[Laser] New optical front ends

[steve kavanagh]

> The 0,47  MF capacitor fixes the dynamic gate 
> voltage of the load FET so the real drain  dynamic 
> load of the input FET becomes 2,2M at is Idss 
> current and at its max  transimpedance.
> The gain becomes very high! 
> With such a load the output  stage must have a very
> high impedance to keep the 2,2 M load value and so
> to  keep the gain.

Thanks.  I think I understand that, Yves.

This (VLF) circuit got me thinking again about whether
the preamp's second stage should have high or low
input impedance.  I did some rudimentary circuit
modelling in SPICE with the following results:

The "input" FET has, of course, more voltage gain with
a high impedance load.  However the power gain of the
input FET is much higher with a fairly low impedance
load.  In my example I got 40 dB more power gain with
100-1000 ohms load resistance than with 10 Megohms (at
frequencies above the -3 dB bandwidth, slightly less
difference at lower frequencies).  Therefore I have
more confidence in a circuit with a
low-input-impedance second stage that the noise
contributions from the second stage and the current
source will be negligible.

However, my modelling also showed that (for the same
overall gain over the two stages) the noise output was
essentially the same for both high impedance and low
impedance second stages, so it appears that in
practice it doesn't make much difference to noise
performance in a good design.  It might matter in a
cooled front end, I suppose.

But there still is a big difference, in the case of
low capacitance photodiodes (on the order of 10 pF). 
In my example the bandwidth of the circuit with a
low-input-impedance second stage (common base 2N5088,
Ic=270uA) was 5 times the bandwidth of the circuit
with a JFET source follower as the second stage.  I
think this is mostly due to the difference in the
"Miller Effect" capacitance at the input FET gate,
resulting from feedback via the input FET's gate-drain
capacitance. There is less effect on bandwidth for
high capacitance photodiodes, of course.

However, since both the signal gain and noise output
curves have about the same variation with frequency
this may not have a large impact on weak signal
sensitivity at any single modulation frequency.

73,
Steve VE3SMA

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com