Compact Power Amplifier from Elektor

[jazzpeter65]

HI!

Has anyone here built the Compact Power Amplifier from Elektor may 1997?

Take a look at the schematic. It's quite special in many ways. It uses current feedback instead of voltage feedback. It also incorporates Toshiba IGTB's for drivers. These IGBT's are made especially for audio and should sound very good indeed.

Any comments?

Best,

Peter

PS: I can mail the whole article as a PDF-file if some of you would like to read it.

[KaLok]

I'm interest, pls email to "chan_kalok@sinaman.com", thanks a lot!

[hugonot]

I have build this amp about four years ago. There is a lot of problems with it. You have to carrefully match input transistors to get a low offset, and amp was unstable. I have burn a lot of these high price GT20D101 and 201, so if you want to build it at first buy a cheap power mosfets to make a tests. I have problems with parts, because I use an elektor pcb. If you want make this amp, I have an proffesional made boards boards ( 3 or 4 pices ). I really don't know how it's sounds. I never ran two channels at one time.
Today I see that output stage of this amp is not good. It is something call here as CFP ( or sth ). A lot of problem with thermal stability ( DC servo is not a good solution ), and has a large amount of feedback ( last two stages ), so going to be unstable.

[janneman]

I can't say anything about the above experiences, but pls note that the drivers are regular bipolars, not IGBT. Only the outputs are IGBTs. The feedback is actually voltage feedback (the output voltage is returned to the input side) and is developed into a voltage and combined with the servo output. Because of the servo, input matching should not be required for DC offset (it may give better performance sound-wise though).

Jan Didden

[Dave S]

CFP output stage using IGBTs is scary. IGBTs have 2 hf poles in their frequency response, when presented with a low impedance load one of them moves downwards low enough to cause issues with loop stability.

I have built follower o/p stages using these Tosh IGBTs and IMO they sound just like bipolars.

CFP can be made to work well - it's fine in the ESP P3A for example (althought that design needs an output inductor).

You should also be aware of the latch up possibility of these IGBTs under high current conditions - 20A or so - could be less if they are hot. I would recommend some sort of current limiting to be on the safe side.

Dave

[Dave S]

BTW the IGBT is already a CFP on its own i.e. a mosfet input with CFP connection to bipolar output (with a parasitic bipolar thrown in for good measure).

Does this make the Elector design a CFT?

[Samuel Jayaraj]

It is indeed a Current Feedback Amp as clearly explained in the Elektor article, not a Voltage Feedback type.

[janneman]

Well, we may have a semantic issue here, but clearly the signal fed back is a representation of the output VOLTAGE, whatever the Elektor article says. Isn't that the definition of voltage feedback? This will stabilise the output voltage.

Now if I have a power amp and put a resistor from the speaker return to ground and feed that signal back, it is a sample of the output current so I would call that current feedback, it stabilises the output current.

Jan Didden

[rickpt]

Well, we may have a semantic issue here, but clearly the signal fed back is a representation of the output VOLTAGE, whatever the Elektor article says. Isn't that the definition of voltage feedback? This will stabilise the output voltage.

No. this input stage is a current feedback one, it has 2 input nodes, one has high input impedance(the non-inverting input) and one has low input impedance(the inverting input) so this amp is controlled by current(that’s why the amplifier uses so low value resistors on the feedback network).

Now if I have a power amp and put a resistor from the speaker return to ground and feed that signal back, it is a sample of the output current so I would call that current feedback, it stabilises the output current.

this creates a high output impedance, and the amp becomes a current output amplifier or transconductance amplifier... it has nothing to do with the input stage...

Best Regards

Ricardo

[janneman]

Yes, I know all this, but isn't the important issue *what you feed back*? If you feed back current, it's a current feedback amp, no?

Anyway, I'll look it up this evening. Maybe I have it upside down.
As they say in CA's Governer's Mansion: I'll be back!

Jan Didden