"flipper" wrote in message
...
On Sat, 21 May 2011 14:16:32 +0100, "Ian Field"
wrote:


"flipper" wrote in message
. ..
On Fri, 20 May 2011 20:00:14 +0100, "Ian Field"
wrote:


"Dave" wrote in message
nternetamerica...

"Ian Field" wrote in message
...

"Dave" wrote in message
netamerica...

"Dave" wrote in message
news Am trying to bias the JFET in the attached schematic to feed signal
to
the following BJTs while keeping Vgs very near Pinch-Off. Works
fine
as long as the batteries are fresh, but as they begin to wane Vgs
surpasses Pinch-Off (which is 4.0 volts for the JFET I am using,
while
I try to hold it at 3.95 VDC.) It has been suggested that I use a
BJT
(configured for Common Emitter) to do this, but I am in the dark as
to
how this might be accomplished. ANY help would be greatly
appreciated...

Thanks,

Dave




Okay, Murphy's Law rules. Was effecting a minor change in the
biasing
of the JFET and when I hooked it all back up I let +12VDC hit the
varactor D4 out of utter carelessness and too much wire where it
wasn't
needed. Now all my varactors are blown and I may have damaged one or
more transistors. Will be back later after I have sorted all this
out.

JT: Thanks for the input on the source resistor, but the NTE451
apparently limits Id to 3mA by itself (much to my frustration). I
don't
need to do that for it. What I do have to do is work with the 3mA
it
allows. Anyway, I appreciate your input and may be back later if
what
I was trying doesn't work to my satisfaction.

If you're having trouble stabilising the bias against reduction in
Vdd
headroom, I'm wondering why you feed the gate with a divider instead
of
a
simple "gate leak" resistor.

If the frequency isn't too high you might get away with a small
signal
MOSFET with a bipolar transistor regulating the bias by sensing
source
current - the gain will be huge compared to a JFET so you can cancel
some
of the higher input capacitance by splitting the source resistor in
two
and only decoupling the resistor that senses DC source current.


Huuuh. Thank you for the intriguing thought on a different approach.
Any
ideas on where I would look for an appropriate part number? Honestly
suspected a MOSFET was maybe a better idea, but didn't know where to
start
with that. And JT has already provided some details on using a BJT to
sense source current that sounds like what you are speaking of.


One trick is to use 2 JFETs, one of which has its gate connected to
source
and passes Idss, that gives you a current source instead of the source
resistor - unfortunately Idss spread between devices of the same type
can
be
pretty wide giveing not quite the resuts you were hoping for.

A more repeatable alternative is the single transistor current source in
the
JFET source lead, bias the base with 2x Si diode or a LED with an
appropriate pull up resistor from Vdd, the emitter goes to Vss via a
(relatively) low value resistor which determines the collector current.

With some JFETs, Vgs-off can have a wide spread which can make
difficulties
if you don't have a lot of Vdd headroom.

You can use a portion of the source current to control a BJT which sets
the
MOSFET bias point, the input resistance is then determined by the
bootstrap
resistor connecting the MOSFET gate to the collector/pullup resistor.
The
source current sensing resistor should be adequately decoupled and it
doesn't hurt to put a small capacitor to smooth the collector.

The MOSFET has higher input capacitance than the JFET but it also has
much
higher gain, you can cancel most of the capacitance by inserting a
degenerative feedback resistor between the current sense resistor and
source, this sacrifices gain and you set this to more like what you'd
expect
from a JFET.


Well, as far as Vdd headroom goes it's not so much the spread as Vgs
max and I'm not sure you're much, if any, better off selecting a
MOSFET than a jFET.

Well I suppose if you used a high voltage MOSFET from a SMPSU that had a
Vgs-thr in the direction of 6 - 8V it might get a bit tricky, might not do
too well at the frequency range suggested by the inductances in the OPs
design either.


The 2N5484 has a max pinch off of -3V, the BF245A is -2.2V max, and
the BF861A is -1V max.

Which MOSFET were you suggesting?

There's quite a few with Vgs-thr around 1V.