Another N-channel

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

I am not soo good at this but very intrested and trying to learn about amps. I found a schema on a HEXFET amp, using IRF540 /9540 as O/P. I´m trying to make this as an n-channel amp, using some small ideas from Zeta (IRF9640 as an inverter). I have not simulated the construction but I have made a real, working prototype.
Well, sound is comming out but is seems to be difficult to set the bias correct. If I set the bias to 150mA it seems to decrease quick to a lower value of 50-70mA with the BD139 mounted on the heatsink of O/P. If i take it away from the heatsink the same happens but slower. I think this circuit is not good optimised.
It works but with all knowledge here maybe someone can give me some advise to make the circuit more stable and correct!


Trying to post picture again....


  • nchannel.jpg
    64.5 KB · Views: 986
Thanks for your reply mr Clausen

So I take away the BD139. I am not so familiar with n-channel but as the original O/P was IRF540/9540 and it was impossible to bias the circuit without BD139 VBE. In this one I am using IRFP260 and the whole circuit seems to behaive "different".

So, I will continue and play music from my ZAP Player.....

I can not have a better source I guess....


Thanks for your answer Mr Clausen

Ok, I took away the 0R33.

I missunderstood you, I think that only in amps with lateral mosfet you don't need "BD139" on the heatsink to prevent thermal runnaway as they have negative tempkofecient (?!). In this circuit BD139 will also compensate for heat and reduse the current to prevent thermal runnaway, did I understand this correct???

I will decrease 2k2 to 1k and increase 100 to 470ohm if I understood correct. BD 139 is needed for thermal compensation, right ?

Learning all the time....

Is it a good choise to use IRFP260 ? I have some of this transistors...

Thanks for your advise. Anyway I changed resistors around BD139, 2k2 to 10k and 100ohm to 2k9 and trimmer to 4k7. I also changed the base connection of BD139. It got little more stable. I will run the amplifier for some hours to check.

I will try to add a 100ohm resistor to IRF9640 tomorrow.



  • nchannel2.jpg
    43.9 KB · Views: 528
hello Bengt

Good work - but could do with some improvement.

5mA input diff'l pair current could go down to 1mA or lower for a big reduction in input bias /offsets. try a 1K in the CCS

Keep the 0R33's and , if you have an oscilloscope, check the 10KHz sq wave, as the lower current input now has less degeneration and MORE LOOP GAIN for lower THD - partic to tame that output stage which is unlikely to be the picture of linearity. You may need some feedback lead comp.

You should have an RC to ground on the output 10ohm and 100nF say.
Hi Bengt,

Your circuit is very good & nice, but it suffers from rail loss of about 4 to 5 volts due to the high Vgs of mosfets. This type of circuit works best with Bipolars but with mosfets it is very inefficient.
The output voltage swing cannot reach upto supply rails.

All the Great members here knew this, but i donot know why they have not said this point. maybe something.......:)

Thank you for all support.

I added a R/C at the output already but did not in the schematic. I have an old philips oscilloscope and a tone generator and I will do tests as amplifierguru suppose later this day and try to post some photos.
Yes, this is a simple construction and inefficient. The rail is +-66v 500va unreg and I use 2x 22000uf. The reason I use only IRFP260 is that I bought lots of them at . I was looking to build Zeta but wanted to try if I could remake an other construction.

I´ll send pictures later,

Regards, Bengt
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.