Clamp on ferrites are not going to help with a sump pump level of noise. In fact most devices won't. I could send you a demo unit to see if it would help, that will work on lower frequencies a sump pump may generate.
Which Freq range do you believe the sump pump may puke noise out at? 60hz or something still in the kHz range?
Based on what I measured with the i-phone APP it seems most of the grunge is pretty high in the Freq band (8-16kHz) during boot up of the amps.
I think the brain box may be contributing more to some of the noise then the sump itself so I may just skip attempting to cure any sump noise because in theory it shouldn't be spinning its motor unless activated.
So are you guessing the change in resistance of the different length cables is the cause?
RLC, three properties that increase some with length.
Something must have been resonating. Crossover & binding posts, speaker wire, and the output indicator may not have been doing well. Maybe it has to do with elliptical crossovers.
ClassD have output filters, they don't need Zobels like other amps as far as I know; so it's probably not the class topology so much as the output filter.
I'm coming late to this thread CJ, but I have KG Kipod Signature passives which as you'll know have aluminium cabinets.and I have never had the problems you describe.
My system is extensively Entreqed and includes amp negative speaker out puts grounded to a Poseidon ground box and it has been one of the best Entreq applications I have used. I really do not know if it would cure your problem but thought I would share that with you in case you wanted to consider it.