Measuring frequency response of an amp with a multimeter?

What device do actual techs use to measure the frequency response of an amp? 20Hz-20khz.
Can you play a 20Hz- 20khz sweep from something like a WAV file or CD and monitor the voltage of the amp's speaker jacks and note any deviations with a multimeter? This would be assuming your source of the sweep has a "flat" output, which I suppose you could measure/monitor before running it through the amp. Would you measure with the amplitude of the source turned up to like 90%. Just thinking 100% amplitude could introduce clipping. Thanks.
 
Administrator
Joined 2007
Paid Member
Yes, you can do exactly that. The quick test would be using a scope. You can use a meter provided you first prove the meter responds correctly to all relevant frequencies.

Its really easy to make an MP3 or WAV frequency sweep file which you can play using anything suitable or you can create a test CDR. If you need to know how to do that then just ask :)

The signal must be a sine wave and normally we test at low levels such as a 1 watt equivalent which in practice means an output of about 3 volts rms. The meter reads rms values, the scope displays peak to peak and would show 8.5 volts from top to bottom of the sine.

Testing at full output or even close to it requires care. I would not advise you to attempt that at this point,
 
  • Like
Reactions: 1 user
A Behringer UAC222 audio interface cost £25 from Amazon and download the very good and free REW.
Apart from some cables, that's all you need.

https://www.roomeqwizard.com/

The AC input even on true rms multimeters is no use because above about 1Khz, the reading will fall giving mislleading results.
I have room EQ wizard. How do you use that to measure electrical output of the amp? I use it and a umik1 to do acoustical measurements.
 
Few multimeters have anywhere the accuracy to measure frequency response even to 20kHz.
Any decent scope with a properly calibrated probe would be much better.

I only have a "pocket" oscilloscope ...
AUKUYEE Updated 2.4" TFT Digital Oscilloscope Kit with Power Supply and BNC-Clip Cable Probe Q15001 (Assembled Finished Machine) https://a.co/d/3votiyV
So I keep an eye on the voltage reading on the oscilloscope as the sweep plays rather than my earlier plan of monitoring that same reading on the fluke multimeter?
 
The most basic Windows laptop or desktop machine with audio inputs and output will do the job better than any meter or scope. The scope would only be better for looking above 20,000 Hz bandwidth of most sound cars. The manual for Arta has schematics for resistive dividers and diagrams and instructions for how to make measurements. You calibrate your interface by connecting the input to output and recording a sweep. That result can be loaded as a calibration to ensure a flat response. Arta can be used for free. All the analysis and graphing work in the free version. You can save the plots to a document. If you wish to store the impulse responses you pay for the software. https://artalabs.hr/
 
Administrator
Joined 2007
Paid Member
I only have a "pocket" oscilloscope ...
AUKUYEE Updated 2.4" TFT Digital Oscilloscope Kit with Power Supply and BNC-Clip Cable Probe Q15001 (Assembled Finished Machine) https://a.co/d/3votiyV
So I keep an eye on the voltage reading on the oscilloscope as the sweep plays rather than my earlier plan of monitoring that same reading on the fluke multimeter?
That should work over the audio band. What you need to do is first check the output of the source component playing your sweep and make sure it is constant. Providing it is then you can do a sweep of the amplifier output but keep the level reasonable.
 
  • Like
Reactions: 1 user
Most small DMMs are worthless for high frequencies, but some bench DMMs are quite good. I like the old Fluke 8050A because it's good to 50 kHz and has a dB range. That means you can do a sweep and read the response directly in dB. Or volts, if that's what floats your boat. The common HP 3478A is also good, with a BW of over 100 kHz. Sound cards are great but there are pitfalls. I like to verify what they're telling me with standard equipment before trusting them.
 
Moderator
Joined 2011
So I keep an eye on the voltage reading on the oscilloscope as the sweep plays rather than my earlier plan of monitoring that same reading on the fluke multimeter?
Don't use a sweep, just a single frequency.
Compare the 1kHz reading with the 20kHz reading (both on same scale with same probe).
You can do the same for the 20Hz rolloff.

Then the 20kHz rolloff in dB is: 20 x log ( V20kHz / V1kHz )
This will be negative, reflecting the drop off in amplitude.
 
Don't use a sweep, just a single frequency.
Compare the 1kHz reading with the 20kHz reading (both on same scale with same probe).
You can do the same for the 20Hz rolloff.

Then the 20kHz rolloff in dB is: 20 x log ( V20kHz / V1kHz )
This will be negative, reflecting the drop off in amplitude.
So if it were to measure the same at 20hz, 1khz and 20khz, we'd say it was flat through the entire audio range without having to measure the entire range? Thanks!
 
Moderator
Joined 2011
Yes.

A power amplifier would be a real junker if it had significant aberrations between the extremes,
with a resistive load. Don't even know how that could happen, aside from tuning/stability problems.

Integrated amps should have tone/filters defeated for such a test.
Phono stages are more complex to measure.

Some quirky tube power amps can indeed have significant midband aberrations, with a reactive load,
due to higher output impedance. Some have underdamped HF response as well.
 
Last edited:
The usual limit for a good true rms dvm on ac volts is ~500Hz. Some sources suggest using that frequency for calibrating a normal audio interface set up for say measuring SPL.

An AC analogue milivolt meter could easily do the entire audio range.
Oh wow. That's interesting. I didn't think the old analog meters had any use anymore. I've probably thrown more than one of them away over the years as technology progressed. I should be able to grab one of those fairly cheaply 👍🏼
 
Yes.

A power amplifier would be a real junker if it had significant aberrations between the extremes,
with a resistive load. Don't even know how that could happen, aside from tuning/stability problems.

Integrated amps should have tone/filters defeated for such a test.
Phono stages are more complex to measure.

Some quirky tube power amps can indeed have significant midband aberrations, with a reactive load,
due to higher output impedance. Some have underdamped HF response as well.

Gotcha. I see solid state amp reviews with measurements of the frequency range. I just assumed they did vary from "flat" which is why they included the graphs.
The vintage amp I'm particularly interested in measuring at the moment is solid state but is like 30 yrs old. No tone controls. Just amp.
 
Moderator
Joined 2011
Yes, just perform the check at 1W output, or 2.8 Vrms into an 8 ohm load.
That is equivalent to 8 V peak to peak, which is easier to measure on a small scope.

Resist the temptation to crank it up to clipping, since the amp may not be in shape for that yet.
Which amplifier is this?
 
  • Like
Reactions: 1 user