Hi again,
Any meter that uses digital averaging will see this same problem because it cant sync to the waveform. Could be the cheaper ones because that's what i was referring to, one Radio Shack and one Velleman.
I don't think that just
any meter that uses digital averaging will see this problem. The common technique is to use a crystal oscillator to derive a reference frequency close to the grid frequency. Maybe the cheapest ones don't do that, although it would seem that nowadays a crystal controlled oscillator isn't particularly expensive.
The cheapest DVM I own is also a Radio Shack. I created a test waveform by connecting a 20 volt DC supply in series with the secondary of a small transformer providing 25.2 VAC. I supplied the primary of the transformer with a variac. This allowed me to create a 20 volt DC voltage with various amounts of AC riding on the DC.
I've attached 3 images showing 3 of the 5 waveforms I measured. The reference level of the scope was offset 2 cm down so that all the waveforms would be on screen.
I started out with 20 VDC and added various AC voltages of zero, 20 V P-P, 40 V P-P, 60 V P-P, and 80 V P-P. Notice that for the last two, the instantaneous voltage goes negative.
I measured the AC+DC waveforms and got the following results with the Radio Shack on its DC range:
Waveform
1.) 19.92 VDC
2.) 19.91 VDC
3.) 19.88 VDC
4.) 19.78 VDC
5.) 19.45 VDC
There was no bouncing of the readings.
I also made the measurements with a Fluke 187 meter on its DC range.
1.) 20.014 VDC
2.) 20.013 VDC
3.) 20.013 VDC
4.) 20.014 VDC
5.) 20.014 VDC
The problem the Radio Shack is having is due to the limited dynamic range of its ADC. This is the same reason that for RMS measurements DVMs have a maximum crest factor for rated accuracy.