Power J-fet + PNP buffer

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
...trying to figure a way to use my Lovoltech JFETs other than The Pass F3 and circlotron, and always wanting to test new small amps for my 110dB compression drivers, I came up with this selfbiased sweetie.

I was thinking of a supply voltage of +-9V, and an idle current of say 100mA-200mA. I plan to include it into a feedback loop of a gainstage, but it could work on its own as a buffer.

The only question is - and this is were YOU clever guys can have a go - how to deal with thermal compensation of the PNP?

NB: the Lovoltech JFET will ask for something like -1.4V gate-source for 100-150mA.
 

Attachments

  • Nrik buffer.JPG
    Nrik buffer.JPG
    9.2 KB · Views: 529
Bigun - yes Power Jfets are nice - they looove the Lovoltechs over at the Pass forum.
But Futterman is a circlotron - and tubes can only be 'n-channel'.

I want complementary topology, and both circlotron and 'all N-channel' (or quasi-complementary) is something else.

But since there are no P-channel Power JFets around, one have to do something else.
 
The only question is - and this is were YOU clever guys can have a go - how to deal with thermal compensation of the PNP?

NB: the Lovoltech JFET will ask for something like -1.4V gate-source for 100-150mA.
Hi Nrik, i have never used a Jfet but i might have a simple idea for you ;)

Why not simply connect the Jfet gate to the bottom of the resistor & select the resistor to allow the bias current you are looking for? Ok so you'll drop a bit more voltage accross the resistor, but the bias will no longer be subject to the Vbe of the PNP transistor.

Failing that put a 1N4148 between the gate & transitor base & thermally connect it to the PNP transistor. I realise both methods involve a higher voltage drop accross the resistor.

You wanted simple though :D
 
Hi Nrik, i have never used a Jfet but i might have a simple idea for you ;)

Why not simply connect the Jfet gate to the bottom of the resistor & select the resistor to allow the bias current you are looking for? Ok so you'll drop a bit more voltage accross the resistor, but the bias will no longer be subject to the Vbe of the PNP transistor.

Failing that put a 1N4148 between the gate & transitor base & thermally connect it to the PNP transistor. I realise both methods involve a higher voltage drop accross the resistor.

You wanted simple though :D

Thanks for your feedback
Your first suggestion (turning the JFET into a constant current source) will still leave the PNP to do its thermal runaway. Besides - the CCS would then transform the circuit into a single-end class A.
Your second suggestion is more or less what I am doing in my second post - I am trying it out right now so Ill get back on that one.
 
You shouldn't get any thermal runaway as long as the PNP transistor has some kind of heatsink. If the Jfet was turned into a CCS then the only way the transistor could turn on more current is if the load accepts it. Yes, the base emitter voltage may drop due to heating but it wouldn't be able to turn on more current.

Having said that i'm not sure about the behaviour of the power Jfet under heating, no idea if this would be positive of negative temperature coefficient so you might be right ;)

As i said i have never used a jfet, let alone one designed for higher power usage.

Hope you get it sorted, looks like you are trying for some kind of SRPP action :)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.