identifying a mystery transistor

[mordechai]

I have a few unmarked PNP germanium transistors in a TO-5 package.  They have no normal part number on them...just a code that reads 6460005-1.  Beneath that there's an insignia...looks like a C or G and a smaller I next to it, followed by another number: 7019.  That latter number varies from piece to piece, but the other number and the insignia are the same.

These have really low leakage and good gain levels.  Anybody know what they are?

[tonyharker]

Pictures?

[mordechai]

Sure, here you go.  Any ideas?

[Mark Hammer]

I recall back in the mid-'80s there was a master semiconductor cross-reference you could buy, thousands of tiny-print onion-skin pages thick, with what seemed like a 4" wide spine.  Not only did it list every semiconductor by number ever in production, but it also listed house numbers and their equivalents.  So if a company asked a fabrication house to make them a special order, you could find its equivalent in there.  For example, if you looked up EH1048, it would let you know that this was really a CA3094.  It also equated all versions of a given semiconductor with different prefixes.  HUGE, requiring a magnifying glass.  Made the Science Citation Index and large OED look like mere pamphlets.

It was a marvelous resource.  I don't know if some sort of functional equivalent is still in production as a digital database, but I gather that the paper version simply became too cost-prohibitive by the late 1980s.  Maybe there is someone in one of the tech shops at your school who still has one of those 30+ year-old ones sitting on a shelf gathering dust.  That's where I encountered it.  I sure wish I had one.

[mozz]

Looks like General Instruments, or also known as General Instruments Corp. The 7019 is the date code, 1970, 19th week.

Really low leakage, like almost none, it could be a silicon transistor. Being 1970, could be either one.

[Rob Strand]

Those numbers are likely to be OEM part numbers which are equivalent to some common part number.  It's very hard to track down OEM part numbers.

The pic actually shows a 1969 date.   So we are talking '69 '70.

It would be best to first work out if it is Silicon or Germanium using your multimeter or by setting-up a test circuit.   You could even work out the gain.

From that, the package, and the pinout you could probably narrow down the choices quite a bit by comparing against the GI data.

The other way to look at it. If you characterize the Si vs GE, Hfe, and maybe leakage for a GE.  You don't really need to know more to be able to use it for audio.  If you wanted to build a 30V or 50V driver stage then it would matter because you don't know the voltage rating or capacitances.

A common PNP silicon transistor in that package is a 2N2905/2N2907.

[mordechai]

Apologies for the length in time in replying.  The transistor is germanium, so that leaves me with a bazillion options still I guess.  I tried it in a Rangemaster and it sounds really good...that's why I want to find more!

[Mark Hammer]

I don't know if I was unconsciously biased by mozz's post, but without reading his post, upon looking at the logo on the transistor, my first inclination was a General Instruments device.

I will reiterate my suggestion that you see if you can find a tech shop on campus that might have the reference manual I described sitting on a bookshelf gathering dust.  The engineering school is probably the first place to try, and medicine after that.  Whoever stands to be most likely to have to deal with archaic equipment, either currently or in past.  Even if you found a 1988 edition - well prior to such things disappearing in print form or becoming subscription-only on-line affairs - there's a good chance you might find reference to it.

Alternatively, perhaps there is a catalogue of General Instruments devices somewhere.

[mozz]

Definitely General Instruments. I ran those numbers but chances are they made them for some certain company or product and you will never likely find out. I do see they made a lot of 2n1307 and 2n1309 so chances are that's what they are.

[Mark Hammer]

The reference book I referred to actually lists house numbers as well (i.e., parts made specifically for a contract), and cross references them to the actual company part number for the original.  That's how, as I noted in an earlier post that one could look up an EH1048 and find out that it was actually a CA3094.  I'm not saying it would be easy - it is a VERY big thing to flip through, the pages are complex, and is a few inches thick despite using onion-skin paper and 8pt font - but the chances are reasonably good it's in there and cross-referenced.  All of that assumes you can find a copy and they let you sit down with it for an hour undisturbed, magnifying glass in hand. 

There are a few big universities in Mordechai's city.  I am naively assuming that one of them has a copy sitting in the corner of a tech shop shelf.

[mozz]

I'd like to find out the name of that book. I use Sams transistor books, they give specs and which companies made them.

I do remember from my dad's TV shop he had a transistor book about 1"thick with lots of specs and part numbers, could have been for Japanese transistors , not sure. I probably have it here boxed up with 200 other books.

[esnabez]

Dug out my NTE semiconductor cross reference catalog from 1995. It doesn't have 6460005 listed, but 6460006 shows an NTE102 (PNP germanium in a TO5 package) as a suitable replacement.

[duck_arse]

I will use my 80000th to say - D.A.T.A. transistor data book.

[Mark Hammer]

Congrats on 8k!  It's a thrill when the odometer rolls over to a string of zeroes.

[willienillie]

my 80000th

uh...

[Mark Hammer]

He simply made a promise about what he'll do 72000 posts from now.  All good things come to those who wait.

[willienillie]

80k / pi =


how long do ducks live?