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

Wait, when you bridge an amp you run:

Amp L  (+)  ---> Speaker (+)

Amp R (+) ---> Speaker (-)

That doesn't sound like what you are describing. 

First, to be precise it sounds like what you are doing is paralleling the two channels of each amp, which is not the same as bridging. (Although the term "bridging" is sometimes misused to refer to "paralleling"). Bridging involves operating the two channels with opposite polarity signals, and connecting the speaker between the two positive output terminals. You are not doing that since you "used a jumper between each of the "positive" taps," and I presume you are supplying both channels of the amp with the same signal. Also, I suspect that this amp is not bridgeable.

An amp which can be bridged will usually provide a greater increase in power capability into 8 ohms or thereabouts than paralleling the two channels, compared to the amp’s power capabilities in stereo mode, but paralleling will usually be better able to handle low impedance loads.

In any event, it appears that the amp has a single output tap for each channel that is optimized for a 6 ohm load. By paralleling the two channels the tap would be optimized for a 3 ohm load. Looking at the impedance curve for your speakers that is shown in John Atkinson’s measurements it appears that the impedance magnitude varies between a low of about 3.5 ohms around 100 Hz and very high impedances in much of the mid-range and lower treble. The impedance being more than 20 ohms between about 2 kHz and 4 kHz.

Given all of that, I’m not sure what might be causing the symptoms you have described. With the amp presumably now optimized for a 3 ohm load perhaps it isn’t happy driving such high impedances in the mid-range and lower treble. Or perhaps the tubes or something else in the amp isn’t well matched between the two channels, and the two channels are trying to put out slightly different signals and are therefore fighting each other to some degree as a result of having their outputs connected together.

Also, just as an FYI, JA measured the amp’s input impedances as 54K and 34K for the direct and variable inputs, respectively. So by applying the same signal to both channels you are presenting the preamp with a load of 27K or 17K, depending on which input you are using. Some tube-based preamps will experience a bit of deep bass rolloff driving such low impedances, but it sounds like that is not occurring in this case.

Those are all the thoughts that occur to me at this point. Perhaps the background I’ve provided will result in additional possibilities occurring to someone else.

Regards,
-- Al

If you are attempting passive bi-amping:


Amp L ( + ) ------ Speaker HF (+)
Amp L (-) .-------> Speaker HF (-) 

Amp R ( + ) ------ Speaker Low F (+)
Amp R (-) .-------> Speaker Low F (-)


In neither case should the (+) amp terminals touch each other, even from matched amps. The slightest (0.001V) variation in output can induce significant current as the sides fight for the right value. 


Best,


E

Sorry, if you are using different output amps, your set up would look more like this:

Amp 1 - L  ( + ) ------> L Speaker HF (+)
Amp 1 - L (-) .------->  L Speaker HF (-)

Amp 1 - R ( + ) ------>  R Speaker HF (+)
Amp 1 - R (-) .------->  R Speaker HF (-)

Thanks guys this helps a bit.

I may be confusing bridging and parallel, depending on how it was described to me. 

The setup is basically diagrammed here, aside from the multiple speaker taps and the fact it is a MQ300 and not an MQ-88uSE (but they should be set up the same way)

https://onahighernote.com/blog/productfaq/connect-luxman-mq-300-mono-block/

As Al mentioned, there could be a number of factors contributing to what I'm hearing. I'll continue to run in the new amp and see how things settle down over the next couple of weeks.

Also, I confirmed before purchasing the second amp from Luxman that they could be run in mono configuration using the above setup so I am confident it is setup properly. 

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 stages. 

Erik, I've seen some cases in which paralleled operation of the two channels of a stereo tube amp has been recommended by the manufacturer.  Examples include the McIntosh MC275 and the old Dynaco ST-70, despite the fact that like the OP's amp those amps have output impedances that are particularly low for a tube amp.  (JA measured the OP's amp as having an output impedance of about 0.6 ohms; the ST-70 and the current version of the MC275 have specified damping factors of 15 and >22 respectively, which in both cases is suggestive of output impedances of well under 1 ohm).

A concern I've always had about doing that, however, is the possibility that if a small signal tube serving one channel were to fail, and music is then played, one channel would be attempting to output a full amplitude signal while the other would be attempting to force the output to zero volts.  Which doesn't seem likely to be healthy for the amp, especially if the listener isn't nearby or for whatever reason doesn't promptly realize there is a problem.

Best regards,
-- Al
 

More than likely the impedance of the speaker is too low for the amps in this configuration. Get yourself a high quality active cross over and bi amp the Rivals. They are set up for it! 

