News:

SMF for DIYStompboxes.com!

Main Menu

The lowly 2N3904

Started by Sparky, August 10, 2022, 05:41:42 PM

Previous topic - Next topic

Sparky

I never really got into 2N3904 [or 2N3906] transistors.   I mostly stuck with the BC clan and avoided the others.   I recently tried a couple 3904's and........they sound good.   They seem to be on the smooth side and have a low-ish  gain. 
What's the general level of experience with these?   Have I been missing out?? 

Rob Strand

As good as any silicon if you ask me.   In apples to apples comparisons I didn't find them to sound much different to my BC5xx's.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

Sparky

Quote from: Rob Strand on August 10, 2022, 06:15:04 PM
As good as any silicon if you ask me.   In apples to apples comparisons I didn't find them to sound much different to my BC5xx's.

I found a bag of them and they so far are around 200 HFE.   Ive been fooling around with some PN2369 [A] at around 70 HFE so it sounds like a good match for FF circuits.   Some of the BC's get up to 600- ish HFE and I need  a break from the ultra high gain stuff.  Have you tried 2N3053's ?   Really grainy / distorted quality.

Rob Strand

QuoteI found a bag of them and they so far are around 200 HFE.   Ive been fooling around with some PN2369 [A] at around 70 HFE so it sounds like a good match for FF circuits.   Some of the BC's get up to 600- ish HFE and I need  a break from the ultra high gain stuff.  Have you tried 2N3053's ?   Really grainy / distorted quality.
You might hear some differences with your BC5xx's.   In effect they look like C grade devices.  They could be ungraded with high-gains.

Off-hand most of my BC5xx's had gains around 300 to 400 (and as I recall number of them were B grades and other ungraded).   The 2N3904 were more like 150 to 300.    The extremes for the BC5xx's were over 400 and the extremes of the 2N3904 were down near 100.   The BC1xx's had gains a little lower on average than the BC5xx's.

I have tried 2N3053, some from the 70's some from the 90's.  Also many other similar parts from that era.  Also tried BD139s.   To be honest I can't remember how they sounded exactly, it's getting too long ago, but not bad that's for sure.

I specifically remember some very old Hitachi NPN silicons sounding very cool and unique.   They were in TO-1 packages like Hitachi germaniums.    I think they had quite a bit of base-collector capacitance.  The hFE's were in a normal range maybe around 200 - so the gain wasn't the contributing factor.

There's a stack of common TO-5 transistors from the 70's and the base-collector capacitances vary widely.  So do different brands of BD139's.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

Sparky

Don't the more modern  BC's have higher gain than the older original ones?

Rob Strand

#5
QuoteDon't the more modern  BC's have higher gain than the older original ones?
For transistors in general there's been a general increase from the between say pre 1990's to post 2000's.   A lot more low gain examples from the 70's.

For the BC's I don't have a lot of data because a good deal of the BC parts I get are graded (A, B, C) so by
definition their gains cannot increase as they have to meet the gain specs of the A, B, C grades in the datasheet.

For ungraded parts like 2N3904's it's easier to see the pattern.

If you buy ungraded parts then you might be seeing that upward trend in the BC parts.   Also if manufacturers sort A,B, C parts and there's an upwards trend in gain maybe there's less low gain parts to put into the ungraded parts.  I don't know anyone who has worked in discrete semiconductor production, so I'm only speculating.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

ElectricDruid

I used to keep BC54x and BC55x transistors around as well as 2N3904/3906, but at some point I decided to simplify things and that I only needed one NPN/PNP pair, mostly because I got sick of having layouts with different pinouts. I somewhat arbitrarily chose the 2N39904/3906.

They're good parts. They do everything you need a basic jellybean transistor to do. I use them for everything, unless there's some specific need for some special part. It's the transistor equivalent of the TL072!!

Mark Hammer

Elektor magazine had their nomenclature for project parts of TUP/TUN -DUG/DUS. Transistor-universal-NPN, Transistor-universal-PNP, and diodes that were silicon or germanium.  I gather this was both an encouragement to use what was available and cheap, as well as a recognition by an essentially "international" magazine that part availability would vary around the world.

Vivek

At university in the early 80s, I would have severe withdrawal symptoms if some student grabbed the latest issue of ELEKTOR and tore away some pages before I could get my hands on that issue.

