Transformer Distortion Project

[PRR]

> maximum level of volts x seconds

Differentiator, transformer, integrator.

[D_Ex_Patria]

PRR, you seem to have a lot of knowledge. Am I close in the calculations I did, or am I still completely missing the boat?

[Rob Strand]

Differentiator,

The constant current drive I mentioned above essentially does this.

Look at some solid-state reverb drive amplifier circuit.
Accutronics used to have an applications note explaining this.

[PRR]

> Am I close in the calculations

I guessed 0.2H. You measured 0.16H. I think we are both right, and some clever engineer figured how to just-barely meet specs with a minimum of winding.

I think you are well into try it and see.

[vigilante397]

PRR, you seem to have a lot of knowledge.

Fixed that for you Seriously, Paul has helped me (and countless others) out on many a project.

[D_Ex_Patria]

> Am I close in the calculations

I guessed 0.2H. You measured 0.16H. I think we are both right, and some clever engineer figured how to just-barely meet specs with a minimum of winding.

I think you are well into try it and see.

 

Thank you so much for your help and insight.

[D_Ex_Patria]

More Progress!

I seem to have figured out my 386->380 problems, the root of which appears to have been the layout of my power supply. I essentially used the approach from ElectroSmash's 1WAMP (http://www.electrosmash.com/1wamp). The nasty screeching is vastly reduced, and I think a little more filtering between the two amplifiers, plus actually soldering stuff down, should take care of the rest.

It's time to put the transformer back in the equation, and add some heat management; those LM380s really like to heat up. Will update once I've had a chance to play around.

[D_Ex_Patria]

More progress! 



After messing around with some cap values, I've built the above on breadboard, and it works.

The final setup is the RLC high-pass suggested in this thread, plus a 330nF cap to ground after the transformer, which is acting as a low-pass. That low-pass has taken some of the high-end nasties out of the signal, which has left things a little darker, but more pleasant. The less-than-scientific approach I was using earlier (running white noise from an mp3 player to get a picture of the frequency response) gave me some insight into what that post-transformer cap is seeing, which is apparently ~240Ohms of resistance before it in the circuit. It seems to correlate with the combined DC resistance of the transformer's primary and secondary coils more than anything else I can see, but I'm suspicious of declaring that that's what's actually happening.

I'm going to have to find a heat sink, and a trim pot, because the LM380s are giving me some clipping and are heating up too fast for me to play for an extended period of time (I've been pushing them for about 30 seconds at a time, which probably isn't healthy for them, but I'm impatient).

Also, I have no idea if my original intention is happening; that transformer may not have enough voltage to saturate. For all I know, I've just built a needlessly complicated filter setup between the 386 and the 380s. A bridged setup, like the Little Gem mkII (http://www.runoffgroove.com/littlegem.html) might give me a little certainty, but I'm just thinking aloud at this point.

Thanks so much for your help so far, everyone! I'm super stoked I got this far. Onward to heat-sinking...