Faking an MPSA13

Dr Tony Balls · Post by **Dr Tony Balls** » 12 Aug 2009, 18:23

Does anyone know what 2 NPNs one would use if they wanted to make their own Darlington MPSA13?

I need something to keep me busy while i'm ordering the real deals.

earthtonesaudio · Post by **earthtonesaudio** » 13 Aug 2009, 02:38

Any small signal Si should be fine. They don't have to be matched or anything.

tabbycat · Post by **tabbycat** » 22 Apr 2015, 04:22

am in kind of the same boat. built an ampeg scrambler and need three mpsa14 or equivalent. only have two until i do my next bits order, but would like to test my build asap.

found the above post by tb re mpsa13 and this post by rg re mpsa14 on diysb "Hook up two 2N3904s in a discrete darlington. Hook both collectors together, hook one emitter to the base of the other. The collectors are the collector of the new darlington, the free base is the base, and the free emitter is the emitter. It's not perfect, but it may work until you can get an integrated darlington." https://www.diystompboxes.com/smfforum/ ... #msg397755

rg suggests 2n3904 to get an approximation as that what what that particular poster had to hand in this case, but am wondering if there are any other transistors that could be paired to get closer to the characteristics of the original mpsa14?

have read that the mpsa14 has 10k-20k typical hfe. i have a dmm that is supposed to read hfe but isn't doing so today, so am limited re working that bit out for myself atm.

i have 2n2222/2n2222a, 2n5088, 2n3904 for high gain options and bc107, bc108, bc109 for lower. also mpsa18. and a few ic opamps in case they can be adapted for this too.

many thanks for any suggestions.

tabbycat · Post by **tabbycat** » 22 Apr 2015, 05:15

btw, related to the above, was reading this (or trying to, tried to get the gist anyway) and wondered if there are any common stompbox applications that the differences noted between darlington and compound pairs would suit darlingtons being replaced with compounds?

Compound Pair Vs. Darlington Pair
http://sound.westhost.com/articles/cmpd-vs-darl.htm

mictester · Post by **mictester** » 22 Apr 2015, 09:12

The original Darlingtons in the Scrambler were "BC" types - not the MPSA13 at all. All the Scramblers I've built (and I've made lots of them over the last 35 years) used pairs of BC549C or BC109C (basically the lowest noise versions with the highest gain) and the pedals were always indistinguishable from the original, except they had proper bypassing and hissed less!

I did do a PCB for the Scrambler, using discrete transistors a few years ago. It's probably on here somewhere.

Incidentally, the "extra two" diodes that were omitted from some of the Ampeg originals are important - I'm fairly sure that their omission was a manufacturing error - because if you leave them out, the octave is much less pronounced, and the character of the distortion changes quite a lot.

One thing I did add to some later builds was a tone control circuit (3-band, using gyrators), with a "tone bypass" footswitch. This was a very useful addition (IMO).

grrrunge · Post by **grrrunge** » 22 Apr 2015, 09:28

For small signal purposes, any two decent SI-NPN transistors will do the trick. MPSA14 have a minimum specified ß of 10.000, so the product of the minimum specified ß for the two transistors should be in that ballpark.
At a quick glance BC550 seems like a suitable candidate for pedal building, just note that for higher current applications BC550 have a max continuous collector current of 100mA compared to MPSA14's 1A

EDIT: I don't know how i got confused with the wole MPSA13/MPSA14 thing, but MPSA13 has even lower ß at just 5000. Apply that to the above

mictester · Post by **mictester** » 22 Apr 2015, 23:33

https://www.freestompboxes.org/viewtopic ... Scrambler! shows the PCB I used for basic versions. The transistors were BC550C or BC549C in pairs for the Darlingtons. I made loads of these on Veroboard, using BC109Cs

tabbycat · Post by **tabbycat** » 23 Apr 2015, 01:28

thanks grrrunge and mictester, alas i don't have any bc550s either, so i will go with the 2n3904 as rg suggested and get some mpsa14 and bc550 in on my next order. there's always one rogue bit.

@mictester thanks for pointing me in the direction of your layout. i did see it last night night and admired its tube-map like symmetry. am using this one by mark at guitar fx layouts but they seem to agree in all important aspects, so all good.
http://4.bp.blogspot.com/-6z4qd6u2mRo/U ... ambler.png

and found this today re the bc550 and 2n3904:

"2N3904 Vs BC547

Its electrical properties compare much more favourably with the BC547, where Vcbo = 50 V, Vces = 50 V, Vceo = 45 V, Vebo = 6 V. This is similar to the characteristics of the 2N3904 where Vcbo = 60 V, Vceo = 40 V, Vebo = 6 V.

