Univibe concerns

[Brossman]

I've been staring down my hand-drawn schemo for a Univibe.  It says all transistors are 2SC828, with the exception of Q1 (2SC529).  As every tranny in the bloody thing is NPN (I count 12 828's), and given the OVERabundance of the wholly INexpensive 2SC828 tranny it calls for (COMPLETE sarcasm), I am wondering what the outcome of a 2n2222 substitution might be...

Would I have to redo the entire circuit? or would the crazy thing even work properly, let alone, at all?

Thoughts?

[brett]

Hi
2N2222A subs for a lot of things.  Rarely fails.  But you might want a lower noise device like a 2N3904 or 2N5088.  Also, the PN100 is designed to be the "universal" SS NPN.
Do you have a link to the 2SC828 datasheet?
cheers

[RedHouse]

I've been staring down my hand-drawn schemo for a Univibe.  It says all transistors are 2SC828, with the exception of Q1 (2SC529).  As every tranny in the bloody thing is NPN (I count 12 828's), and given the OVERabundance of the wholly INexpensive 2SC828 tranny it calls for (COMPLETE sarcasm), I am wondering what the outcome of a 2n2222 substitution might be...

Would I have to redo the entire circuit? or would the crazy thing even work properly, let alone, at all?

Thoughts?

You'll be fine with 2222's or 3904's, better quality build with 5210's and 5088's.

[Brossman]

Brett,

Here's a link to a couple of dif datasheets for 2SC828...let me know if they don't work and I'll see what i can do...

http://www.datasheetcatalog.org/datasheets/700/489181_DS.pdf

http://www.datasheetcatalog.org/datasheets/185/489180_DS.pdf

Just for kicks, I thought I'd post the datasheets for the other NPN's y'all have recommended...

2n2222a -  http://www.datasheetcatalog.org/datasheet/SGSThomsonMicroelectronics/mXyzzyw.pdf
2n3904  -  http://www.datasheetcatalog.org/datasheet/fairchild/2N3904.pdf
2n5088  -  http://www.datasheetcatalog.org/datasheet/fairchild/2N5088.pdf
2n5210  -  http://www.datasheetcatalog.org/datasheet/fairchild/2N5210.pdf

NOTE: I'd recommend not using Firefox, as my FF browser doesn't let me open these pdf's.  If you have no trouble, disregard this, haha.

[R.G.]

I use Firefox. I'm using 3.6.8 right now, and it opens all the PDFs just fine.

I agree with using 2N5210 and 2N5088.

[RedHouse]

Just a heads-up about Fairchild 2N5210's.

I bought a load of them from Mouser about 6 months ago (I don't have the Mouser part number handy) and they are overly sensitive to base current input.

We had to raise the 22K input resistor to 28K to stop a blatty distortion problem when you hit whole chords hard on the guitar.

[jasperoosthoek]

That sounds a lot like the problem I had: Everything sounds ok but loud chords are distorted. Only in my case it was caused by oscillation of the input stage... I solved it by installing transistors with a hFE lower value. I wonder if what you did would have worked in my unit.

[Mich P]

Hi same problem here oscillation of the imput stage
I had three 2N5089 Hfe around 560 same batch
i have changed Q1 for a 2N3904 Hfe 195
and it's perfect now !
Mich P.

[jasperoosthoek]

Mine were about 280 (2n5088s maybe, not that important) and I changed them to about 190 (2n3904). But I've just bought some 2n3904s with an hFE of about 290... The input buffer was designed very poorly: Not one which would have worked, not two which might have been better but three transistor connected in a weird cross linked way.

[RedHouse]

That weird way was called a boot-strap arrangement like the FuzzFace used.

[Brossman]

Looking into more detail of the datasheets for 5088/89's, I've discovered they may be low noise but can have a gain of 350-1400...

It seem to me this is much too high.  Better quality build, maybe.  But will this produce a better effect?  And, come to think of it, who's to say what's "better", except the beholder...

Rather than go through the painstaking task of trying to decide which ones sound better (having to buy a dozen of each kind and swap them out in tranny sockets), may I ask a recommendation?

Based on the fact that 5088's have a noise factor (NF) of 2dB, and median gain ~800, while the 2222a has a NF of 4dB, with a median gain of ~200, I would think less gain is better overall...

Reverting to n00bage, would the fact of having quadruple the gain factor affect the audio signal too terribly much (i.e. more/less distortion)?
OR
Is it better to take a hit building with slightly noisier components?

Thanks for all your help guys!

[RedHouse]

...Rather than go through the painstaking task of trying to decide which ones sound better (having to buy a dozen of each kind and swap them out in tranny sockets), may I ask a recommendation?...

Been there, done that, that's why I recommended the 2N5088.

