Are Silicon Transistors Fairly Consistent?

[hughesj]

I'm getting parts together for a Colorsound Power Boost clone and i came across these transistors:

https://wellgainelectronics.com/transistors/38157-micro-electronic-bc169c.html?gclid=EAIaIQobChMI1uK-vdyQ1QIVl7bACh0A9Q_2EAYYASABEgKhXPD_BwE

I know that you can never rely on new-production germanium transistors (fakes, low-gain, etc.), but does the same apply to silicon transistors? Also, has anyone used the transistors in the link before?

[Plexi]

Never... but still better than old germanium ones.
All that inconsistencies are with lower variations: but are there.
Almost every electronic compononet, have their inconsistencies: voltages, gain, hfe, etc.

You can check all the datasheets for min and max values intervals.


BTW: welcome! 

[greaser_au]

Modern silicon devices are a lot more consistent generally, because the manufacturing processes are more mature, and the material is not subject to deterioration like germanium is.

As for fakes, that's not just a germanium transistor problem, it affects many device types because of the end of production of certain device types (as I'm sure you're aware). There is a lot of traffic about this here and in other electronics groups. Most complaints here seem to have been with delay ICs and through-hole FETs. I guess if it's vintage, it's a potential minefield. The trouble is even some bigger suppliers have been bitten with supplies of fakes or faulty parts.  The supplier you've linked say they specialise in vintage/obsolete components, and their customer list is interesting (assuming it's accurate) - you woudn't last long making some of those names miserable!

Heading into off-topic here, sorry...     The nice thing about silicon transistors is that they are readily substitutable. The device you have linked is a 100mA, 300mW, 20V, very high-gain (400-800), high frequency (150MHz), NPN device.  Not sure where you are, but me being Down Under I'd suggest you might substitute a BC547 thru BC549 - in this case a Fairchild/OnSemi BC548C or a BC549C is quite a good match to the linked device and is current production. Watch the pinouts though!

david

PS  welcome to the forum!

[ElectricDruid]

Welcome!

I hate to say it, but the answer is "it depends...".

Yes, in many ways, modern silicon transistors are extremely consistent. I do a lot of analog synth stuff, some of which requires matched transistors. In the old days, this was often done by hand, measuring many transistors until suitably similar candidates were found. People here have done similar things with FETs for phasers and such like. Often for modern synth designs, it is enough to pull two transistors from the same batch and use them. Further matching might give an improvement, but you might not care that much for the minor difference.

That said, they're not consistent at all. Take the BC547 Plexi just mentioned:

https://www.sparkfun.com/datasheets/Components/BC546.pdf

One (very) significant figure on this datasheet is the DC gain, hfe. For this Fairchild Semi version of this transistor, they give the gain as min=110, max=800. Further than that, they won't commit themselves. They won't even go as far as giving a "typical" value within that range. It could be anywhere. That's not exactly my definition of "consistent".

The reason this is possible is that the parameter that might need to be matched might not be one with such a wide variability as the hfe. So it depends a bit what you're after.

HTH,
Tom

[greaser_au]

https://www.sparkfun.com/datasheets/Components/BC546.pdf

One (very) significant figure on this datasheet is the DC gain, hfe. For this Fairchild Semi version of this transistor, they give the gain as min=110, max=800. Further than that, they won't commit themselves. They won't even go as far as giving a "typical" value within that range. It could be anywhere. That's not exactly my definition of "consistent".

At the bottom of page 1 it shows these are binned depending on gain,  A: 110 ~ 220    B: 200 ~ 450    C:420 ~ 800  (each still about a 2:1 variance across the range - still not tight, but better than the 110-800 suggests!)

If I was to buy a bag of  Si devices and put them in a drawer in my home office today, and pull them out to use in 10 or more years time, there is unlikely to be much drift in characteristics.  Same goes for the NOS devices in the OP link. The same can't be said for a bag of AC128s, unfortunately

david

[thermionix]

their customer list is interesting

Smith & Wesson, Beretta, and...Kent State.

Hmmm...interesting indeed.

[dschwartz]

I'm not really sure, please correct me if I'm wrong, but i think for gain stages, you can make the gain of a silicontransistor pretty much independent of their hfe or temperature variations with feedback bias.

[balkanizeyou]

I'm not really sure, please correct me if I'm wrong, but i think for gain stages, you can make the gain of a silicontransistor pretty much independent of their hfe or temperature variations with feedback bias.

that's true in most cases, but it does not mean that silicon transistors are consistent - negative feedback is just a way of coping with the inherent inconsistencies of Si transistors

[bool]

I'm not really sure, please correct me if I'm wrong, but i think for gain stages, you can make the gain of said GAIN STAGE pretty much independent of the silicon transistor hfe or temperature variations with feedback bias.

A slight anally retentive iternet wordplay.