Maxcom MX 620 Multimeter with "odd behaviour" in "20A Range / AC"

[kblue]

Hello there!

As my first text got deleted pressing "preview" I'll try again

I recently repaired a Maxcom MX 620 Multimeter with missing digits (pins from main IC got detached) and wrong measurements (burned 900 Ohm SMD Resistor). Now everything seems fine, everything except the 20 A AC Range.

The results are always kinda low. 1 Amp AC will be reported as "0,85 Amps AC". AC mA or µA is okay, so is DC Amps, mA and µA.

Is this some kind of "design flaw"? Because in other pictures of the same multimeter I've found, they are using the usual shunt ... in my one it's a double-sided big PCB trace. I don't have any schematics nor a datasheets for this multimeter (and I know it's not the best, I just like to repair stuff). I try to understand what the exact problem is or am I chasing ghosts?

50 Hz sine
1,00 A | 0,85 A | 15%
2,00 A | 1,85 A | 7,5%
5,00 A | 4,85 A | 3%

400 Hz sine
1,00 A | 0,90 A | 10 %
2,00 A | 1,90 A | 5%
5,00 A | 4,90 A | 2%

I don't think it's the Quad OPamp, as it's inputs are used for other measurements too and they work quite well.

Thanks in advance for any hints and best wishes to ya all!

[Kleinstein]

The error looks odd, a bit like some "offset" in the AC range. Are the AC mA and V working OK at the low end ?

An odd point in the picture are the missing diodes and no visible diodes for protecting the current ranges.
The protection could still be on the under side.

Using a copper trace as a shunt is a bit odd and can cause erros.

[kblue]

Mayhaps the white picture with some measurements does not pop up/is not shown? AC mV AC mA and µA are not affected. That's what's boggling me ... it's only 20 A AC Range. Ohms okay, DC all ranges okay, AC all ranges ... except 20 A ... is okay.

It's driving me nuts

[coromonadalix]

you wont like me,  but you know  that's  a cheap meter ... they use the pcb as a shunt,  it's even cheaper,   with shunts  you had  slight possibility to add or remove some material to adjust it ...

now  you need to figure out  how to calibrate that thing,   could be some rebranded Hycon chip or else ?

or simply use  clamp meters ??   sorry  i would simply dump it,  you don't push  20 amps in such meters for that kind of design, even at 10 amps ...

i always use  external shunts and the meter is always in volts,  when you bust expensive fuses (they where at 11$/ea  now at 30-40$)   on better meters, you change the way you do things ... 

20 a  is a crude range to begin with,  for sure the lower ones could be more precise ?  never expect precision on this one,  such deviation  for sure is not good  BUT it's cheap once again

[kblue]

I know it's cheap and I don't intend to use it for anything serious. It's more about understanding why it's behaving like that ... my best bet would be AC and cheap PCB-Shunts just don't work quite well enough. I've found a TI-paper regarding PCB-Shunts and their advantages/disadvantages (like accuracy).

I just wonder why the same shunt, OPamp and so on work "ok" with ... say ... 1A DC but do not with 1A AC. So is it a design flaw? As said, other units have through-hole components and the usual shunt.

It would rather be a "sacrificial" meter. Its just about understanding why things are that way ... like my mentioned theory: PCB Shunts and frequencys dont go well and that's why it has no problems measuring DC Amps (20A range) but serious problem with AC ... or whatever. All lower ranges are SMD Resistors and no "PCB-Shunts", that could be the reason, why everything below the "20 A range" in AC works okay.

Regarding alignment ... it has 3 pots VR1 (Volts) VR2 (Amps) and VR3 (Farad). Anyway, I guess I'm chasing ghosts here and it's just a design flaw/cheapness.

Best wishes!

[Kleinstein]

It could be that the 20 A range uses less burden voltage than the lower current ranges. This would need more gain and for AC one has the problem that one has more noise BW and this way a kind of background from the noise. There is a chance that the meter tries to subtract this background in some way, and if this goes wrong (e.g. lower noise OP-amp than in the ref. design, or wrong cal. constant) the reading could be off.
Some DMMs have some issues with AC at a low fraction of the full scale.
Modern DMM chips sets may use digital RMS and this normally should work OK to low values, at least of the background is subtracted correct.

The numbers look like 150 mA too low. It may be worth also testing at some 500 mA and 200 mA. Similar one should test the other AC ranges also at the low end like 5 % of FS and below.

[kblue]

At first, the usual AC Volts ranges (always 50 Hz sine):
[Range]
[other meter] // [MAXCOM]

200 mV
10,0 mV // 10,0 mV
50,0 mV // 50,0 mV
100,0 mV // 100,0 mV
180,0 mV // 180,2 mV

2 V
0,100 V // 0,097 V
0,500 V // 0,497 V
1,000 V // 0,997 V
1,800 V // 1,797 V

20 V
1,00 V // 0,97 V
5,00 V // 4,98 V
10,00 V // 9,99 V
18,00 V // 18,00 V

200 V
10,0 V // 9,7 V
50,0 V // 49,9 V
100,0 V // 100,0 V
180,0 V // 180,2 V

Now the AC Amps ranges (still 50 Hz sine):
[Range]
[other meter] // [MAXCOM]

20 µA
1,00 µA // 0,95 µA
5,00 µA // 4,95 µA
10,00 µA // 9,95 µA
18,00 µA // 17,95 µA

200 µA
10,0 µA // 9,7 µA
50,0 µA // 49,7 µA
100,0 µA // 99,8 µA
180,0 µA // 179,8 µA

2 mA
0,100 mA // 0,098 mA
0,500 mA // 0,498 mA
1,000 mA // 0,998 mA
1,800 mA // 1,802 mA

