How to figure out what transistor is used?

[bblackwood]

Someone who knows about such things recently told me that there are devices made that can basically measure an unknown transistor and give you all the necessary info for matching it's manufacturer and part number.

So, I have an unpotted, unused Sontec HS1000 opamp. I'm thinking that if someone knows where to hire one of the pieces out, we can possibly figure out what the transistors are on the HS1000 - everything else is easy to figure out by simply looking at the PCB and following traces.

Any ideas?

 

[NewYorkDave]

There are devices that can measure all the relevant parameters, but it still takes a lot of old-fashioned human effort to pore over the databooks to find a particular type number that corresponds...

 

[Peter Simonsen]

[quote author="bblackwood"]Someone who knows about such things recently told me that there are devices made that can basically measure an unknown transistor and give you all the necessary info for matching it's manufacturer and part number.
Any ideas?[/quote]

Brad,

That someone..imho would surely be the same someone who would say that just buy a XXX-rated machine and you´ll get the true sound of some of the most famous preamps, Eq´s, and Compressors in the world !!! Or for that matter just buy a software "mastering analyzing program" with presets, and you´ll never need another mastering enginner in your time..

Now people talk to much these days... :grin:

Kind regards

Peter

 

[bblackwood]

Thanks, Dave, that's what I thought.

Actualy, Peter, this 'someone' is a designer that most here would cut their legs off to be able to clone. He knows more about circuit design than anyone I am aware of, but thanks.

 

[DrFrankencopter]

Unfortunately, even if you can curve trace a transistor and get all the other relevant parameters you still don't know what the selection criteria is. I think that Mr. Massenburg himself mentioned this the last time he was here.

As far as Sontec's go....I'd sub a Forssell JFET 992 or 993 for the op-amp and be done with it. Probably end up with better sonics too! I've looked at Fabio's sontec clone page and the Forssel looks like a great sub in that circuit.

Cheers,

Kris

 

[Peter Simonsen]

[quote author="bblackwood"]Thanks, Dave, that's what I thought.

Actualy, Peter, this 'someone' is a designer that most here would cut their legs off to be able to clone. He knows more about circuit design than anyone I am aware of, but thanks.[/quote]

Brad,

I´m sure this "someone" is good at what he does, thats not so much the isue..*S* (I´m not into "cloning" myself ;-))). What I dont understand is how that "someone" can tell you that devices are being made that can match everything in an unknow transistor..even down to manufacturer and part number. ..!!!


I´d really like to know about this "New machinery" your fia (Friend in audio) talks about then !!! Would it be to much if you asked him the brand of this revolution..I´m sure a lot of people would like to know as well.

Kind regards

Peter

 

[bblackwood]

I didn't say you'd know the part number, just that it could give you the info needed to cross-refernce and find it.

Yeesh, never-mind. Figured you guys would be interested in building Sontecs that actually sounded like Sontecs.

 

[NewYorkDave]

The best way I know of to characterize a transistor is with a real curve tracer, like one of the deluxe Tektronix models with high-current and high-voltage adaptors. But even then, you end up with a lot of raw data that must be sorted through and cross-referenced by a human.

I'm not saying it can't be done; it's just that it's not an "automatic" thing.

 

[Peter Simonsen]

[quote author="bblackwood"]
Yeesh, never-mind. Figured you guys would be interested in building Sontecs that actually sounded like Sontecs.[/quote]

Brad,

Please..and I mean this with good heart..dont let my posts stand in the way..I´m sure others around here would be very interested in cloning a sontec..that sounds like a sontec..*SS* It´s just that..as NewYorkDave said..a lot of raw data needs to sorted..you have to know the acturel design/selection criteria as well..its not an eazy task I´m affraid *S*

And again..please dont hold all the others around here against my posting..Best of luck with your cloning.

Kind regards

Peter

 

[DrFrankencopter]

Figured you guys would be interested in building Sontecs that actually sounded like Sontecs.

I'm interested in what transistors were chosen for the original, but as far as construction is concerned I'd rather start with the Forssell op amps, and would definitely choose different servo IC's. I'm pretty sure that most of the magic in the Sontec lies in the EQ structure and circuit; how the bands interract, not really in the filter amps and buffer/output amps. I guess I'm not really interested in direct cloning, but rather evolving classic designs.

Cheers, and good luck

Kris

 

[SSLtech]

>that can basically measure an unknown transistor and give you all the necessary info for matching it's manufacturer and part number.

I think that the answer is essentially "yes"... though we do rather repeatedly come back to the fact that you have to do a search once you've got the data.

-Sounds to me like Brad is aware of this; from his post: "give you all the necessary info"... Not like the test unit spits out "BC182L"... just that it gives you some numbers that -after a lengthy search- you can match to the numbers on a page on the same line as "BC182x" and several other similar device numbers...

