DIY Source-Measure Unit SMU Chapter 1

Welcome to my Source-Measure Unit project, DIY-SMU. 6 current ranges up to 100mA, and 3 voltage ranges up to +/- 150V. Open source and Arduino-based. / For more design details, see:
www.djerickson.com/diy_smu

[@reps]

Wow, fantastic work! I feel your pain with the off the shelf subrack enclosures, been looking for something similar for quite a while too.

[@ericksonengineering7011]

Thanks Marco. I'm trying to get an ME buddy interested in designing a 2U, half-rack box. Love your work!

[@VinisLab]

@@ericksonengineering7011 I could help you with that, I have one designed for OSMU actually.

[@hullinstruments]

What ever happened to 0SMU? I actually mentioned that in a comment on his most recent video regarding this meter. I mentioned your design and was curious if anyone knew about it… or had any updates? It would really be a great contribution to the community… Not many folks that have your skill set and willingness to design something like that to share. I watched your new video probably five times tonight. It’s your fault I didn’t get a new truck this year. Been saving and shopping around for two years. Instead… I have an empty truck fund… And several fiber lasers and a few beautiful DPSS heads. All for pennies on the dollar of course… But the shit was still expensive. And I’m hoping by this time next week… I’ll have the yellow DPSS module back together. For years I’ve been proud of my trio of 34401A meters. And my mountains and mountains of test equipment. There’s absolutely nothing in the world I could imagine needing for my lab. I don’t even have a good excuse to on most of the stuff I’ve got. But it’s your fault that I have a freshly repaired/re-calibrate 8.5 digit Beauty. She’s a dirty little whore (just the way I like em) she makes me cry but her 8 1/2 digits don’t lie. She was a bit promiscuous in her early years… But she’s just now hitting her prime… And i know how to make these dirty girls behave. You and her are the reason my truck phone is gone. Probably be my wife next.

[@vmiguel1988]

@hull guitars USA we are working on osmu, but the people involved has quite limited time to the project and Marco does not feel comfortable to advertise it while we don't have something minimal functional.

[@hullinstruments]

@@vmiguel1988 i’m glad to hear it still being worked on! There’s almost nothing referring to it on Google. What’s with the big bird on his shirts and mugs? I thought that was the logo for the OSMU? That’s what it’s labeled on the website I think. It’s been advertised an awful lot in videos for years if that’s the logo… and it’s all over his merch...Just thought there might be some more details available I appreciate you getting back with me. I look forward to hearing any more news or updates.

[@robertw1871]

Looks great, for the cost of them, Keithley wants your firstborn, definitely a worthwhile project.

[@petr56321313]

Great project!

[@purplesim1364]

This is great, thank you for sharing.

[@user-my7bn4nc9m]

I'm watching you. That's amazing!

[@movax20h]

This is amazing. I was hunting for SMU, and was thinking about Keithley 2400 series, or maybe building one on my own once I am more proficient with design, but that your design just looks fabulous. The fact that it is open source, and can be easily hacked and programmed, could be a huge selling point for me. Sure, on Keithley you can use Lua (TSL) to do programming, but with open source SMU like this, I can do even more.

[@Alpine_flo92002]

I am in no way smart enough by any meaning of that word to even just grasp the concept of that thing...and you sir got me to learn. Cant wait to maybe replicate that thing...or atleast understand it

[@patrickmagee774]

Fantastic work!!

[@tubeDude48]

Very Impressive!

[@bfx8185]

Good job so far , Thx.

[@yngndrw.]

This is great, the higher current that you mentioned on your site would be great for measuring the IV curve of some of the larger LEDs. You the processor, you might want to consider an ESP32. They have 500kB of RAM, a dual core 240MHz processor and most of the Arduino libraries work on them. In addition to this they also have Wifi which is great if you want remote control or a web-based front-end.

[@aliens1192]

Very great job! I am going to build it! :D

[@t1d100]

A great project! A small matter = @~6minutes Clamp Voltage symbols are both "-.," instead of "+" and "-."

[@ericksonengineering7011]

Got it. Thanks

[@Spirit532]

I highly recommend trying to go with the STM32H7 series, using a generic 480x272 touch screen with a parallel(RGB565) connection, and LVGL for the graphics. One of my recent projects had an STM32H743ZIT6 with an ER-TFT043A3-3. The bringup was painless and easy(the code and hardware just handles all the display stuff for you), even for someone who _really_ doesn't enjoy writing code. I can share bits of code for prettier graphing on top of LVGL, since I needed that too.

[@VinisLab]

I guess a pretty UI is the last of his concerns. those Nexion displays are pretty handy and easy to interface with too.

[@movax20h]

Another huge benefit of STM32H7 would be that it is relatively easy to hook it up to Ethernet. But in a high speed and high precission design, it might be also be beneficial to have two micros, one for the UI, communications, loading calibration data, data logging, etc, and the other one for just the raw processing and time critical stuff. For example one might want to create fast controlled sweeps, or also limit the power, either due to the requirement of the measurement, or to extend the range of voltage and current in a more flexible way. Sure, it can be probably done with 4 quadrant multiplier and some comparators and extra DACs, but that is just an example.

