[nikongod]

hello board.

i know that this has been posted before. contarary to the standard policy of not dealing with repeating threads, i encourage a moderator to edit this thread, and merge it into a relivant one...

anyways.

i bought 72 (36 each npn/pnp) transistors for my dynalo. being a lazy bum, they have sat in baggies untill jsut recently when i went to match them.

NONE of them match. all of my pnps are like 270ish, and all of my npn's are like 250ish. i am usign a transistor tester built into a dmm.

someone please tell me i dont have to buy more, but i suspect i will. that would suxor.

 

[Pars]

I would assume these are all the same gain class (PNPs and NPNs)? The readings you got, while obviously not 1% stuff, really aren't that bad. Short of buying more, you could try contacting Dan Gardner... he (IIRC) had a lot of these and may be able to either swap with you or sell you PNPs or NPNs that more closely match what you already have. Assuming of course that he would want to get rid of some.

The ones I have, the PNPs run from 239 to 360 (GR gain class) while the NPNs are much more tightly grouped at 270 to 350. If I build another one, I'll ave to get some more PNPs to get tighter matches.

Chris

 

[Garbz]

Take the top most of the PNPs and the bottom most of the NPNs. The end result matches to almost a Hfe difference of 20. The purposes of matching was more to ensure that you didn't end up with transistors with a hfe of 200 and others of 350.

Small deviances like yours will only ammount to very small offset providing the well matched transistors are in the cascode and front stage. The output stage isn't too sensitive to poor matching. The end result can be compensated for at the bias stage.

 

[nikongod]

i shoud have been more specfic perhaps.

those are average numbers, and there is NO overlap.

does it matter if the averages for the "buffer" section" are as closely matched? tyhen i could just buy a few matched pairs, and go.

 

[SnoopyRocks]

This is being made into a bigger issue than it really is. Match them as well as you can and be done with it. I built two dynalos with transistors from different gain classes. My matching betwen npn (~250) and pnp(~170) wasn't even close to what you've measured. Both have low offset (~6mV with servo off and less than 1mV with it on) and sound great. Use similar transistors in the same spots on each channel (R/L) and put the closest matches in the current sources. Use what you have and stop worrying. This really isn't that big of a deal. I'd be shock if you could hear a difference.

 

[nikongod]

Quote:

Originally Posted by SnoopyRocks This is being made into a bigger issue than it really is. Match them as well as you can and be done with it. I built two dynalos with transistors from different gain classes. My matching betwen npn (~250) and pnp(~170) wasn't even close to what you've measured. Both have low offset (~6mV with servo off and less than 1mV with it on) and sound great. Use similar transistors in the same spots on each channel (R/L) and put the closest matches in the current sources. Use what you have and stop worrying. This really isn't that big of a deal. I'd be shock if you could hear a difference.

you are hereby certifyed to be awsome.

thank you.