Pro/Forums - View Single Post

Cathar · 07-05-2004, 06:53 PM

AngryAlpaca covers most of the glaring anomalies.

The testbed can be seen here.

Another issue of note is the uninsulated heater area which can be seen here:

There is basically as much surface area open to be passively cooled by air as a small heatsink. Smaller sized waterblocks that don't block off the socket as much will allow for greater secondary cooling of the heat die.

As AA points out, the testbed chokes flow quite dramatically. I've been trying to get a hard figure on it from browsing various European forums, with language being the biggest barrier for me. The figures I found bandied about were around 1.0-1.4LPM (0.25-0.35GPM) for many blocks, but this was just "heresay".

The one-mount-only thing is a fairly dramatic methodology error. I'm sure that even Phaestus who is getting more confident with his mountings would be extremely reticent to claim that he gets it right first-time every-time. For my own testing I get the same sort of repeatability/variation that Phaestus gets, but every now and then I do get a bad mount that's 1-2C worse than anything else. I would not like to claim to rank waterblocks on a first-time-only basis.

Edit: On a more personal level, I was rather upset that the Cascade that was tested was a user-modified and user-lapped version with non-standard fittings. In my own tests and in independent tests, the Cascade performs significantly better at very even at low flow rates than any number of other blocks that WCP showed to outperform or match the Cascade. Aside from the testing anomalies mentioned above, I must also conclude that the user-level modifications and lapping were largely responsible for that. Am fairly positive that most manufacturers would not readily agree to random people randomly modifying their blocks outside of the manufacturer's control prior to a block being sent for review, nor do I think would people looking to buy a block accept that such was done prior to a block being reviewed, but this is exactly what happened.

07-05-2004, 06:53 PM	#4
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	AngryAlpaca covers most of the glaring anomalies. The testbed can be seen here. Another issue of note is the uninsulated heater area which can be seen here: There is basically as much surface area open to be passively cooled by air as a small heatsink. Smaller sized waterblocks that don't block off the socket as much will allow for greater secondary cooling of the heat die. As AA points out, the testbed chokes flow quite dramatically. I've been trying to get a hard figure on it from browsing various European forums, with language being the biggest barrier for me. The figures I found bandied about were around 1.0-1.4LPM (0.25-0.35GPM) for many blocks, but this was just "heresay". The one-mount-only thing is a fairly dramatic methodology error. I'm sure that even Phaestus who is getting more confident with his mountings would be extremely reticent to claim that he gets it right first-time every-time. For my own testing I get the same sort of repeatability/variation that Phaestus gets, but every now and then I do get a bad mount that's 1-2C worse than anything else. I would not like to claim to rank waterblocks on a first-time-only basis. Edit: On a more personal level, I was rather upset that the Cascade that was tested was a user-modified and user-lapped version with non-standard fittings. In my own tests and in independent tests, the Cascade performs significantly better at very even at low flow rates than any number of other blocks that WCP showed to outperform or match the Cascade. Aside from the testing anomalies mentioned above, I must also conclude that the user-level modifications and lapping were largely responsible for that. Am fairly positive that most manufacturers would not readily agree to random people randomly modifying their blocks outside of the manufacturer's control prior to a block being sent for review, nor do I think would people looking to buy a block accept that such was done prior to a block being reviewed, but this is exactly what happened. Last edited by Cathar; 07-05-2004 at 07:02 PM.