• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Elektor AXL V2 with Lateral MOSFETs

Hi metanastis and HerrFlick,

My progress so far and I have a question. How hot is the driver transistor? The PCB is quite small and I had to find suitable heatsink for that. Will the heatsink in the picture be sufficient? And if I put a rail capacitor in front of it (touching the heatsink) will it be a problem? Otherwise I need to find/make a heatsink that will extend off board to give room for the capacitor.
 

Attachments

  • AXL_photo1.jpg
    AXL_photo1.jpg
    146.2 KB · Views: 329
Hello JOHNEGO,
It all depends of the quiescent current you are using. I have a infrared temperature probe so I measured the temperature of my amplifier using a quiescent current of 0,2 A and all four driver transistors are about 25 ºC. The heatsink is 12 ºC/W.

Hello HerrFlick, thanks. 200mA is below the optimum stated by the document which is 350mA and metanastis uses 0.6-1A which he said is much better; so there will be big possibility that I will try that class A approach...

May be I will start with 200mA first and then learn from that. Or, I can use bigger heatsink but put it (with the transistor) hanging on the bottom site of the PCB :up:
 
What kind of power mosfets do you use?

Many. Mostly K135/J50. But the ones on the picture are BIPOLARS. Just there for part placement planning. Latfets are sensitive to electrostatics so I used bipolars for this purpose.

BTW the Rout are too close to heatsink. I may lift its position to avoid heating (if this is an issue). Currently it's 0.15 Ohm but later I may reduce the value as I have down to 0.005 Ohm.
 
Ok, I'am using ECF10N20/ECF10P20 in class AB, the application notes from Profusion tells me to use 0,1-02 A. There is a very interesting application note from Texas Instruments AN-1645 -Driving a MOSFET Output stage- I can recommend that to you. 27 pages of info about using 2SK1058/2SJ162, BUZ901/BUZ906, 2SK1530/2SJ201 and IRFP240/IRFP9240.
How to calculate a Snubber network, Bias etc.
 
Ok, I'am using ECF10N20/ECF10P20 in class AB, the application notes from Profusion tells me to use 0,1-02 A. There is a very interesting application note from Texas Instruments AN-1645 -Driving a MOSFET Output stage- I can recommend that to you. 27 pages of info about using 2SK1058/2SJ162, BUZ901/BUZ906, 2SK1530/2SJ201 and IRFP240/IRFP9240. How to calculate a Snubber network, Bias etc.

Yes, I think you have put a link to that document here somewhere. Latfets have zero temp coef. around 100ma+ so without heatsinking and class A consideration 100mA-200mA is about right. But it doesn't mean that it is the best approach. But it is strange that higher bias with latfets has usually little benefit if any. My preferred device for class A is actually HEXFETs where in class A their capacitance is not a big issue but there is clear benefit from their higher transconductance. May be an array of latfets is as good but that will be expensive.

None of these latfets pairs you mentioned are balanced between its channels. So balancing seems logical. But I have never observed improvement in sound quality with balancing the junction capacitances so I don't go the trouble of adding extra resistances or capacitances (the snubbers).
 
Last edited:
Try to adjust the quiescent current more than 750mA and listen to the different sound of that great amplifier. I have and I can assure you that there is a great difference between class AB from class A. If you use 2SK1233/34/135 & 2SJ48/49/50 you must match them. Otherwise you will not be able to increase the current more that 700mA t both channels.
 
Try to adjust the quiescent current more than 750mA and listen to the different sound of that great amplifier. I have and I can assure you that there is a great difference between class AB from class A.

Unlike Pass Labs' amplifiers, IME amps like this requires preamp to sound good. Otherwise it will sound rather "dull". But I don't usually use preamps. I have only DCB1 (buffer) here. What preamp do you use?
 
Mine has just an outstanding performance using NOS american tubes .It uses1 and half tubes per channel. I have copied many tube preamps like Audio Note, Cary, Audio Reseach etc. Marantz 7c is my best so far!

