Hows to select Darlingtons? I'm scrambled,

[Mark Hammer]

Several years ago, I made a couple of units that combined a modded Distortion+/DOD250 front end feeding an Ampeg Scrambler.  This combination produced some exquisitely beautiful and some very sick tones.  Because the Scrambler lacks any native sensitivity control, the Dist+ served as an excellent front end to precondition the signal to achieve a wider range of sounds than normally achieved, including some terrific Tychobrahe-quality octave sounds, as well as some Neil-Young-like tones if Neil was on a smorgasbord of psychotropic drugs and plugged one tweed Deluxe into another.  I could not stop touting the many virtues of this combination.

Flash forward a few years, and it has been absolutely impossible for me to replicate what I was able to with those first two.....and I have TRIED.  Lord, I've tried.  So, I'm in between doctor appointments right now, and thought I'd give it another whirl, swapping these transistors for those ones.  I simply cannot replicate those other ones. 

Now, when people build discreet fuzzes, transistor selection is often part of the magic recipe for getting the desired tone.  For the more plebian among us, we turn to our trusty DMMs and check the hfe of the transistor.  That is often quite enough to sift through the parts drawer and find the winning combinations.

But what about Darlingtons?  The Scrambler uses a trio of darlingtons.  I've used 2N5306, MPSA13, and MPSA14.  Nada, rien, garnicht, zip.  I certainly get lots of savoury overdrive and sizzling fuzz with any of them, but I simply cannot get the octaving and imploding sounds of my earlier attempts.  Unfortunately, every single Darlington is overrange for my meter.  So I can't even take the trannies out of the earlier ones and check them because I'll still get the overrange reading.

So how on earth does one select Darlingtons for such a purpose, or when the hfe is that high does it even matter what you select?  I need some guidance, here.  I know what awaits me at the end of the rainbow and I'm getting a little exasperated trying to reach it again.

[EATyourGuitar]

darlington slew is a lot more than BJT slew. I think the magic is not in the gain since the gain is Hfe 10,000 anyhow. run square waves through it at different frequencies and measure the slew of the original darlingtons from the working unit. try to get info on the manufacturer, batch number, and date code if you can. this happens to me too and it is really frustrating. all my fairchild PN2222 sound nothing like the last batch of some other maker that has now been forgotten.

[R.G.]

If I had to guess, I'd say it's not because of the darlingtons - at least not the darlington gain.  It's probably some side effect not normally associated with how the darlingtons and rest of the circuit are acting. Now that I think about it, it's probably the actions of the four diodes connected to the two last darlington bases that make it go when it goes.

I've not messed with that for over a decade, but wasn't the Scrambler supposed to make an octave as it sat? The only one I ever actually saw had two of the five diodes missing, looked to be from the factory. I was not impressed with that unit, but possibly that's because it was never giving out the sounds you got.

I'm going to be terminally busy for a week or two, but if you don't get it solved by then, poke me and I'll do some sim work.

Darlingtons can be measured by many of the same tricks as germaniums. It's just the DC gain that's tough, because you have to come up with some microscopic base currents and voltages. The leakage of a darlington is a big issue, as any leakage in the first device is magnified by the gain of the second, so you're seeing leakages of maybe 100-200 times a bare silicon instead of germanium's 1000 times. A back biased diode can make a good nanoampere current source, as the Millenium shows, so it's possible (I think...) to gen up some kind of measurement/matching setup. I have some old semiconductor test textbooks that may give me some insights as well.

[LucifersTrip]

But what about Darlingtons?  The Scrambler uses a trio of darlingtons.  I've used 2N5306, MPSA13, and MPSA14.  Nada, rien, garnicht, zip.  I certainly get lots of savoury overdrive and sizzling fuzz with any of them, but I simply cannot get the octaving and imploding sounds of my earlier attempts.

not quite clear on the question...are you saying you can't get any octave out of the Scrambler or you can no longer get the type of octave tones you were getting with the Distortion+/DOD250 front feeding a Scrambler?

The Scrambler, by itself, produces some of the most brutal octave tones and worked with any of the 4 or 5 darlingtons I tried (I didn't have 2N5306). There were small differences (degrees of octave) between the lower & higher gain ones I tried, but I settled for some random MPSA14's. I did not match Q3/4 and don't know if it actually helps to do so...

edit: I also did not match any diodes.

BTW, the Peak Atlas gives darlinging hfe's.

[Mark Hammer]

I'm a fair bit closer, though not all the way there, in terms of reliably getting both octaves from the Scrambler itself, and the implode-then-bloom tones I got from earlier builds.

I am using the Tonepad Revolcador layout, and had long forgotten that one of the caps is shown backwards on the schematic.  Reversing the polarity orientation of the cap helped a lot, but like I say, I'm only 95% of the way there.  I'm beginning to suspectthat a 2N5089 was likely a poor choice for the "normal" NPN, and will try out one with lower gain later this evening.

I still have to say that, even without the "magic tones", there is just a heckuva lot of flexibility to be achieved by using a clipping front end that can also vary input signal to the Scrambler.  Every little knob-nudge gets you something different and interesting.

[garcho]

The only one I ever actually saw had two of the five diodes missing, looked to be from the factory.

According to Francisco over at Tone Pad, that's the standard factory practice from the original models, despite the original schematic.

I was not impressed with that unit

I've never been impressed with any Scrambler I've breadboarded or built. It always sounded 'shriek-y'. Slightly less shriek-y with the 5306, but that might just be placebo effect. I'd love to hear how it sounds with Mark's front end driving it. How about that for suggestive, ha!

[CynicalMan]

that might just be placenta effect.

That sounds unpleasant.

[garcho]

HA! OMG what a typo.

[brett]

Hi
I don't know much about this, but if you ran 1 uA through a 10,000 hFE Darlington, you'd get 10mA, which it will tolerate and is easy to measure across a resistor.
Or you could piggyback them (no emitter-to-emitter resistor) and use a normal DMM tester. (Two BJTs with hFE~300 read about 10 like this)
cheers

[amptramp]

One property of a darlington is that it turns on rapidly but turnoff time depends on the recombination time of the second transistor in the chain.  Where people connect two transistors in a darlington configuration, they often use a resistor from the second base to the second emitter just to provide a path for the stored charge.  The quick turn on / slow turn off time of a darlington tends to turn a square wave into something more like a sawtooth, which contains strong even and odd harmonics.  Turn off time is generally not specified for a darlington, so you may get different results from two transistors of the same type.  It pays to design using specified parameters if you want repeatable results.

[blackieNYC]

The idea of attaching a gain stage or DOD250 front end as an I out to the scrambler - did anyone nail this down?  About to build a tonepad scrambler. Also, is the consensus that it is better without the omitted diodes?  Easily tested, just taking a little poll.

[Mark Hammer]

Once you push that sucker hard enough, I don't think the omitted diodes would do anything particularly noticeable.

And, just to follow up, since the thread was resurrected, I did manage to trace the problem with the pedal eventually, and the (self-made) board was bad.  Luckily, I had several stuffed-but-unwired units sitting around, and when I finally resorted to one of those, it fired up just fine.  Had nothing to do with the Darlingtons.