NPN transistor that consistently comes in around 200 to 230hfe?

[skiraly017]

I'm looking for a transistor than consistently falls into the 200 to 230hfe range. I ordered a bunch of ones that were supposed to be 220-ish from Mouser but they all seemed hover between 140 and 180. Any recommendations would be greatly appreciated, thanks.

[John Lyons]

I think it comes down to batch. Buy a small amount and then if they
falls into your gain sweet spot buy a ton of them as they would likely
be from the same batch. I have 2N2222a that are 250 and then some
that are 50sih hfe consitently but from different batches and makers.

[Chugs]

I find all the 2N3904's I buy fall exactly in that range.

[skiraly017]

I think it comes down to batch. Buy a small amount and then if they
falls into your gain sweet spot buy a ton of them as they would likely
be from the same batch. I have 2N2222a that are 250 and then some
that are 50sih hfe consitently but from different batches and makers.

Agree and usually do that but I'm having trouble finding one(s) that fall into that range and I'm slowly accumulating a nice stock of NPN's that fall elsewhere. 

[skiraly017]

I find all the 2N3904's I buy fall exactly in that range.

Really? I ask because the last large batch of 3904's were consistent but not in that range. May I ask how you're testing them? I'm using the Peak DCA55.

[Hides-His-Eyes]

That's really not a surprise mate; chances are all the transistors made on the same day will be about the same. Asking for a model of transistor with that hfe is just silly; nobody would ever go to the effort of producing a transistor with such a tight range.

Depending on how many you need, your best bet is just asking nicely if everyone will measure the handful in their stocks and hoping someone can sell you a handful, although I do wonder what they're for to be after such an odd range!

[R.G.]

Order from different sources (to get different batches) and test for gain. No other way.

It's that "variability" thing again.

[skiraly017]

Order from different sources (to get different batches) and test for gain. No other way.

It's that "variability" thing again.

I hear you and thanks for chiming in. I was hoping another board member had stumbled across some helpful info so I could avoid having to look at umpteen datasheets but I think that's what it's going to come down to.

[Gus]

You also need to know different transistor checkers can/will give different readings.  If i measure a transistor on a Sencore meter and a cheap DMM transistor checker I often get different readings.

Hfe varies with current some transistors more than others.  Do you know the current the meter you used works at?  Look up the transistor number spec (type it in google etc.)and look at Hfe vs Ic curve to see what I am posting.

Maybe the transistors do measure 220 with another meter.

Also keep track of the temperature in the room when you measure them.

[Hides-His-Eyes]

Order from different sources (to get different batches) and test for gain. No other way.

It's that "variability" thing again.

I hear you and thanks for chiming in. I was hoping another board member had stumbled across some helpful info so I could avoid having to look at umpteen datasheets but I think that's what it's going to come down to.

Even the datasheets will only tell you if it's possible that the transistor is in the range specified. Can we ask again, what are they for?

[skiraly017]

Order from different sources (to get different batches) and test for gain. No other way.

It's that "variability" thing again.

I hear you and thanks for chiming in. I was hoping another board member had stumbled across some helpful info so I could avoid having to look at umpteen datasheets but I think that's what it's going to come down to.

And give away my mojo?!  Surely you jest!
Even the datasheets will only tell you if it's possible that the transistor is in the range specified. Can we ask again, what are they for?

[Toney]

 Yeah I have had 2n3904's that sit right in that range and other fall above and below by considerable margins. I was chasing some ~ 200HFE a while back and pretty much did the same as you.
Eventually I bought four(more) test sets from three different stores and went home to test them. Even though they were the same brand (two of them) they all had different Hfe ranges although I realized the batch numbers were the key. The ranges were very consistent within the batch numbers. So thats it. Went back and bought up the "right" batch.

Another way to go- some Japanese manufactures sort their trannies into gain ranges with letter codes. I remember from the Superfuzz. I was trying to have it original and it was all about getting 2SC828's with the "Q" suffix ~Hfe200.

The European trannies are roughly matched too aren't they? the BC types are usually A,B,C.

