What does an impedance-matching issue "sound" like?

BUT occasionally I am hearing what sounds like a bit of breakup deep in the mix of some songs. Like the whole channel isn’t breaking up, but suddenly a rhythm guitar part sounds noisy like it’s being played through an AM radio or something. On some songs, the entire mix may sound reduced in scale and less dynamic.

Read this above that you posted then this below from Eli. I see a similarity.

"Very few amps work well, with simple paralleling. Unfortunately, heat and distortion are the typical result."

As well I would have thought 2 separate channel negative feedback’s suddenly linked together via the transformer secondary tap could also be a problem, and maybe a source of this distortion.

I could see if there was no global feedback as in some amps, then paralleling could work.
But Luxman tube amps all use global feedback from the circuits of their’s I’ve seen.
This is one that would be almost the same as yours, with the feedback highlighted. https://ibb.co/pRpGtPQ

Just found the 88 circuit https://ibb.co/RhG8hJj

Cheers George

I have two questions for you. 
  1. Is the "Problem", repeatable? 
As in your having the ability to re-play a specific song/s or passage which gives you these mid-range errors "Every-time"? 
Or is "It", an inconsistent problem? So that when the same song is played over again, the same errors refuse to show themselves? 
  2. Have you tried using any software/hardware test and evaluation systems made for trouble-shooting and/or tweaking a system? You can purchase the software and hardware, "Usually just a test mic, software DVD for your PC and an audio adapter", very cheaply these days and it is now very user friendly AND exponentially more powerful compared to those available even just a few years ago.  And it is the exact same gear many professionals utilize today to Install, test, fix and tweak all of your components including cables! I have owned a few over the years.
"Dayton" has a system named, "DAT's V2", for around just a "C-note"! There is also a "True RTI" system, "Which I still use", for a bit more. 
"The (True RTI's) cost is dependent on the (Octave Set), ordered".     
Either of these packages will let you very simply and easily go through your system. "I am pretty sure others here will chime in on the different (makes/models)", of software/hardware packages and kits.
   You could learn to use and then completely test out all of your gear in a single weekend. There should also be someone locally you could hire to come out with their gear and do it. But "They", can get quite expensive very quickly. OR, You could just keep "Guessing"...…
    I personally? Do NOT think it is an "Impedance" problem. Alas, I am too far away to hear it to test anyway! "Possibly a bad transducer or crossover though". But with the cost of your gear? And a problem with a plethora of variables that would keep, "At least ME", up at night!! I think that you, "may" want to give this a try! 
I did, and never looked back!
   

I've never been a fan of anything paralleled or bridged. IMO, the sonics always suffer.

Looking at the mq88 schematic, there is a complex output filter and a multitude of pots for bias, gain and balance. It is possible that the particular combination of all the parts moves the channel symmetry enough at certain frequencies to cause an imbalance. Tube non-linearities/sub-par tube could also contribute.

Transformers like to see the load for which they are designed. The best advice here is to Bi-Amp. Doing so could help to isolate a problem in one particular channel.

After reading the Stereophile article and viewing the measurements, well... let's just leave it at that.

Interesting diagram, and that means I’m out of my experience level. I’ve never seen an amp "bridgeable" like this before, but the high output impedance of the transformer probably makes it less frightening than a SS amp with feedback coupled output stage

@erik_squires , @georgehifi When using tube amplifiers you can just parallel the outputs (and inputs) regardless of whether feedback is used or not. The Dynaco ST-70 instruction/assembly manual details how to do this.
Solid state amp are usually bridged (so the input to one channel is inverted from the other) as the output section of one amp is otherwise loaded by the other- leading to bad smoke.

For the former, ’monostrap’ is the correct term, for the latter, ’bridged’. The terms get conflated.

@psqt I take it then you have enough channels (four) to do stereo. Al makes a good suggestion as the inputs of the amps are also paralleled; if your preamp isn’t up to it you might hear some distortion, but I would expect that to be accompanied by reduced bass performance. I think something else is afoot- check all your tubes and make sure they are OK. If not the amp will lose delicacy and detail and may even distort in unpredictable ways. IOW I think you have a bad tube somewhere.

Thanks to everyone, and @atmasphere I am glad you chimed in (was hoping you’d be lurking here :) I am using a Backert Labs Rhumba Extreme preamp with 2 unbalanced outputs (L 1 and L2 going to the left Luxman, R1 and R2 going to the right Luxman - pictures in album linked below).

Also, all 8 KT88s have held bias steady around 485-ish since I first checked them last week.

I have had some time today to disconnect all connections, run in a single stereo configuration, then reconnect everything and give it some time to warm up, etc and make comparisons.

What I mentioned hearing before - I think it could have been my ears/imagination playing tricks on me, or simply something not properly connected or loosely connected. Everything actually sounds as it is supposed to in the parallel configuration.

What I am hearing now vs the stereo comparison is a more spacious sound field, more power in the bass in particular, and generally a stronger, beefier signal. I’m listening at about 90dB.

On the downside, it seems there is slightly more distortion in the sound running the 2 parallel amps. If you didn’t do a direct comparison it’s not likely you’d hear the difference. But there is slightly more "noise" around things like vocals, acoustic guitar plucks, etc. Again, it could be a difference in the age of the tubes, etc.

What I did was put together a small gallery of documentation for this process.
https://jsqt.smugmug.com/Luxman-MQ-88uSE-Diagrams-Measurements/n-cgz5rM/

It includes the actual block diagram of the MQ-88uSE (these amps), the specs, the instructions from Luxman on running the "C" version of this amp in parallel, as well as some additional measurements from HiFi News (UK) on this specific amp which aren’t exactly matching JA’s measurements in Stereophile.

Anyway I wanted to put all of this info together as a reference for this thread and if anyone else has ideas or questions about running 2 stereo tube amps in parallel.

Thanks again for all the info!