Your Thoughts: Phase drifts in 2CH system as Freq increases

[Maxpower]

- 2 Channel HiFi system, the left channel has a dull high frequency response that is audible
- the SPL response doesn't show the problem (interesting) but the Phase response does
- I physically swapped the speakers (L<>R) and remeasured. The phase response measurements follow the speakers

So I cautiously conclude, this is a speaker problem and the rest of the system is OK

The Twist: after physically swapping the speakers, I did a listening test and the fault still appeared to be the left channel.
This was a real surprise The data does not support that reality.

Any ideas?

System Summary
Speakers - Martin Logan Motion 40 Floor standing
(3-Way, crossover points: 500Hz and 2600Hz)
Amplifier - Cambridge Audio CXA80
Measurement Mike: U-MIK2
REW version V5.30.9

Screenshots below to save members loading files (REW files attached if you want to though)


Phase Response

SPL Chart

 

[sm52]

If you keep in mind the location of the speakers as when measuring, try changing the polarity on the left speaker. Possibly only on the tweeter, but maybe on the left speaker input. I also see a big hole from 240 to 500 hertz. This is also significant.

 

[Maxpower]

Thank you @sm52
I did take a measurement of the left channel with polarity reversed. My system is bi-wired, so that made the task more straightforward. See below and attached.

Additional Thought / Question: from my original post above, which channel has the expected phase response? I assumed the Left channel had the issue. However, I think phase response is likely to drift as frequency rises?
Thanks again for any help members can provide!

Phase Response (Inverted Left CH (Green), Normal Polarity Left CH (Red) overlay)

 

[sm52]

You have reversed the polarity of the entire left speaker. I looked at what happened. This is not the right decision. It is necessary to return the polarity of the left speaker as it was before. Moreover, the tweeter of the left speaker does not need to be inverted either. It sounds like you're dealing with reflections near the left side that add up to the main signal to produce a dull sound on the left side.

 

[sm52]

which channel has the expected phase response?

If you place the microphone exactly between the right and left speakers, the phase response on each channel at the end of the frequency range should move around zero degrees. This may start at 10kHz, or 2kHz, or some other frequency. Your measurements are not synchronized, so it is not immediately clear whether the phase response looks correct.

 

[Maxpower]

You have reversed the polarity of the entire left speaker. I looked at what happened. This is not the right decision. It is necessary to return the polarity of the left speaker as it was before. Moreover, the tweeter of the left speaker does not need to be inverted either. It sounds like you're dealing with reflections near the left side that add up to the main signal to produce a dull sound on the left side.

Thank you, your advise is greatly appreciated. I did only reverse the the mid/tweeter (the bi-wiring makes that straightforward). The bass was left at normal polarity.

If you place the microphone exactly between the right and left speakers, the phase response on each channel at the end of the frequency range should move around zero degrees. This may start at 10kHz, or 2kHz, or some other frequency. Your measurements are not synchronized, so it is not immediately clear whether the phase response looks correct.

OK. I can try that and post the results here. My original measurements above were done individually, on axis @1M. I configured REW to only drive and measure one channel at a time.

Thanks again, this is not my area of expertise, I appreciate the guidance.

 

[sm52]

This is right. First one channel, then the second. Two channels are only sometimes measured at the same time, for example to evaluate how the right and left influence each other. But the same thing can be done by having measurements of each channel separately and using arithmetic in REW.
At this point, I don't see any obvious reason why the left channel is dull. There is an interesting thing that happened when you reversed the MF-HF polarity in the left channel. The hole at 250-500 hertz disappeared. But the correct polarity is the same as in the first post.
Why are you taking measurements at 1 meter? If you are not going to rebuild the internal crossover, then take measurements from your listening position. Each channel is separate.

 

[Maxpower]

This is right. First one channel, then the second. Two channels are only sometimes measured at the same time, for example to evaluate how the right and left influence each other. But the same thing can be done by having measurements of each channel separately and using arithmetic in REW.
At this point, I don't see any obvious reason why the left channel is dull. There is an interesting thing that happened when you reversed the MF-HF polarity in the left channel. The hole at 250-500 hertz disappeared. But the correct polarity is the same as in the first post.
Why are you taking measurements at 1 meter? If you are not going to rebuild the internal crossover, then take measurements from your listening position. Each channel is separate.

Thank you! Brief responses below

"But the same thing can be done by having measurements of each channel separately and using arithmetic in REW" I did not know that, I will learn how to do this.

"The hole at 250-500 hertz disappeared. But the correct polarity is the same as in the first post" Yeah, and I really don't understand this. The MF-HF section of the crossover is completely isolated from the LF section. It looks like the midrange driver may have be outputting below its 500Hz cut off point? (perhaps an unexpected outcome when reversing polarity?). Not sure on this.

"Why are you taking measurements at 1 meter? If you are not going to rebuild the internal crossover" Honestly, that's where I thought this was going. A shorted-turn or open circuit inductor, bad capacitor, etc. I will take the listening position measurement as you have suggested.

I also shifted the physical location of the Left CH speaker to check for any environmental / reflection issues. HF dullness was unchanged.

At this point, I am assuming I made bad measurements and/or came to the wrong conclusion. It is starting to look like a left channel signal issue. These speakers do double-duty as Home Theatre and 2CH Audio devices via a 'Home Theatre Bypass' function in the 2CH integrated amplifier. When the Home Theatre AVR is on, the power amp in the integrated 2CH amp is driven directly by the AVR (2CH preamp is completely bypassed). When the AVR is off, the 2CH Amp operates normally with its analogue/digital inputs.

I'll try an amplifier substitution first and if there is no joy there, speaker cabling is next. Wish me luck!

 

[sm52]

Earlier you wrote that after changing the right and left sides between each other, the problem moved to the right side. If the speaker wires remained in place, then they have nothing to do with it. And the reason will be inside the left speaker.

 

[Maxpower]

“Little by little, one travels far.” – J.R.R. Tolkien

TLDR version: JRR is right. The many small things were: loose binding post nuts on the cross-over PCB (Left speaker), Left channel volume of the integrated amp is 5-7% lower across the spectrum, end-to-end speaker cable resistance for Left Channel approx. 15% higher than Right Channel, Speaker placement issues (Left Channel), discovered the Left Channel midrange driver has been replaced prior to my ownership

In case it helps others; I will post some pics, measurement screenshots and useful reference data in the next few days. I will share what I was able to fix and what may not be practical to fix.
This was a LOT of work over many days . . . I haven’t had a chance to re-measure with REW, but I will, and share here.

Thanks again to @sm52 . . .

 

[sm52]

You did a good job. Happy for you.