I deleted my earlier post as I have not been able to repeat the results, which I now doubt. I do not know what is going on.
I am back after evening out the wiring after the Y split, to getting fractions of an amp on the controller I put the ammeter on, whether the pos or neg wire to the battery or on which controller.
I think maybe what is happening is that with the 10a mulitmeter on as part the path on one side, this sends almost all the amps along the other controller's fatter wires (than the meter's) since it now has the better path. No clue, of course :(
This could show why you might need a clamp- on ammeter to measure the amps on each side of a Y, so it does not disturb the thing being measured.
( I do remember in science class that in Lab, they said you can't measure anything without changing it so your measurement was dubious. Something about Empirical something or other, IIRC. This was in 1960, so sorry I can't remember it better! :)
Anyway, never mind my earlier report---all is in doubt, and I have no measurement of how many amps go on each side of the Y. All I know is the overall result is working great, and I have not fried any of the two 20a controllers.