Vertical scaling defect on FNIRSI-5012H portable oscilloscope

[Tetris_BR2]

I have recently bought this portable oscilloscope because i am learning electronics and it would be very useful, sadly mine started this strange behavior while I was testing a psu PWM. I couldn't diagnose the defect because I am a beginner with basically no experience and no one knows how to repair it where I live, so I made a video of the problem (link below) and I hope you guys can help me, any help would be appreciated.
BTW, calibration only works in 10V/5V (1x) or 100V/50V (10x), in other scales I get a "Failed!" message.

https://youtu.be/pKvbly4wNl0

[ataradov]

This is a really strange failure mode. Horizontal scale affects the ADC sampling rate, and it is strange that this introduces a vertical offset.

Does this happen on different (all?) vertical scale settings?

I would at least disassemble it and see if everything looks fine visually. Take a few high res pictures.

[Tetris_BR2]

I couldn't upload the images to this topic so I did it on GDrive:

https://drive.google.com/drive/folders/1v6h4areFM4AiSdD4VdOs90GqRPwvrhSh?usp=sharing

[ataradov]

That ADC rework looks pretty bad. I'd at least use some alcohol and clean up the area around ADC.

[Tetris_BR2]

That ADC rework looks pretty bad. I'd at least use some alcohol and clean up the area around ADC.

Did it came like that from factory? I didn't touched it.

[ataradov]

Well, it is China. They do crazy stuff. They overclock the ADC like crazy, so it is possible that they assembled the device and it did not work. It is also possible that it had this issue and they detected it, and replaced the ADC, but never properly retested, and the ADC is not the problem.

This scope is not great in general.

[Tetris_BR2]

It was working normally while I was measuring the psu output voltage, so I don't understand how measuring 12V (max) could do any harm to this. Maybe changing the MCU solves the problem? Then I'd have to flash the new one, right?

[ataradov]

Was it plugged in into the power supply at the time of measurement?

It is not the MCU problem. There is also no official firmware binary available.

Here is a big topic on that scope https://www.eevblog.com/forum/projects/reverse-engineering-fnirsi-5012h/ There are schematics of the front-end, you can try to do some diagnosis based on that. Especially pay attention to the reference voltage level. It should remain constant when you switch time base.

[Tetris_BR2]

50s - 6ns = 1,648V measured from ground to probe connector;
1,8mV difference between OPA356 Out and +In;
6,3MΩ forward polarized "red" diodes / +200MΩ (couldn't measure) reverse polarized diodes resistance;
I couldn't find the two optocouplers LEDs in series, they all have one transistor to them;
156,5nF (brown) capacitor in parallel with the solid state relay and 0,370nF (white) capacitor in parallel with the 470kΩ resistor.

[ataradov]

So with no probe connected input voltage (==reference voltage) remains stable. Is OPA356 output (ADC input) also stable when you switch time scale? If it is stable, then it just looks like some ADC issue where its input is sampled differently depending on the clock.

[Tetris_BR2]

Measuring OPA356 Out pin to ground and changing Voltage scale:
10V = 1,55V
5V = 1,48V
2V = 1,31V
1V = 1,16V
500mV = 0,98V
200mV = 0,80V
100mV = 0,80V
50mV = 0,73V

[ataradov]

In your original video you changed time scale, not voltage scale. Or have I missed something?

[Tetris_BR2]

I changed both, first time scale to 500ms, then I changed V scale to all values getting strange readings (like -30V) only by changing V scale

[ataradov]

Ok, but why did you measure Vref over different time scales, but not voltage scales.

Your goal here is to figure out what changes on the outside when you make changes to whatever setting that is broken. In all cases, opamp output should remain about the same. And the difference may come wrong the wrong Vref, or damaged signal path to the opamp.

[Tetris_BR2]

You said it should remain constant while changing time base, so I changed Time base, but these are the values while changing V scale:

V scale x Vref:
10V = 1,645V
5V = 1,648V
2V = 1,601V
1V = 1,605V
500mV = 1,629V
200mV = 1,619V
100mV = 1,551V
50mV = 1,351V

[ataradov]

And the output of the opamp at the same time?

The idea here is that in order for the trace to remain stable, ADC input must be stable. Varying Vref would translate to the ADC input, unless this variability due to calibration that accounts for inaccuracies in the resistor dividers.

If ADC input changes  proportionally to Vref, then it appears that calibration parameters are incorrect. And I have no idea how to fix that with the stock firmware.

[Tetris_BR2]

And the output of the opamp at the same time?

V scale x Out Pin to ground V
10V = 1,55V
5V = 1,49V
2V = 1,33V
1V = 1,16V
500mV = 1V
200mV = 0,80V
100mV = 0,82V
50mV = 0,73V

If ADC input changes  proportionally to Vref, then it appears that calibration parameters are incorrect. And I have no idea how to fix that with the stock firmware.

It was working, but after the psu test, all went wrong. It is possible that the firmware has corrupted?

[Tetris_BR2]

Changing the Time scale, the Voltage changes too.

[ataradov]

It was working, but after the psu test, all went wrong. It is possible that the firmware has corrupted?

No idea, I don't know what stock firmware is doing or how it works. I ran it for 10 minutes to realize it was borderline unusable.

[ataradov]

Changing the Time scale, the Voltage changes too.

Does Vref change too? If it does change with time scale, then something is definitely screwed up.

[Tetris_BR2]

Does Vref change too? If it does change with time scale, then something is definitely screwed up.

No, 1,648V at 5V scale

[ataradov]

Ok, so Vref does not change with time scale? Does opamp output change with time scale?

If Vref stays the same and opamp output stays the same, then it is some sort of ADC issue.

[Tetris_BR2]

Does opamp output change with time scale?

OpAmp Out pin and ground Voltage is 1,475V to 1,485V while changing time scale. It isn't keeping a steady voltage even at the same scale.

[ataradov]

Minor variations are probably fine.

I don't understand though. How is it possible that Vref is changing, but ADC input is not. What happens to that changing voltage? I'd trace it out and look at it at various stages of the signal path.

[Tetris_BR2]

I don't understand though. How is it possible that Vref is changing, but ADC input is not.

Ok, so Vref does not change with time scale?

They don't change in Time scale, only in V scale. The trace on display only changes with the time scale in 500nS-200nS(7V), 200nS-100nS(4,6V) and remains the same in 100nS-6nS(4,6V), "calibrated" at 5V.