Aksa Lender P-MOS Hybrid Aleph (ALPHA) Amplifier

Hello xrk/Hugh,

I opened up alpha for checking since I did not get any sound. It seems to be OK except the offset is around - 100mV. I know the was some guidance on how to reduce the offset but I am not able to find it in this large thread. Can you please help and let me know how to reduce the offset?

Thanks
Balaji
 
Yes, Anand, the BC547CBU subtypes are just fine.
They are just the same transistor, low voltage, high beta, high speed.

GnuB,
There is a string of resistors supplying current to the long tailed pair (R2, R3 and pot).
Try increasing R2 from 3k3 to say 4k7. Then measure the offset, meter the increase of the offset.
Then readjust to remove the remaining offset to within 5mV.
The stage current sets the offset.


HD
 
Last edited:
Anand,

Good to see you embarking on the build. The ALPHA 20 is a fine sounding amplifier.

As X has already mentioned, no start up adjustments it is a relatively simple build.

What chassis are you planning to use? And what speakers will you be using? Depending on the speaker impedance load you need to adjust the bias resistors a bit.
 
Zman & X,

Yes, I've read Danny99's comments about various source resistor settings. My impedance nadir between 100-200 Hz is about 4.8 ohms. Everywhere else it's above 6 ohms. Speakers are 96-97dB@2.83V sensitivity.

It's a 5U/400 chassis, with 800VA Toroidy/4 secondaries and dual SLB supplies.

Currently I am finishing the build with Aleph J boards. I'll listen to that first and then switch to Alpha 20. Then switch to Alpha Nirvana. Then make comparison of all 3.

Best,
Anand.
 
Zman & X,

Yes, I've read Danny99's comments about various source resistor settings. My impedance nadir between 100-200 Hz is about 4.8 ohms. Everywhere else it's above 6 ohms. Speakers are 96-97dB@2.83V sensitivity.

It's a 5U/400 chassis, with 800VA Toroidy/4 secondaries and dual SLB supplies.

Currently I am finishing the build with Aleph J boards. I'll listen to that first and then switch to Alpha 20. Then switch to Alpha Nirvana. Then make comparison of all 3.

Best,
Anand.

Hi Anand, I’m interested in the comparison you refer to above. Anything to report yet? Cheers.
 
Hi Raj,

Thanks for the post. Yes, it seems a bit high, I don't like to have more than 6nF but there are mitigating factors here.

#1 Class A amps by definition do not switch their outputs and this changes the usual rules. It is the transitions of on and off which tax the gate drive; but if the device is always on the gate drive is hugely reduced because the charge flow into the gate is reduced. In RF mosfets operating at very high currents the gate currents can approx 2A! To charge and discharge such gates you have to use very low impedance drive and gate stoppers are very, very low.

#2 Where outputs are used as source followers the main issue is Cgd, which rapidly drops with increasing voltage across the drain source. That is, at low audio output the Coss is very low, because the Cgs is bootstrapped since the source follows the gate potential and the Cgd is low since the Vds is high. This is a happy situation since at low output the Coss is very low since there is high voltage across the mosfet, and this maintains resolution.

#3 As a rule, you can halve the drive current with a Class A over AB; in fact, some designers use bipolar followers to move the charge into and out of the gate to achieve fast switching with AB amps. This prevents cross-conduction at high frequencies, which can destroy many output stages.

I think Anand will be OK although I would prefer to use outputs with gate capacitance of about half what he is proposing. But it should be fine; I have heard a AN 39 set up for 4R with large mosfets (AndyR) and the resolution and accuracy is wonderful, one of the best sounds I've ever heard.

Cheers,

Hugh