An ammeter is just a voltmeter in disguise.
Not if it's an analogue meter.
An ammeter is just a voltmeter in disguise.
Not if it's an analogue meter.
In that case it's the other way around.
My main concern with the ammeter version of the DPMs would be the stability of the current shunt resistor over temperature. Heating in the shunt resistor when fed lots 'o amps could be a significant source of error.
My main concern with the ammeter version of the DPMs would be the stability of the current shunt resistor over temperature. Heating in the shunt resistor when fed lots 'o amps could be a significant source of error.
I have a few of these 3A, 5 digit meters and they are excellent. The shunt does warm up at 3A and causes a few counts of error compared to my 34401A (but that has a pretty mediocre 12 month accuracy spec of .14% at 3A). I can't remember exactly how much it drifted but I'm pretty sure it remained well within the .5% + 2 digits accuracy specification.
[EDIT] Actually I did post my measurements later in that thread in reply #125:
https://www.eevblog.com/forum/projects/$5-voltmeter-with-5-digit-(0-1mv)-resolution/125/Maximum error relative to my uncalibrated 34401A was 0.2% (meter 1 at 2A) with most < .1%
Note there are a few variations of these 5-digit panel meters. Look at the pics before buying.
One uses a "22 bit" A/D SOIC-16 marked "BX5815" possibly a remarked Titan Micro TM7707 or 16-bit TM7705, delta-sigma.
The other variant another uses SOT23-6 for the A/D, possibly MCP3421.
The voltmeters come in 3-wire or 4-wire versions.
I think the ammeter's shunt looks suspicious with lots of copper colour, so I'm not sure how that alloy drifts with temperature.
That shunt looks about the same colour as you see in DMMs (although their high current reading do drift a bit), probably innocent until proved guilty.
I found
Manganin resistance wire is 84% copper which explains the colour. Constantan is also advertised, but chinese alloys I find all over the place for mix. Probably the tempCo would have to be measured and it is not a 4-wire part, so the accuracy might be limited.
For shunts Manganin is the better choice than Constantan: it is easier to solder and the thermal EMF better matches copper. Not having a dedicated 4 wire shunt is normal in the low cost section. It needs some care with the layout and may be a little sensitive to solder quality, but its not a real problem. The usual 4 digit handheld meters use very similar shunts.
It is interesting to see variations with different ADCs. It makes sense so save on the ADC in this price range - though the MCP3421 is already a rather cheap one.
As a followup here's some pics of the one I bought for $USD 4.80 including shipping from China to Australia.
Seems to work ok and the least significant digit does seem to usable.
The shunt has some very sharp 90 degree bends. Not sure if that is normal or desirable.