Keep in mind most of us using those giant TECs are running them at roughly half their rated max for much higher efficiency and less waste heat. Also, chillers work better with multiple smaller TECs while direct-die TECs need to be much larger and more powerful to keep from stalling...just the nature of the beast.
Your build using a shared cold side block for each pair of TECs is a neat idea. Having that much centralized cooling though will need some pretty serious coolant to keep from turning to syrup or freezing - as you have noticed, since it sure sounds like the coolant is just freezing inside the blocks. Easy fix
FWIW a 655/D5 has no seals to mess up with the factory top, only the large body o-ring that does not come in contact with any coolant. The rest is just a heavy impeller with a graphite bushing riding on a single ceramic bearing, with the whole casing magnetized. I ran a pretty nasty Ethanol mix @ 0c through one for over a year and the pump didn't even flinch...until the pump ran dry and the ceramic bearing ate the bushing (the only moving part and wear item in the whole pump, and the 655s only real failure point). Keep in mind the pump was around 7 years old and was ran 24/7 and was still kicking up until that point!
IMO as long as you keep the fluid thin enough to not cause excessive stress on the graphite bushing I do not see any reasonable temps being an issue with a 655. I also trust their reliability a LOT more then any other DC pump because I have had at least 5 of them and they all lived at least 6 years before starting to rattle or vibrate from bushing wear. They usually give tons of warning before dying, and will go from a noticeable vibration to a full on death rattle, giving you plenty of warning to change it vs a pump that just quits one day.
I'm sure if swiftech released a replacement bushing or impeller then we would see plenty of 10yo 655s still running flawlessly.Edited by Puck - 6/30/16 at 6:35pm