How to Solve High-Frequency Suckout in Room?

Okay, I tried testing a series of frequency tones without any load / speakers. The voltage of the right channel seems to be slightly higher than the left channel, when playing tones from 800Hz or so through to 12,000 Hz. The difference in voltage is greatest between 4,000 Hz and 6,000 Hz with the maximum difference in voltage between left and right channels being 7mV at 5,000Hz. Across the 800Hz-12,000Hz band, the right channel is always tad HIGHER in voltage than the left side. Outside of this band, the voltages between channels are identical.

Is this difference enough to matter? Interestingly, my ears tell me the problem is centralized around the upper mid range, so that corresponds with where I measured the greatest variance in voltage between channels, at 5,000Hz. But, given the missing frequencies on the right, I would have expected to see LOWER voltages on the right, not higher! Hmmm.

I had my amp set to a moderately high volume setting (20 on the Diablo 300) which is which is about 3-5 higher than my typical listening volume. Here’s the data, with the frequencies with the largest variance between left and right highlighted in bold and underlined:

250 Hz L 694mV R 694mV

400 Hz L 694mV R 694mV

800 Hz L 695mV R 696mV

1000 Hz L 690mV R 692mV

1250 Hz L 684mV R 686mV

2500 Hz L 647mV R 652mV

3150 Hz L 627mV R 633mV

4000 Hz L 600mV R 606mV

5000 Hz L 567mV R 574mV

6000 Hz L 533mV R 539mV

7000 Hz L 497mV R 502mV

8000 Hz L 459mV R 463mV

9000 Hz L 419mV R 422mV

10000 Hz L 382mV R 384mV

11000 Hz L 349mV R 350mV

12000 Hz L 319mV R 319mV

13000 Hz L 287mV R 287mV

14000 Hz L 252mV R 252mV

Is this enough data to tell Gryphon I have an issue?  Or are these differences in voltages between left and right too small to matter?  If the differences too small to matter, well I do find it a bit coincidental that the largest variances in voltages are in the same frequencies where my ear hears the problem (centralized around upper mids).

Play the  1 kHz test tone into  speakers don't play so loudly that it damages your speakers. Set your voltmeter for AC Volts. Measure the volts across the speaker terminals. Power is: E²/R where E = volts RMS and R is resistance of load. Both channels should be the same at the same volume setting, maybe slight variation. 

Your left and right are perfectly matched. No issues

I thought there had to be a load on the amp to measure this  with multimeter but I might be wrong.  

@theaudiomaniac I think you might be right, in terms of AC voltage output being close enough. That said, there is a 1.2% variance at 5,000Hz, and close to a 1% variance around this frequency. While small, wouldn’t that be audible

@djones51 , I think the voltage output should be sufficient to compare the differences (without needing to convert to power which should give the same readings. Unless, of course, you are thinking there might be differences in the resistance between left and right and I should measure that too? As I said the problem was evident to a lesser degree on my old speakers, and I also tried swapping the speakers with each other and the problem remained on the right.

I highly doubt testing with the speakers connected will show much difference (but I will try later today). This is because I won’t be able to play the 1kHz tone at the same volume as I did without connecting the speakers, and therefore I am guessing there will be an even smaller difference between left and eighth channels.

I did read an old forum post that indicated that subtle differences may not be detectable using a voltmeter. The ultimate test is to use a dual channel oscilloscope, inverting one of the channels phases, then using the “sum” function of the scope. Because you’ve inverted one of the phases, the channels should cancel out each other when you sum them, and any variance from 0V that the scope shows would indicate an imbalance. I used to have an oscilloscope that hooked up to your PC but it’s so old it won’t work anymore.

Thinking my next step could be to rent a dual channel oscilloscope.

And, I’m going to book a frequency response hearing test just to totally rule that out! Although my wife also detected the problem without being told exactly what the issue was first.

I will also try to measure the frequency response playing the test tones, 1m from each speaker playing one side at a time.

Finally, I will also try rotating my entire system 90 degrees in my room.  Just as one final “room test”.  This will require me to dismantle my couch again so might be a while…

I do have a good relationship with my local dealer who I did not buy my amp from.  I could potentially ask them to borrow one of their high end Moon amps for troubleshooting.  I’ve tested Moon gear before and I found their stuff to be very highly resolving, so if the issue is not present with a Moon integrated, I’d think that would prove definitively that my amp is at fault.

The things we do for our hobby!