[DougH]

 

This thread sure makes the case for designing circuits that don't depend on transistor hfe. Kind of the obvious elephant in the room to me...

[Toney]

No doubt Doug, but I assume Skiraly is recreating a classic whose nature changes dependent on Hfe.
Some of the simple two tranny jobs have much to offer sonically but are quite Hfe range picky.

I built the Zvex Mammoth and it absolutely sings with the "right" Hfe for the 3904s.

[skiraly017]

No doubt Doug, but I assume Skiraly is recreating a classic whose nature changes dependent on Hfe.
Some of the simple two tranny jobs have much to offer sonically but are quite Hfe range picky.

I built the Zvex Mammoth and it absolutely sings with the "right" Hfe for the 3904s.

Toney's got it right, I'm building something that has a terrific sweet spot with the gain range I'm trying to find. I've got a batch of 3904's that all tested between 140 and 170 and then a batch of 5172's that all fall between 240 and 280, it's that middle range that's eluding me.

[DougH]

No doubt Doug, but I assume Skiraly is recreating a classic whose nature changes dependent on Hfe.
Some of the simple two tranny jobs have much to offer sonically but are quite Hfe range picky.

I built the Zvex Mammoth and it absolutely sings with the "right" Hfe for the 3904s.

Toney's got it right, I'm building something that has a terrific sweet spot with the gain range I'm trying to find.

That much is obvious, but if the circuits were designed correctly to begin with you wouldn't need to sort transistors via hfe to get the "sweet spot".

[JDoyle]

That much is obvious, but if the circuits were designed correctly to begin with you wouldn't need to sort transistors via hfe to get the "sweet spot".

Not to mention, on stage, under lights, in a packed club - the increased temperature will cause the hFE to move out of the 'sweet spot' and you will no longer have the tone you want. Especially considering how small of a range you seem to need.

Additionally, the circuit will be nearly impossible to replicate. That which cannot be controlled must be made irrelevant.

The temperature issue is, I think, the main reason why I gave up on Fuzz Faces. I could get them to sound great at home but once on stage, where stage lights can easily cause the temp to reach the upper 90's, and a packed club can bring it up to triple digits, none of my FFs ever sounded even adequate. It might have been due to the difference in stage versus 'bedroom' levels, but I'm certain it was the temp.

Regards,

Jay Doyle

[DougH]

And biasing for temp compensation is such a fundamental concept, been in use for 50 yrs at this point.

[R.G.]

A bit of "amplification" on that variability thing.

A quick look at datasheets tells you much of this. The range of hfes in datasheets is never less than 3:1, and often 5:1 or even 10:1. That covers the range of yields from a given process. On a specific wafer, which gives several hundred transistors, the hfe spread is quite small. That's why analog circuits inside ICs can be different from discretes - in an IC, you are certain that the device gains will be nearly identical, because they all had the same processing and conditions. But from wafer to wafer, the gain may vary a lot. With discretes, same thing happens, but a wafer is cut up into several hundred transistors. All of those will be quite close, but the next wafer will be different. So if you get 100 transistors from the same bunch, from the same supplier, chances are they may mostly have come from the same wafer in the lot of 10,000 the supplier bought. They get bunched up in characteristics. But the maker will sell ANY device that fits in the min to max on the datasheets.

[amptramp]

I thought I was going to have to be the skunk at the garden party here, but DougH and JDoyle beat me to it.  Hfe is generally dependent on temperature and collector current as well as having the initial values vary all over the place.  In addition to stage lighting, you may be transporting the unit to the gig in an unheated van then bringing it inside and sujecting it to high temperatures from overcrowding (yeah, that's a nice problem to have) or stage lights as has been mentioned.  It is quite possible to design a differential stage fuzz that wouldachieve the same effect without such large variations.

If you look at the National Semiconductor Application Notes, AN-115 shows a function generator (which would make a good LFO) using an overloaded differential amplifier to "round off" the triangle waveform into an approximate sine wave.  This circuit could be used, with some biasing changes, to create a more temperature-invariant fuzz.  Unfortunately, the early app notes no longer show up on the National website, so you have to have the book.