Kirkus - one of the problems with class AB is required gain to get rid of nonlinearity. Class AB amps have gain (before feedback) of couple thousands while class A couple hundred. Huge gain (before NFB) and delays just invite TIM if input slew rate in not limited. It might be possible to bias amp just a little higher (class A has about 150% of max current) to move the "kink" a little further away and set minimum gain to get minimum spects like 0.1-0.5% THD and IMD and bandwidth of 50kHz. The key, I believe, is compromise without going to over specifying it.
TIM is one of the reason of class AB sound, but it wasn't known until 1972. "experts" in denial claimed then that all parameters of SS amp are as good as tubes and therefore they must sound the same while average person could hear otherwise. |
Spatine - Not the smaller speakers but more efficient once. |
Bombaywalla - not only bias but also gain (before feedback)is different between class A and AB. Class AB requires much higher gain (usually about 10x) to linearize output transistors with deeper negative feedback. Large gain causes amplifier to overshoot (and even choke) on fast changing input (pulse) since it is unable to feed signal back in the same phase (delays the signal). This is called Transient Intermodulation Distortion and was dicovered in 1970. Before that first SS amplifiers had excellent THD and IMD but unpleasant sound (overshooting introduces odd harmonics).
Using different classes of amps for woofer and tweeter is fine if we can guarantee that phase shift is the same. |
Spatine - I've never had bi-amping setup but noticed that people, most of the time, use identical amps for that. Absolute phase should be the same unless amplifier inverts the phase 180 deg - easy to fix by reversing speaker wires. Problem starts when you get amplifier from different family. My amplifier, for instance, is a class D creature based on Icepower 200ASC module from B&O. Data sheet for this module shows +40deg phase shift for all frequencies (0-20kHz). I'm not sure what to think about it. Rule of operation is different and this amp shouldn't perhaps be bi-amped with an amp from another family.
It would be cool though, to use class D with its excellent power, dynamics and bass control with small class A amp for the highs. I'm afraid that harmonics would add improperly and sound would be strange.
Standard (class A, AB, tube) amps might have slightly different absolute phase shift (and different interaction with the speaker) and I wonder about using different amps (or different family of amps) for bi-amping.
Had anybody done that? |
Kirkus - we're not talking about absolutely best amp in existence but about one that is about 100W/ch and less than $3k. Do you feel that class AB amps would have certain advantage here over class A (less estate = better components)? |
Kirkus - I read article that shows harmonics of output stage (class AB) with underbias, proper bias and overbias. It's a little eye opening since many people believe that increasing bias will make class AB sound like class A. It is not so simple - otherwise everybody would have done that.
What is typical bias these days? I remember vaguely something around 100mA.
I also read on the subject of complementary pair distortions that you mentioned once. It pretty much confirms what you said but I just couldn't believe that, according to it, some companies use all NPN transistors because PNP transistor is a little more expensive. Crazy. |
Kirkus - Thanks, it makes sense. |
Or like a gorgeous chick that knows how to have fun when the lights are out but doesn't go all the way. |
Kirkus - The magnitude of TIM is highly dependent on the open loop gain (everything else being equal) up to point where output transistors go to momentary saturation and stay there for a moment (having charge trapped at the junction). We cannot hear it (brain fill the gaps) but it make us tired.
TIM can be easily shown with just sum of two signals and the scope but it doesn't show in normal measurement of THD IMD etc. That was the problem in 1970 and is still now.
In an article in Stereophile "A future without a feedback"
http://stereophile.com/reference/70/index.html
Maritn Colloms claims that sound of 700 amps he reviewed was inversely proportional to amount of global negative feedback. One amp he mentions is a CARY monoblock with a strange feature of negative feedback adjustment. It sounded best at the lowest feedback.
In order to guarantee that amp would be free from TIM designer has to limit input slew rate (or frequency) to levels that output has (slew rate or frequency) before feedback is applied.
The issue here, I believe, is not a lack of resources but lack of discipline. I wouldn't buy class AB amp that has 0.0001% THD - that would be insane. At certain point of open loop gain very low level THD distortions (mostly odd harmonics) will be traded by for higher level TIM artifacts (also mostly odd harmonics). In both cases there will be also more (than in class A) harmonics of the higher order.
Yes TIM is a stability issue - when somebody decides to put gain of 10000 into audio amp and publish perfect spects.
Todays output stages are much faster than at the time Otala published his paper but desire to make class AB amp that is as good as class A amp - still exist. |
Kirkus - Yes, saturation can appear without any gain in the circuit but it has nothing to do with the issue we're discussing.
High slew rate input signals come back to summing junction thru negative feedback delayed because of signal path delays. For a moment amplifier has no feedback and overshoot appears at the output (or earlier dependent on design). This will happen to any amplifier if slew rate is not properly limited at the input.
Amount of this overshoot is a function of amps open loop gain and in really bad case will take output stage to momentary saturation.
Let forget what is causing it, I agree, and look what to do to fix it. Class AB amp exhibits higher order of mostly odd harmonics at very low signal levels while THD and IMD is measured at substantially higher levels and doesn't show it. In order to lower it - either components have to be very linear or feedback has to be deeper. Careful selection of transistors and better circuit will help to a degree but will never eliminate big "kink". Local feedback will help as well but most of the linearizing will be done in the global NFB (global vs. local is a separate discussion). Bandwidth has to be limited at the input to bandwidth of the amp without the feedback (open loop). That's all. It is tradoff between low level THD and bandwidth on one side and TIM on the other. |
Kirkus - class A amp at very low levels has both output devices conducting simultaneously doubling their voltage gains. It creates wobble in output linearity - no escape from that. It is known as "gm doubling". Increasing bias won't help since overbiasing creates higher order of odd harmonics (because of gm doubling)as well as underbiasing. |
Kirkus - I don't do much with discrete stuff so I remember wrong definition of class B (as a one without bias current). Now I realized that it has to have bias current in order to break Vbe voltage of output transistors. |
Shadorne - tests show that underbias and overbias causes increase in distortion (especially higher order odd harmonics) for all signals (in wide range of output power). Increasing bias above optimal level increases non-linear area (of non-constant gm-doubling) and makes it worse.
Proper way to do it is to bias it into class A so nonlinear area would never be reached and gm doubling would appear constant for all signal all the time. Unfortunately this requires (typical) bias of 150% of anticipated max output current. At least that's how I understand it. |
Shadorne - careful topology is always important but I think that as long as transconductance in biased area is non linear extending this area doesn't alleviate the problem but makes it worse (extends nonlinear area).
This problem is different from crossover distortions or switching distortions. Of course huge amounts of negative feedback will make everything nice and linear but for the price (TIM) and even then still not as good as class A. The only strength of class AB is that it costs less and therefore better quality amp (power, parts etc) can be made. |