More than likely the impedance of the speaker is too low for the amps in this configuration.

That would probably be the case if the amp were in fact bridged, as was initially indicated, since as you probably realize bridging causes an amp to "see" the speaker’s impedance divided by two.

But note that as indicated in the subsequent posts the amp’s two channels are being paralleled, not bridged. Which has the effect of converting its single 6 ohm output tap into a 3 ohm output tap. 3 ohms being less than the speaker’s impedance at all frequencies, and much less at most frequencies. So in this configuration the amp is being lightly loaded, perhaps even too lightly to be optimal.

Regards,
-- Al

Al,

So if I am running the amps in this configuration to the 6 ohm speaker, you are saying based on JA's measurements and the information about the Luxman amp that it could be having a negative effect on the performance? Is it possible there is any damage being done to the amps or speakers? I was told by both the dealer that sold me the 2 Luxmans and the speaker manufacturer that this would not cause any issues with the equipment. 

Thanks!

jsqt OP

Bridging amps in solid state,
Doubles the output impedance
Halves the damping factor
Increases the distortion
Has have the current ability.
Makes the amp less stable into lower impedance’s to what it was in stereo.
So everything takes a hit, except you gain higher watts. I believe the same happens in tube amps also.

https://www.stereophile.com/images/617Voltifig1.jpg
If so where you may have had a amp that could "just" sort of drive the Volti Rivals 3ohm load at 100hz with 50 degrees of -phase shift in the bass. Now you have no chance bridged.

Along will come the suggestion of autoformers, they are just a band-fix don’t get sucked into that.

Solution go back to Stereo.
You could vertically bi-amp the two stereo amps (one channel bass, the other mids and highs. (same for the other side)
Change speakers or change amps.

Cheers George

@jsqt

I don’t envision any possibility of damage. But whenever the impedance of a speaker is significantly mismatched to the load impedance a tap is designed for there may be compromises to both maximum power capability and distortion performance.

As I indicated, paralleling two channels each of which has a tap that is optimized for a 6 ohm load in effect creates a 3 ohm tap. Applying a 3 ohm resistive load to such a tap would typically result in the amp being able to supply approximately twice as much power as it could supply into 6 ohms when running in stereo mode. With the impedance of your speakers being much higher than 3 ohms at most frequencies I’m pretty certain that the amp is falling significantly short of realizing that 2x increase. And although I’m just speculating perhaps the mismatch is contributing to the adverse sonic effects you described.

Regards,
-- Al

@georgehifi 

As I've indicated multiple times, the amp is not being bridged!

Regards,
-- Al

@almarg
Ooops sorry, just saw what he said in the first sentence.

added a 2nd Luxman MQ-88uSE to my 2 channel system, bridged to mono so I am running a Luxman to each of my 6ohm Volti Rivals

In any case that phase/impedance load around 100hz is nothing that tube amps want to be seeing.
https://www.stereophile.com/images/617Voltifig1.jpg
Change the amps if you want to keep the speakers. or visa versa


Cheers George

Thanks guys, this is all a little confusing to me as both the speaker mfg and the dealer who sold me on running 2 Luxmans said there would be no issues. 95% of the time this sounds amazing, and I do hear the benefits of a more powerful sound and better separation between channels. But that fuzzy sound in some of the midrange on some material is what brought all this up. Hate to think that I'm stuck with these amps now if all I've done is decrease the quality of my sound. 

Also regarding tube amps and the Rivals, they are voiced with Border Patrol SET amps which are used to demo the speakers at all the shows, etc. So I don't think tubes are the issue, but it's the 3 ohms into a 6 ohm speaker part that has me concerned. 

So I don’t think tubes are the issue, but it’s the 3 ohms

That 3ohms at 110hz with -50’ phase angle at 75hz to me is an issue, especially with SE tubes, this is where all the amps power is needed for bass repoduction.

I find it hard to believe a speaker manufacturer would voice a speaker with this kind of load graph with "feeble" SE Border Patrol amps, that are probably good amps with the right speaker, easier loads than this..

Cheers George

@georgehifi this Luxman is a push pull triode amp ... and to be more accurate I don’t want to make assumptions for the manufacturer so I cannot say they are voiced with the BP amps but they are definitely known for being paired together at shows.

@georgehifi

George, I understand what you are saying about the speaker’s impedance being challenging for a tube amp in parts of the bass region. But note that the OP has indicated that the addition of the second amp, with the two channels in each amp paralleled, has improved the bass (which is to be expected given that paralleling the channels in effect creates a 3 ohm tap), and he has not indicated that there are any problems in that region.

