Looks like the Alpha Pal 8045 has been dethroned, review of the new Swiftech MCX-462

[Barnaby W. Füi]

link



<<

The biggest difference between the MCX-462 and the MC-462 are the ribbed helicoid-pins that are used on the MCX-462. As you can see, these pins have an increased surface area vs. polished pins and therefore should be able to dissipate heat better. The pin count, at 371, is up from the count of 261, and this, coupled with the ribbed pins has prompted Swiftech to claim a 50% increase in surface area for the unit. The pins have been directly mounted into the base. The other major differences are the bevelled corners of the base that serve, to my knowledge, to allow for the heatsink to fit into spaces with capacitors that may be in the way. This created the need for the fan mounts that protrude out from the heatsink. The bevelled area around the corners was where the fan previously screwed into, and since it was removed, the fan mounts now serve as an attachment point for the fan. As you can see, each side's bracket runs the width of the heatsink, so it is properly braced and secure.

The fan bracket is riveted to the base.

>>

 

[ElFenix]

ah crap the rheostat doesn't come with it!?!

so should i get the rheostat or go get some nice panaflo for this sumbitch? (arrives on wednesday)

 

[Agent004]

How heavy is it?

 

[Antisocial Virge]

The alpha placed 8th at this site but look at the weight of the heatsink tested.
http://www.ipkonfig.com/Reviews/Heatsinks/CoolerMaster_HeatPipe/

 

[ToBeMe]

There was another review out last week about the MCX-462, but, I can't recall or find back where I saw it.......... It was nearly identical but, found the cooling results nearly identical to each other with the 8045 edging out with the Delta. It also sited the MCX-462
as being heavier, harder to install, and quite expensive also. Nice combo, but, think I'll stick with the 8045 in either case........

 

[VBboy]

Ya, but have you seen the PRICE of that thing? If you were considering getting an Athlon XP 1700+ and overclocking it with this, you might as well just get the 1900+ with the cheapest heatsink (which is guaranteed to work with that CPU) - something like a $15 cooler.

They have to stop charging so much money for a piece of metal.. Think about it, you can buy a nice Duron CPU for less than this price - and the CPU has a little more components in it

 

[odog]

yeah with a $20+ price difference between the bare heatsinks i don't think i'd think twice about buying a MCX. i have an Alpha so i might be biased, but the performance is amazing for a $33 heatsink. some people unlike me want the ultimate in performance. they will pay $55 for a heatsink just for 2C with the loudest fan currently in use, and 1C with another really loud fan.

from an engineering point of view, the MCX is hands down technically superior. beating alpha at being alpha was hard but they are doing it..

 

[Pabster]

You're just now realizing PAL8045 has been "dethroned"? heh.

MC462-A outperforms PAL8045 now. MCX462 is even more superior.

 

[odog]

the MCX462 outperforms the 8045 on some cpu's but not on others according to kyle. it loses by a degree F on the 1466mhz tbird but wins by a degree F on the 1650mhz tbird. midget loving



i did consider the MC462 dethorned, will they regain the crown with the MCX462... yeah but they bought the election.

 

[sharkeeper]

Using the thermistor beneath the cpu as a metric to overall heat sink efficiency is bad! Review sites should immediately start using the built in thermistor of the Althon XP/MP when a mainboard that can accurately and reliably interpret its data is available! I have a 8045 and MC462A and the Swiftech allows me to overclock the CPU higher AND with lower voltage using the same fan. This tells me the Swiftech is more efficient. The mainboard thermistor showed temps within a degree of both heatsinks, the the readings are lower on the Swiftech!

Cheers!

 

[woodie1]

It seems to me that 1-2 degrees either way doesn't prove much when you consider the testing method used. There could be that much error IMHO.

 

[Pabster]

Thermistors are terribly inaccurate.

I use a thermocouple drilled in to the base of a heatsink to measure temps. Some argue this is only "slightly" more accurate, but I find it to be considerably more representative of core temperature than a socket thermistor.

Using this method, I've found MC462-A to be superior to PAL8045. I'll have a couple of the new MCX462 next week.

 

[Mikewarrior2]

It is considerably better than socket-thermistors, if for no other reason because socket-thermistors are terrible. They offer too much "user-control" for my liking, but they do work fine. They are still compressed "external" reads, though, therefore, as far as total accuracy, are somewhat lacking. This can be seen with the P3/internal diode and heatsink-thermocouple results as well as DIE simulator testing at overclockers.com.

But this is precisely what several of us mean by socket-thermistor compression. A 1C "tested" difference could be a much bigger "core temp difference".



Mike

 

[Seeko]

What makes the MCX462 tops in my books is that it will fit on any motherboard because of its scaled down size as well as being the best cooler. No more worries of, "Will it fit?", when shopping for a new motherboard.

 

[odog]

<< is that it will fit on any motherboard >>

i know they conform to the AMD design.... but complete compatibility(many motherboards don't)... i'm gonna go check the website now.

 

[odog]

hmm.... they say they can fit on all socket A mobo's with the 4 mounting holes...... that would mean compatibilty with the soltek KT266A board...yummm

 

[ToBeMe]

<< It seems to me that 1-2 degrees either way doesn't prove much when you consider the testing method used. There could be that much error IMHO. >>


LOL! That's exactly what I saw in all of this! I'd bet within days there will be tests which "prove" the 8045 is still better! At any rate, as long as I can O/C my 1900+ to 1800 with my 8045 stablely whenever I want..........I'm not going to run out and by any new heatsink much less a more expensive one!

 

[John]

Glad to see the SK6 is still a top dawg. I considered lapping the surfaces on both of mine.

 

[Seeko]

<< they say they can fit on all socket A mobo's with the 4 mounting holes >>



I forgot to add that. Thanks for doing so.

 

[kgraeme]

I'd be really interested in comparisons of the Alpha and the Swiftech with slow, quiet fans. The review above used a Sunon, but it was still a fairly noisy 50cfm fan. I want to see how they compare with a Panaflo L1A. The Swiftech may come out ahead, but I've seen results flip/flop before.

 

[RC5Bri]

It is an interesting heatsink, but overclockers.com has an article testing this to the MC462A. They rate the MC462A with a lower C/W than the MCX462, 0.30 vs 0.31. They drill a hole into the base of the HS, and insert a thermocouple.