Power amp testing

[Dimitrov]

What would be the best way to set up a listening test to test for differences between let's say an AVR, and an external power amp connected to it? What little I know about amps is that different amps have different gain levels. AVR internal gain can vary compared to external power amp gain. So this needs to be accounted for to obtain a more valid outcome.

But the designing of the test seems challenging to me. How would one level match the AVR and the power amp and then listen to both? That would be a very tedious exercise in my mind. I'm assuming you would have to set a fixed level on the AVR. Listen. Then connected the power amp, adjust for the level variance then listen and minutes could have gone by during listening sessions.

Most people don't have two separate set ups. One AVR to compare against another with an external power amp. Unless I'm missing something super simple about a comparison like this, it seems like a challenging undertaking.

 

[peng]

What would be the best way to set up a listening test to test for differences between let's say an AVR, and an external power amp connected to it? What little I know about amps is that different amps have different gain levels. AVR internal gain can vary compared to external power amp gain. So this needs to be accounted for to obtain a more valid outcome.

But the designing of the test seems challenging to me. How would one level match the AVR and the power amp and then listen to both? That would be a very tedious exercise in my mind. I'm assuming you would have to set a fixed level on the AVR. Listen. Then connected the power amp, adjust for the level variance then listen and minutes could have gone by during listening sessions.

Most people don't have two separate set ups. One AVR to compare against another with an external power amp. Unless I'm missing something super simple about a comparison like this, it seems like a challenging undertaking.

If you compare them in mono mode, you should be able to level match them easily.

 

[Dimitrov]

Can you give me a rundown on how you would conduct the test? Like a step by step?

 

[Dimitrov]

Any experienced members here that have done testing like this? Can you share your thoughts? How would I go about setting up this test? Would an SPL meter be fine to level-match?

 

[solderdude]

You will need a voltmeter and test signal (sine of around 400Hz) to level match as it should be within 0.1dB (1%).
This should be done with the speaker connected and at a level of around 1V (so the meter should have a 2VAC setting).
This means you need to be able to measure and adjust in the range 1.010 and 0,990 V.
Cheapest meters around usually are not suited and will have a 200VAC setting or maybe 20VAC these are not accurate enough
This will result in a quite loud 400Hz tone (around 90dB at 1m) for each speaker which for a continuous tone is unnervingly loud.
That said a dB meter (with 0.1dB resolution) is better than no measurement at all when not owning a multimeter.

An SPL meter is not accurate enough even when it seems to display tenths of dB.
Especially when you want to switch outputs of amps to one speaker.
Also keep in mind that if you need to switch both the - and + outputs of each amplifier.
Also you should remain below clipping levels of the least powerful amp.

You will need to use an amplifier that has more gain than the one in the AVR as you need to attenuate (both channels separately) the signals to the external amp.

 

[Dimitrov]

So basically as I understand it, in order to do this properly I need two AVR's. One without power amp connected. The other with power amp connected.

I have to use a voltmeter. If I set the volume on my AVR to a comfortable listening level, let's say 70 dB with an SPL meter. I measure what the voltage at that level. Then on the other AVR + power amp, I set the voltage the same as the first AVR without power amp.

Am I getting this? This seems awfully involved. I've never used a voltmeter before.

 

[solderdude]

70dB on an SPL meter means the amp will put out around 0.1V to 0.2V meaning you must be able to discriminate between 0.101V and 0.099V (assuming 90dB/V speaker at 1m distance, around 0.2V for say 85dB/V speakers)
You would need to have a meter with 200mV AC measurement range.
it needs to be a continuous tone, not music.
The volume control position might be in around the same position as when listening to music at around 70dBA average.

When you have 1 AVR which has an output for an external power amp and you want to know if the sound actually differs/improves you will need to have a power amp that has higher gain than that internal to the AVR.
The amp output will need to be attenuated (using a stepless attenuator).
As you cannot attenuate the internal power amp section using an external amplifier that has lower gain can not be used.

When the external amplifier has variable input (a volume or level control) it will be easier than when the external amp is 'fixed' in gain.