However, the BC550 is an even closer match with Vcbo = 50 V, Vceo = 45 V, Vebo = 5 V.

For non-critical low voltage applications, it is often possible to make a direct replacement on some circuits; if you have a hobby project to light a couple of LEDs, or drive a small relay, then it is generally fine.

However, please note, that the pinout orientation is different between the two transistors. The 2N3904 pinout is emitter, base, collector, whilst for the BC range, the pinout is collector, base, emitter. The BC replacement transistor would therefore have to face in the opposite direction."

http://www.petervis.com/electronics/2n3904/2n3904.html

which makes me feel that a discrete darlington 2n3904 might not be so far from a discrete darlington bc550 as a substitute for the mpsa14 as i thought it could be. i only want to get the scrambler rolling so i know the board is fit to be boxed up. will socket and sub the proper mpsa14 at a later date.

btw, have you or has anyone you know tried a scrambler with germanium diodes or opamps for the transistors? could be interesting for extra-nasty?

grrrunge · Post by **grrrunge** » 23 Apr 2015, 05:20

No problem

The cool thing about darlington pairs is, that you can combine different transistors to achieve desired behavior.
Let's say you need a transistor capable of handling 8A continuous collector current with a ß of at least 10000. Most power BJT's have rather low ß, so let's say the one you have has ß=50. You would then need a ß multiplication of at least 200 times before the criteria has been met.
Lucky for us: that's exactly what darlington pairs do. Multiply ßQ1 with ßQ2, and you have ß of your darlington pair. A small signal transistor with ß>200 capable of handling the given supply voltage conditions would then be a fit candidate for Q1, boosting ß to specified level, while the power BJT as Q2 will handle the large current draws.
We just need to remember that the voltage drop is two diode-drops across the BE junction for darlington pairs, compared to just one diode drop for a regular BJT. Otherwise them darlington pairs is one big mix and match party

grrrunge · Post by **grrrunge** » 23 Apr 2015, 05:25

By the way Tabbycat: I suspect cool things will come from your 2N3904 adventure. In my stash of transistors it's those and BC109 i like the best for fuzzy stuff

tabbycat · Post by **tabbycat** » 24 Apr 2015, 02:05

grrrunge wrote:No problem
The cool thing about darlington pairs is, that you can combine different transistors to achieve desired behavior.
Let's say you need a transistor capable of handling 8A continuous collector current with a ß of at least 10000. Most power BJT's have rather low ß, so let's say the one you have has ß=50. You would then need a ß multiplication of at least 200 times before the criteria has been met.
Lucky for us: that's exactly what darlington pairs do. Multiply ßQ1 with ßQ2, and you have ß of your darlington pair. A small signal transistor with ß>200 capable of handling the given supply voltage conditions would then be a fit candidate for Q1, boosting ß to specified level, while the power BJT as Q2 will handle the large current draws.
We just need to remember that the voltage drop is two diode-drops across the BE junction for darlington pairs, compared to just one diode drop for a regular BJT. Otherwise them darlington pairs is one big mix and match party

big thanks for this explanation grrrunge. normally an explanation like that would just leave me stunned. but you managed to pitch it just above my level of competence so that i didn't know completely what you meant, but i understood just enough to appreciate that if i made the effort to fill in the blanks (by reading up on what i didn't get) that i would know something really worth knowing. so i did.
and fixed my dmm so i can read hfe now. took my first hfe readings ten minutes ago:
2n3904 = 250ish, 2n5088 = 450ish, bc109c (metal can) = 500+ish. just a handful of randoms on my desk.
couldn't read the hfe for the mpsa14s i have. think they went off the scale of what my dmm can read. have read that they can have hfe of over 1500 so maybe 1000 is my limit. may be time to upgrade dmm.

but now that i can read hfe and have your equation i can get into doing the maths to give me an a x b = 1500ish total hfe for the pair. so feel a bit empowered instead of confused, which is a novel place to be.
there may well be more to faking an mpsa14 properly than just hitting hfe, but if i get that right then i can think about other factors.
and am learning v useful stuff in the process that applies elsewhere in stompland. so maybe being one mpsa14 short has been a blessing in disguise. necessity being the mother of invention and learning.

many thanks for schooling me on darlingtons.

grrrunge wrote:By the way Tabbycat: I suspect cool things will come from your 2N3904 adventure. In my stash of transistors it's those and BC109 i like the best for fuzzy stuff

have 2n3904 in my univox superfuzz clone, and that is intensely bad-ass. takes the paint off the walls.
but on just discovering that the bc109c have an hfe of double the 2n3904 i get a feeling i am going to find a use for them very soon...