[@NaudioElectronics]

Great project! Respect. Waiting for the polished DIY project :)

[@hullinstruments]

Same here. I check back regularly to see if he’s updated it. I’ve seen a lot of other people mention it and be interested in it in forums and such. The KZbinr “Marco reps” had planned and designed an open source measure unit several years ago… called the OSMU and with the silhouette of an emu/ostrich as the logo on all of the boards and enclosures. He designed everything and even had the boards and was in the beginning stages of testing it from what I understand. But he never picked it back up. He continued using that bird logo on all of his PCBs and they can pretty much be found on each one of his builds. His knowledge and command of electronics theory and implication are second to none. hopefully one day he will revisit it. He mentions it in comments and blog posts from time to time but since now he is spoiled with a lot of high-end equipment and precision metrology gear for a little or no cost from sponsors… He really doesn’t need an SMU now… But still shows interest and hopefully we’ll tackle it. He understands that most folks out here could really use something like that and don’t have sponsors casually tossing $20,000 meters and calibrators our way. His open source DIY 8.5 digit precision multimeter video and project are incredible. And I even had the boards made and parts purchased to build one of my own when I get the time. Incredible that engineers at CERN will release designs like that. And so thankful Marco made a well documented video discussing the build and the process.

[@paulwong6262]

厉害厉害

[@johnshaw359]

Could you use 2 identical boxes, one with the controls and one with the power like a pre-power amp? It might make upgrades easier with 2 sections.

[@ericksonengineering7011]

For a long time I thought virtual instruments were the way to go: no front panel, use a web page/PC/Tablet for the user interface. At Analogic T&M and Teradyne, that's pretty much all we built. Like PCI, PXI, VXI.... But most instruments need immediate controls, and the touch displays, CPUs, and knobs and buttons have gotten smaller and sooo much cheaper. Some $200 power supplies have color displays now. So for now, one box it is. Can always remove a UI later.

[@hullinstruments]

I feel dumb for asking this… I understand the importance of characterizing base components. But I was always a bit confused regarding smu's Most smu's I've seen only put out a very low amount of power usually under 20 Watts Obviously helpful for characterizing components but wouldn't it be more useful to have a source meter capable of higher powers? So that other pieces of test equipment and devices needing higher power could be tested and characterized? I feel really dumb asking this but any input would be greatly appreciated.

[@johannessturz8694]

It is a amazing circuit, but I don't understand the voltage sensing circuit with the opamp U23. There is no reference to ground. I think the circuit will be clamped to node R65(2). Higher input voltage R68(2) means higher load.

[@nicolasjonasson4820]

How's the plotter doing, and what is the sample rate of the ADC? Fantastic project! Is it still alive?

[@ericksonengineering7011]

Plotter basics are working now, but changing pages deletes the old data. That needs some work. You can plot externally in Python, that works very well. The sample rate of the ADC is a few samples per second currently. ADC speed is currently just a constant. Needs a SCPI command to set the ADC rate / NPLC. I'm always measuring both the force and measure, and toggling ADC channels slows the ADC down about 3x. I need a mode to select and display just the measure channel. I still need to make the ADC reads non-blocking and come up with a proper SCPI trigger scheme. There hasn't been a whole lot of interest, so improving the software hasn't been my top priority. Thanks, Dave

[@kunalgupta3353]

Can i have the gerber file for all the boards ??

[@cedrichirschi8036]

Hello David, I'm currently working on improving the current clamping circuit of my own more-like four quadrant PSU. Are you willing to share your LTSpice simulations with me? Cheers

[@ericksonengineering7011]

Sure, I just added the PSPICE simulation .ASC (schematic) file to the .zip file at: www.djerickson.com/diy_smu/files/DIY_SMU_Files.zip It simulates constant voltage with current clamping. You'll need to change it for Constant current. The amplifier and crossover models are pretty complete. The voltage - sense amplifier has its own floating power supplies and I simulated it as a behavioral model: ideal voltage buffer. Let me know if you need any other files, and good luck! Dave

[@cedrichirschi8036]

@@ericksonengineering7011 Thanks a lot! I wrote in your EEVBlog thread too, don't bother answering me there :D

[@kefler187]

10 nanoamp resolution on current ?!

[@ericksonengineering7011]

10nA is conservative. I'm aiming for 1 to 2 nA, 1/1000th of the 1uA current range. Operating this low will require slowing the ADC down to get better noise reduction, maybe with further averaging. And shielding.

[@francoisgrenier8530]

Hi, Where is the source code ?

[@ericksonengineering7011]

Thanks for the reminder. I just updated the .ZIP file with the latest design files: www.djerickson.com/diy_smu/files/DIY_SMU_Files.zip Schematics in .PDF and Diptrace .DCH format PCB source files in .DIP format Amp board Rev 3 Main Board Rev 2 Front and rear panels BOM in Excel .XLS format Current Arduino / Teensy / Nextion application code, up to and including functional SCPI The Nextion .HMI file Main simulation schematic in LTSpice .ASC format Enjoy! Dave