I've tried most of the usual suspects: 2N2222, 2N2369, 2N3904, 2N4401, 2N5088 and 2N5089, 2N5210, 2SA1015, 2SC1815 etc etc etc.

[Brossman]

Well, the only reason I ask is that people have said something about getting oscillation from higher gain tranny's...

Why did you pick it over the rest? Did anything in particular stand out from the rest? Or is this your preference?

[R.G.]

It seem to me this is much too high.

Can you expound on *why* it seems too high?   

Based on the fact that 5088's have a noise factor (NF) of 2dB, and median gain ~800, while the 2222a has a NF of 4dB, with a median gain of ~200, I would think less gain is better overall...

Can you expound on *why* less gain is better overall?   

would the fact of having quadruple the gain factor affect the audio signal too terribly much (i.e. more/less distortion)?

The rhetorical questions above were a tongue-in-cheek way of saying that almost all transistor designs rely on making the variation between parts not matter by using circuit techniques that make the gain variations immaterial as long as you have enough. Having to select parts means either that (a) you can't get good enough parts consistently no matter what you do or (b) you are for some reason unable to design one which eliminates the sensitivities to part variation.

Put it another way: if you have enough gain, it's easy to reduce the gain by feedback and get more predictable performance. It is rare that having too much gain is a problem in a transistor if you have enlightened design techniques.

Is it better to take a hit building with slightly noisier components?

It is never better to take a hit building with noisier components if you can get higher gain, quieter ones unless there is some very special purpose.

The Univibe circuit after the first three transistors is practically immune to transistor variation as long as you have enough. In fact, the two-device cascades are there because single device gains were not high enough. You can replace the two-transistor darlington connected pairs with single darlington packged devices like the MPSA12 or MPSA14 at very little change to operation. The three transistor front end has some quirks because the design had to be done with transistors available at the time. Replacing all with 2N5088 has been done many times, and quite successfully.

[R.G.]

Well, the only reason I ask is that people have said something about getting oscillation from higher gain tranny's...

Why did you pick it over the rest? Did anything in particular stand out from the rest? Or is this your preference?

Keen's Second Law is "When in doubt, use a 2N5088.".

This is a distillation of
(1) modern transistors are very, very good, and mostly any NPN will work for any other NPN as long as
(2) the gain is high enough, and over 300 is a good place to start
and
(3) the 5088 is a high gain, high frequency, low noise device which does almost everything well. And you can get them for $5.00 per hundred. What's not to love?

Hence Keen's Second Law.   

The exceptions are where current handling or voltage handling is too high for the 5088; this is almost never a problem in effects.

[RedHouse]

Well, the only reason I ask is that people have said something about getting oscillation from higher gain tranny's...

Why did you pick it over the rest? Did anything in particular stand out from the rest? Or is this your preference?

There has got to be something else going on with the oscillation problem. I've built well nearly 50 now and a greater portion of them have been with 2N5088 and haven't had a single one have oscillations problems in the front end. Maybe it's my layout I don't know, sometimes layout can be a problem.

I know that it's real common to mis-place the 47K and 4K7 resistors, and have had many try to use a 33pF cap where the 330pF one goes, and substituting a 1k2 where the 1M2 goes gives one fits. Another thing to check is if you have a ground plane under the board, sometimes with your board on the bench w/o a ground plane under it things can get goofy. I mounted a proto-board into a 6" x 9" peppermint tin and jsut toss the PCB into that when experimenting on the bench.

Who was having oscillation problems in the front end?

[Brossman]

WOW, great! Thankyou very much...

now to get my hands on a batch of 5088's...as I lack a commonly necessary tool for almost anything these days...a credit card.

BAH!

THANKYOU!

[Brossman]

Hi same problem here oscillation of the imput stage
I had three 2N5089 Hfe around 560 same batch
i have changed Q1 for a 2N3904 Hfe 195
and it's perfect now !
Mich P.

RedHouse I believe this is what your looking for...

[RedHouse]

Ok thanks

[R.G.]

Hi same problem here oscillation of the imput stage
I had three 2N5089 Hfe around 560 same batch
i have changed Q1 for a 2N3904 Hfe 195
and it's perfect now !
Mich P.

RedHouse I believe this is what your looking for...

Reducing forward gain does stop oscillation if the circuit has an oscillation problem to begin with. I've posted a couple of times about the 330pF to ground from the collector of Q2 causing a blip up in response at high frequencies, and changing this to 30pF from Q2 collector to Q2 base being much better theoretically.

That being said, I've built and seen many built with 2N5088 exclusively that worked fine. There are differences in batches and between manufacturers of 2N5088 - as there are with all transistors. I've measured hfe400 2N3904s, too. I heavily suspect the implementation and wiring in that example. But there is always variation to be coped with.