You could see a lot of action (vapor bubbles) escaping through the mounting holes and out from under the edges of the board as the vapor builds up under the board it makes for an interesting show but obviously this is not the normal configuration. If it were a vertical setup you would have that rolling "boil" look I suppose. It just bet the SP10 board so well designed that in comparison to a smaller pcb of a higher density with 'older' hotter chips is comparing apples to dragonfruit with the SP10. They don't really need to be in Novec with those Hammer chips, but if you had a much smaller blade maybe that would be required. It be great to get some numbers on the SP10 especially can you overclock the snot out of the Hammer chips in the Novec even laying flat? Or do we all have to wait for Inside Bitcoin Hong Kong to learn more about the potential? I can't wait to get to Allied Control and check all these things out.
Curiosity!
Note that the FW pushes jobs to ASICs at the rate of ~600 jobs per second. If some ASICs run too fast (I think about 630Hz or more) they will finish job before the next job is ready, so they will run on idle. This gives you theoretical limitation of about 1.8T. You will see that the FW claims 1.9Th, but in pool you will see 1.8Th. If ASICs run faster then 630Hz, I will need to compile FW that pushes jobs at faster rate.