ExtremA, class-A strikes back?

[jacco vermeulen]

Glen,

most here do not build such an amplifier design.
Most don't have the parts, and most certainly will not have had the necessary exercise to start with something like this.
The ones who have find the fully balanced operation more than pleasing.

[powerbecker]

Hello Sander,

thanks for the near perfect looking design!

May be it is possible for you to post the LTspice file here for those lounger like me, which like to test it but are to lazy to "build it

Regards
Heinz!

[LineSource]

Hi Sanders,

I would appreciate your design thoughts on the DC offset circuit from T25-T26 diff amp controlling the cascode pull-up bias.

1) How well does this circuit work at holding 0V offset?

2) Did you try other circuits like input diff pair current control, or the traditional summing an offset voltage at the negative diff input? Does your circuit function as well in terms of true 0V offset control?

3) your circuit includes a RC filter on the output bias, but does not include the traditional integrator capacitor on the diff-amp to filter the output signal. How stable is your offset control at high output swings?

Thanks for the education.

[unclejed613]

i'm guessing that since this design resembles an op amp, and has global feedback, that this amp (unlike many class A designs) has a pretty good damping factor.

[Pasi P]

Very interesting amplifier. Low power version could work well with active speakers (Orions) M&T, woofers driven by UCD
Some novice questions if you don't mind.
I suppose it can be used unbalanced as well?
How about input impedance, slew rate and power bandwidth?

[SSassen]

Heinz,

Shoot me an email (ssassen<at>hardwareanalysis<dot>com) and we'll talk LTspice, I'm not putting files up here for everyone to download for obvious reasons.

Best regards,

Sander Sassen
http://www.hardwareanalysis.com

[SSassen]

LineSource,

1) How well does this circuit work at holding 0V offset?

Very well, provided you give the amplifier a chance to warm up prior to setting the offset. Overall, even with a non-matched input pair, DC offset should never be more than about 15mV anyway.

2) Did you try other circuits like input diff pair current control, or the traditional summing an offset voltage at the negative diff input? Does your circuit function as well in terms of true 0V offset control?

Yes, we looked at a few solutions, this just works fine, mostly due to the fact than when warmed up the delta-T in the various stages will not change much.

3) your circuit includes a RC filter on the output bias, but does not include the traditional integrator capacitor on the diff-amp to filter the output signal. How stable is your offset control at high output swings?

DC offset at high output swings is not an issue, remember we're driving both the frontend and the output stages from two different power supplies. The output stage draws a constant current which will only affect the rail voltage when you're pushing it beyond class-A into class-B territory.

Best regards,

Sander Sassen
http://www.hardwareanalysis.com

[SSassen]

Unclejed163,

i'm guessing that since this design resembles an op amp, and has global feedback, that this amp (unlike many class A designs) has a pretty good damping factor.

Output impedance is about 20milli-ohms up to 200kHz. Because damping factor is calculated by dividing the load resistance by the output impedance we arrive at about 8/0.02 = 400.

Best regards,

Sander Sassen
http://www.hardwareanalysis.com

[SSassen]

Pasi P,

I suppose it can be used unbalanced as well?

Well yes, but that negates some of the benefits of using a balanced setup in the first place.

How about input impedance, slew rate and power bandwidth?

As per the schematic both inputs are driven into a virtual ground, as is normally the case with any balanced opamp (remember this is basically a power balanced opamp by design). Hence input impedance would be 2K2 per input. As for amplifier slew rate, gain is flat out to 200kHz, so I'll leave for you to calculate the slew-rate, same for power bandwidth.

Best regards,

Sander Sassen
http://www.hardwareanalysis.com

[lgreen]

Nice amp, thanks for posting it.