Sensor not reading test voltage [SOLVED]

[Garagehermit]

In my blind eagerness of getting inverter assembled and tested, i realized resistor R5 was missing from voltage sense board. I didnt notice until after running voltage configuration test with 320v connected. JOHU had doubts a missing R5 would destroy the board.
With R5 installed on volt sense board, the opamp output (JP2 pin 2) reads 4.97 regardless of test voltage (0...72v...330V etc) which might explain why i got the following values during configuration:
set udcofs 0
get udc
4092
Turns out 4092 is the max offset. Regardless of what i set for udcofs, the read voltage never changes with varying test voltages applied. This was still the case before I added the R5 resistor. Without R5 I think i measured 4.67V on opamp output. I would need to remove R5 to verify. But i still have nightmares of what i did to PaulHolmes' inverter board from several rounds of desoldering. I will probe around the voltage sensor board to look for irregularities, but I wouldnt know what i am looking for since opamps and general EE mystifies me.

[Garagehermit]

With 78v test voltage applied, measured 3.77V between isolator/Si8920 (IC1) pins 8 & 7. With no test voltage applied, measured 3.89V between isolator pins 8 & 7.

I measured 95.8 ohm across R4, a 10k resistor, which connects to opamp pins 2&3. When i removed R4, i measured 96.9 ohm across opamp pin 2 & 3, which are power supply, Vcc, and non-inverting input. I'm getting ready to dig through my cache of old pcb boards to test similar opamps to see if that resistance is 'normal'. Im going to keep playing this random troubleshooting game until someone with actual intelligence drops some knowledge .


Next day: put some solder paste on opamp, hit it with hot air and some small nugget shot out from under it like a cockroach. Unfortunately i had already bent the ground pin and broke it off in my usual brute fashion. I put in a different opamp and getting a reasonable udcofs value. But still not getting udc to recognize a test voltage. Hopefully ordering the correct opamp will sort it out.

[johu]

Yes the opamp is the R version with funny pinout.
Please check voltage between 6 and 7 not 7 and 8 if you continue to have issues

[Garagehermit]

I replaced the opamp (with mouser part no. 595-OP344NA/3K) and measured voltage between Si8920 pin 6 & 7 was .035v and 0.268v with 0 test voltage and 78V test voltage, respectively. J2 pin 2 is reading 4.97V always. I am not testing shorts on any opamp pins. I will use rework station to remove opamp, cleanup and try again.
((Some few hours of interrupted sleep later))
Confirmed the pinout for the OP344 i ordered is wrong for the board...pin2 & 5 are GND & Vcc , opposite of board layout. Lesson learned, no more reading spec sheets half asleep at 1am.

[johu]

Yeah sorry using an SMT part might have been a silly idea.
Output voltage of Si part sounds good.

[Garagehermit]

Finally able to get the voltage sensor board to recognize test voltages. My new problem: udc decreases when I increase test voltage. Is there an easy fix? Swap resistors R7 and R8 for potentiometers?

With udcgain=1, udcofs set to zero-out the udc reading for test voltage disconnected, then when i apply a 91V test voltage, udc reads negative 3V. So i would need to set udcgain = (-0.025), which is not allowed and seems too low even if it was +0.025. I' m still getting good readings between Si8920 pins 7(+probe) & 6(-probe), 305mV.
The opamp on the board was originally installed with Vcc and GND pins reversed. Would that ruin the op amp? Do i need to replace the opamp to see if problem persists or are there other things to try? Would software edit fix this issue?

I put a rough estimation of the si8920 output in online spice model and got a similar trend for the opamp output. Im playing around to see what i can try next.

[johu]

3 digits at 91V is waaay too low of a change. I'm guessing your OpAmp is toast.

[Garagehermit]

New opamp took care of it. Voltage sensor is reading test voltage accurately. Thank you Johannes!