few here would disagree that using an ESR meter is the fastest and surest w ay to find bad electros. If the meter reading is several times or more high er than normal, that electro must go.
This week, I saw a Fender tube amp with pretty obvious filter electro troub le that almost caught me trusting my (Bob Parker) ESR meter too much.
Under a steel cover were five, "IC" brand 22uF, 500V axial electros - from the symptoms I figured ones immediately following the rectifiers diodes mus t be bad. The ESR meter agreed, giving open ( >100ohms) ESR readings for tw o and good readings for the remaining three.
While extracting the bad pair from the PCB, I realised they were wired in p arallel. The electros looked in good condition so my suspicions raised, I p eeled back the case with nippers to loosen rubber bung and take peek inside .
Turned out there was NO metallic connection between the positive lead and t he rest of the capacitor. The bung and lead stub simply fell off. I opened the other three and found two more in the same condition while the last was perfectly OK with all its connections in place.
The electros concerned are exactly like the ones in this pic:
The ESR meter was fooled by the fact (except for one) the caps had NOT gone high ESR, but had suffered severe corrosion of the plus leads where they e ntered the cap roll.
I have to agree with the author of the linked page, Illinois Capacitors hav e produced a bad batch and full replacement is the only smart option.