20 mA
1,00 mA // 0,98 mA
5,00 mA // 4,98 mA
10,00 mA // 9,99 mA
18,00 mA // 18,00 mA

200 mA
10,0 mA // 9,7 mA
50,0 mA // 49,7 mA
100,0 mA // 99,7 mA
180,0 mA // 179,9 mA

20 A
0,10 A // 0,02 A
0,20 A // 0,06 A
0,50 A // 0,33 A
1,00 A // 0,83 A
2,00 A // 1,83 A
5,00 A // 4,80 A

This time i did some "sanity checks" against other, older meters without true-rms (which sh.ouldn't matter, right?) and with some Fluke handheld DMMs. All of them agree with each other with the exception of the Maxcom meter. As seen above, Volts is alright compared with other older meters (with lower counts). Everything BELOW the "20 A" range seems fine too.

As for the amps measurement, i used two meters in series ... mayhaps thats not "ideal"? Even though if i use any other meter with another one in series for amps, both of them always agree with each other. My source for everything is a restored Fluke 760A ... i know it's old as hell.

/edit: It does not matter if i have the MX-620 in series with another meter for comparison or standalone. If i set my Fluke 760A to 1A Output AC 50 Hz and the "pointer" reaches 10 ("search") on the analog scale i get around 1,000X AAC 50 Hz with my Fluke Meter. So far, everything is alright. If i couple the MX-620 alone to the 760A and reach my dialed-in output, it is still wrong and shows the above mentioned value for 1A (0,83 or 0,84).

Best wishes!

[Kleinstein]

The other rangs look better, though they still have a bit low current readings when at low currents. There is the posibilty that the 20 A range is actually internaly from the voltage at the shunt more similar to a 100 A range and just scaled up.
The result for 0.2 A and 0.5 A in the 20 A range are really bad - still too low by about a constant value.

It may make sense to look at 1 mA, 2 mA and 5 mA in the 200 mA range.
It is nice to know at what point the AC ranges fail with small readings. Not just for the 20 A problem, but in general for low readings.

2 meters in series in the amps more is perfectly fine, not issue there at low frequency.

Are there parts on the lower side of the PCB ? 

[kblue]

Ye, they are bad i know

I wonder if the "other designs" i saw online from the very exact meter with "usual shunt" and couple more diodes (but everything through-hole) are earlier or later revisions. Don't know Still everyone of them is a kinda cheap build, but this behavior is interesting and it's always good to know "why something behaves as it does" for other repairs or oddities.

Anyway, here are some more values from the 200 mA range:

50 hz sine
source // maxcom

1,00 mA // 0,7 mA
2,00 mA // 1,7 mA
3,00 mA // 2,7 mA
4,00 mA // 3,7 mA
... and so on, it's always x,7 mA ... see posting before. Until close to the end of the selected 200 mA range. Here are some more "high-res" pictures! Oh and please don't mind "R2" ... the two SMD resistors making some sort of pyramid. I don't have a 900 ohm SMD resistor, so i selected two 1% parts which are in sum 900,0XX ohms (actually measured). R2 was the "burned" one from the repair so far and as said ... everything works rather "well" for an instrument with whatever %-tolerance. Except the very low 20 A range (which should never be used for 20A).

Thanks once again!

[Kleinstein]

I don't think the shunt would make a difference when it comes to the AC problem. For the reasone behind the fail I have to likely candidates:
1) A wrong calibration constant in the form of some "offset" to subtract from the AC readings. This would especially be the case with digital RMS to measure AC, like some chips sets support.
2) If there is a analog rectifier it is also possible to have an offset problem, like an OP-amp with rather high offset, so that the rectification is not working well for low voltages. The copper shunt may need an additional gain stage to compensate for the TC (ideally with a very thin copper trace close to the shunt, that I don't see). For this reason only the 20 A range may be effected.


It looks like the 4 missing diodes are for the protection of the lower current ranges. I would consider adding them (like some 1N400x or similar SMD ones).

Having 4 diodes for protection would make sense, but the layout does not look like this. 2 seem to be for protection as the transistor test.

[kblue]

Some more observations ...

Every "amp" input, doesn't matter if ac or dc switch is engaged, connects through a 110k resistor to pin 5 (none-inv) of the NJM062D. If i measure the "output" with my meter while in any mA or µA range (AC), the output voltage represents what's shown on display. Ye, i know it's not the best way to measure but if i have ... lets say ... "65,0 mA shown in display while in 200 mA (AC) mode, the output pin 7 of this IC NJM062D has something like "650+/- mV" output (which will rise to the correct value, once i remove my meter ... so far, so okay). The output pin 7 is linked to the negative input from a rubycon cap (4,7 / 50), while the positive side is linked to two MELF-diodes acting like a rectifier and some more resistors.

Once i switch into the 20A range, the output is rather low for lets say ... "1 Amp AC in" i have around 244 mV from the JFET showing "0,84 A" in the display.

Anyway, thanks so far!

I tried to make the "20A path" a bit more visible.

[kblue]

I guess i'll take a break from this, as it seems (from first glance) bit more complicated. I don't think it's the cheap pcb shunt and i don't think the problem is in the dmm chip section but thats just a "feeling" and feelings can be wrong big way ... but because everything else works. Everything except 20A range and only in AC mode, not DC. I would guess it must be something "in between" the line. But mayhaps it is the DMM Chip itself and whatever is "set internally" what i can't change.

Tried to make up "sections" and what i would suspect, even if everything seems okay.

But oh well ... i'll repair some other meters first, before i go back to this "project" as it is just for fun.

Many thanks once again Kleinstein and best wishes!