What is your ultimate solution to performance & noise?

[Seyeklopz]

I thought a thin radiator with open fins on the underside could go under the pc case. In the center would be a thin hamster wheel fan which blows in every direction, sucking the air down through a hole in the case bottom. Could be a very large rad with a low speed fan, and it would take up zero room in the case.

I guess this sounds like a giant GPU air sink.. hehe

[bobkoure]

Originally Posted by PieEyedPiper

If this is correct, and asuming it is, can we not all achieve our high performance low noise with this solution?

Hi - It's Mr "HPPS broken record" again :-)
If you get a chance, go check out the Innovatek/Eheim HPPS:
12VDC, quieter than a 1046, better temps than a 1048 with Swiftech 5000 or 6000 (maybe even better than with a 1250 - haven't tried one of those with either of these waterblocks).

[PieEyedPiper]

sound like the ticket to me, too bad they're sold out and have funky fittings.
Glad I was on the right track.

[bobkoure]

Yep, they seem to be sold out.
Fittings were just the same as on a Eheim 1046 - came supplied with an innovatek compression fitting for the outlet, which might be the "weird" part you're thinking of.
I hope they re-appear as a couple of friends have heard my system and want the same...

[Ruiner]

Just throwing this out there, but comparing the 1048 to a muffled D4/mcp650, when does the gain from improved head become negated by the added pump heat?
I suppose this is only an issue if radiator capacity is marginal to begin with.

Despite the relatively small temp improvements (at the diode, FWIW) of a 1048 vs. mcp650/D4, the OC improvements are supposed to be more significant (if the block can take advantage of the extra head). I was able to get a nearly 100MHz better overclock with that upgrade, as it would allow me to use a slightly higher Vcore and maintain stability.

As to whether 100MHz is worth the noise of the stronger pump.....


bob, do you have links to curves for that pump? The 12v convenience mod aside, what do they do to it to improve its head?

[HammerSandwich]

FrozenCPU has the HPPS.

Originally Posted by Ruiner

when does the gain from improved head become negated by the added pump heat?
I suppose this is only an issue if radiator capacity is marginal to begin with.

Cathar cleverly pointed out that the pump heat should be kept low relative to the load. IOW, 50W from the pump is a lot worse when cooling a 50W CPU than a 100W CPU plus a 50W GPU.

[PieEyedPiper]

interesting, frozencpu claims that the only difference in this pump is that its 12v...not enhanced like other sites mention..and like we've seen noted in this thread.

[bobkoure]

Originally Posted by Ruiner

bob, do you have links to curves for that pump? The 12v convenience mod aside, what do they do to it to improve its head?

I'm sorry to say I do not have curves - though not for lack of searching.
About the only thing I've turned up was a spreadsheet at watercoolplanet.de - and it wasn't curves, just three resistances. They got flow rates for each pump pushing through a 2m length of hose - and tried it for three different hose IDs.
When I pointed out this spreadsheet here, several folks pointed me to their waterblocks rating section, which seemed a bit suspect - which in turn made me suspect the pump spreadsheet data (although it might have been fine). Anyway, I'd made my own copy of the sheet (translated to English, deleted pumps uninteresting to me) which you're welcome to here.

The housing and impeller look very much like the 1046. Whatever they're doing, they're doing by electronically controlling the motor, not by a fancy impeller or housing. Actually, I haven't had a HPPS impeller and a 1046 impeller side-by-side and I think the HPPS impeller is a bit larger in diameter, but that's about it. No Iwaki-style directional impellers, even - even though, with a DC motor and a control board, I'm pretty sure they know which way the impeller's going to be rotating.
Now, there were two HPPS pumps - the first being a 12VDC mod for the 240V 1046, which added no additional head, and the second (maybe current and maybe something else is being changed, which would explain them being out of stock everywhere) was 12V from the ground up. this is the one with claims to additional performance - but all the claims are from Innovatek and are in terms of flow in their restrictive systems - and so can't be compared to anything else...

[threeputt]

$105..ouch...

[PieEyedPiper]

haha, yeah not to mention the absurd pricing.