true SILENCE, not QUIETNESS

Cooling Processors quietly

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
hwtf
Posts: 16
Joined: Tue Oct 08, 2002 5:41 pm

true SILENCE, not QUIETNESS

Post by hwtf » Tue Oct 08, 2002 5:48 pm

what is the best heatsink (s370) for operation without a fan?
i have been seeing zalman, swiftec, and thermal right heatsinks around in these forums a lot. however, i have not seen any posts that claim the best heatsink for fanless operation. as you may know, the best hs running with a fan will not necessarily be the best without a fan.

what is the area of the slk800? i cant quite find the specs. i am suspecting that either this or the zalman 6000 is the best. btw, i am getting an unlocked p3 1.4ghz soon :P . i intend to lower the multiplier down while retaining the 133mhz fsb :D

quokked
Posts: 333
Joined: Sun Aug 11, 2002 3:26 pm
Location: Melbourne, Australia
Contact:

Post by quokked » Wed Oct 09, 2002 4:01 am

ummmmmm depends on what type of setup you're going for, the Zalman 6000 would probably work best if u were going for the fanless optionand u'd need to have SOME form or airflow going over it whenever it was some ducting for a fan in the bottom of your PSU
or maybe even the ordinary fan :)
Personally I'd go for the SLK-800 it has a few advanatages over the Zalman personally I like the 3 Lug clip on it and the hardware sites rate it better cooling, as well.
but if you're going fanless the flow resistance of the Zalman will help u heaps if you're going passive.
with the swiftec heatsinks I've talked and read a lot about them and they only seem to perform well when they've got a decent RPM fan with a lot of pressure, just look at those pins :) anyway's that's my advice/ramblings have a hunt round the forums :D

hwtf
Posts: 16
Joined: Tue Oct 08, 2002 5:41 pm

Post by hwtf » Wed Oct 09, 2002 5:49 pm

it is possible to run it without a fan. a friend of mine from the hardforums ran his 1.13ghz tualatin p3 on an slk800, at 45c full load. not too bad.
i think the swiftecs would perform badly for passive cooling; the pins are thick, and fins are preferred in passive cooling.
ima email thermalright and ask them for the surface area of the slk800- it looks like its around the same area of zalman.
any other comments? :D

Red Dawn
Posts: 169
Joined: Fri Oct 04, 2002 11:46 am
Location: Stockholm

Post by Red Dawn » Wed Oct 09, 2002 6:08 pm

Yeah... if you are into the extremes you can always custom mount some HUGE heatsink from, say, alpha?

They have a number of 'out-of-spec' heatsinks that seem hard to fit; socket-wise, and will probably give you a lot of troubles and headaches until you have it locked down nice and tight. :)

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: true SILENCE, not QUIETNESS

Post by MikeC » Wed Oct 09, 2002 6:14 pm

hwtf wrote:i am getting an unlocked p3 1.4ghz soon :D
Tell us more about that!

quokked
Posts: 333
Joined: Sun Aug 11, 2002 3:26 pm
Location: Melbourne, Australia
Contact:

Post by quokked » Thu Oct 10, 2002 12:25 am

yeah how'd u manage to score one of those?

hwtf
Posts: 16
Joined: Tue Oct 08, 2002 5:41 pm

Post by hwtf » Thu Oct 10, 2002 7:55 pm

i found an auction on ebay last week at night. it was for a p3 tualatin 1.4ghz unlocked, 10.5mult max. i think intel made the engineering samples such that they cant be run at a higher multiplier than the retail version of it.

at the time, it was going for 120$. at school the next day, i decided to get it. when i came home, THE AUCTION WAS OVER! boom, i felt like crap. one of the rarest chips, and one that i really need for a silent pc, and it was sold already. btw, it sold for $180.

i checked ebay for other currently opened auctions, but to no avail. i checked the already sold items, and i noticed that the guy who sold the 1.4 also sold two other unlocked p3 tualatins.

i registered for an ebay account, and emailed the guy asking him where he got so many of them. he didnt reply to where he got them from, but informed me he still had 1 1.4 left. he offered to sell it for an incredible price of 150$. i immediately agreed to buy it. it should be coming in 1-2 weeks.

of course, ill post pics as soon as i get it. unfortunately, i wont be able to run it for an even longer time. i dont have a fc-pga2 motherboard yet, and i wont be getting any more money for quite some time.

can you believe it? a stock 1.4 costs 212$, and i got an unlocked one for even cheaper! :D :D

Old Dude
Posts: 47
Joined: Fri Oct 11, 2002 10:59 pm
Location: Finland

Post by Old Dude » Fri Oct 11, 2002 11:13 pm

I have both Zalman 6000AlCu and Thermalright SLK-800 and tested their cooling effect without a fan by using an XP 1700+ processor @1000MHz with V-core of 1.2 V. It turned out that 6000AlCu was slightly better than SLK-800, the temps being respectively 72C and 74C. With fans on, the order of the coolers is the opposite.

