[morzh]

Today I extracted all 4 other electrolytics, all axial.
One of them, the 10uF 350V, is RIGHT ON THE MONEY. I almost regretted taking it out. But......it is 62 years old. Time to go.

I ordered the 6 pieces I need from Mouser.

The cap from the crossover is an enigma. It is the same thing as we had in 38-690 with that 1uF cap.
The cap IS a polarized electrolytic, 20uF 30/40V. With "+" clearly shown.
I am not sure how this works with no problem but this is one of 2 or 3 electrolytics that is still tests well. And, yes, I am changing it - I am leaving no cap unturned except the polyesther (very few of these, the rest is paper and all are bad) and tubular ceramics.

So, I unsoldered this cap, that is unusually large for its parfameters, and measured it on my LCR meter (Standford Research) in "C+R" mode to see the ESR.
It is very surprising, the ESR drops to 0.06 Ohm at 10kHz, being about 3 Ohm at 100Hz.
The high-ripple 20uF 450V cap by Panasonic (almost 700 ma ripple) I usually use for my restorations as he rectifier filter, shows about 1 Ohm with little variations across the frequency range. Which made me decide against it in the crossover.
I only had one other electrolytic (new) actilng like this, namely Non-pol from Parts Express, the 100uF; the rest have much higher ESR.

So, I chose to use a film cap, 20uF 250V. Those can take high current and have very negligible ESR, and I do not have to think up the theory how a polarized cap works in pure AC curcuit. An expensive SOB, they start at about $8. I mean, there are those $4 apiec from China....But I prefer Illinois cap from Mouser: wherever it is made, it is a reputable name.

So.....meantime I have to cut open the big can for which I ordered two 100uF 400V caps to restuff. Hayseed will quote but I imagine the price, especially the very special one like this, with the raised panel, will cost too much, and me being accustomed to restuff big cans, that us a sheer waste.