(NOTE: STILL EDITINGSEE BOTTOM OF PAGE FOR VALIDATION RESULTS (FOR SOME OF THE TESTS)
This page will go over the test plans for the solar sense board and all the unit/integration tests for each subsystem implemented. It will be written out based on what to populate, then subsequently what to test so we populate only the subsystems we need to.
...
Make sure you have the same parts populated as in the above SPI testing, as it includes the parts needed for the voltage sense.
Use the following space to record findings:
<insert test table>
Temperature sense
We take our own thermistor measurements from the panels for each section. If you have a spare 10k ntc, that would be ideal to connect, otherwise we can just use a resistor and make sure that the voltage divider and internal ADC work. Make sure you connect it from pin 1 and 4 on P15, and populate C42 and R46.
Use the following space to record findings:
...
LV isolation reverse bias
HV isolation reverse bias
Voltage drop across diode all cases
Thermals (shouldnt be anything significant)shouldn't be anything significant)
The following will outline my testing of the ideal diode:
I populated everything as instructed above
I first wanted to ensure the forward bias works. The following table outlines the drop across (everything without load btw) the diode based on the output and input. Note, the table only has forward bias testing, where no power supply was loaded onto the cathode.
Voltage on the anode (V) | Voltage on the cathode (V) | Drop across the diode (V) |
|---|---|---|
4.19 | 3.908 | 0.282 |
9.02 | 8.95 | 0.07 |
12.99 | 12.94 | 0.05 |
16.55 | 16.51 | 0.04 |
29.95 | 29.93 | 0.02 |
31.33 | 31.31 | 0.02 |
Now I want to test the reverse bias protection to make sure that when i put a supply on the cathode, it will not conduct to the anode. The following table summarizes my results, and note that once again there is no load.
Voltage on the cathode(V) | Voltage on the annode(V) |
|---|---|
2.256 | 0.172 |
9.68 | 0.360 |
19.95 | 0.475 |
29.99 | 0.571 |
31.36 | 0.583 |
31.38 | 0.592 |
The ideal diode works!! that was the smoothest validation so far → ideally id test with higher potentials, but considering that this testing was above the rating of the zener diodes already, it’s safe to say it operates as intended. Of course, double checking with a potential of over 100V would be ideal before integration testing (in which case I can finish filling out these tables).
The following will outline my testing of the relay and relay driver:
I have populated everything necessary and rerouted the relay as per the schematic issue of placing it low side instead of high side.
When I toggle a GPIO for enabling the DRV120APWR, I notice that if it is a low or a high value of GPIO, the contacts on the relay still close. Need to still investigate this, but it seems like the relay is always pulling the output low. This is also probably a fault with the rerouting because it seems to just turn on no matter the code and just when 12V is present. Also, the LED doesnt ever turn on.
I can also hear a high pitch hum as noise from the driver → might want to adjust the resistor/cap to adjust the PWM.
So turns out, I also forgot to move the flyback diode pins, and therefore there was always a potential across the diode since it was still connected to ground. What I ended up switching for fixing the relay configuration was putting pin 3 to the 12V rail, and the pin 4 to the output of the relay driver. I then put a flyback diode directly across the leads of the relay coil, and changes the values of Rhold and Ckeep (holding current and time for initial latching sequence). I found that in some combinations of high Ckeep that the relay would fully latch, but have a constant 100% PWM thereafter. In other cases, ckeep would be lower and I would increatse Rhold and here rapid clicking of the relay and NOT a constant PWM.
It seemed like there was something going wrong with the current control loop, and the IC was not behaving as expected. I had just replaced the IC, and did not think there was something wrong with it.
After troubleshooting for a while longer, I noticed that because of having to rework the board for the proper configuration of the relay coil, I had a bunch of fly wires soldered from the relay to the board. Since the relay coil feedback was part of the current loop that was regulating the power supply, this was destroying the signal and power integrity of the feedback. I removed all the wires, and pushed the pins on the relay slightly so they would just fit into the slots of the incorrect footprint (big oof). I then fixed the routing board side for the wrong nets by cutting one of the traces and jumpering it to 12V. When I tested the relay this time, it latched and then had a constant PWM on the output! It did not go crazy either (did not latch and unlatch rapidly) so the holding current (100kohm) and the time for latching (2.2uf) was good enough.
This relay driver troubleshooting was annoying but has a good takeaway (esp because imma have to validate a whole lot of other boards) → even when prototyping, trying to stay true to the layout is important, especially when dealing with a power supply (which now that I am saying out loud, sounds obvious, big oof).