BTW, I've been lurking here for a while but this is my first post in this forum :D

hwtf
Posts: 16
Joined: Tue Oct 08, 2002 5:41 pm

Post by hwtf » Fri Oct 11, 2002 11:58 pm

thanks, ill go with the zalman. its too bad that they dont make the 6500cu for the s370 platform

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Sat Oct 12, 2002 12:07 am

You know, if there is only 2C dif between the 2 and it is at 72 vs 74, I'd have to vote for the SLK800: it is definitely the better HS with minimal airflow. I haven't finished conclusive testing yet, but it beats my old Swiftech MC462A for T-bird/XP cooling with a Panaflo at 5 or 7v by several degrees C. I don't have all the data compiled, and my test platform has been tweaked more, so have to run some of the tests again, but it looks like about as good as it gets among straight aircooled socket a/370 right now. I definitely prefer the retention clip on the slk800 - the zalman 6000 cost me an xp (though that was partly due to a worn cpu socket lug).

Old Dude
Posts: 47
Joined: Fri Oct 11, 2002 10:59 pm
Location: Finland

Post by Old Dude » Sat Oct 12, 2002 1:51 am

MikeC wrote:You know, if there is only 2C dif between the 2 and it is at 72 vs 74, I'd have to vote for the SLK800: it is definitely the better HS with minimal airflow.
That 2C difference is only indicative because I did not write down the room temperatures. My point however was that those heatsinks are practically equally good on their own even though the SLK-800 is more effective with the fan on. I think that the SLK-800's supremacy with the fan on is based on the heatsink's small bottom area but also on the fact that the fan is attached to the heatsink while the fan of the Zalman HS is blowing from a farther distance.

The temperature distance between the fan on/off is huge: with a Panaflo blowing at 12 V the temperature drops to 38 C! BTW, all the temps I gave are taken by the processor's thermal diode and should thus be as accurate as they can be.
MikeC wrote:I definitely prefer the retention clip on the slk800 - the zalman 6000 cost me an xp (though that was partly due to a worn cpu socket lug).
My preference would also go to the SLK-800 if I had your experience :wink: I however consider Zalman's retention clip at least as good if not better than the one of SLK-800, and their installation tool is really good (you don't have to look for a suitable screwdriver). One advantage of the Zalman over the SLK-800 is that the HS goes automatically on the right place while the SLK800's front edge may go on top of the socket's "threshold" if you position the HS too far forward. On the other hand a definite plus of the SLK-800 is it's three attachment holes per side. So there are pros and cons in both retention clips.

I used the 6000AlCu with a shimm but now use the SLK-800 without it because otherwise the pressure of the heatsink would likely go totally on the core and I don't like that (the retention clip causes a very strong pressure). The use of a shimm in the case of the Zalman may have contributed by a couple of degrees to it's cooling effect without a fan.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Sat Oct 12, 2002 8:37 am

old dude, points taken. i actually did my tests with a panaflo simply perched on top of each HS. no fan distiance issue there. the other disadvantage of the zalman 6000 clip is its length (and the stiffness level of the material used) allows the clip to be bent quite easily. i have 3 zalman HS that use the same clip. none of the clips is quite the same; the one that slipped was more "splayed" (longer) due to slight bending. took very little effort with hands o bend it back... you're right about positioning with the slk800, but 3-lugs per side makes it so secure.

going back to the question of how much temp dif there is with a Panaflo at 5 or 7V - the slk800 beat the mc462A by ~4C; the mc462a beat the zalman by ~2-3C. I won't say those are absolutely correct, I will confirm later. I do recall the excitement of finding a HS that finally beat the swiftech so convincingly; prior, it had been tops against anything, including the zalman 6000cu.

