Q?: Designing with Darlingtons

[thehallofshields]

Hey guys,

I've got some mpsa13's I'd like to have some fun with. I've read that Darlingtons require something like 2x the voltage difference from collector to base that a single Si NPN would require. I'm not sure if that's the whole story.

What I'd really appreciate would be some examples of Darlington's being used as buffers or boosts, so I could get an idea of how to bias them for simple common-emitter or common-collector setups. It would be really nice to see how to adjust something like an LPB-1 circuit to take a Darlington, but any advice would be appreciated.

Are there any other tonal differences or pitfalls to watch-out for?

[smallbearelec]

I purposely designed the board for the Ursa Minor to accommodate either a single transistor or a made-up Darlington pair. Numerous possibilities! Please do check it out:

http://www.smallbearelec.com/servlet/Detail?no=1441

Regards
SD

[R.G.]

A darlington is literally two bipolar transistors cascaded, emitter of the first to base of the second. This setup was developed back when single transistors had low-ish gains, to get high gains with the available devices.

The idea is that the current gain of the first one is multiplied by the current gain of the second one. This works, but it's not as great as once thought. Unless you use dissimilar devices, the first transistor is probably working at low (for it) collector currents, which also gives low gains. Still, it's not hard to get current gains well over 1000 with discrete devices, and much, much higher with monolithic devices designed for the current levels.

A single transistor can have its base driven hard enough to saturate the collector-emitter to a fraction of a volt. A simple darlington connection can't. About the best you can do for a darlington is to get down to 1.2 -1.4V from collector to emitter. This is probably what your friend meant. Darlingtons don't saturate to as low a voltage as ordinary bipolars, because the base current for the second device must be passed from the collector of the second device through the collector- emitter drop of the first one.

The leakage of the first device is also multiplied by the gain of the second device. This means effectively that only low leakage devices need apply for darlington use. Monolithic darlintons like the MPSA13s should be fine.

Darlingtons do not give the huge increases in voltage gain one might think, because in most designs, gain is determined by something other than hfe. They DO offer really low loading on whatever drives their base, so they're well suited to buffer jobs. But the base of a darlington is not the 0.45-0.7V above the emitter, it's typically 1.0-1.4V above the emitter, as befits the base to emitter being two junctions in series.

As to tone, that's more an issue of the circuit you put them in; whether there's enough power supply voltage, whether the 1+ volts of saturation makes a difference, and whether the higher current gains make any difference at all or set you up for some of the tone-changing of simply having a high gain.

Still, not a big deal.

[amptramp]

Just to add to what R.G. said, darlingtons are not very fast in the turnoff direction.  There is no place for the base current of the second transistor to go, so getting the second transistor to turn off depends on recombination time.  Some people make discrete darlingtons out of two transistors and add a resistor to the emitter of the second transistor or some lower voltage to speed up the turn off.  Some monolithic display drivers that used darlingtons added this resistor to speed up turnoff and reduce the sensitivity to leakage current in the first device.  You rarely see darlingtons used above audio frequencies and some are at their limits within the audio band.  The ones you have should be OK.

[thehallofshields]

Okay. I'm doing my best to understand. So how can I bias a Darlington to take advantage of its gain but falll short of clipping.

I think what I'm really after is something simple like the LPB-1 Circuit adapted for the MPSA13.

[PRR]

> something simple like the LPB-1 Circuit adapted for the MPSA13.

With straight transistor the output can swing from 0.3V to 8.2V.

With Darlington the output can swing 0.8V to 8.2V.

The difference is NOT worth worrying about. 7.9V, 7.4V.... both are much too much for your amplifier, you get a Volume pot, you will trim that for the right result.

So WHY do you want to use a Darlington??

The real question is: what "improvement" do you hope to find with a Darlington? Lower output impedance? Higher input impedance? More consistent self-biasing? Lower distortion? Higher headroom? (You won't get either of the latter.)

Traditionally Darlingtons cost more. (In DIY economics, not so much.)

The bias used on the LPB was dubious enough with well-choosen simple transistors (every one sounds a little different). When you double the voltages and multiply the gains, bias is more likely to be wrong than right.

If you are stuck on a desert island and ONLY have Darlingtons.......

First change the bias so that it biases-up predictably. Within limits, a Darlington can bias very predictably. Study the original and get the output loop (B+, 10K, C-E, 390, Gnd) to the same levels as the original. Then scale the base loop to get those levels. Collector-base-ground voltage divider works against the relatively large Vbe, can set collector voltage, which is a more direct and consistent way to bias than LPB's 4-R scheme. It works out that the input impedance is a bit higher than LPB, which may not be bad; if you want the treble-suck off the guitar you can add a 100K or 50K de-pop resistor before the input cap.

[thehallofshields]

PRR you're great. Thanks for always whipping up little drawings late-night. This is probably the 3rd or 4th time you've done that for me. Don't forget to check out the Eclipse.

Anyway, This is just what I was missing. I've been frustrating myself trying to bias them, mostly by playing with the collector and emitter resistors. I wasn't expecting any improvements, just mad that I couldn't figure it out myself.

I don't understand what the purpose of a higher-gain transistor with no increase in headroom is... Just the Impedance difference?

Either way, I'm satisfied. Thanks again.