Balanced vs Unbalanced?

Herman, just for the record, I like SETs a lot myself and have built a number of type 45 and 2A3 based amps.

But before we go into that,

One unmentioned difficulty is that for everything to work as designed you have to have a circuit capable of producing 2 signals that are exactly the same except for their polarity. I agree that using a differential circuit makes it easier but there are still a lot of things that can cause an imbalance. If there were not the CMRRs achieved would be much higher than they are.

The above comment is not exactly correct. It only applies to the use of balanced circuits **that are not differential**. IOW, if you have an imbalance in a differential circuit, it will be corrected in the next stage. In addition, a differential circuit will absolutely insure that the inverted and non-inverted phases of the signal will be absolutely 180 degrees out of phase but otherwise identical, even if their amplitudes are slightly different. So it does not have to be as exact as you describe in order for the CMRR to be retained.

I have found in the execution of differential circuits that the constant current source used in the circuit is a big deal. Most CCS circuits only have a single device (tube or transistor) and so are ineffective. They have to have at least 2 stages to work right. My point here is that doing balanced right is like anything else: a matter of execution. We get CMRRs well in excess of 100 db, even without perfectly matching signals.

Although its a different subject, you did mention the multiple tube issue in OTLs. That is no different from using multiple tubes or transistors in other amplifiers, and just like in other amps (including SETs) there are ways to make those tubes work quite well together! I agree that SETs have a lot of linearity, especially in that first watt where everything is so critical. It might interest you to know that the only other type of tube circuit that has the same behavior is an OTL- if designed right they have no crossover or notch issues, so like an SET the distortion vanishes as power goes down. Unlike an SET, they don't make nearly as much distortion at higher power, and so offer less coloration and smoother sound. It is easy to demonstrate. PM me.

Good source + triodes + good horn speakers = bliss

This is a formula that has worked well for me for a very long time :) I add to it: balanced operation- this ends the sonic role that interconnects traditionally play.

Ralph which would you prefer your electronics with a electrostatic or a horn system with all else being equal;I am just curious.I am a soundlab owner and have been watching the horn thread that has been going on with great interest as I am not familiar with that technology.

Rleff, it really depends on the speaker rather than the technology. Its the 'all else being equal' part that prevents an answer beyond that. Either one can be awesome or suck really bad.

even if their amplitudes are slightly different.

I thought one requirement for a high CMRR was that an equal input was amplified by an equal amount so they cancel? How does that get corrected in the next stage?

That is no different from using multiple tubes or transistors in other amplifiers

I completely agree, and all of those would be better if they could avoid doing so.

Unlike an SET, they don't make nearly as much distortion at higher power,

Totally agree, that's why I use 107dB efficient horns and usually listen at moderate levels. If I crank it up I'm not too worried about distortion as long as it isn't gross.

BTW I have heard your amps at the Analog Room in San Jose hooked up to a pair of Avalons and it sounded very, very good. Too bad the proprietor is so condescending and the place reeks of cigar smoke. I readily admit it looks like you have turned the OTL from a fireworks display into a fine sounding piece of equipment.