Old Dude
Posts: 47
Joined: Fri Oct 11, 2002 10:59 pm
Location: Finland

Post by Old Dude » Sat Oct 12, 2002 11:54 am

MikeC wrote: going back to the question of how much temp dif there is with a Panaflo at 5 or 7V - the slk800 beat the mc462A by ~4C; the mc462a beat the zalman by ~2-3C.
How high were the absolute temperatures? It's a fact that the higher the temperatures the greater the differences.
MikeC wrote: I do recall the excitement of finding a HS that finally beat the swiftech so convincingly; prior, it had been tops against anything, including the zalman 6000cu.
Unfortunately I experienced only a small temp diff between the SLK-800 and the 6000AlCu. I wonder if the installation of the SLK succeeded at all.

I did some further temp measuring today. The left panel of my Lian Li PC-60 case was off and only the 2 front fans were rotating at the low speed. My XP 1700+ was running at 1000 MHz and the voltage used was 1.15. I measured the CPU temp with a thermal diode, the socket temp and the temp of the HS just above the clip on the rear end with a thermal sensor. I took the temps first with the fan off and then on (Panaflo@12V). I took all temps in BIOS. Here are my results (fan off/fan on) in degrees Celsius:
CPU 76/36, socket 59/30 and HS 49/29. The room temp was 23C during the whole test.

I don't consider the CPU temps too impressive. I suspect that the connection between the core and the heatsink is not like it should be. The big temp differences between the 3 measurement sites when the fan was off indicate that to me. What do you think, MikeC? Could you suggest another test that would tell you if my CPU temps are like they should be?

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Sat Oct 12, 2002 12:51 pm

Those temps are for Z6000al-cu? They don't seem that bad. What motherboard are you running?

I went back to find my notes from when I was trying out the Thermalrights. I couldn't find the data for the Z6000cu; maybe I had already put it off to the side because the Swiftech clearly beat it anyway.

Here's the data I dug up:

*Fan: Panaflo 80mm "L" on an open test system (not in a case). Simply placed on top of the HS.
*Ambient: 23C
*CPU: T-bird 133 x 7 = 933 MHz @1.5V -- obviously very modest. (~35W)
*Test: CPU Stability, 100% CPU load for 15 mins. (reaches max temp within ~10 mins)
*Temp monitor: In-socket thermistor read by MBM5 -- add 8~10C to get approximate core temp.

With Panaflo first @12V, then @5V
-- Thermalright SLK800: 34C, 44C
-- Thermalright AX-7: 35C, 45C
-- Swiftech MC462A: 38C, 48C

The above temps taken with CPU at 100% max for 15 mins.

Again, this test is NOT conclusive. Sooner or later, all the HS on hand will be tested in one long session to start a serious HS review database.

Old Dude
Posts: 47
Joined: Fri Oct 11, 2002 10:59 pm
Location: Finland

Post by Old Dude » Sat Oct 12, 2002 3:45 pm

MikeC: The temps in my previous message were for the SLK-800. My motherboard is Soltek 75DRV5. The trouble in comparing my results with those of various SLK-800 reviews is that they are usually conducted by using a much more effective fan on the HS and/or the temps are usually not measured by a thermal diode.

hwtf
Posts: 16
Joined: Tue Oct 08, 2002 5:41 pm

Post by hwtf » Sun Oct 13, 2002 8:28 pm

decisions decisions :D
i still havent gotten a reply from thermalright on the surface of the slk800, though. anybody know this?

quokked
Posts: 333
Joined: Sun Aug 11, 2002 3:26 pm
Location: Melbourne, Australia
Contact:

Post by quokked » Sun Oct 13, 2002 9:29 pm

hmmm I'd be guesing the SLK-800 has a larger surface area but surface area isn't the be all and end all, u also have to consider how much air gets to the actual fins and the area.
I did find the dimensions on it though :)
Dimension:
L87 x W56.4 x H48 (mm) - Top, without fan
L57.5 x W56.4 x H48 (mm) - Bottom
L25 x W56.4 - Base

Ausone
Posts: 35
Joined: Sun Aug 11, 2002 3:26 pm

AX-7 better than SLK-800 for fanless operation

Post by Ausone » Mon Oct 14, 2002 11:01 pm

For fanless cooling, i.e., cooling without a fan directly on top of a heatsink, a heatsink with more space between fins is better than the one with dense fins even if the latter has slightly better performance with a fan on it. So, I'd rather suggest AX-7 over SLK-800 because both perform similarly but AX-7 has more sparsely placed fins.

Post Reply