AXL amp

This amplifier was mentioned several times on this forum, but never had his own thread, and that is a pity. AXL amp was described in Elektor 03/1985, I find French issue, if somebody have English or German issue please post. This amp has all features a good amp should have:

-Symmetrical layout, to suppress even harmonics;
-High gain to keep distortion inaudible;
-wide open-loop bandwidth to ensure enough NFB well beyond audible spectrum.

Unlike Crescendo and Mini-Crescendo, both of them much more popular at the time, this amp has emitter follower before output MOSfet-s, so non-linear input capacitance (bootstrapped) and reverse capacitance do not influence upper end distortion figure.
Moreover, the Miller capacity of voltage amplifier section T11/T12 can be neglected, together with Early effect, due to the low output impedance of emitter followers T9/T10, so dominant pole is determined only by C7/C8, which can be silver mica or at least polystyrene type. Input section is a double cascaded differential pair, again for wide bandwidth open loop.


  • AXL.jpg
    102.3 KB · Views: 1,259
Last edited:
I modified original amp replacing lateral 2SK/2SJ with easy available (to me) IRF540/IRF9540 VMOS, because of their higher transconductance. Thermal stability was not problem, I increased value of source resistors, and set idle current to 1.5A.Output MOSfet-s are doubled to stand power dissipation. Fairly large input capacitance is doubled, so I increased driver stage current just a bit to maintain output stage speed.From a cold start, idle current was much less than 1.5A, and it takes too long to reach desired value. So I added Vbe multiplier from the copper side to touch heatsink, and that was neither over neither undercompensated, just ok.Among others serve me well for 10 years and it's my favorite.Just want to share, and maybe receive some feedback.Greetings to all.


  • axlrev.jpg
    401.2 KB · Views: 1,196
Last edited:
I still have two fully assembled boards of the L-Mosfet version on boards designed by me. They work well. But I have never been able to bias them beyond about 200mA. Elektor had suggested an Iq of 1.5A for 15Watt Class A operation.

The Vertical Mosfet version looks interesting. Any comparison in sound quality with other well known amps? Thanks.
In fact, 1.5A quiescent current is enough to work up to 30W in class A at 8ohm load. With1A will work in class A do15W (8ohm), or 7W (4ohm). difference in sound immediately noticeable. The upper part of the spectrum, although not sharp like crescendo, for example, becomes softer. Bass somehow gets fuller and better defined both, which usually does not come together.
I meant that the maximum possible Iq was only 200mA at max. setting of the trimpot. There was no more scope for increase in Iq with my board and using the exact values shown in the Elektor diagram.

For trimmer maximum resistance, the voltage between the gates should be about 5V.With 2sk134/2sj49 should be about1.6V.


  • 2SK135_HIT.pdf
    117.7 KB · Views: 187
  • 2SJ50_HIT.pdf
    119.5 KB · Views: 106


2013-06-01 2:33 pm

this is my first post, I recently purchased a new AXL Akita AT30 amp, excellent clean channel but lost a bit on the overdrive when gain past 5. I replaced the stock chinese tubes (preamp unknown chinese make and Shuguang 6L6GC Power Tubes) with RFT (Telefunken) ECC83's and Tung Sol STR Power Tubes, the AXL really cranks it out now, excellent tone reckon I got lucky and bought the right tubes 1st try
Hello there, I am using this Elektor class A amplifier fore more then 30 years now. Today I want to upgrade it a little bit. First thing is the power supply. I made a pi filter with 2x 12.000 uF and resistor 0,47 Ohm. Two resistors 0,22 Ohm between power transformer AC output and rectifier input.Hugh difference in sound quality. I ordered a small power voltage regulator with LM317 and LM337 to separate the power between the preamp and final stage (2SK134+2SJ49). Can'NT wait to try it out....Last thing will probably be new Mosfets ECF10N20 and ECF10P20.