Vintage Transistor Replacements

[jmcilkley]

There are a number of sites such as "alltransistors" which list alternatives for transistors and I also have compiled a short list of my own from what I have found on forum sites. It's a hard job using sites like "alltransistors" as they often give a lot of alternatives. You trawl through to find the closest match only to find that is also not available or else only at an impossible price - then you have to try another possibility and so on...
Is there a site where sensible alternatives are listed with consideration given to availability and price?
As an example, I'm currently about to look for alternatives for 2SC871 (maybe 2SC1815?), 2SA572 and 2SC671 - so any advice there would be great!

[stevehertz]

Good points you make there. Anecdotally some time ago I located a faulty transistor in a Sanyo receiver that I was working on. I can't recall details but on investigation the transistor had a spec with lots of specific parameters etc. I didn't have anything remotely like it so just for fun I soldered in a common or garden PNP transistor of an industrial type that I had from my days working at Thorn Automation. It worked perfectly, or at least it sounded perfect. My point being that it's surprising how much leeway you have with transistor replacements. Oh, I did replace the transistor with a 'proper' equivalent later on, more to satisfy my brain than my ears.

[Craig Sawyers]

I've been resurrecting a Tektronix 125 power supply. This has a 135V regulated supply based on an 85V reference, a -90V regulated supply, slaved to the 135V supply - both lines valved.

And a -6V heater supply, using Ge transistors and slaved to both the 135V and -90V supplies. Two TO5, a TO3 and a final power transistor in the weird and long obsolete TO36 package. All PNP of course. The TO3 had very low hfe - 4, whereas the spec sheet was 35 to 70. I replaced it with a rough equivalent (also Ge) which measured hfe of 135.

In the end, 135 line measures 135.5V , -90 measures -88.9V and -6 measures -6.03V. All of which have sub mV of ripple.

Why have I done this? To supply a UHF heterodyne wavemeter (for which the -90V is redundant).

Anyway, the point is that I shoved a near equivalent into the -6V regulator and it worked just fine and dandy.

Craig

[RichardGM]

I do sometimes wonder why there are thousands of types of transistor when many circuits will work just fine with almost anything that has adequate voltage, current and power ratings and appropriate fT.

[Boulevardier]

As a postgraduate I had a Saturday and summer hols job in the local TV and radio shop, and I picked up a little book from the Willow Vale rep on one of his visits: TVT-6 Transistor Equivalents (1975-76) -

I haven’t seen this book anywhere before or since then. The introduction made much the same point as RichardGM.
It gives as equivalents for 2SC871 - BC109, BC169, BC239, BC173, BC184,and BC 209, - sometimes with different pin-outs.
Mike

[Boulevardier]

Picking up on the other two the OP asks about, the book doesn't list 2SA572, but it lists as useful equivalents for 2SC671 - BD438, BD204, BD236, BD242A and BD188.

Mike

[retailer]

I've found the NTE (Nearest To Equivalent) book very helpful in identifying semis and at least to be able to point one in the right direction to find a suitable replacement - I don't think it is available any more but there is a mobile phone app called NTE QuickCross - I haven't tried it but I believe it does the same thing.

[jmcilkley]

I tried the NTE app but it just gave ERROR for all the transistors I tried!
The transistors I need to replace are on an old Pioneer quad receiver control amp board - so probably need to be good quality. So far I'm thinking that BC550C might be a good replacement for the 2SC871 and for the 2SA572 maybe the KSA992FTA.
I completely agree with RichardGM - what is needed is a cross-reference to the most commonly available alternative. Somebody must have that sort of information accumulated over the years.

[G6Tanuki]

Originally Posted by jmcilkley

I tried the NTE app but it just gave ERROR for all the transistors I tried!
I completely agree with RichardGM - what is needed is a cross-reference to the most commonly available alternative. Somebody must have that sort of information accumulated over the years.

Isn't that what T.I.T.S. - Towers' International Transistor Selector - provides?

In practice, for most things I only find I need to stock a few transistors:

TUP and TUN [as Elektor/Elektuur would call them].
BFY51/2N3053
2N3055
BU208
BD139/140
AC128/153/178
AC127
OC35/NKT404
AF139

For anything more-esoteric - like RF power devices - you need to be a bit more precise with your substitutions.

[radio69]

If I have to replace a small signal transistor, tbh, I just tend to chuck in whatever I have to hand. Even replacing Germanium with Silicon. You can blag quite a bit....

My trusty Levell voltmeter died recently, traced to a s/c transistor, I replaced with a 2N222 I had to hand, it looks a bit odd (!!) but it works just fine.
I've also recently replaced Germanium with Silicon in a couple of radios, without altering the bias resistors, and again, no problems.

A different matter, ofcourse with output transistors...

I find checking equivalent transistors on the web quite a faff. I have an ancient copy of Towers, and an old WW equivalents book, which are much easier to use.

