true RMS meter vs. oscilloscope

[Robert Nichols]

In article <KsGdndg2Jf3tA7DeRVn-gg@comcast.com>,
<normanstrong@comcast.net> wrote:
:
:Which waveform has the greatest discrepancy between average and rms values?

Presuming that you're thinking of unipolar waveforms, that would be a
train of narrow rectangular pulses rising above a zero baseline. The
ratio of average to RMS is the square root of the duty cycle, so the
ratio Vrms/Vavg goes toward infinity as the duty cycle approaches zero.

--
Bob Nichols AT comcast.net I am "rnichols42"

 

[Robert Baer]

mike wrote:

Robert Baer wrote:

mike wrote:

Robert Baer wrote:

mike wrote:

Robert Baer wrote:

phillip.liu@fphcare.co.nz wrote:

Hi, I am trying to measure the rms voltage for a slightly
distorted AC
sine waveform. I used a "True RMS" multimeter as well as an
oscilloscope. Expecting both rms readings to be the same, but
different. Any idea why? Thanks.

Using the scope, one can only measure the peak or peak-to-peak
voltage, then convert to RMS.
Not any different than the way AC scales on most DVMs are
implimented.






I must be confused...cause he said he used a "True RMS" meter. A
True RMS meter, by definition, measures the RMS value of the
waveform. If the waveform is only "slightly distorted", a TRUE RMS
meter should measure within the specs of the meter.

I don't know about your scope, but mine has a button that
calculates the
RMS value of the waveform. I didn't bother to look up the accuracy
spec.

It would be interesting to know the exact equipment used and the
magnitude of the discrepancy.

One common source of such problems is DC offset of the AC waveform.

mike

You are fortunate to have the "read in RMS" option on your scope.
But, i bet that it works by measuring peak or Peak-to-peak and
converting that to RMS.




OK, what are you betting? Send it to me.
It's a TEK TDS540. Suppose one could look up the spec. It can do FFT;
it's hard to imagine they'd not do a proper RMS calculation.
mike

NOW i know something i did not previously: what make and model of
the scope.
Since it digitizes, then it could calculate energy of each sample
and thus the equivalent RMS value.
If one waveform was converted this way, and if there were (say) 8
bits per sample, then about 3 bits would be used for the + half and 3
bits for the - half and one for zero; this is a bit (no pun inended)
better than measuring a peak (or Pk-Pk) and converting.
Higher bit resolution would tend toward better accuracy for complex
waveforms.
Still, no where as good as using the thermal method...


This is going from funny to sad. Half of 8 bits is 7 bits.

Nowhere as good as the thermal method at the NIST...but probably on par
with a typical cheap commercial thermal rms converter.
mike

The best "cheap commercial RMS converters" are ICs that do a superb

job, and i think some get fairly close to what NIST does.