Transistor substitutions

[CBD]

 Is there a general rule regarding transistor substitutions?
Here are a few that I have.
2n2222a
2n3904 -2n3906
BC182
2n4401 -2n4403
2n2907
Here are a few that are listed on the various projects
bc108
bc109
2n5457
2n5088
bs170

[frank_p]

You can start by trying to use the substitution engine by NTE.
http://nte01.nteinc.com/nte/NTExRefSemiProd.nsf/$$Search
Group those who fall under the same code.  Then check the characteristics on standard datasheets to compare.
Use the NTE system as a guide or rule of thumb, not as absolute truth.

[LucifersTrip]

sadly, using that NTE thing was a mistake I made early on. Loads of different transistors will sub to the exact same NTE transistor.

I don't build much with silicon, but unlike germanium, where leakage has a considerable effect, I found the biggest thing you have to worry about is simply the gain. Yes, different si transistors with the same gains will sound different, but not nearly as much as with ge's...sometimes unnoticeable.

With silicon, my suggestion is to just pick transistors of the same type of course (pnp, npn, mosfet, etc) and with similar gains (check out the datasheets).


good luck

[frank_p]

2n5457  NTE457  (JFET)

bs170  NTE490 (Mosfet)

They are all NTE123A (npn) NTE159M (pnp) (standard) anyways, except for the JFET and the Mosfet.
So go with the gain as Lucifer say.

[nomorebetts]

Check this out 

http://www.aronnelson.com/gallery/main.php?g2_view=dynamicalbum.PopularAlbum&g2_itemId=11084

[LucifersTrip]

Check this out  

http://www.aronnelson.com/gallery/main.php?g2_view=dynamicalbum.PopularAlbum&g2_itemId=11084

that's pretty cool, but can also burn you with vintage fuzzes. For instance, on the first line "if you need BC108, then 2N5088 will usually work". This is also a mistake I made at the  beginning. It is very common for people to recommend plopping a 5088 in there to substitute for any mid - high gain vintage transistor, but I've had very little luck at all. A 5088 can have a gain of 600+, while a BC108 will rarely be that high. I wound up with a very harsh sounding fuzz and pulled the 5088's immediately.

Though, I have found they are good for non-fuzzes, like a Green Ringer

[CBD]

Thanks to all.
This forum is a wealth of information.
This will help me see if a particular effect I am building, is working.... even if it does not sound quite right....at least I will know if it works.
Then I can simply order the correct part and replace it. (maybe even install a socket from the start to interchange  different ones and compare.

[R.G.]

This is also a mistake I made at the  beginning. It is very common for people to recommend plopping a 5088 in there to substitute for any mid - high gain vintage transistor, but I've had very little luck at all. A 5088 can have a gain of 600+, while a BC108 will rarely be that high. I wound up with a very harsh sounding fuzz and pulled the 5088's immediately.

I'm probably the source of that advice, through Keen's Second Law: When in doubt, whip in a 2N5088.

There is a difference between the cases where you don't know which of seventeen (or seventy) devices to use, none of which are the original part number, and when you are sifting through for the best possible tone. In the first case, you're trying to get it to work at all. In the second, you're adding a dash of cumin and oregano.

The vast majority of NPN transistor amplifer applications where the voltage and current is not too high have the gain determined by the circuit, not the transistor. You just need enough gain. The 2N5088 nails most of these applications. There are a few circuits which are (mis)designed to rely on the characteristics of the transistor, notably the fuzz face and its variants; there are others. When you're listening to the transistor making the clipping, the exact device matters in the tone.

Even then, in almost all cases a 2N5088 will get the thing working well enough to check out that the circuit works, the wiring is OK, the soldering is right, and so on. For beginners, these are the majority of the problems. Once you know it's working,  you can go sifting through the world of transistors for the perfect device.

So I would call a 5088 not a mistake, but a good if not always perfect first step. Once it works, there is time to go find the perfect device. Any transistor you can actually get and hook up, even if it's not the perfect tone you were looking for, is 100% better than the perfect-tone device that is unfortunately not available at all.

So - when in doubt, whip in a 2N5088. If that's not good enough, go looking. But you get to use it while you look, and in the majority of cases, it'll be all you ever need.

[seedlings]

Thanks to all.
This forum is a wealth of information.
This will help me see if a particular effect I am building, is working.... even if it does not sound quite right....at least I will know if it works.
Then I can simply order the correct part and replace it. (maybe even install a socket from the start to interchange  different ones and compare.

Socket is a must.  Actually, breadboard the circuit first to get the tone how you want it, then still build with a socket just in case.

CHAD

[jafo]

Socket is a must.  Actually, breadboard the circuit first to get the tone how you want it, then still build with a socket just in case.

+1 to this. Quantum weirdness, experimental error, general random cussedness of the universe, phase of the Moon, whatevs -- weirdness happens, so be ready for it. (Another option is to over-engineer to put a stranglehold on random variation, but it's not nice to BDSM Mother Nature.)

Also, +1 to Keen's Laws.