Finding a ASY28 equivalent.

[Diabolical Artificer]

I have a dead ASY28 out of a Wein bridge oscillator, see schematic, it's the only germanium NPN Q in the whole circuit. I've looked in my spares stash, and have nothing I can see that'll do the job as a replacement.

Firstly, do I need to replace it with another germanium transistor, or will a silicon type be ok?

Second, I presume apart from a similar or above voltage rating I want a similar hfe (min of 20 on datasheet). I've got a BSX32, which has an hfe of 20-30, might that do?

TFL, Andy.

[paulsherwin]

The hfe is unlikely to be critical. I would fit whatever you have and see if it works - it should do.

I have a few ex computer NPN Ge types in the junk box if you don't get anywhere.

[G6Tanuki]

Hfe of 20 is deeply unspectacular; you could probably get away with something like an AC127.

[Loads of such spectacularly-underachieving transistors were used in the 1960s/70s in telephone exchanges, computer switching-circuits and the like]

Alternatively, new-old-stock exists: http://www.buy-transistors.com/transistors/asy28.html

[Restoration73]

Towers lists 2N1304 as u.s. equivalent.

[paulsherwin]

The 2N130x series was used in a lot of 1960s American computers and industrial kit.

[Boater Sam]

The thread https://www.vintage-radio.net/forum/...ad.php?t=93231

may give you some help. Several equivalents given.

BC108 suggested as usable despite being silicon.

[Radio Wrangler]

The diagram's spectacularly unhelpful.

It tells us that the incoming DC voltage is either 24v or 50v but does not give the value of the zener regulated supply to the actual circuit, nor does it show the zener voltage.

Bit if the incomer is at least 24v then the rail to the circuitry ought to be plenty to support silicon.

VT7.... should get away with murder. BC108, 2N3904 any general purpose NPN Si ought to be fine If you add some resistance in series with the bias diode MR1 to set a comfortable bias current.

I'd add low value resistors in the emitters of VT6 and the 'emitter' of the VT7/VT8 compound pair, the increase the bias a bit further. THis circuit clearly dates from the age when germanium and wide thermal drift were the norm.

Plenty of volts, use them!

David

[Refugee]

I would avoid using a silicon NPN transistor as it would interfere with the bias diode.
It would work if you replaced all three transistors and the bias diode with silicon types.

[Diabolical Artificer]

Thanks all, don't have any of those, but it's a start.

Thanks for the offer Paul, will see how I go.

David, actual voltage is 12v ish, so zener is 12v.

The OP stage is quasicomplementary isn't it? With Vt5/6 and 7/8 connected as Darlington's. How about just replacing the lot with two silicon Darlington's? I suppose I'd have to alter the bias, FB R and metering shunts etc. actually I could just replace the lot with an opamp, no, bit drastic that. Just thinking out loud.

Back to the OQ, is transistor replacement just a case of finding a similar Vmax and hfe/gain rated part? Oh, and frequency of coarse, but that's only up to 20khz here. how essential is that I use a ge tranny?

Andy.

[paulsherwin]

It certainly depends on the circuit, but bipolar transistor substitution is rarely the critical business that is sometimes assumed. You do need adequate voltage and current handling, and a decent amount of gain at the application frequency, but the hfe doesn't usually need to be closely matched. Transistor manufacture historically produced very variable components, so it would be a very unwise designer that devised a circuit that was reliant on a precise figure. Ge types in particular were all over the place unless carefully selected (which to some extent they were).

Nowadays manufacturing tolerances are much closer, so manufacturers produce parts which can satisfy the ratings of a large range of parts and just mark them differently. For example, if you buy anything from the BC54x family it's quite likely to be a BC546 with the noise characteristics of a BC549, whatever it says on the outside. A generic 2N3904 is also likely to be identical.

Ge and Si parts do have different characteristics but you can usually get away with a Ge - Si sub.

[TrevorG3VLF]

The output transistors are connected as Darlingtons as said. They will need about 4 times 0.2V between bases = 0.8V.
Replacing these with single high gain silicon transistors will need about 1.2V between bases so MR1 should be changed, possibly to a pair of silicon diodes.

[paulsherwin]

As I said, I'm happy to send you a couple of ex equipment Ge NPNs FOC if it would make life easier.

[Diabolical Artificer]

I've tried a few si NPN's with mixed results. PM sent Paul.

Andy.

[John_BS]

PM sent.

John

[buggies]

Bit late to the party - but stores are in loft which is a mess.
If you are not already sorted, PM me an address I will acouple of ASY28 NOS in the post FOC.

[paulsherwin]

I've sent Andy a couple of NPN Ge transistors, an ASY74 and another with an obscure house code. Hopefully one or the other will be close enough to work.

[Radio Wrangler]

Because this is in an output stage biased by the drop in one diode versus the drops of a couple of Vbe's without any resistors, it really will be rather critical.

Needlessly critical. It could have been done with greater bias voltage and dominated by some resistors.

David

[Diabolical Artificer]

Thanks George, but I'm sorted for replacement's, thanks everyone. Will hopefully have a spare too for the future.

I'll get the thing working first, and will maybe have a look at improving it later. The OP in the schematic is coupled to quite a large OPT, max OP is a tad over 5v RMS. This is sufficient really as I intend to use it as a low distortion signal source, increasing the OP will probably increase THD, so counter productive. Anyhoo, an OP of 2v RMS max is all that's needed to drive most amplifiers into clipping.

Thanks again for all your help and advise, Andy,.