On Wed, 26 Dec 2007 19:22:43 +0100, Joop wrote:

On Wed, 26 Dec 2007 13:54:47 GMT, "John" wrote:

Hi, Robb -

I used SPICE to simulate the important parts of the circuit composed of the
bridge and IC1-C and IC1-D. I used a pulse generator rather than IC1-B. I
used voltage-controlled-voltage-sources rather than op-amp models. I looked
up an arbitrary capacitor's ESR for this simulation.

I got the following output "indications" vs capacitor (arbitrarily 100uF)
ESR:

ESR (ohms) Indication (percent)
0 100
.018 100 (specified value)
.18 99
1.8 92
18 52
180 7

A 10uF with 0 ESR also indicated 100%.

So, the circuit appears to work, but it depends on what you are expecting
from it as to how well it works. If you decide that anything below, say, 50%
is bad, you may not catch the capacitor that causes equipment to fail when
it has more than 1.8 to 18 ohms of ESR.

I think the amount of ESR that can be tolerated will depend on the target
equipment's sensitivity to ESR. In some cases, you may find that even 1.8
ohms causes malfunction. In other cases you may find that 1K ohms of ESR
causes no problems.

To verify the simulation, you could build the meter and test various
capacitors known to be new and good. Then add some series resistance to each
and record the readings. Then all you would need is experience to know when
equipment will fail due to high ESR.

Cheers,
John

Funny, I did the same thing for the Ludens-type meter.
It seemed it could benefit from a better opamp in order to work at
100KHz. So I took an TLC272. Also the detector diodes were replaced by
1n5818's. Shunt resistors were lowered from 10 to 4.7ohm to improve
headroom a bit in the low-ESR range.
Simulation then gives:
scale ESR
100 0
97.74 0.1
95.59 0.2
89.56 0.5
80.93 1
67.55 2
44.18 5
26.82 10
13.74 20
4.21 50
1.34 100

Although this is not a binary group, I took the liberty to append a
small image of what this would do for a meter scale.

Joop
The image did not seem to be passed on.
So repost here.