# Why does an AC Voltage reading on a DMM decrease as the frequency is increased?

I'm doing an assignment for a lab and I've found that frequency doesn't affect the voltage reading of an oscilloscope while it does for a DMM. Online, it says this is because an oscilloscope displays peak voltage instead of RMS voltage like the DMM. But even then I don't understand why RMS voltage reading would decrease with higher frequency on a DMM. From my understanding, finding the voltage RMS on a sine function would be to divide the peak voltage by root square of 2. So if the frequency increases, then theoretically the DMM reading should stay the same since the peak voltage doesn't change. Can someone explain to me how this works? Thank you.

• The AC Volts accuracy of my Fluke 111 is specified from 50 to 500 Hz. You are using the DMM outside its designed range. I don't know why they chose to limit it to 500 Hz. Sep 23, 2023 at 20:57
• For extra credit, borrow an RF signal generator from the lab that goes to ~1 GHz and repeat the experiment with the oscilloscope - you'll see the exact same kind of roll-off on the oscilloscope. Sep 24, 2023 at 13:01

The oscilliscope is designed to display a wide range of frequencies accurately - from DC to 50 MHz or more, while a common DMM will usually only cover a much smaller frequency range - mine shows the AC accuracy as +/-1.5% from 50 Hz to 500 Hz.

There are specialized meters that will cover wider frequency ranges.

On a cheap meter the signal is AC-coupled, then rectified using a precision rectifier and then low-pass filtered yielding an average voltage, which is then (for display) (effectively) multiplied by ~1.11 = ($$\\frac{\pi}{2\sqrt{2}})\$$.

The precision rectifier will have a bandwidth associated with it so you'd expect it to roll off at the high end. Yours appears to show some peaking, which is also not too surprising. It should also roll off at the low end for most meters (due to the AC coupling), well below mains frequency.

Most such meters are designed to measure mains frequency reasonably accurately (given sine wave input) and perhaps up to aircraft/military 400Hz (because it's easy) and not much else.

For example, a typical (and decent, but for low voltages/currents only due to low-cost fusing) hobby meter is the Unit-T M890G which has specified performance:

Frequency Response: 40Hz - 200Hz (40Hz-100Hz for the 200V and 700V range).

A more expensive meter such as the 34401A has true-RMS and is reasonably accurate up to 100kHz, which allows accurate measurement of lower-frequency periodic signals that are far from sinusoidal (such as audio) as well as high-frequency sinusoidal signals.

• My Fluke 111 true rms meter is limited to 500 Hz, so there are other reasons. Maybe a filter before a sampling circuit. Sep 23, 2023 at 21:10
• @Mattman944 Maybe some lower cost way of doing the true-RMS than HP/Agilent/Keysight used. Sep 23, 2023 at 21:12
• @Mattman944 The 500 Hz is probably just to ensure that it can work on the full range of common line frequencies (50 Hz, 60 Hz, and 400 Hz are the most common). They presumably thought it wasn't necessary to go any further, so they didn't spend the extra effort designing and specifying it for a wider bandwidth. Sep 23, 2023 at 23:30
• @SpehroPefhany FWIW I believe the peaking shown in OP plot is an artifact of the excel spline fit. The points look to be centered on the "bins", and if this is true both the 500 Hz and 1 kHz look to be right on the money Sep 24, 2023 at 8:15