scalable to boards that have no Analog Inputs at all. I envision one of the add-on ADCs to be a full-scale ADC used for digital potentiometer setting to balance bridge legs and the other ADC will be the high-sensitivity ADC for magnified display. The full-scale ADC levels will only be displayed when DEBUG_BRIDGE_BALANCING mode is defined. The inputs of both ADCs will be in parallel, so both must be high impedance. Note that the ADS devices are indeed full-scale and inexpensive, but they are not truly high-Z.

In my datasheet research, I am finding that inexpensive ADCs like the ADS1x15 in general undesireably make the source (the plant in this case) charge a S/H capacitor.[This point is moot if you just use an AD8244] I'll investigate the ramifications of that after I get the digital potentiometers working to track and autobalance.

I continue to be encouraged as I learn that most people I talk with pseudo-randomly accept the premise on which this project is based. I really didn't use to think the acceptance level for telempathic horticulture was that high. I'd say the time is finally right for this!

15 June 2018

Received four MCP4162-103E/P instead of four MCP4162-104E/P...I've made the same mistake again!!! Ordering again.

Using the MCP4162-103E/P that I had, I learned how to address two of them in case I need a second one in circuit for finer adjustment.

08 June 2018

Had to re-order the programmable potentiometers due to the resistance value of 100 kOhm not being available from Chinese site, forcing me to order a 10 kOhm value that will require a series resistor and limit the adjustment range. I must have gotten interrupted during that shopping effort and forgot to continue. Will do what I can with the one I got and ordered from a stateside supply house for $5 shipping instead of free or pennies. Upside is that I may receive the order in a week or two.

05 June 2018

HX711 now supported...you'll have to balance the bridge yourself, though since my programmable potentiometer hasn't arrived. Use a SAMPLE_TIMES of one. I only tested with PGA A at default gain of 128. This is a sensitivity breakthrough, since this little ADC converts a swing of a mere few millvolts into 24 bits compared with the ADS1115 converting rail-to-rail swing to 15 bits. The ADS1115 sensitivity doesn't hold a candle to any ADC with a full-scale conversion range in the millivolts. The library to use with HX711 is https://github.com/bogde/HX711

I haven't completed constructing a satisfactory electrode subsystem yet...

04 June 2018

Home maintenance issues are taking much of my time away from this project. I am trying to develop a method of holding the electrode assembly against the leaf and keeping a good seal. In a shopping trip to Menards-type and Walmart-type stores, I put together a few items to try various methods.

My programmable potentiometers haven't come in as of yesterday. They are important for getting the HX711 going. Otherwise, the only differential input devices I can support are those that have a full-scale range of near-rail-to-rail.

29 May 2018

Considering supporting AD7780. Looks like a good value 24-bit ADC for full-scale differential input... ADS1242 as well.

29 May 2018

Decided I will need to support the ADS1231 as well. The HX711 will be difficult due to max scale occuring at 0.08 volts or less, and I feel that will be wishful thinking at best to insist on that close of a match between A0 and A1 legs in the lay person world besides dealing with the noise from all sources and junction contacts at that amplification. Considering LTC2400, AD779x and MAX112x0.

28 May 2018

I realized the advantage of reading differential when using the LM334s. When I get differential reads working it will allow use of 24-bit load cell-type ADCs. I have two 24-bit HX711 on hand, so the project will eventually additionally use HX711. I expect first or second week of June will see this project working with differential while still keeping the capability for single-ended. I have secured domain names for "reviving civility" and "telempathic horticulture" which point here where they will remain pointed until I get a better page built.

19 May 2018

I received the LM334 and tested the noise level with and without it. I am pleased to say that the noise level IS lower when that part is used in the circuit as the source of the current going through the plant signal electrode. However, in my efforts to investigate into more detail, I discovered that the ADS ADC I was using is exactly what I ordered - an ADS1015. That means it is only a 12-bit ADC, and I want the 16-bit ADC for greater resolution. Work on this project will therefore be stalled until my newly-ordered ADS1115 arrives in a couple weeks (I paid $2.35 extra for expedited delivery this time). 24 May - ADS1115 arrived, but unable to work on it for a few days due to being out of town for an extended family gathering.

When I was testing with the LM334, which alternately is a temperature sensor, the temperature drift was too overwhelming to work without temperature compensation. Temperature compensation is provided by a silicon diode and two resistors (instead of one) setting the current. Those two resistors must be of an exact ratio to one another to cause the proper portion of the diode's temperature coefficient to exactly offset the LM334's opposite temperature coefficient. Since I'm not going to force myself to buy the specified diode that provides exactly -10 times the LM334's temperature coefficient, and with programmable potentiometers available (AD5242 and MCP41100) I don't need to, I'm just going to use a common diode such as 1N4148/1N914/1N4000 and write Arduino code that will proportion the resistances correctly for whatever diode is in the circuit and store that as a calibration in EEPROM. Writing that code will add another week or more because I'll need to have the right conditions drop into my lap to get the time and conditions to work on it. (Note to myself: the 1N457 has/had a tempco of -2.5mV/C while a 1N4148 might be anywhere between -2mV/C and -2.5mV/C, a 1:7.8 vs. 1:10 resistor ratio.) UPDATE - I intend to forego the diode idea in favor of software compensation ujtilizing a second LM334 circuit sending a temperature driven analog signal to a second analog input.

I am unable to find programmable pots of sufficient resistance to directly utilize the zero tempco circuit as presented in application notes, so I will have to try variations on that theme. Add more days/weeks to the tempco effort with an as yet unpredictable outcome.

When I've finished these tasks, I hope to offer a good estimate of how much better the LM334 is than just using a resistor to supply the current to the plant. I expect to find that the use of an LM334, two programmable potentiometers, and a diode or so will actually be indispensible to use in place of a simple series resistor to the signal electrode.