Spehro Pefhany wrote:

Judging by the patent someone referenced, the impedance
measurement should be more than good enough for repair purposes,
when carried out @ 100kHz. The Xc of a 170uF perfect cap is 0.01
ohm.

Most ESR testers use a 100KHz square wave or a current pulse, so the
capacitor impedance is not relevant.

We can compare the voltage due to the ESR with the triangular ramp
from charging the cap. We can use the bridge circuit to eliminate
the inductive spike.

Voltage due to ESR:

E = I * ESR

Capacitor charging ramp:

I = C * dv / dt

dv = I * dt / C

Set these equal:

ESR = dt / C

From your example:

C = 170e-6
dt = 5e-6

then

ESR = 5e-6 / 170e-6

= 0.0294 ohm

With the above parameters, the capacitive ramp voltage equals the IR
drop at an ESR of 0.03 ohm.

It would be difficult to claim these circuits can measure ESR down
to 0.01 ohm.

You could increase the test frequency or decrease the charging pulse
width to reduce the effect of the charging ramp, but you then run
into problems with the leading and trailing edge inductive spikes.

In a perfect world, the bridge ESR circuit would eliminate the
inductive spike, but it still has implementation problems trying to
measure very low ESR. A two-port measurement with a network analyzer
can follow very low values of ESR, but is impractical for in-circuit
test due to multiple cable and contact problems.

As we ask higher performance from faster circuits, we will
inevitably need to use lower ESR caps.

The current crop of ESR testers probably has reached its limits. We
need a better way to measure ESR for tomorrow's technology.

Regards,

Mike Monett