CFA Topology Audio Amplifiers

Jump to Latest
K

kgrlee

Member
Joined 2011
O/P base R

Here are some 'real life' tests on this subject http://www.diyaudio.com/forums/solid-state/234408-stability-analysis-ef-output-stages-2.html

Note this is just a pair of O/P devices and 2 resistors.

If you use the evil Ccb cure like Cherry (and now also promoted by Cordell & Self) they have to be AT the output device pins. Sticking them on the VAS has minimal effect.

The 'real life' commercial amp was only 50W 8R so there's no investigation on the thermal effects that Bob goes on about with multiple O/Ps.

I've never used ferrite beads on the bases.
___________________________________

If you go through ASTX's thread on his 2stageef-high-performance-class-ab-power-amp-200w8r-400w4r there's a bit where we reduce the series resistance from the Vbe bias multiplier to the drivers to increase stability of the overall amp. :eek:

The cap across the Vbe multiplier and also between the O/P bases also has an effect on stability & THD. Don't assume making it a zillion uF is best.

This is not just SPICE world but are also 'real life' effects I've encountered in Jurassic times.
 
Last edited:
ostripper

ostripper

Member
Joined 2008
RNMarsh said:
Clipping (hard or soft) is a non issue IMO. Just dont do it. Scale and match system componenets so it doesnt happen. It should be an accident if it clips and should not blow up if it did get accidently clipped.

How good can you make an amp sound while clipping isnt a worthy subject.... worthy if it involves circuit/speaker protection etal.

THx-RNMarsh
Click to expand...

80-0-80 V rails and and 5-pair OPS's / 60Kuf (w HS's) is a 200-250$ USD
endeavor.

I can drop objects off the shelves with the cheap 150W H. badger , no clip.
If I clipped ,the wife would say "turn it the #$%# down" ....
She is my "input limiter" (all too clever- clipper) :D.
OS
 
Kindhornman

Kindhornman

Member
Joined 2012
The way I have always looked at clipping is if the amplifier has enough power to run at full output without clipping and the amplifier is sized so that it has more power than the speakers can handle you should never clip the system. If you have to much power output for the speakers and blow them up, you obviously have exceeded the output of the speakers. By the time you are approaching the limits of the speaker if you can't hear that then you have some other problems and it isn't with the system!
 
ostripper

ostripper

Member
Joined 2008
Your jaw is dropped by the grandeur of a powerful presentation.

Then a tweeter fails (aw shucks :( )
Get some powerful tweeters ... objects fall off shelves (for a while)...

Ooops - click.. X-max has been reached.

Very powerful amps have an appetite for drivers. Amp tech is
way ahead of driver tech ..
PS - so far ... Dayton drivers have survived the full onslaught of the 150W AB amp.

OS
 
Last edited:
3

31697B

Disabled Account
Joined 2012
I'm not referring to selling a commercial 50W amp. I'm talking to you guys here and in the context of DIY build/design for your home use.

We used to have meters and clipping/distortion indicator lights on amps to 'educate' the user. That is a worthy subject of design.

-RM
 
Last edited:
Max Headroom

Max Headroom

Member
Joined 2013
RNMarsh said:
I'm not referring to selling a commercial 50W amp. I'm talking to you guys here and in the context of DIY build/design for your home use.

We used to have meters and clipping/distortion indicator lights on amps to 'educate' the user. That is a worthy subject of design.

-RM
Click to expand...

Good point, and a feature that is all too commonly absent.

I used to run a modded/tweaked/improved old school high quality Aus built 500/4 + 500/4 PA amp into 4 modded/tweaked/improved large old school Phillips 12" 3 way cabinets (stacked inverted pairs).

I ran a modded/tweaked/improved CDP direct into this amp, and with the input volume controls at max, maximum peak level encoded on CD would only just induce momentary clipping according to the front panel indicators (1 to 5 seconds between clip events on the loudest passages).

This amp was for all intents perfectly clean all the way up to indicated clip, and well behaved coming out of indicated momentary clip.

The result was that these momentary clip events were quite inaudible !.
1000W of big, clean rock'n'roll sound at full tilt was serious fun !.

I do not know the attack/decay characteristics of the above clip indicator display, but it might be conceivable that a frequency contoured, specified attack/release processing circuit to drive an indicator would be the ticket.
Perhaps such an indicator threshold that is variable to give warning of approaching limits would be even more useful.

Clipping indicators are uber useful when pushing a system to the limit.

Dan.
 
W

Waly

Member
Joined 2011
scott wurcer said:
I observed it at the -140dB level by sticking a pair of diagonal cutters into an air core inductor. The coil was part of a 600 Ohm passive LC filter at 10kHz and the levels were ~7V rms level.
Click to expand...

A typical situation for an audio amplifer. We all know that diagonal cutters are flying around amplifiers, amplifiers are full of large air core inductors and, of course, -140dB is as audible as a gun shot in a church :rofl:.
 
J

jcx

Member
Joined 2003
jcx said:
Hofer found some distoriton in his AP sine generator when the steel cover was put in place - the production fix was to add patch of Al sheet to the steel cover over the oscillator region

the Eddy Current "shielding" of the Al sheet knocks down the B reaching the steel

a inverted application of what Ott, others show for severe EMI shielding by layering high conductivity and high mu materials - the Eddy current in the conducive material "shields" (or reflects) anything more than a few skin depth in frequency content, the iron can short out lower frequency that gets through

so maybe audiophiles need to line their steel boxes with thick Al
Click to expand...
 
W

Waly

Member
Joined 2011
john curl said:
It is best to avoid using steel chassis if you can. Richard Marsh is correct, and mumetal is even worse in this regard.
Click to expand...

This (steel chassis) was not the topic.

Mr. Marsh stated that MIL components (with heavy copper plated iron core terminals) are to be avoided, since they add distortions. As long as this statement cannot be substantiated, it remains what it is: another piece of anecdotic in the delusional audiophile toolbox.
 
dadod

dadod

Member
Joined 2006
I think you can't avoid ferric component terminals in some cases. It has to do with thermal material expansion, if component is encapsulate in glass then it needs terminal with similar expansion factor and there it come steel or chrome(I am not sure). Most glass encapsulated diodes or zeners show magnetic behavior terminals, and tube(in glass) terminals are magnetic. For most musical tube amplifiers there magnetic terminals, isn't it interesting??
BR Damir
 
ostripper

ostripper

Member
Joined 2008
Dang ! I might have to upgrade my CF returns to silver. But how could I
convince my Indonesian friends of the merits for this endeavor ?

They say they can hear no noise and the sound is wonderful. Could they
discern the difference's over the sound of the erupting volcano's ... :D

A HARD sell ... and I'm no salesman.

My thread has a little technical info injected and gets "slammed" ,
here we go from CFA ---> to materials ... all is good. :confused:

We should discuss basic CFA "stuff".
On the VSSA variant that Bimo prototyped , the input pair heats up. :eek:

He uses very low values for the FB divider and at very high OP levels the
IP pair heats up. I suggested a more "robust" IP pair (2sb647 / 2sd667).
With the smaller VSSA's this might not be problem.
Is this the "current on demand" phenomenon stressing the IP pair ???

PS - even with the heating , the amp was still thermally stable.

OS
 
You must log in or register to reply here.
Top Bottom