🚀 go-pugleaf

In article <1990Oct25.080455.3910@undeed.uucp> dwoodwar@undeed.uucp (Duncan R Woodward) writes:
>hi Netters,
>
>Does anybody out there know of a neat circuit for a voltmeter with an
>input impedance of approximately 10 Giga ohms ? I am particularly
>interested in the input stages. The meter must be able to measure
>input voltages in the range 0 to 50v DC.

	Many MOS input op amps can reach 10 Gohm, but few if any can
take 50V input range with respect to their power supply rails.  So,
there's an old trick; one bootstraps the regulated power supply
from the op amp output to make a wide-range follower.


         +70V
           |
           |
           +-----+
           |     |
          R1     |
           |    /
           +---|
           |    \   (emitter of NPN transistor)
          CR1    |
           |   _____
input      +---|     \_________ output
   --------|---|+    /   | 
           |   -----     |
           +------|------+
           |      |
          CR2     | (emitter of PNP transistor)
           |     /
           +----|
           |     \
          R2      |
           |      |
           +------+
           |
          -20V


	Choose zener diodes CR1 and CR2 according to the power supply
rating of the op amp; probably two 5V zeners are best.  The transistors
will need to take up to 80V.  For the op amp, I'd think of CA3140 (Harris)
or LF441 or LM356 (National Semiconductor).  The benefit of this circuit is
that the common mode voltage is constant (so the bias current fluctuations
of the op amp are not related to the input voltage).  Choose any op
amp with sufficiently low bias current, preferably with low power supply
current requirements.

	John Whitmore
	whit@milton.u.washington.edu