UT139C AC range jumping all over

[standal]

Hi,
I have UT139C multimeter, and after 2 years is not working properly anymore. These are the symptoms.
1) all AC ranges are jumping all over. For example volt measuring plus minus 100V, mV +- 100mV, current the same.
2) NFC constantly beeping
3) mV DC range, drifting from 25mV to minus infinity.
4) clamp meter range also jumping as AC ranges, but also DC range
5) in mV AC range is measurable 50Hz (50% duty cycle) frequency without using probes if near power line, if far away from power line, than it jumps from 30 to 80Hz (duty cycle also jumping)
It looks like, that multimeter is somehow adding amplified signals from NFC antenna to all these ranges.
Do you have any idea what is wrong with that multimeter?
Because now it is ready to go to garbage. It is only DC A/V, temperature, and resistance range working.
I bought the multimeter in germany in conrad, so it should not be some not approoved piece from chinese seller on aliexpress.
Thanks for any ideas
standal

[standal]

Oh god, I found this.
https://www.eevblog.com/forum/testgear/uni-t-ut139c-multimeter-some-functions-stopped-working/
So it is a piece of crap.
For others, never buy anything from this shitty UNI-T company! They produce crap.

[J-R]

At least take it apart and check the rotary switch and contacts?  Post some pictures.

[standal]

Before posting I already took it apart. Rotary switch looks ok. I measured all caps if they are not shortet. All resistors if they have approx the value which is written on them. Diodes also. Transistors only with diode check, but some are so connected, that it is not possible to check everything. From smd code is also not clear if it is fet or bjt, but transistors with same smd code behave under diode test the same.
I measured with the scope on uC pin 34 and 35 (or 36 and 37 not sure anymore) because I hope I tracked it well that there is NCV and current clamp range connected, but did not see anything. Without schematics to track mA range or mV range to uC is almost imposible (it goes too much over rotary switch).
I just do not understand, how is it possible that DC range works and AC is jumping. It uses the same input on uC or not?

[Kleinstein]

There are different possible reasons for such a malfunction. One is that some part of the circuit (e.g. amplifier or voltage regulator) is oscillating and this way causes superimposed AC in many places. The reason could be a bad capacitor / solder joint and a marginal design.
Another plausible reason could be an open input (e.g. cold solder joint or broken resistor / open trace) in the AC ranges. The internal amplifier is likely very high impedance, and if it for some reason looses the connection, it can pic up all kinds of hum and parts of the signal.
A bad switch contract can also cause an open input.

For the AC RMS function, there are 2 possible implementations:
1) with an analog RMS function (in the main chip or a separate chip like AD8436). This signal than goes to the same ADC input like DC.
2) with a seprate faster ADC input and using digital RMS
So the AC function may use a largely different path.
The UT139 has a rather limited bandwidth for the AC function and this points to the 2nd option.

A first test would be to see if the batteries are still OK. Old classic zink carbon can be an issue and not detected by the low bat symbol. For some strange reason they are sometimes still shiped with new electronics. 

[standal]

Hi,
amplifier and voltage regulator is integrated in uC. On the board is only uC and eeprom.
Before posting I already resoldered all components, so hopefully no cold joint.
Caps are not short, resistors have right value.
I would say it is the first possibility, there is only main chip.
I tried new heavy duty (znc) batteries and cleaned contacts and the same behaviour.
Here photos from both sides.
I hate this multi now, even though I liked him before
PS:
I just remembered now, I have seen garbage swiching pulses on the ncv antenna from switching display segments. But I do not think that it was not there before, becouse the antenna is just next to the caps from display.

[rsjsouza]

Since others had the same issue, I can't help but wonder if the PCB developed minute cracks on it. That or residual flux or the Lead Free solder started to develop problems.

Since you have nothing to lose, I would give the board a bath with Isopropyl Alcohol:
- Remove the buzzer;
- Put enough alcohol to wet the board just above the top layer (so you don't risk ruining anything else by soaking them)
- Leave for 5~10 minutes
- Using a static-free brush or a lint-free paper, scrub the board
- Wash carefully on water, avoiding soaking the parts

Apart from that, I would start suspecting the EEPROM with calibration constants somehow started to decay and miss its data. Obviously that wouldn't explain the NCV behaving erratically, but it might explain the sudden loss of stability.

