开云体育

It appears that all it really does is compare the test cap to a 66M resistor (three 22M in series), using a couple of 2N3904's configured to be very high impedance.

It uses 27V (three 9V batts in series), so that may be enough to catch 'most' leakage situations (but note that it exceeds the V-rating of some small caps, which is not good!).
I still don't understand how the "forecast" feature might work, and the schematic doesn't appear to have a switch (etc.) for that feature. 'Predicting' a possible cap leakage failure with such a simple scheme seems like snake-oil to me.
In the cases where an old magic-eye machine fails to do the right thing, I believe that can be attributed to poor sensitivity in those old designs, and/or components in the tester that are out of spec.

Pete