>CFA exists till the early beginning of the audio electronic
All my power amps built in the sixties (a lot of them) were CFA's.
Probably valve amps?
These amps were single input device CFAs and begun to be replaced by differential input stages circa 1970 giving them much better overall performances. Then the push-pull input CFA topology appeared (circa 1990 ?) but did not overcome the market.
As far as I recall, the advantages quoted for CFAs through this thread and some others only concerned those having a push-pull input.
There are not many solid state amps having a single transistor input enjoying high reputation : Quad Current Dumping series, Nytech, but what else ? (the Renardson Mosfet amp can also be listed but it has a Sziklai pair input giving it exceptional performances).
Last edited:
>CFA exists till the early beginning of the audio electronic
All my power amps built in the sixties (a lot of them) were CFA's.
Was there anything else at that time ?
EDIT : no, in fact, CFA had a different meaning.
Everybody thought them as VFA, there is the highest probability that you too.
And many still do.
And then in the seventies, eighties, nineties... ?
Last edited:
Edmond, I edited my previous post.
http://www.diyaudio.com/forums/solid-state/240712-cfa-topology-audio-amplifiers-77.html#post3662482
http://www.diyaudio.com/forums/solid-state/240712-cfa-topology-audio-amplifiers-77.html#post3662482
Well, I am not too much interested in complex NDFL either.
Only the simplest case, of Output Inclusive Compensation.
But your earlier work on NDFL should have provided some information about what maximum ULGF is practical around the OPS?
Thanks for the link. Fun for him and educational, but I prefer more simplicity.
Reduced distortion from more feedback, obvious way is to increase the ULGF.
Where is the limit?
I asked in another thread and have quoted the start here because you and Richard Lee probably missed it (for reasons that will be clear if you follow the link
). Best wishes
David
You are still not providing any authoritative reference about those commercial amplifiers with ULGF or 6MHz. Interesting and twisted way to protect the intellectual property: you are saying you provided the schematic, but you are not disclosing the manufacturers. Usually it's the other way around.
The Cyrus amps that use current feedback have ULGF under 1MHz. Check for example the attachment (it's in the public domain) and note the (no less than) 300pF shunt compensation. If you are aware of any Cyrus amp with an ULGF > 3MHz, please post a reference to the schematic.
Late 70's and early '80's.
As was pointed out and described in the beginning of this forum... feedback to a cathode or emitter (signal into grid or base) is not usually a CFB Amp design.
Thx-RNMarsh
Mosfets allow higher ULGF's to be used, but, I note that Bob Cordell believes the upper limy there is also about 3 MHz ( please correct me if I am wrong).
You cannot just pile on the feedback when you have a big fat pole sitting at 30 MHz, and in all practicality, much lower than this because the driver stage impedance a are not zero either. Further, you must consider the layout parasitics.
You cannot just pile on the feedback when you have a big fat pole sitting at 30 MHz, and in all practicality, much lower than this because the driver stage impedance a are not zero either. Further, you must consider the layout parasitics.
I'm running a simulated (using tian probe) 3.7Mhz ULGF in my CFA amp. Any higher than 4Mhz and there is a background oscillation of around 6v pk-pk. In sims there is no reason for this oscillation but in reality it is there. Seems more than coincidence this is close to Bonsai's limits.
Last edited:
Not my limits mcd99uk. Mother Nature's!
liquid nitrogen may help... or may be not.
It seems to me that the output stage pole is the problem once again. My attentions have turned to Vertical MOSFETs using separate drivers for each pair and maybe distributed output zobels. Do these Verticals hold the key to speeding up the output stage?
Hi David,
As far as the PGP amp concerned, I don't think the figures are of much value with regard to the present discussion. This is because this amp is a rather special case: a HEC output stage with lateral MOSFETs, which is not the preferred configuration for maximum ULGF. As a matter of fact, it was relative low, only 700kHz, and there wasn't much room for making it higher.
I did a quick and dirty sim with 2SJ201 & 2SK1530 as OP devices and 2SA1407 & 2SJ3601 as drivers. The front end contains only idealized components (two VCCS that kind of stuff). I used OIC plus a lead-lag shunt compensation to stabilize the OIC loop. The max. ULGF was 23MHz, the phase margin was 67 degrees and THD20k was 0.1ppm. For a real life amp however, all these figures are unattainable, of course. Besides, it started already to oscillate with 2nF || to the output.
Indeed, I missed it. So which thread is it?I asked in another thread and have quoted the start here because you and Richard Lee probably missed it (for reasons that will be clear if you follow the link
Best wishes
David
Cheers, E.
You probably saw my comment that Richard Lee's BJT OIC proposal had a Spice ULGF of 13 MHz (and PM < 60) so I am a bit sceptical
But much of the compensation is done by the poles from transistor capacitances and roll offs.
He says he did lots of work to tweak it(and perhaps had some luck) but the result is very neat.
Need to study it to work out what controls the poles and zeros.
That's what the little link after "Posted by Dave Zan" is for, it takes you to the post.
Best wishes
David
Last edited:
Some (including me) pretend to have noticed sound improvement with high slew rate amplifiers. Far from the theoretical needed limit to reproduce highest frequencies musical components.
Please, don't fire-me, i was very interested (and intrigued) by this and made a lot of tries for real on this aspect.
In the same time, an improvement in transparency is noticed when you set a strong low pass filter to the input signals. (200KHz ?)
It is kind of a contradiction.
So, what can be the reasons ?
Better rejection/immunity to the RFI/RMI/HF parasitic components ?
Less IM products ?
Something else like 'precision' (overkill sample rates in digital) ?
About the closed-loop bandwitch, according to the X100 margin law i believe we can target 2MHz for a power amp. Easy.
I Just want to precise that i'm not looking for incredible HF components for the pleasure. My personal enclosures are two ways, using a 1" driver for medium/trebles with a 16KHz limit. And any attempts to add a tweeter (with 40KHz of FC) failed.
Lack of coherency and 'natural'. It was like only distortions and disagreeable parasitic sounds were added. My system sounded suddenly 'hifi', see what i mean ;-)
This said, one of the important point in my enclosures is the spherical wave horn is designed to have the same emitting 'surface' at the crossover than the bass speaker, and the same directivity. Impossible with any tweeter. May-be only the reason ?
Please, don't fire-me, i was very interested (and intrigued) by this and made a lot of tries for real on this aspect.
In the same time, an improvement in transparency is noticed when you set a strong low pass filter to the input signals. (200KHz ?)
It is kind of a contradiction.
So, what can be the reasons ?
Better rejection/immunity to the RFI/RMI/HF parasitic components ?
Less IM products ?
Something else like 'precision' (overkill sample rates in digital) ?
About the closed-loop bandwitch, according to the X100 margin law
i believe we can target 2MHz for a power amp. Easy.I Just want to precise that i'm not looking for incredible HF components for the pleasure. My personal enclosures are two ways, using a 1" driver for medium/trebles with a 16KHz limit. And any attempts to add a tweeter (with 40KHz of FC) failed.
Lack of coherency and 'natural'. It was like only distortions and disagreeable parasitic sounds were added. My system sounded suddenly 'hifi', see what i mean ;-)
This said, one of the important point in my enclosures is the spherical wave horn is designed to have the same emitting 'surface' at the crossover than the bass speaker, and the same directivity. Impossible with any tweeter. May-be only the reason ?
Last edited: