Measuring electronics (e.g., amplifiers) with REW

[ElArte]

This topic is gaining popularity, yet there is still a fair bit of uncertainty out there, so I thought I would seek clarifications.

First, the set-up:

1) Akitika 1 kHz oscillator as the source, connected to the amplifier using a single RCA cable. The other amplifier input is shorted.

2) A cable from the amplifier output to the 3-pin XLR input 1 of a Scarlett 2i2 audio interface.

There is no volume knob leading into the amplifier.

There are 3 knobs on the 2i2: 1) A big knob marked "Monitor", 2) a gain knob for input 1, 3) and a gain knob for input 2.

I calibrated the audio interface in REW and the preferences pane looks as follows:

During the calibration process, I had to turn up the "Monitor" knob so that input level would match output level, as indicated by the Out/In gauges.

Does that mean the "Monitor" knob should be left at the same position for the calibration to remain valid?

Now, in the RTA window,

Should there be an extra calibration in place for the input?

and

Should I uncheck "Get fundamental from generator"? Since the Akitika oscillator generates the signal instead of the REW generator?

Should "Ref resistance" be adjusted to match the value of the dummy load?

Finally, here, how to use this "FS sine Vrms" and the associated "Calibrate level" button, when the source of the signal is the Akitika oscillator?

Some people say it should be the Vrms value being output by the Akitika oscillator, some people say it should be the Vrms value being output by the amplifier itself, across the dummy load, before attenuation.

Arguably, not everyone can be right.

I wouldn't mind getting the facts and writing a guide for this.

 

[ElArte]

Another theory on the Internet: “FS sine Vrms” should be the maximum Vrms the input of the 2i2 can accept.

 

[trobbins]

Please clarify what the 2i2 output channel is doing, as you seem to want to use an independent signal source ( Akitika 1 kHz oscillator) ?

The Akitika 1 kHz oscillator and the amplifier are the 'input' to your 2i2, which should be attenuated and then connected to the 2i2. Any 'reference' for REW measurement is then the amplifier output signal, which would be measured by a separate meter, and if you wanted then you could insert that measurement (or some related value) into REW as a calibrate level or FS level if that helps you relate the REW plots etc to the amp output level better.

 

[ElArte]

Please clarify what the 2i2 output channel is doing, as you seem to want to use an independent signal source ( Akitika 1 kHz oscillator) ?

Hi Tim,

In my set-up, the left and right output of the 2i2 are unused.
As you said, the oscillator is the signal source for the amp and the (amp + load + attenuator) output is the input to the 2i2 input 1.

Based on your reply, I should measure the RMS voltage coming out of the (amp + load + attenuator) and use it to calibrate the input in the RTA window.

Correct?

Isn’t that the same as taking the maximum input level in the 2i2 specs sheet and entering it in the box without clicking the Calibrate button?

 

[trobbins]

Ok, so your comment below is meaningless, as you aren't using the output of the 2i2.

During the calibration process, I had to turn up the "Monitor" knob so that input level would match output level, as indicated by the Out/In gauges.

Wrt to 'Correct?', then no. It is likely that most will use the voltage across the load to align the signal level you assign within REW, but you can use the attenuated level (ie. the input level to the 2i2) if you want, or whatever (I rarely assign a Ref level when making REW tests on a power amp, as I describe the conditions that any save plot refers to). What you have to do however is add description to whatever results you generate for yourself, so that you can look back and say - eg. that 0dB FS sine Vrms is the max level when clipping starts on the DUT amp with so-and-so resistive load, or whatever description best describes your setup so that you can re-create it again without lots of question marks. Some people use dBV, some use dBW - perhaps look around at some amp tests/reviews that present REW results in a way that you can understand and appreciate better.

Apart from setting relative scale values that make sense, you need to appreciate how to operate the 2i2 so that when you send it signals, the 2i2 won't be internally clipping or distorting (in a significant way) throughout the test process (eg. for sweeps or steps), so perhaps try and follow xrk971's advise.

 

[DanDan]

Why not do normal sweep measurements. I did this with a vintage Radford STA15 Valve Amplifier. The Frequency Response Graphs were stunningly good. Distortion was reassuringly very low. This data help sell it for a small fortune. It was simple Line I/O on an iMac