Because the OP requested it in post #16

 

 

Because the OP requested it in post #16

I'm gonna get in trouble here.
I had seen your post #15, and the OP's reply. It just seemed like it had nothing to do with the original post. That's why I asked the question.
God knows I have gone off on plenty of tangents here, so I should just shut up and hang on.

 

Though it is slightly off topic, I think it is a very important characteristic of FETs that more people should be aware of.
Thanks #12, for reminding me.

 

Personally, the idea of a temperature stable constant current source in 2 parts gives me a bit of a chubby. In comparison, a transistor, a zener diode, and 2 resistors seems clumsy, and it still has temperature drift.

 

Personally, the idea of a temperature stable constant current source in 2 parts gives me a bit of a chubby. In comparison, a transistor, a zener diode, and 2 resistors seems clumsy, and it still has temperature drift.

Try to design a current source (at some selected output current), with just a JFET and a resistor, which is repeatable from one unit to the next. Now try to design one with zero TC, at that same current level.
You will find that current is a function of Idss and Vgs(off), both of which vary widely from unit to unit within the same part number.
Look at the datasheet of 2N4416, which I picked solely because its a part number that's permanently stuck in my head. Vgs(off) varies from -2.5 to -6V. Idss varies from 5 to 15mA.
There may be some correlation between Idss and Vgs, but I don't think they track so as to make current source design predictable.

http://www.vishay.com/docs/70596/70596.pdf

The following is a good article about zero-TC JFET current sources, but apparently scribd is not allowed in a hot link:
www-dot-scribd-dot-com/doc/74052127/47/Using-the-JFET-as-a-current-source

There is a reason why people use zeners, transistors, and resistors (and sometimes op amps) to make current sources/sinks.