 

[Dimitrov]

Stepless attenuator? This is getting too complicated. Clearly this type of test I think was not meant for the layman to do. This is more for a professional. I was thinking (hoping) it would be simpler.

If I have a AVR and I loan a 2 ch power amp. At the minimum I can detect if the SPL increased using test tones before and after. Keep the knob pointed to -20. Run test tone. Add power amp, run test tone. See if the SPL changed.

It might be crude, but if it's gone up 1 or 2 dB then that tells me ... that can definitely influence the outcome of the test. Of course then I have no means of level-matching.

Plenty of people online will say after adding a power amp to their AVR that it made major differences. But it's also possible the differences were due to gain. I'm beginning to realize that performing a valid test probably not feasible unless you are an engineer/scientist type.

 

[solderdude]

Stepless attenuator? This is getting too complicated. Clearly this type of test I think was not meant for the layman to do. This is more for a professional. I was thinking (hoping) it would be simpler.

Yes, not easy to do and is why most of the 'this or that amp sounds better' reports come from. (Small) differences in loudness.

If I have a AVR and I loan a 2 ch power amp. At the minimum I can detect if the SPL increased. It might be crude, but if it's gone up 1 or 2 dB then that tells me ... that can definitely influence the outcome of the test. Of course then I have no means of level-matching.

Yes, to do a proper test 1 or 2 dB difference is to much. Audibly louder (if only slightly) is already too much (>0.5dB)

Plenty of people online will say after adding a power amp to their AVR that it made major differences. But it's also possible the differences were due to gain. I'm beginning to realize that performing a valid test is just not feasible unless you are an engineer/scientist type.

Yes, very likely or even a higher output power and the lower output amp of the AVR already clipping some peaks where the (usually more powerful) external amp is still 'clean'.
Hard to say what the actual reason is, could even be a too low (or too high for some class-D) amps.

 

[Dimitrov]

Thank you very much for the advice. Here I thought it wouldn't involve lab testing and it turns out, it kinda does. Well, I guess don't need the stepless attenuator if I have two identical AVR's. One with and without power amp attached. I guess that concludes this thread.

Again, thanks.

 

[solderdude]

When you have 2 identical ones and one is connected to an external amp and you can feed both with the same signals and adjust the volume so they match within 0.1dB (requires the voltmeter + test tone) and can safely switch between both amps you are set to do a 'valid' comparison. Assuming you stay away from any clipping points.

 

[peng]

So basically as I understand it, in order to do this properly I need two AVR's. One without power amp connected. The other with power amp connected.

I have to use a voltmeter. If I set the volume on my AVR to a comfortable listening level, let's say 70 dB with an SPL meter. I measure what the voltage at that level. Then on the other AVR + power amp, I set the voltage the same as the first AVR without power amp.

Am I getting this? This seems awfully involved. I've never used a voltmeter before.

It depends what you really want to achieve, or prove. A quick way could be done as follow:

Connect the ext amp to the front right of left channel, then you can compare them quickly, in mono only obviously.

Level match:

- if you don't have a multimeter, you can just use the test tone and an spl meter, then you can use the level trim to equalize them to within about 0.5 dB, or
- use REW to do the same, assuming you do use REW and have a mic such as the popular Umik-1.

To get within the ball park quickly, find out what the external amp's gain is, say it is 26 dB, then right from start you can set the level trim for the channel it is connected to, +3 and you should likely be within 0.5 dB matching the AVR's amp. To me that's pretty much good enough but if you can score better than 50% consistently then you may need to fine tune your level matching procedure to get within better than 0.5 dB.

I would strongly recommend figuring a way to minimize the delay between switching from amp A to amp B. I did it with a toggle switch as shown below:

3e audio A7/A7 Mono Amplifier Review

EPDR is a fictional number that assumes a class B rail-to-rail amp. That is a fictional device but the calculation is a decently representative worst case scenario for class AB amps as they are basically class B at high power. Class D amps don’t have EPDR at all. At high phase angles they have...

www.audiosciencereview.com