Inexpensive Dual NPN Transistor in Metal Can Package?

[Chris56000]

Hi!

I'd like to try experimenting with the circuit for the Audio Sweep Generator here:–

https://www.americanradiohistory.com/Archive-Poptronics/70s/1973/Poptronics-1973-10.pdf

. . .but I'm having trouble finding a dual NPN TO–78 transistor that doesn't cost an arm and a leg–can anybody recommend one?

Chris Williams

[Wolfgang]

Some question:

- Whats your budget ?
- an alternative to a metal can dual transistor could be a transistor array IC.

[Chris56000]

Hi!

Ideally one figure – < £7.50 (or less!) Including P & P.

The original article didn't discuss the theoretlcal aspects of the circuit, so not being exactly sure why one was specified, I can't think up substitutes!

Chris Williams

Would two of the transistors from a '3046 do? There was no pcb artwork given as the article said "these can be bought from. . .", so if a '3046 would do, I'd be more than happy to make up a design to use one in that section of circuit, or even wire it on "proto–board" to try it – when a home constructor calls for an odd transistor type, there's obviously a reason for it!

[Wolfgang]

I looked at your circuit, and I am sure that transistor arrays would be suitable.
Examples: CA3046, TPQ2222, CA3018 all on eBay.

[soldar]

Also you can probably just take a couple of paired transistors (with very close gain and other characteristics) and join them thermally with paste so they are thermally coupled.

[Zero999]

Matching V_BE is more important than gain. Shot the collector to the base, test using a multimeter's diode test function and use two transistors with as similar readings as possible. Glue them together so they're in good thermal contact.

[David Hess]

A pair of 2N3904s or even better, 2N5089s, matched for Vbe within a few millivolts and glued together will work.

[Wolfgang]

... of course this will work after buying a bag of them and fiddling around for an hour or two depending on required precision.
a good dual transistor costs a few € and a transistor array is a few ten €Cents. Your time - priceless ?! 

[schmitt trigger]

The Nexperia BCM847 is available at Digikey for US$ 0.45 in single quantities. It is available only in SMT format, but since very likely you will have to design your own PWBA, this should pose no problem.

On the other hand... I love the 1970s Poptronic magazines.
From this vantage point, looking back at the "future" of proposed Quadraphonic FM broadcasts, and the discussion of whether to open a portion of the HAM radio band to CBers, a quick smile comes to my face.
 
EDIT: The DigiKey ad on page 121 is priceless.

[David Hess]

... of course this will work after buying a bag of them and fiddling around for an hour or two depending on required precision.
a good dual transistor costs a few € and a transistor array is a few ten €Cents. Your time - priceless ?! 

I never buy jelly bean transistors less than 100 at a time anyway.  I have done this many times and it has never taken more than a few minutes to find a match when using the birthday paradox.

There are some other relatively inexpensive matched pairs from Diode's Incorporated and NXP but they are in surface mount packages.  The THAT parts are a good option if you can find them but they are not cheap and the Analog Devices MAT parts are so expensive that I consider them unavailable.

Mouser does not list any transistor arrays anymore.

[Wolfgang]

... eBay is full of transistor array offerings. I like the classic ones best, and their matching is a lot better than discrete single transistors.

[markce]

Hi Chris,

BCY89 is BP 5.50 incl P&P on ebay.

[spec]

Hi!

I'd like to try experimenting with the circuit for the Audio Sweep Generator here:–

https://www.americanradiohistory.com/Archive-Poptronics/70s/1973/Poptronics-1973-10.pdf

. . .but I'm having trouble finding a dual NPN TO–78 transistor that doesn't cost an arm and a leg–can anybody recommend one?

Hi Chris56000

The image below shows the short-form data for the Fairchild 2N49555 matched NPN BJT pair.

There in nothing very special about the 2N49555, going by its data and the application in the sweep generator. It is simply being used as a pair of matched super diodes, where the exponential VBE/ICB characteristic provide a logarithmic transfer function, and hence a logarithmic sweep.

There are many modern, low-cost, plastic-case, matched NPN BJT pairs that would do the job (a metal case is unimportant). One of the most popular NPN matched pairs is the BCM847xx (SOT457 is the largest case). But make sure you buy from a reliable source, or you may just get two transistor dies stuck in the same case.

UPDATE#1: just noticed that schmitt trigger has already suggested the BCM847 in reply #8 (I said it was popular )

https://assets.nexperia.com/documents/data-sheet/BCM847BV_BS_DS.pdf

[Kleinstein]

That circuit is full of EOL parts. The dual transistor is just one of them and relatively easy to replace with 2 matches transistors.

The ICL8038 is also hardly available anymore and here there is not easy replacement. The original XR2206 is also difficult to get an the Chinese copy is slight different (lower max supply,...).

[spec]

That circuit is full of EOL parts. The dual transistor is just one of them and relatively easy to replace with 2 matches transistors.

The ICL8038 is also hardly available anymore and here there is not easy replacement. The original XR2206 is also difficult to get an the Chinese copy is slight different (lower max supply,...).

It is a brave man who tries to build and test that circuit. I can only imagine that the OP is ding it for nostalgic reasons.

[Wolfgang]

Right, the ICL8038 is getting rare. If you are already on the trip to nostalgia, you could "emulate" the 8038 by a triangle generator and a diode/transistor waveshaping circuit to get a reasonable sinewave. 

[schmitt trigger]

I do find merit in a hobbyst attempting to build a vintage circuit, at least while most parts are somehow available.
It is a challenge, like properly restoring a classic car, and careful substitutions will require to be made here and there. It also requires a lot of elbow grease.
Godspeed to them!

Fortunately, the OP does not have PWBs, in which he would be forced to use a certain footprint for a component. Even then, one could still manufacture a small adapter board for those instances.

If it were me, the 8038 would be my biggest concern, as it is indeed the heart of this circuit.

[JPortici]

Matching V_BE is more important than gain. Shot the collector to the base, test using a multimeter's diode test function and use two transistors with as similar readings as possible. Glue them together so they're in good thermal contact.

i like this transistor matching technique
http://www.dragonflyalley.com/synth/images/TransistorMatching/ianFritz-transmat0011_144.pdf