ElectricDruid

I like TUP/TUN. Definitely describes my use of 2N3904/6.

Incidentally, a tup is a dialect word for a ram, a male sheep, where I'm from in the UK, and a tun is a vessel for brewing beer (a "mash tun"). DUG is the past tense of dig, but DUS I can't think of any meaning for!

Quote from: Vivek on August 11, 2022, 07:39:10 AM
At university in the early 80s, I would have severe withdrawal symptoms if some student grabbed the latest issue of ELEKTOR and tore away some pages before I could get my hands on that issue.
Tearing pages out?! That's terrible misbehaviour! Anyway, that's why libraries have photocopiers!

Mark Hammer

"Tup" is also, from what I understand, an older colloquialism for having intercourse, perhaps by virtue of its association with rams.  I would imagine that when Canadian "father of Confederation" Sir Charles Tupper visited England, the smarmy jokes were many.

iainpunk

i have always been a fan for using power transistors in non-power applications due to large capacitances, low gain, even worse linearity than signal transistors and their really low prices.
especially the BD139 keeps giving me great results, from a heavily modded fuzz face, to a solid state amplifier, its a great jelly bean

cheers
friendly reminder: all holes are positive and have negative weight, despite not being there.

cheers

duck_arse

Quote from: ElectricDruid on August 11, 2022, 07:41:52 AM
DUG is the past tense of dig, but DUS I can't think of any meaning for!


DUS is a term used in the broom-making industry.


I grew up with BC. when I was doing production, australia was built with BC's. I have some of the 2N's in the drawers, but [they are backwards pinned and] the numbers confuse me. 3906 and 3908? or 2906? 3904?
" I will say no more "

Phend

QuoteAnyway, that's why libraries have photocopiers!
Wouldn't work, we had mimeograph machines, loved the smell of damp tests being handed out !


  • SUPPORTER+
Do you know what you're doing?

PRR

Quote from: ElectricDruid on August 11, 2022, 07:41:52 AM....that's why libraries have photocopiers!

Do you not remember a time before Xerox?

Of course Phend's school and mine could not afford Xerox for tests. They did enough duplication to buy the Mimeograph machine. Poorer folks used hectograph, a cookie-sheet of gelatin with a purple-ink master, uck. I thought that technology was forgotten but apparently it lingers?
  • SUPPORTER

Rob Strand

QuoteWouldn't work, we had mimeograph machines, loved the smell of damp tests being handed out !
That's what I remember most as well.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

Rob Strand

#16
FWIW, in practice for high speed circuits I find the 2N3904/2N3906's to be slightly faster than the BC5xx's.  Without looking at my notes it's probably something to do with CB capacitance being slightly lower.   Perhaps a subtle difference like 3.5pF vs 4.0pF, just enough to nudge performance.   Unlikely to be noticeable for most audio stuff.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

Sparky

Quote from: Rob Strand on August 11, 2022, 10:37:13 PM
FWIW, in practice for high speed circuits I find the 2N3904/2N3906's to be slightly faster than the BC5xx's.  Without looking at my notes it's probably something to do with CB capacitance being slightly lower.   Perhaps a subtle difference like 3.5pF vs 4.0pF, just enough to nudge performance.   Unlikely to be noticeable for most audio stuff.
So far it seems to me that the 3904  "sounds"  clean and smooth.....good even tone  [not treble-heavy].  Maybe it's just because I've been listening to the 3053's a lot lately.  I also dug up some 2N2484 and 2N3903 but haven't compared them yet.

StephenGiles

Quote from: Mark Hammer on August 11, 2022, 09:14:44 AM
"Tup" is also, from what I understand, an older colloquialism for having intercourse, perhaps by virtue of its association with rams.  I would imagine that when Canadian "father of Confederation" Sir Charles Tupper visited England, the smarmy jokes were many.
In the Shepherds Play, performed as part of The Mysteries at the National Theatre London, they did a song called "Who's tups are these?"  - it's all on youtube!!
"I want my meat burned, like St Joan. Bring me pickles and vicious mustards to pierce the tongue like Cardigan's Lancers.".

Sparky

I think I should be getting my transistors from Mouser instead of Amazon!  These knock-offs probably are "off spec".