Good luck!

[Kleinstein]

I would not expect a bad EEPROM to cause unstable readings. With a seriell chip the content is usually read once on power on.

Some problem with flux or maybe tin wiskers / solder balls is possible. So a good clean around the main chip could be a thing to try in desparation.

If something is coupling from the NFC part / LCD display to the AC input, one should see this trouble also at the input in the mV AC range.
There would also be a difference between an open and shorted input.

To a lesser degree one may get AC problems if there is a shield that is no longer connected.
However normally the AC current ranges, especially the higher current part should be relative insensitive to a coupled signal.

[standal]

The flux residuum is from resoldering by me. After that, nothing changed so I did not bothered to propper clean it. But it is only colophonium, so it does not eat tracks.
Before posting I also made a backup of eeprom, and yes it is read only once on the starting of multimeter. After backup I tried to flash three other backups found here on forum, but nothing changed. I thought, that it could be some bit switched in eeprom, but if all other three eeprom backups are the same behaviour I do not think anymore.
Here some pictures of the garbage (blue) on the nfc antenna in different ranges. Red is some driving signal, which match to that garbage and goes to display.
I also made the video so you can see how it behaves.
https://youtu.be/LpvBksPUUPY
In the meter is of course no shield. In the upper ranges behaves the same, it looks like it added calculated value to measured value in the uC. Because in V range it jumps about 100V. Interesting is that on A range only the more precise range jumps, the less one jumps after while to zero. But at V range there is not such a thing, there it jumps in all manual ranges.
I hope that this topic helps others, and warns them to not to buy anything from UNI-T.

PS: Now I have seen something. There has to be some leakage from buttons. If I press and hold any button, than the value jumps much more higher.

[mqsaharan]

Hi standal,
I am sorry that your meter is not working normally. But before you throw it try a few things. And I would suggest you to still keep it even if it doesn't work; you can repurpose some of its parts.
The good thing with this meter is that it is based on a commercially available all in one chip so you can try to use its datasheet for troubleshooting. Because it will be using something similar to the proposed design from the datasheet.
I don't own this meter so what I am writing here is from the pictures and the datasheet that I have found on the internet. So, there could be some mistake.

Try the following:
Remove C7 (its under the display, below the main IC) and see if meter's behavior changes. It looks like a bypass to ground cap in the main V (input socket) line. If it is leaking it could cause such behavior, I assume.
If that doesn't help, start disconnecting the main IC from the outside world. This will help you determine if it is OK or not.
As a first step, remove R20 (820 ohm, I guess), it is right above C7 and see if it makes any difference with NCV, Hz and Cap. It is the input to main chip for these functions.
If current ranges are giving any problem, remove C3 and check again. After removing C3, remove L2 to disconnect that input completely.
C4 and L4 are also input to the main chip for R function, I guess. It works along with R20 line.
Then as last resort, you can remove either any one of R1a to R1d resistors, or L5 inductor (along with C7). It will disconnect the V input to the meter.

I hope it'll help.

[standal]

Hi mqsaharan,
it is on interesting way, thanks for the datasheet.
C7->mV worse, C7 not possible to measure in component tester (open?)
R20->mV xx.xx range stops jumping but xxx.x range is worse, temp jumps but before that was stable 29degree, R is 680ohm
C3->A range stops jumping, C3 measured as 21pF
Do you know what values schould be the C7 and C3? Because they are not in the schematics.

[mqsaharan]

Wow, that was quick.
I am sorry. As I wrote before I don't own this meter. I took the guesses from the datasheet and pictures of its circuit board. I don't know these capacitors actual values.
I used schematic on page 17 for my last reply.
After watching your video and reading your post again, I have understood that it is throwing garbage reading in almost all functions when you select AC. I think I didn't realize it before.
As per the datasheet, what I have understood, the AC function does not use external circuitry (page 22). If it has gone bad, there isn't much we can do.
I don't know if this chip is available somewhere.

As per your first message, if DC V, A and resistance functions are working, then you can use it for these functions only. But I guess it is not trustable anymore.