Audio amp

Hmmmm, the way you apply the feedback directly to the input is what i do not like, the this is the job of the input OpAmp here with two PNPs. Check this and you have also a better tuneable overall gain. BTW, the 6.8Ohm for the sziklai is a little to low, i also change some more things e.g. the input transistor needs some DC bias, in your circuit it just gets AC and no DC op. http://everycircuit.com/circuit/5147928403116032. Why 4.6Ohm i have only 4.7Ohm in my sortiments.

correction you get a DC op, its coming from the output, hmmmm

Thanks for the input, hurz. I looked into your circuit last night and plan on perhaps modifying mine to match, today. Not all the features but several. You are correct about 4.6 ohm. Not sure why I recalled that value. They are indeed 4.7 ohm. Managed to find my 1 and only 1000uF cap so I included that in the circuit. Also took your recommendation and increased resistance at the output pairs. I needed to add 4.7 ohm emitter resistors there too because the thing kept trying to runaway. It doesn't sound too bad. I'm happy with the results so far. I plan on getting more involved in audio stuff. Have you read "Audio Power Amp Design Handbook"? Being an audio newb myself it seems really good.

It still needs work. Probably going to experiment with different emitter resistances to find the lowest I can that prevents runaway.

Changed the emitter resistance to 2ohm for now. No more updates for a bit lol.

The time i had big interest in audio power amplifier is 40years ago. At that time i never had enough money for a complete book. So i got a paper about a design here or i got some schematics and little text from electronic magazins. But there where never enough money for a nice book. Later as student i used Tietze & Schenk which is not for audio amps only. Its more the broad view of electronics. Later the internet was coming and helped a lot to get information cheap, today i have enough money, but not this fire of interest in audio amplifier. I anyway observed a change in designs. Class AB which was over decades state of the art will be replaced by cheap powerful and high quality Class D. I see an end of these traditional class AB designs, you even dont find them in todays consumer electronics anymore. All devices, Radios, Mobile Phones, TVs make use of Class D.

You have choosen the sziklai topology instead of standard darlington, were the emitter is virtual were the collector is of the power transistor. Or the emitter is presented by the emitter from the first stage transistor. Which is your choice NPN and PNP for the power transistor, 2n3055?

I chose the sziklai for several reasons, namely didn't want 4*0.7V biasing transistors, and my hope was to have the two output transistors share the load. I wanted them to share the load because all I have on hand today is 2n3904 and 2n3906 and they have a max sustained current of 200mA. In the future I will find a more appropriate transistor for the output stage. If you know of any audio quality transistors that you really like and would like to share, I'll probably buy some.

Plus, I built several amps back in college that used Darlington and I never got around to using the Sziklai in a project until recently.

I suppose I could have put some in parallel to break up the load but, eh I'm happy with it as is so far. Listening to Tool with it right now, and it sounds pretty damn good.

the idea behind sziklai is not to share the power onto two transistors. Its better in linearity and as you already said it needs a lower bias voltage. Read here a little about thermal stability is also mentioned http://everycircuit.com/circuit/6262141636837376

Great info. Bookmarked that page. If you find yourself curious of audio amps again, check out that book I mentioned before.

BD139 + 140 should do the job. Anyway, i like the idea to emulate a power sziklai with multiple TO92 2N3904,06. Lets see how that works. What you need is an external thermal compensation, your 220mOhm. And you need for each 2N390X an extra emitter resistor to put them in parallel. More resistors shoulded be needed, your extra 4.7Ohm i would put.

Going to order those for sure. The previous version, with two output drivers in parallel worked well for about 10-15 minutes on full volume (~2W) then one of them would blow. Sometimes it was the NPN, other times it was the PNP. Going to try this final version with three in parallel and see if it can last an hour or so. It should be better.

Going on 30 minutes and still working. Wouldn't consider this an awesome set up (some work to improve it) but it is functional, and a lot of fun. For those that are interested and want to build it, tweak it, etc. Give it a go!

why not adding little BJTs to get a bigger one. I remember we did that same thing with 2n3055 parallel them upto 10. I know even more is possible. If they are equal enough from production its possible BTW, mosfets do have a positive temperatur coefficient and can be parallel mounted without any aligement resistors. While multiple BJTs do need multiple more base current, Mosfets do add its Cgs with each instance. This makes the requirements for the driver different when putting them parallel.

I noticed your input signal level is 300mV, standart line levels are 1V RMS or 1.41Vpk how about in reality source voltage is the same as in EC simulation?

Hallo I don’t understand how to simulate a input and output jack!