Power amp under development

Hi anatech,
I set 15v across 100R resistor, when i replace 100R with 10R, the voltage across 10R resistor is 2v.
actually it must be 1.5v.
when I set 15v across 100R resistor, everything is ok.
but when I replace 100R with 10R, I must reduce Votlage across 10R resitor from 2v to 1.5v, and this problem occurs and when i touch input, bias goes down to 0.6 volt, and when i connect input to GND, bias goes to 1.5v again.
I'm confused.
Thanks.
 
Last edited:
Hi chico67,
I connected MJE's heatsink to GND and changed pots, now that problem fixed.
Grounding the heat sink is what likely fixed it.
but I set 15v across 100R resistor, when i replace 100R with 10R, the voltage across 10R resistor is 2v.
actually it must be 1.5v.
why it happens, can anyone explain me why it happens?
Probably because you changed the voltage drop by a large amount. Stick the resistance you expect to have in there and set the bias with that value installed. Turn it down to 1.5 volt right now though.

-Chris
 
Hi chico67,
Wide bandwidth audio designs often need input filters to prevent high frequency signals from reaching the amplifier proper. The individual parts are faster than the entire circuit is, so you can have both output devices on at the same time. That tends to draw a lot of current.

Try ferrite beads in the input circuit if an R-C circuit at the input doesn't work.

-Chris
 
hi guys !

this is my first post in this forum 😱

i want to build nmos200-TO247 version but before building i have a few questions :
1.can i use 2 pairs of irfp260 fets in this amp instead of one pair ? (for driving 4Ohm speaker) how should i connect them ?
2.what was the final decision about using an amber LED instead of r21 in this circuit ? does it improve sound quality ?
3.can i use 2sd669/2sb649 instead of mje340/350?
4.can i use 2n5401 instead of bc556 ?

forgive me for my poor english ..

cheers :cheers:

Welcome to the forum on your first post. The answers to your questions are yes, with serious warnings. Be aware that the substitutions you suggest have different pinouts and different voltage ratings and any changes may have unexpected consequences. For sure you will be restricted in voltage and must use a new PC layout and bigger heatsink. From a quick look, this design already has a supply voltage limit of +-70v based on parts limits, but you would have to drop that to say +-50v if you make the changes listed. That cuts power in half, to 8 ohms, but is probably safer on 4 ohms anyway.

One by one, here goes. It appears the IRFP260 is a fine upgrade, no problems, but for 4 ohm you must use about 3-4 pairs, and keep your parts cool with a bigger heatsink. Be aware that people recommend adding zener protection to the gates, but it's tricky because you don't want to mess with the gate drive much, keep it very close and don't add loads. About the LED, I suggest keeping with the proven original design. The 2SD669/649 has a different pinout and lower voltage rating of 120v vs 300v for the MJE350, so you must limit your supply voltage to under say +-55v, 50v is safer still. The 2N5401 has a higher voltage rating, good, but a different pinout.

So bottom line, you are messing with a proven design so you will have all the risks of making new changes, which can give odd parasitic behavior not seen in the original design. I'm not sure it's wise to take those risks unless you are an experienced designer. It seems from your question, you may not be experienced, so is this wise? You can try it, but for sure, you have new lower limits on power supply voltage, and you'll need a bigger heatsink for higher power dissipation.

To be frank, in your shoes I would not make so many changes. For 4 ohms, I'd keep it simple, maybe add a third pair of the irf250s, beef up the heatsink, drop the supply voltage and call it done.
 
Hi chico67,
Wide bandwidth audio designs often need input filters to prevent high frequency signals from reaching the amplifier proper. The individual parts are faster than the entire circuit is, so you can have both output devices on at the same time. That tends to draw a lot of current.

Try ferrite beads in the input circuit if an R-C circuit at the input doesn't work.

-Chris

Hi.
Thank you for your answer.
actually it is an uncommon problem i encountered.

Any touch by body and metals to input pin (1mF cap) causes the bias decreases.
I've try to fix it all day long, but i couldn't fix it 😕
 
Hi chico67,
So far it's very hard to say. You really need to look at the output on an oscilloscope to say anything definitive.

Bias and DC offset are normally set with no load and connected directly to the mains (no lamp in series).

-Chris
 
Hi.
Actually this is first time that I make Nmos500.
I've made very amplifier with no Problem, but this one made me angry!
As you say I set bias without bulb.

Thanks for your reply.

I have a question at the end :
After some minutes when mosfet's heatsink warm up, the bias goes down about 22ma per pair fets.
is it ok?

Thank you.
 
Last edited:
....................
I have a question at the end :
After some minutes when mosfet's heatsink warm up, the bias goes down about 22ma per pair fets.
is it ok?.................
this is termed "overcompensation".
neutral compensation would arrange for there to be no change in bias current for small changes in amient temperature and heatsink temperature due to variable power delivery.
Undercompensation is when bias current increases as temperatures rise.
But you have not said start and finish values.
522mA changing to 500mA would be a reduction of 22mA and could be ignored as insignificant.
25mA changing to 3mA is also a reduction of 22mA would be very bad.
 
I think i explain it badly.
When heatsink is cool the bias is 30ma per pair FETs.
After some minutes the temperature of hetsink is 40°C.
on 40°C the bias is 1.16V across 10R resistor and when i try to cool it by fan or anything like hardryer, the bias increases.

I want to know if there is something wrong with it, how to handle it.

Thank you.