Keith

 

[Consul]

Although that would be pretty cool, and not impossible to do. It would just need to contain its own database of components and their specs.

... hrm...

Okay, I'm thinking too much again. :green:

 

[Scodiddly]

I could easily imagine a machine with a database of the NTE replacements. You'd still have to guess/infer/reverse-engineer the dissipation limit (or sacrifice a sample or three) and such.

But I can't really see NTE or a similar company building such a thing. Most NTE replacements are sold to replace burned-out components.

 

[Consul]

You could still build a web-based database full of components and their specs. Type in the important specs you need to match, and it spits out the part numbers for in-production transistors.

...

Okay, I really, really need to stop this thinking stuff. :green:

 

[rafafredd]

No need to build it. There´s some databases available with more than 130000 transistors.

 

[Mark Burnley]

I've seen the Atlas Transistor (and other 3-legged beast) analyser in Farnell and always wondered if they be any good:

Atlas Component Analysers

...check out "Atlas DCA".

It will supposedly allow you to attach any BJT/JFET/MOSFET to it with any lead connection, and it will tell you what type of device it is, which lead is which, and of what polarity.

...but can it??

And if it can, I want one!

:grin:

Mark

 

[hodad]

I think Brad's idea is a good one--not just for Sontec cloners but for Sontec owners. I am not quite sure what Peter Simonsen is whining about.
There are three key issues with Sontecs: first, I have yet to see a Sontec opamp with a standard 990 footprint. Not only that, I know of at least 3 different footprints for Sontec opamps. Second, there is zero support available from Sontec. I think actually their support measures in the negative numbers. Third, Sontec opamps are prone to failure. Very prone to failure.

I don't see any big problem with analyzing a Sontec opamp for cloning purposes. There are many owners of broken Sontecs (including myself--I have a nonfunctional mic pre) who'd love to have a decent replacement for the original opamps. I don't see what the big problem is with wanting a Sontec that a) works, & b) sounds like a Sontec.

And, although all part #'s are removed from the transistors, they are color coded. With a big enough sample of the transistors, it should be possible to get a good idea of what characteristics match with which color code. For my part, I would be willing to donate a mostly depotted HS 2000 (same footprint as the 1000) to be dissected--hopefully the knowledge gained could bring back to life some rather beloved pieces of recording equipment.

Tom

 

[Bauman]

hodad, I volunteer to reverse your opamp. Very interested on this matter.

 

[Peter Simonsen]

[quote author="hodad"] I am not quite sure what Peter Simonsen is whining about.
Tom[/quote]

I´M not sure either Tom...I´ll let you know when I´m really Whining..

It mightjust be the case that...Brad and I..err..hmm misunderstood each other..It sounded to me like some friend of his know about some devices that would "out of the blue" tell you all about transistor so you could clone..the sontec very eazy... I just tried to tell Brad that..no way !! could this be done very eazy.. to Quote what Brad said that let me to belive this !!!

I'm thinking that if someone knows where to hire one of the pieces out, we can possibly figure out what the transistors are on the HS1000 - everything else is easy to figure out by simply looking at the PCB and following traces.

Kind regards

Peter

 

[PRR]

> It will supposedly allow you to attach any BJT/JFET/MOSFET to it with any lead connection, and it will tell you what type of device it is, which lead is which, and of what polarity. ...but can it??

Fairly trivial. A BJT is two diodes. An ohmmeter will tell you which leg is Base and if you know the ohms polarity, if it must be NPN or PNP.

To tell Collector from Emitter on a BJT, put a low-current high voltage pulse across the reverse junction. Emitter junctions usually break down at 5V-8V, Collector junctions usually higher. --- hmmm, this box seems to only put 5V across things. I suppose you can measure Alpha and Beta both ways, the one with the higher gain is probably the right connection. There used to be symmetrical BJTs for chopper-amps which would confound this test, but those are probably all gone by now.

A JFET will have conductance from Source to Drain at very low voltage (a BJT won't conduct anywhere at very low voltage). The Gate looks like a diode to both other pins. Its polarity tells you N or P type. You can not tell Source from Drain because in most circuits it does not matter, and many devices are perfectly symmetrical. Only difference is some devices have a few pFd less capacitance from Gate to the Drain.

MOSFET will be a lot like JFET except the Gate will not conduct at any safe voltage. Source, Drain, and polarity might be inferred from the parasitic back-diode found on most but not all MOSFETs. If no back-diode, then you can get the polarity and enhancement/depletion info by curve-tracing, which could be just injecting a little hum in the Gate and checking if the output offset is high or low.