P-channel version of the BF862 JFET

[Nemo]

Why there are some nice n-channel JFETs with low input capacitance,
Idss, gate threshold voltage etc but p-channel types are rare?

There seem to be the J177 types, and possibly a device from Central
Semiconductor ("contact factory" - looks expensive for just messing
around) and that's about it. Am I missing some obvious device, are they
extraordinarily difficult to make, or is there a way of using n-channel
types in some sneaky way so p-channel types aren't often needed?

(I'm looking at bootstrapping a photodiode with positive bias instead of
the more usual negative bias, based on the classic circuits in Linear
Tech app notes like DN399 and their LTC6244 datasheet.)

 

[Nemo]

I'm not a fet connoisseur, but have you tried Interfet?

Yes, I flailed around on their website for a while, but it's like
National Semiconductor's - they assume you know what you want before you
go, and divide things into their own categories which aren't necessarily
how I think, and the part numbers are meaningless without some kind of
key 8) So it's not searchable if you don't really know what you're
after. I did burrow around for about 40 minutes before deciding they
probably didn't have anything better than a J177, though I grant there
may be something in a corner of their site I didn't visit.

 

[JW]

Yes, I flailed around on their website for a while, but it's like
National Semiconductor's - they assume you know what you want before you
go, and divide things into their own categories which aren't necessarily
how I think, and the part numbers are meaningless without some kind of
key 8) So it's not searchable if you don't really know what you're
after. I did burrow around for about 40 minutes before deciding they
probably didn't have anything better than a J177, though I grant there
may be something in a corner of their site I didn't visit.

Try http://www.linearsystems.com/

They have some real nice parts and a cross reference search engine.

 

[Nemo]

The fundamental properties of "holes" (e.g. mobility) are different
than those
for electrons. There may also be some differences due to the
particular
elements used as dopants. These differences occur in bipolar
transistors
(where the interesting part is the minority carriers) as well as FETs.
(majority
carriers). This means that the tradeoffs for gain, breakdown voltage,
and the
like cannot be the same for N and P type devices.

Sometimes there can be good reasons for using a particular biasing
scheme,
though it can be inconvenient.

Well explained, I see your point. Alas I have to use this polarity of
bias due to other constraints, but I'll see if I can get an N-channel
JFET to work in some configuration.

In reply to JW, yes, I checked out Linear Systems, but thanks for
checking I knew about them 8)

Nemo