25W Class A Amplifier

I really love my Pass Aleph 4's. It is probably the best amplifier that I have ever heard. This is purely an exercise using spare bits in my spare box.

Many many years ago I did listen to a Class A amplifier that was only 3W but did sound stunning. The Aleph 4 does have breathtaking dynamics with its 100W headroom.
 
Last edited:
Wow this beast gets hot - very hot

I finally completed the two amplifier modules and put one of them on test.

For such a simple design they do sound very good.

With the monster heatsinks that I have got they still run at about 56C biased at approx. 2.2A.

I've got no detectable hum with 15000uF caps per rail.
 

Attachments

  • 20200227_145900.jpg
    20200227_145900.jpg
    974.6 KB · Views: 574
Before anyone comments, there were a few mistakes in my PCB design which resulted in two of the resistors residing under the board.

I managed to fit the 2 x 100uF rail decouplers into the original design by drilling an extra hole in one of the tracks.
 
To stay in Class A, Iq must be 2.5A DC
No, not really. The amplifier has to deliver 2.5A peak, but when one O/P transistor is fully on and the other fully off. When they are idling, the current should theoretically only need to be 1.25A.
But older transistors as has been mentioned many times (in the JLH thread) have a non-flat gain:current characteristic. At lower currents the gain is higher, so the Iq tends to be more than half the peak output.
JLH and L Nelson Jones discussed this. Nelson Jones missed this point. JLH suspected this I think, but did not quite say so. He erred on higher Iq than theoretical, correctly.
As a guess I would suggest the Iq only needs to be 1.5A to maintain low distortion.
 
No, not really. The amplifier has to deliver 2.5A peak, but when one O/P transistor is fully on and the other fully off. When they are idling, the current should theoretically only need to be 1.25A.
But older transistors as has been mentioned many times (in the JLH thread) have a non-flat gain:current characteristic. At lower currents the gain is higher, so the Iq tends to be more than half the peak output.
JLH and L Nelson Jones discussed this. Nelson Jones missed this point. JLH suspected this I think, but did not quite say so. He erred on higher Iq than theoretical, correctly.
As a guess I would suggest the Iq only needs to be 1.5A to maintain low distortion.
I had them running at 2.2A and they did sound good, they did get rather hot though. My Aleph 4 runs at 4A and doesn't get quite as hot.

The Aleph 4 runs 6 x pairs of MOSFETs to give 100W / Channel. Maybe the heat is better dissipated through the greater number of devices.
 
Last edited:
No, not really. The amplifier has to deliver 2.5A peak, but when one O/P transistor is fully on and the other fully off. When they are idling, the current should theoretically only need to be 1.25A.
But older transistors as has been mentioned many times (in the JLH thread) have a non-flat gain:current characteristic. At lower currents the gain is higher, so the Iq tends to be more than half the peak output.
JLH and L Nelson Jones discussed this. Nelson Jones missed this point. JLH suspected this I think, but did not quite say so. He erred on higher Iq than theoretical, correctly.
As a guess I would suggest the Iq only needs to be 1.5A to maintain low distortion.
Not really. This is a single ended class A design - not push pull. The upper transistors form a constant current source and have to run at > 2.5 A to support full output power. Your 1.25A Iq is valid for a push-pull design where both sides can go up or down. That's the reason why single ended class A designs have such a lousy efficiency (125W DC power in, max 25W output)
 
The design is good but not as electrifying as it is promoted to be.

If anyone would like the modules and the heatsinks, please let me know. They are very heavy and located in UK. This is a completely working prototype but is unfinished and needs a case. 240V transformer included with PSU.
 
Not really. This is a single ended class A design - not push pull. The upper transistors form a constant current source and have to run at > 2.5 A to support full output power. Your 1.25A Iq is valid for a push-pull design where both sides can go up or down. That's the reason why single ended class A designs have such a lousy efficiency (125W DC power in, max 25W output)
John Ellis is quite right, the amp is designed non-switching, therefore it is designed so that neither output transistor ever switches off. :whazzat: Here is a simulation of the current through each output device. The bias is set where the curves cross, about 1.3 amp. This amp sounds good because it has zero cross-over distortion, at any volume level.
 

Attachments

  • current.jpg
    current.jpg
    28.8 KB · Views: 478
Last edited:
I've spent over 50 years listening to Hi-Fi and the Aleph 4 is by a LONG WAY, the best amplifier that I have ever heard.
The Aleph 4 spec. is also a huge 1% THD, though that may be a nominal figure to indicate that it doesn't pretend to be a hi-fi design.

According to his articles, JLH was actually trying to improve the technical performance with later versions of his original class A design. At least, he wasn't about to further sacrifice it in the pursuit of sound quality - a feature that has only emerged as a popular interest with a strong following in more recent times.
 
The Aleph 4 spec. is also a huge 1% THD, though that may be a nominal figure to indicate that it doesn't pretend to be a hi-fi design.

According to his articles, JLH was actually trying to improve the technical performance with later versions of his original class A design. At least, he wasn't about to further sacrifice it in the pursuit of sound quality - a feature that has only emerged as a popular interest with a strong following in more recent times.

JLHs 80W MosFet amp wasn't a bad amplifier. It was basically modelled on the datasheet of the lateral MosFets that it used.