Self regulating Class A

[Mark A. Gulbrandsen]

Circlotron's idea would work beautifully for pre-caluulating the bias needs and I'm surprised that Krell hasn't marketed that sort of thing as of yet..... That would make all the difference in the world.

Its very similar to how record cutting lathes use the preview head on the playback deck to vary the groove pitch needed based on level and frequency.
Mark

 

[Christer]

Circlotron's idea would work beautifully for pre-caluulating the bias needs and I'm surprised that Krell hasn't marketed that sort of thing as of yet..... That would make all the difference in the world.

It is a quite obvious idea, I think. I have had that idea too, and
I am sure many many other have as well. There are some obvious
problems with it though. First, it works only for digital sources.
Second, it cannot be implemented in the power amp, unless it
has digital in and contains the DAC, but must be implemented
already before the DAC(s). That is, it must be implemented both in
CDP and other sources and these must also provide an extra
output for the "presignal". It is not impossible to do. To the
contrary it is quite simple, but it is likely to be highly unpractical
unless there is some standard for it or all of the equipment are
designed to work together as a whole. A remaining possibility
would be an analog delay line, but that would probably be much
harder to implement with sufficient delay and without affacting
the sound negatively.

Come to think of it, it could be implemented for turntables by
using two tonearms, which I think would be something similar
to what you said about record cutting. Well, we might also
consider tape recorders with a tape loop, but most would
probably consider that good enough regarding sound quality.

 

[jewilson]

Sliding Bias

This could be done with a sample and hold circuit. This also could be done with a DAC and a look up-table; of course, that means a small processor and some logic. In addition, this could implemented with a comparators, threshold detector and a RC time constant.

 

[Mark A. Gulbrandsen]

A manufacturer like Krell that also produces their own CD players/DAC's could easily pull this sort of thing off. The CD player would simply have two analog and two digital outputs...one set of outputs would be delayed a tad bit from the other to operate a bias control input on the amp. The attack time of the bias control circuit would be fast but the release time would be much slower to take into account the DAC and preamp being used on the second delayed output. the amp wopuld also have the standard sounds bad dynamic biasing sceme available to work with other brands of equipment.

Mark

 

[Christer]

Sure they could. As I said, it is not particularly difficult to
implement if you control the whole chain. The question is, though,
how many customers have all Krell equipment, or all brand
whatever. It also precludes analog sources like turntables, unless
they are equipped with alternative methods to achieve the same
effect. Of course, the power amp could sense whether it does
get two versions of the signal and use sliding bias only when
it does get both a "presignal" and a signal, but then the benefit
is lost when listening to the "wrong" source.

 

[Nelson Pass]

It still seems that you're going to be faced with a lag problem. Take something like the Gladiator sound track that starts off on the soft side of normal and just keeps getting louder and louder and louder. Unless you purposefully build in a little overshoot on the assumption that louder is to come, you're still going to be playing follow-the-leader when the music exceeds your previous set point.

You are assuming that I was referring to a music-variable circuit,
which I was not. I was referring to a circuit which holds a
constant bias A or AB, and slowly adjusts the bias voltage to
keep that figure.

If you want variable bias, the original dynamic bias concept (mine)
is probably still the best because it tracks instantly - it's really
part of the gain stage, and as I have said before, Cordell and
Hawksford have used to reduce distortion numbers in the
output stage of AB circuits. From the schematics, I would say
it's the central mechanism providing the distortion numbers
for the Halcro.

 

[GRollins]

OK...lemme think on that one. A couple of ideas come to mind, but I'll need to ponder on 'em for a bit and see if they're practical.

Grey

 

[Mark A. Gulbrandsen]

"If you want variable bias, the original dynamic bias concept (mine)
is probably still the best because it tracks instantly "
____________________________________________________

Nelson,
Doesn't thermal stabilization and overall temperature have a big effect on overall sound quality in alot of amps including your designs? The Alephs don't really sound their optimum till they reach normal operating temperature which takes about an hour. While your circuit may act instantly on the bias level or mode, the thermal end of things will still take a while to achieve so there is that lag......
Mark

 

[Nelson Pass]

These are two separate issues. The characteristics of the
output devices are dependent on temperature, and typically
an amp should be allowed to warm up before being adjusted
for optimal use. In my amps, we reach higher transconductance
in the Mosfets at the higher temperatures, so there will be less
distortion. This is also where we do any adjustment.

Of course warming up usually also means a shift in bias, and
here we have been talking (I think) about means of automatically
adjusting bias, which implies temperature independence.

 

[jcarr]

Don't recall if US patent 4,229,706 by James Bongiorno has been mentioned, so here's the link:

http://patft.uspto.gov/netacgi/nph-Parser?patentnumber=4229706

regards, jonathan carr

 

[andreas w]

I've made it easier !!!


In my design I use an conventional emitter follower for output.
This works convetionally in the positive half. On the negative
half the PNP mirrors the load impedance, ie changes its sign.
The upper part NPN, sees the load as a constant current sink
connected to the VEE. Therefore the NPN side continuous to
control the load, that is not 8ohm or what ever but that current
sink connected to VEE. Simple as hell and it works, I'm listening
to it right now. The cirquit is stable as hell. nice bandwith and all.
Golden ears say it sounds class-A, and some say it to pure, LINN people. I would be glad to hearing from other experimentists.


Andreas