[paulsherwin]

You are right, it's amazing what you can get away with. Just make sure the voltage and current ratings are OK.

I also reach for Towers if I need to substitute something, particularly for Ge types. General purpose silicon bipolars are pretty much all the same unless the circuit is particularly demanding.

[Radio Wrangler]

Some circuits do push devices near to one limit or another, and in this case they may need that specific transistor or one even better in the critical direction. You can usually get away with murder, but it helps if you can read the schematic do a couple of sums and spot the ones which are going to be difficult.

Voltage and current ratings are the ones you can't play games with. It might go close only once in a blue moon, but a device only gets to die once.

David

[paulsherwin]

It also depends on what's being repaired. A transistor radio with whiskery AF117s can take a lot more bodging than a hospital defibrillator or a flight control computer.

[G6Tanuki]

The one place I *am* a bit paranoid about when substituting transistors is "Current gain when Vce is low" - quite a number of transistors with seemingly-good current-gain at normal Vce see it die away dramatically when current is high and Vce is only a couple oif Volts.

Important both in series-pass voltage regulators and the power-output stages of audio amps, where you can expect the transistors to spend some time turned almost fully 'on' and can easily run out of base-drive if the gain falls away.

[commie1]

Originally Posted by G6Tanuki

The one place I *am* a bit paranoid about when substituting transistors is "Current gain when Vce is low" - quite a number of transistors with seemingly-good current-gain at normal Vce see it die away dramatically when current is high and Vce is only a couple oif Volts.

All bjt's have a sweet spot, e.g., a bc107 is 2mA whilst a 2n3904 is 10mA etc, if you under bias by a decade or so, the gain will drop, similarly if you over bias then the gain will drop off also.If under biased, then on the run up to the bjt's sweet spot means the gain increases as bias current is increased, ideal for a.g.c. operation.

[G6Tanuki]

Originally Posted by commie1

All bjt's have a sweet spot, e.g., a bc107 is 2mA whilst a 2n3904 is 10mA etc, if you under bias by a decade or so, the gain will drop, similarly if you over bias then the gain will drop off also.If under biased, then on the run up to the bjt's sweet spot means the gain increases as bias current is increased, ideal for a.g.c. operation.

I was thinking more of situations involving power-transistors with emitter-currents of more-than-a-few-Amps, in audio-amplifier stages or the sorts of series-pass regulators where you've got a little LM723 IC driving a 2N3055 which in turn drives an array of 2N3055 for the heavy series-pass stuff.

Even 'good' 2N3055 transistors will struggle to provide the needed gain in such applications when there's only a couple of Volts between collector and emitter.

Same happens with audio amps; when you've got several paralleled 2N3055s either side of a classic 'totem pole' output stage with the speaker connected to the centre-point, you can rarely get the voltage-swing to anywhere less than "supply-voltage minus two volts" because the typical drivers in these amps just can't deliver enough drive to the bases of the output-transistors to ensure they are fully-saturated when Vce is only arouynd 1.2V and the transistors' current-gain has become something pitiful like 4.

[commie1]

Originally Posted by G6Tanuki

I was thinking more of situations involving power-transistors with emitter-currents of more-than-a-few-Amps, in audio-amplifier stages or the sorts of series-pass regulators where you've got a little LM723 IC driving a 2N3055 which in turn drives an array of 2N3055 for the heavy series-pass stuff.

Ah, I see, the 2n3055 has been around for donkey years, off the top of my head, the 2n3055 in particular is a notoriously a poor performer, I think the only use for this device is 2A psu.?, there are better available now.I would never use 2n3055 bjt in an audio amplifier, heck if memory serves me correct, its ft is around 10kHz, crap!

[paulsherwin]

I don't think this is correct. 2N3055s were used in the output stages of some very well regarded HiFi amps in the late 70s and 80s.

[G6Tanuki]

2N3055 has always been a 'fun' transistor; available in Epitaxial and Hometaxial variants. Frequency/gain varied between different manufacturers and technologies.

Some were distinctly 'slow', others provided useful gain up to 2.5MHz - that's kinda the point I've been making - a designer in the 1960s may have produced an amplifier/regulator that worked just fine with transistors that met the original spec; but changes in the underlying technology may mean that a 'modern' replacement transistor supposedly of the same type can cause problems.

Replacing a series-pass regulator's "first generation" 1965-production 2N3055 whose characteristics were a bit soft, by a modern one which would work just fine as a power-amplifier at 2MHz, can lead to some problens when the new transistor turns your regulator into a power-oscillator!

"Same part-number" doesn't mean universal substitutability.

[Cruisin Marine]

I tend to mostly (if not totally) rely on old book references, the info. on the web is at times thrown together and not to be easily trusted.

Then there are the modern fakes, that is another subject that draws out my venom quite easily.