Yes, many liked the 7C but I didn't. I guess because my speaker was not suitable with tube stuffs. Or I just don't like any apparent extra noise added by any extra stages... Or I just don't like distortion... (but I have one tube stage that I like, after a DAC)
 
AXL Amp Review

AXL is an old amp, almost as old as Crescendo. But it took me time to try the AXL because many people said that it is comparable to the (mini) Crescendo. But now I know that it is to my taste better than the Crescendo, and especially when biased high, it is absolutely better.

AXL is better than I expected, but cannot compete with my other amps.

So I modified the AXL using all small signal transistor to copy the idea from NRDS amp by J.M Plantefeve who "modified" the Crescendo also using small signal transistors. The feedback return point is changed as used by common amps and there are many other modifications.

I attached the FFT of the AXL, my AXL mod and NRDS. I believe that most of the amplifiers sound quality can be seen from simulation. And from simulation, most of the quality can be seen from FFT.

The NRDS distortion is higher but it is in the style of the Hiraga super class-A that even though high distortion but many people like it and consider it as one of the best sounding amplifier. I have another amp that is better than the NRDS and have the same distortion spectrum but like the Hiraga, the bias is twice as high (and when I reduce the bias to 1.1A it is not as good as the NRDS).

My AXL mod doesn't look special in the FFT (even though it is better than the original) but it has something else that cannot be seen from the FFT.

Note: No simulation file will be posted.
 

Attachments

  • AXL.png
    AXL.png
    49.3 KB · Views: 250
Not that i like to criticise but i simulated extensively Elektor s Crescendo and most of its later iterations including the AXL and frankly that s quite poorly implemented designs.

The crescendo particularly, where they explain that they use some kind of shunt compensation by the virtue of the input capacitance of the lateral fets loading the "VAS", but it simply cant work due to the high standing current of the driving stage, hence an hasardous stability...

Also the AXL is an inverting amp and contrary to what is generaly thought this has an importance and require to invert the speakers terminals, that is, to connect the plus of the speakers to the the amp s ground.
 
Not that i like to criticise but i simulated extensively Elektor s Crescendo and most of its later iterations including the AXL and frankly that s quite poorly implemented designs.

The crescendo particularly, where they explain that they use some kind of shunt compensation by the virtue of the input capacitance of the lateral fets loading the "VAS", but it simply cant work due to the high standing current of the driving stage, hence an hasardous stability...

Also the AXL is an inverting amp and contrary to what is generaly thought this has an importance and require to invert the speakers terminals, that is, to connect the plus of the speakers to the the amp s ground
Hello,
If you read the old Elektor article, there is no reference in any part of it that the output is inverted.
When I built that amp, I used an audio signal generator and the signal at the output was at the same phase as the input.
Have you tested that amp and have seen the inverted signal at the output?
 
hello

I would like to build this amp with ECX10N20S and 10P20S. ECF10n20 and10p20 are End of life and little expensive.
The AXL has been a great amp in its era. That was my 1st amp I built when I was a student and it has worked perfectly in all of my dj activities( I had to lower the QC to get more power). I have read a lot of good comments for the Exicon mosfets and as HerrFlick has written, they are better than the old 2SK and 2SJ ones. It's up to you.
 
  • Like
Reactions: 1 user
Hello,
If you read the old Elektor article, there is no reference in any part of it that the output is inverted.
When I built that amp, I used an audio signal generator and the signal at the output was at the same phase as the input.
Have you tested that amp and have seen the inverted signal at the output?

Here the schematic of the AXL, that s a phase inverting amp, the input signal is sent on the inverting input and the non inverting input is grounded AC and DC wise, to retain the real phase in respect of the audio signal the speakers connection should be inverted, guess that you confused it with the Mini Crescendo wich is not an inverting phase amp.
 

Attachments

  • AXL-v2-schema.jpg
    AXL-v2-schema.jpg
    294.1 KB · Views: 41