P-channel JFETs

[Andreas]

I'm looking for some P-channel JFETs for switching and I can't find any of the standard ones suggest by RG on GEO. So are there any others that you use which I might find here? Basically my choice of 2N transistors is quite limited as I live in Germany, so B transistors would be better, but any suggestions are welcome. Thanks a lot.

[Andreas]

I did some research and found a 2SJ103 P-channel JFET. Does anyone know this one and could comment on how well it performs as a switch?

Any suggestions for other P-channel JFETs would be still very much welcome. Is it just my perception or are the P-channel ones really that rare?

Or are there any other alternatives to P-channel JFETs. I need them to electronically switch small AC (to ground for starters, but I'm also interested in using these as bypass switches), so an N-channel JFET wouldn't really be an option, because I can't get negative gate voltages (biasing the source up is no option really, as I want to switch pure AC) and MOSFETs would also need to be biased up into the positive region because of their parasitic diode (btw, thanks RG for explaining this to me).

[Rob Strand]

You don't get much selection with P-channels.  I just pulled-up the Farnell catalogue and there were no B numbered P-channel devices only: J176, J177, 2N5461,  2N5462, 2N3820.   I think it's going to come down to what you an get in your region of the world (eg. like those 2sK devices you found).

The specs of these varies a bit so might want to check which ones are suitable for your task.  What exactly is the application (or ckt from RG's site)?

[Andreas]

Thanks a lot for your reply. Seems like I can get the 2N3820 and the 2N5461 here too, so now I at least have the choice between 3 transistors instead of just one, feels a lot better   . Tomorrow I will go through the datasheets and try to find out which one is best suited...

I want to build a clean booster around a MOSFET (I guess I'll use a BS170, cause that's what RG uses in his booster article) which I want to vary between gain 1 and some higher gain (adjustable with a pot) by making a gain 1 common source amplifier (Rsource = Rdrain) and put a big cap, the pot as a variable resistor and a switch (these three in series) in parallel to Rsource, so that when the switch is closed the gain goes up. As I want to do all the switching electronically (i.e. use a momentary switch and a flip flop or two inverters) this switch in the source path would have to be a JFET (or any other suitable electronic switching component, if there is any).

[Rob Strand]

2N3820 and the 2N5461

Both those have high Vgs_off which means they might not work so well as electronic switched on a 9V effect.

Given you are using these as JFET switching for the audio you would be better off using an N-channel JFET, there's no real need to use a P-channel and the N-channel gives you more options to find parts.  Have a look at how Boss do it with N-channels.  If you are using a flip-flop output you don't need to change anything (if you have LED circuits amongst your logic you will have to drive the LED off an opposite sense signal).

[R.G.]

The real, screaming advantage that a p-channel has is that you can run it at a DC bias of ground and pull the gate UP to +9V to turn it off.  

You could do about the same with an Nchannel, biasing it to +9V and pulling the gate to ground to turn it off, but ground is where everything is referenced, so any noise on the power supply is nominally on the +9V line, so you induct noise. That's why all the effects makers bias their N-channels at half the +9V supply - they get to do some noise rejection by bypassing the bias voltage.  This does restrict the choice of JFETs, as only JFETs with Vgsoff of 4.5V or less need apply. Hence the 2N5292, 2N5485, 2SK30a, etc.

R.G.