**Simple Power Amp** - suggestions are welcome

[mac]

While I was re-caping my amp I build this little power amp. It is a push pull amp that I guess can put about 1w to 2w on the speaker with higher vcc. Idle current is about 4ma on the BDs. They keep cool all the time, no need for heat sink.
By itself it amplifies a guitar to a good volume, and can be pretty loud for room use with a stomp box.
It is not spectacular but useful. Tips on the schematic.

I welcome ideas to make it beter sounding.



mac

[soulsonic]

How about maybe setting it up with a bipolar supply so maybe you can get rid of the output coupling cap? Overall it looks good to me for what it is.

[Rob Strand]

I can't comment on the sound.

The main thing I see is that the emitter resistors are quite high so you are throwing away a lot of voltage swing/volume (about 1/3).

Unfortunately if you drop them down too far the bias point won't be very stable unless you couple the transistors and diodes together thermally.

You should be able to fix one of the variable resistors then adjust the other to get 4.5V.   The bias arrangement isn't the greatest but if you put in some DC feedback it might change the sound.

I haven't analysed it but you could try connecting the 1MEGs to the output terminal instead of the collectors.

[mac]

I took the idea from:

http://www.techlib.com/electronics/audioamps.html

Second circuit.

I just added the input fet to increase Z-input, the BC559C and changed bias. First I tried the Big Muff stage for both BCs, it worked but a little less louder.

How about maybe setting it up with a bipolar supply so maybe you can get rid of the output coupling cap? Overall it looks good to me for what it is.

It's a good idea. Since BDs stay cool the center point, 0v, should be stable. I'm going to split the 9v to -4.5v, +4.5v.
I considered the idea of an OT but I do not have a decent one at hand.

The main thing I see is that the emitter resistors are quite high so you are throwing away a lot of voltage swing/volume (about 1/3).
Unfortunately if you drop them down too far the bias point won't be very stable unless you couple the transistors and diodes together thermally.
You should be able to fix one of the variable resistors then adjust the other to get 4.5V.   The bias arrangement isn't the greatest but if you put in some DC feedback it might change the sound.
I haven't analysed it but you could try connecting the 1MEGs to the output terminal instead of the collectors.

I'll try reducing gradually the resistance between BDs emiters, and connecting the BCs feedback to the output point and, why not, between both diodes. Also maybe a little cap from BCs collector to base to tame highs.

Thanks guys.

mac

[Rob Strand]

>I took the idea from:

If you look at the original there is a resistor from the base to the emitter, the resistor usually has a lowish value in relation to the base current.  That scheme will keep the DC output voltage at 4.5 in reasonably stable manner, without adjustments.   With the two transistor arrangement, like on yours, I have a feeling the two halves will fight against each other with this scheme.

[mac]

If you look at the original there is a resistor from the base to the emitter, the resistor usually has a lowish value in relation to the base current. 

I don't see a base to emiter resistor on the techlib link, maybe from base to collector or base to gnd?
The schematic there works fine, but the 1k to gnd shunts a lot of signal, so I changed bias for that reason. Then I tried the symetric BCs arrangement and it boosted the signal.

I did what you suggested, taking the feedback from the ouptput point to the base of the BCs. It works better IMHO. Anyway once you get 4.5v at the output, the same resistors values works if the feedback is taken from the BCs collectors. I set them at about 1M. It is not tricky at all, and stable. I also tried one feedback from the output and the other from its own collector, just to break the symetry. It works also.
I also decreased the emiters resistors to 2 ohms. The BDs are now a little warm. And more juice. Less resistance and the BDs get to hot with no more output to justify the excess heat.
Maybe the fact that the BDs are cool helps the biasing schem?

Should work on a clip.

mac

[Rob Strand]

> or base to gnd?

Yes, same thing (the emitters go to ground).


>but the 1k to gnd shunts a lot of signal,

Yes, the 1k is going a little overboard and it will do that.  You can use much larger resistors but it becomes less stable.  I'm thinking try >100k.  At the moment your circuit is like an infinite resistor there, so it has highest impedance but worst predictability (the pots compensate for that).

> feedback from the ouptput point to the base of the BCs. It works better IMHO.

Great. 

> BDs are cool helps the biasing schem

There's two possible things happening:

- The first is the lower emitter resistors increase the bias current through the BD's.  This is a result of the simple diode bias scheme.  I like to use diodes like this but in many instances the fact you cant vary the voltage across them causes this large emitter resistor issue, which affect s swing.  A better solution is to replace the diodes with a VBE multiplier (search web).  An alternative is to replace one diode with a schottky or a resistor (not always that great an idea).  Yet another alternative is to adjust the bias current of the previous stage, what that does is decrease diode drop and hence reduce the BD's bias current - this is only good to salvage a marginal situation.

- What happens is when a transistor heats up the it's Vbe drops and in that biasing scheme that makes more current flow through the transistor.  What is usually done to combat this is to connect the diodes (or VBE multipliers) physically together which makes the BD's heat the diodes, which in turn makes the diodes drop (tracking the VBE) and this reduced the heat-up effect.  With 2ohm emitter resistors I suspect this effect is minimal.  The larger the emitter resistors the easier it is to stabilize the bias current in the BD's regarding temperature effects.

[mac]

Finally I removed the emiters resistors. The BDs are still cool at 9v. Feedback from the output point.
I quiclky recorded two samples:

http://files-upload.com/342022/samples.zip.html

The output cap makes the difference. I used the Samwha 100x500 I recently replaced in my amp and two 6.8uf poly in the clean demo. A Chemicon 150x450 in the other.
guitar+pedal+breadboard+HH 12" speaker --> mic to mac.
No need to say what the effect is in the pedal+amp demo...

mac

[petemoore]

  Great rythm and tones on the clean, nice shimmer sound, sound pretty dern 'open', good bass.
  Pretty darn good.
  Amp seems to do Fuzz Face very fine also, I think the 'wool' is the nature of the FF with guitar volume up and during bassy chording.
  It's nice to see amplifier Diy-ing, this looks like a 'friendly build.
  I built a couple nice simple Mosfet amps, uses a source follower to drive the speaker.
  "Little buddy' amps like this are really cool !

[km-r]

for 2watts, i use LM380N which is think would be very convenient...

[mac]

Thanks pete. Going to try a source or emiter follower in the future.

I wonder how it would sound with power Ge, ie AC187k/188k.
I have some 2sd352 and some rescued ACs.
Sadly I do not have a higher voltage PS to experiment. Maybe with a 18v and no emiter resistors the BDs do not need heat sinks.

for 2watts, i use LM380N which is think would be very convenient...

Maybe the lm380 and even the lm386 give more power than this one at 9v.
But this is more fun, and have the advantage that caps and all the rest can be tweaked to taste.
My goal was to have it working.

mac