Super Orb Performance, May Surprise People :)

[Mikewarrior2]

Hey Everybody..

No, not my testing yet(still completing it), but a well done cooling review over at legion hardware. here's the link: Legion Hardware PEP66 Review, Includes Other Heatsinks

Not my preferred method of loading the cpu, but since all chips were tested the same way, it should suffice. Notice how the Super-Orb barely manages to outperform the retail 1ghz p3 heatsink(bear in mind, this is a much heftier retail heatsink on the 1ghz).

If a Super Orb can't quite handle 35W of power, how can it possible handle a 60W processor at reasonable temps?


Mike

 

[Viperoni]

hey, it's a super orb, what CAN you expect

 

[conlan]

no surprise here , orbs are over-rated , but they sure are purty MikeWarrior....... what is your preferred method for loading the cpu? Prime 95? Quake3?

 

[Mikewarrior2]

Actually, Its no longer rC5 or prime95.

its called cpu burn 2, available here: Overclockers.com Utilities

Gets my cpu about 2C warmer than wtih prime95 or rc5. But loading the cpu isn't the problem when running a comparison as long as its the same "loading" program.


Mike

 

[conlan]

thanks mike ill check it out

 

[Jovack2]

Sweet cooler

 

[Usul]

mmm, I wander why anyway.
It has a lot of surface...
Is it the radial design?




Jovack2you kidding, right?

 

[conlan]

usul...thats my guess , but thats just a guess.

 

[BradS]

The orb was not designed to transfer and remove heat. I think it was designed to look good, and sell big. TC is now releasing a more conventional design cooler which doesn't look as good as the orb, but will keep a cpu cooler.

 

[miniMUNCH]

I had never seen an ORB until I bought one for 15.00 and slapped it on my 900 TB blue core...and I never thought about the design until I read Mike's blurb about orbs. I'm pretty new to OC'ing and all, having just got my system up and running about 3 months ago.

Here a run down from a guy who got an "A" in heat transfer in college (that's me by the way):

If you look at the actual design of the orb the "heat transfer path" is very long in comparison to designs like the alpha's. By path I mean the distance from the heat source (cpu) to the surface over which air flows (fins and top-side heat sink). If you look closely at an Orb the distance from the cpu contact surface to the fins is over 8mm, closer to 10-12mm really, all this through solid aluminum.

On the otherhand, the alpha PAL6035 has a path of perhaps 2-3mm to the rods/surface directly above the CPU. And the alpha implements a copper spreader which conducts heat much better than aluminum to "spread" the heat out over the base of the HS and increase the effectiveness the heat transfer rods NOT directly above the CPU.

Also, again...if you look at a C-orb, the effective heat transfer surface is very small when compared to an alpha's heat transfer surface.

One more thing: Orb's are forced draft HSF's - they blow air at the HS. Alpha's are induced draft - they pull air from the sides of the HS and draw it up through the HS rods.

The induced draft design is much better for your system as a whole, pulling air away from the the motherboard/mem instead of blowing heated air onto your motherboard surface/chipset/mem.

Also, Mikewarrior reported that orb's seem to cause bizarre chip temp reading in athlon systems. Alpha's, GlobalWin's and the like don't seem to have this problem. This is most likely caused by the forced draft of the orb's.

Mike, it would be interesting to test...flip the fans around on an alpha/althon combo to have a forced draft system and see if and how much the readings are jacked up. This isn't a definitive test but would ofter strong empirical evidence in favor of my idea.

 

[conlan]

now THAT makes sense , thanks mini-munch . i tried turning the fan on my FOP-32 around and it didnt make any difference though , any thoughts on that?

 

[miniMUNCH]

I dunno. Doesn't necessarily rule out this idea entirely. Actually, the particular geometry of HSF (circular or rectangular), the placement of the fans with realtion to HS base, and the type of air flow (laminar, turbulent...and how turbulent) created by the system can greatly influence the dynamics and flow pattern of air around the socket. It may very well be that the conditions produced by ORbs cannot be produced by HS's with non-circular geometries. Again...I dunno. This would require some computational fluid dynamic calc's...definitely not worth it for us.

It is also possible, I suppose, that the placement of the fan motor on the Orbs (closer to the chip and therefore the thermistor) could actually creating some interference (inductive) in the thermistor circuit. But that's a long shot...and I'm no electrical engineer. I'm a ChemE! I don't do E&M on a regular basis.

Anyway, love my alpha...and I have a Chrome Orb FS cheap!

 

[Mikewarrior2]

MiniMunch,

I posted this in General hardware, but hte nature of my general tests are as follows:

P3-Diode

Heatsink W/Arctic Silver
Heatsink W/Radio Shack Grease
Heatsink W/PCTC(if applicable)
Heatsink Lapped W/Artic silver

Heatsinks are used for only one particular test to avoid grease contamination.

Socket-A tests:
Sealed socket-thermistor(not sure how i'm gonna seal it yet)
Unsealed Socket-thermistor
Side-mount thermistor

Same tests as above.

My cpu loader is CPU burn 2, availabe at overclockers.com


Mike

 

[BradS]

miniMunch - Your heat transfer description is exactly what I mean by the orb was not designed with heat transfer in mind. The fins are to far away from the heat source. That, coupled with the low fin surface area, results in poor heat removal. Having said that, I own one too.