Note also that he has said that "fuzzy sound in some of the midrange on some material is what brought all this up." And as I indicated in my first post in this thread the speaker’s impedance in parts of the upper mid-range and lower treble is in the vicinity of 20 ohms or more! Resulting in more of a mismatch to the tap on the amp than when a single amp had been operated in stereo mode. Which leads me back to the first of the two hypotheses I proposed in that first post, with the second hypothesis also being a possibility especially given that the amp has an output impedance that is very low for a tube amp (0.6 ohms in stereo mode, as I mentioned earlier, which would mean 0.3 ohms in paralleled mono mode):

With the amp presumably now optimized for a 3 ohm load perhaps it isn’t happy driving such high impedances in the mid-range and lower treble. Or perhaps the tubes or something else in the amp isn’t well matched between the two channels, and the two channels are trying to put out slightly different signals and are therefore fighting each other to some degree as a result of having their outputs connected together.

Best regards,
-- Al

With the amp presumably now optimized for a 3 ohm load perhaps it isn’t happy driving such high impedances in the mid-range and lower treble. Or perhaps the tubes or something else in the amp isn’t well matched between the two channels, and the two channels are trying to put out slightly different signals and are therefore fighting each other to some degree as a result of having their outputs connected together.

This is more than likely contributing to the issue since one amp has about 300 hours on it (and the tubes) and the other new one has about 15. 

I certainly do not have the electronics background that everyone else has on this thread but I find it interesting that the issues seem to be only in certain instruments and certain recordings to paraphrase. Is it possible that the problem could be in the recording itself? When you come across a track with an issue, does it repeat with subsequent listenings or is it totally random?

jsqt OP


I believe Eli on diyaudio is quite clued up, from pass posts I’ve read, combining the feedback of both channels becomes an issue beside other things, here is what he says about paralleling tube amps.

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

Cheers George

Again I have to come back to the fact that while there is a lot of discussion about what should and shouldn't be done bridging or paralleling tube amps in general, I have specific instructions directly from Luxman on how to do this, and the Luxman distributor specifically approached me with this idea.

So while there may be many amps - even most amps - that this is not recommended for, I have specific instructions from Luxman for this amplifier, so I'm not clear how the overall configuration could be an issue. Nonetheless, this thread does bring up some interesting questions.

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!

@jsqt, thanks for providing the additional info.

It appears that your preamp has very low output impedance, as the non-Extreme version is said to have a nominal output impedance of 75 ohms. So the somewhat low input impedance of two paralleled channels of your amp shouldn’t be any problem for it, even though the two RCA output connectors the preamp provides for each channel are almost certainly just jumpered together internally and are supplying the same signal. And in any event, as Ralph (Atmasphere) and I indicated if impedance issues were present at that interface you most likely wouldn’t be getting the good bass response you’ve described.

Also, as George and Ian (ieales) suggested trying a vertical biamp configuration could very well be worthwhile.

Enjoy! Regards,
-- Al

Thanks Al, I am curious about the vertical biamp configuration just for the sake of trying out all my options to see which sounds the best. This configuration would have benefits by running twice the output power split between the top and bottoms of the speakers?

The Volti Rivals are 3 way speakers with an external crossover (currently using jumper between the woofer and mid-horn). So it sounds like I’d use the Left channel of one amp to power the mid horn and tweeter of the left speaker and the Left channel of the 2nd amp to power the woofer of the left speaker.

The benefit here could be less distortion and more power to the individual speakers, but the trade-off between the parallel config would be re-introducing crosstalk between the 2 speakers and in theory reducing the stereo separation of the mix (vs parallel config)?

@jsqt, no, what you’ve described in the second paragraph of your post above is a horizontal biamp configuration.

In a vertical biamp configuration you would dedicate one amp to powering the left speaker and the other amp to powering the right speaker. One channel of each amp would power the low frequencies of the corresponding speaker, and the other channel of that amp would power the mids and highs of the same speaker. When identical amps are used vertical biamping is considered to be preferable to horizontal biamping.

There are several potential advantages of a vertical biamp configuration, in addition to the overall increase in power capability compared to stereo operation of a single amp:

1) In contrast to a horizontal biamp configuration, in a vertical configuration both channels of the amp are processing the same signal, and therefore interchannel crosstalk that may occur within the amp is essentially eliminated.

2) The amps can be located close to the speakers, reducing the length and potentially the sonic effects of the speaker cables.

