[mwoods]

Has anyone tried using a JFET as cathode bias for a signal stage?  I have seen diodes used and seen CCS is the plate load, but not much mentioned about an active device used in cathode bias.  Any opinions?  What does the plate load become in this configuration? 

[JoshK]

I can't say I've seen precisely what you are asking.  I've seen JFETs used as the lower element of a cascode in phono preamps for instance.  I've seen LEDs for bias (a diode).  I've seen talk of using zeners, but usually results in "its noisey" response. 

CCSs are typically used in the cathode of a long-tailed pair, for gain with phase splitting.  Otherwise, the more logical place for a CCS is the plate load as it maximizes gain and lowers distortion of the stage.   No reason why you couldn't use the CCS in the cathode but it'd seem a bit suboptimal.  A resistor would be the plate load in this case. 

A JFET doesn't make for the best CCS however.  I am not expert on these matters but I've seen a lot of data suggesting the typical ring of two (transistors) does a far better job.

[mwoods]

I thought it would be a good idea to try as an alternative to battery bias.  I'll be trying this soon with the 417A driving 2A3.  At least in this position an active device does not need to dissipate as much heat as it would on the plate side.

[JoshK]

As a ccs, the key to performance is the bandwidth over which impedance is kept constant and for that I am not sure JFETs do as well as cheaper bipolars.

But if you are talking about replacing a battery bias scheme with JFETs then I am now perplexed.  The battery is a voltage source, not a CCS.  In biasing, it is used in a very different way like a bypassed resistor, high DC resistance but short circuit for AC.  CCS is high AC impedance.  I've generally only seen battery bias in fixed bias mode, CCS in cathode bias. 

[mwoods]

Sorry I didn't mean a true alternative, I just meant "something to try."  I will keep bandwidth in mind as well and look at using the BJT's.