openinverter forum

This is a handy procedure from Johannes Patreon video

I confirmed that the process works. As you can see from screenshots it is very handy for troubleshooting new "hand-soldered" boards. I am still missing NAND gate and some other components.

Test Firmware: https://github.com/jsphuebner/stm32-test

Wow, you've actually added it to the web interface! Is that running locally or on esp8266?

EDIT: never mind, the picture gives it away. So could you run it on the esp8266 with your SWD thingy?

I made some progress with SWD over ESP8266 ...promising but not yet. Just finished soldering v3, took 2 month

Wish I could get that web interface to work:)

Is it possible to make this test software a .bin file so it can be uploaded to the STM32 in the same way as the main software for consistency and help us non programmers?

Thanks

The picture in the post and the picture on github have different resistor values (220R vs 500R) wondering if the value of the resistor is important or if its just to prevent a short to ground? I want to make the testing circuit (judt finished soldering a version 3 as well and want to start testing bit I dont have any 220R or 500R reaistors on hand, but I do have plenty of 10K and 20K available in my parts bin

davidwagstaff wrote: ↑Fri Jun 05, 2020 5:40 pm Is it possible to make this test software a .bin file so it can be uploaded to the STM32 in the same way as the main software for consistency and help us non programmers?

Thanks

This seems like a good idea. The toolchain to compile this can be pretty picky to set up on some systems.
If the author of the software isn't going to provide a compiled release and has no objections, I'll post link to HEX / BIN files here.
Archive contains both .hex and .bin compiled binaries

Hi!
What is a Window Comparator and Inhibit Circuitry?

The 2901 comparator that checks the AC currents are within a given window (min/max) and the NAND tree that checks for that window comparator, the desat and the mprot signal to be all high.

Thanks for the tip. I checked the test board, everything is fine. The resistors correspond to the nominal values, all the necessary pins are connected.
The inputs of the comparator 2901, have the following readings:
5 - 0.26v
4, 6 - 0.82v
7 - 0.26v
8 - 0.26v
9, 11 - 2.46v
10 - 0.26v
accordingly, the output is 0.0v
Also on the 3 input NAND (PA3) - 0.0v
Processors are bought in China, maybe they are fake? Or do I misunderstand something?

5, 7, 8 and 10 must be above 0.82V and below 2.46V. Check your input resistor network.

If you apply 5V as indicated on the test board, then the pins P5, P3, P4, P8 of the test board will be 5V*(10k /(47k+10k))=0.88V. But when we installed the inverter, then we get the following diagram: The chart shows that at the input of the inverter, there is one more voltage divider and the lower shoulder of the divider will not be 10K, and (10K*(3.3 to+4,7 K)/(10K+(3.3 K+4.7 K)=4.4 K and, respectively, the voltage at point 1 (at the output of the test board) will be the same 5V*(4.4/(47k+4.4 k))=0.43 V and then consider the voltage divider on the inverter board. 0,43*(4.7/(3.3+4.7))=0.25V
This is the voltage from the dividers if the PA5 of the microcontroller is installed as an input.
If I swap R1 and R2, then PA5 should get 1.18 V.
Rlower shoulder=(47к*(3.3к+4,7к)/(47к+(3.3к+4.7к)=6.8к
U1=5V*(6.8/(10k+6.8к))=2.02в
Upa5=2,02*(4.7к/(3.3+4.7к))=1,18В
Do you think I'm thinking right or am I wrong somewhere?

Ah, so Dimas diagram is perhaps from the time when there was no on-board divider. Yes you are right, the values must be changed. In fact R2 should be no longer necessary.

Ok, i swaped resistors in the divider and now i have 1.2V at PA5, PB0 at the comparator inputs. Output of the comparator is 3.7V. But on the 3rd pin (PA3) NAND I have 0V. PA3 on the microcontroller in the test mode working as input or as output? If it's works as input, maybe i need connect SIG_MPROT at the external connector to 5V? It have internal divider on the invertor board.
I added a 510OHM resistor for SIG_MPROT, as on other pins, and now all the tests are passed. As a result, the following scheme of the test board turned out. Is that correct?