3) Quoting a statement made a while back by GeorgeHiFi in a thread about biamping:

... the biggest advantage of vertical bi-amping is that the whole power supply joule storage of one amp, is dedicated to just one bass driver and not shared, so the bass/upper bass should be better on big dynamic transients. (unless the amps have true dual mono power supplies) which are rare.

Regards,
-- Al

Thanks Al, so this vertical biamping actually sound preferable to running the 2 amps in parallel, if I am reading correctly.

So the connection from the preamp is the same as what I am using described in my post above (both L outputs to one amp both R outputs to the other amp) only no jumpers - I use both sets of speaker taps to go to the mid/hi inputs on one speaker and the bass input on the same speaker, then do the same on the other amp.

So this also effectively doubles the power going to each speaker - how does it affect impedance and is it safer for the amp and speaker?

So this also effectively doubles the power going to each speaker - how does it affect impedance and is it safer for the amp and speaker?

Regarding safety, earlier I described a scenario involving failure of a small signal tube that could conceivably result in damage to an amp that has been monostrapped. (Thanks, Ralph, for citing that term). However that scenario obviously has a relatively low likelihood of occurring, and so IMO it would be reasonable to consider both approaches as being safe.

Regarding impedances and sonics, IMO it's probably one of those situations where you have to try it both ways to decide which is preferable. Since paralleling/monostrapping would in effect provide the amp with an output tap optimized for a 3 ohm load, as I had indicated, that would presumably result in the best match in parts of the bass region where both the impedance of the speaker and the content of a lot of music tend to be especially challenging.   But it would result in a worse match at mid and high frequencies, compared to both stereo operation and vertical biamping.

Regards,
-- Al
 

For BiAmping, you really should have crossover, otherwise the amps see an infinite impedance out of band.

A simple 6db HP crossover can be effected with a single capacitor for the upper and a LP for the low with a choke.

It would be better to use slopes similar to the in-built crossover.

For a quick cheap & cheerful XO, pick up a dbx 223 or similar. It's 4th order or 24db/octave. 

@almarg - thanks, I have now experimented with the vertical bi-amping. It’s probably too early to have a firm opinion but as you noted, I noticed a little weaker bass, if not a smoother mid/hi top end. Also strangely enough it sounds like the stereo field is a little smaller than when I was mono strapping. Funny thing, the song I noticed the "breakup" from my OP was the arpeggiated rhythm guitar part in Tears for Fears’ "Woman in Chains," which now in the bi-amp configuration is NOT mostly placed in the left stereo sound stage but is now spread out across the entire sound field. Very weird as I didn’t change the output config of the preamp at all.

@ieales - the Rival has an external passive 3 way crossover. I’m sending one channel now into the mid-high input and the other channel into the bass input. I’m not sure how I would implement an active crossover into that configuration as the external crossover is built specifically for the Rival and from what I understand must be used (you cannot connect a direct signal on the speaker inputs).

Here is a photo of the bi-amp connection into the Rival passive crossover

https://jsqt.smugmug.com/Luxman-MQ-88uSE-Diagrams-Measurements/n-cgz5rM/i-dFFnJ5v/A

It's a shame they don't give the poweramp (blue) distortion figures when it's in "strapped paralleled" mode also, this "could" give your problem.

Cheers George

Based on about 5 hours of listening tonight in vertical biamp config, I’ve gone back to running the amps in parallel.

The bi-amping sounded a little smoother I think in the higher frequency. But there was an odd effect on the stereo image, seemed more like 2 separate speakers than a “disappearing” stereo field. 

At first I just couldn’t figure out what was causing the effect on the stereo image. Voices seemed to be part of the mix rather than having their own center showcase. And other elements of the sound, especially the bass, seemed more a part of the stereo field than having their own pinpoint placement in the mix.

Overall it just seemed like a “stereo” versus a staging of a live in the room placement.

i finally figured that the issue was that since the bass was compromised by this configuration, the deficiencies in bass caused the sound to be less anchored overall in the stereo field.

Anyway, that’s my best guess.

As I mentioned before, reconnecting everything seemed to correct most of the initial OP issues I had with the mono strap configuration and now that I have gone back to that config, the stereo field is back to normal. Voices are better anchored in the center of the mix. Bass is tighter and more focused. Stereo field extends beyond the speakers again, and the speakers have disappeared more for better imaging.

So it’s been quite an adventure. In many ways I am back to where I started but the Fidelity has been restored and damned if I don’t know a whole lot more about this whole process than I did in my original post.

Thanks to all who helped sort this out