Xigmatek HDT-S1283 & SD964 "heatpipe direct-touch&q

Want to talk about one of the articles in SPCR? Here's the forum for you.
NeilBlanchard
Moderator
Posts: 7681
Joined: Mon Dec 09, 2002 7:11 pm
Location: Maynard, MA, Eaarth
Contact:

Post by NeilBlanchard » Thu Apr 03, 2008 7:02 am

Hello,

Well digging into the KingWin pages on this HSF:

http://www.kingwin.com/product_pages/rvt12025.asp

I found in the PDF manual that they are indeed 8mm heatpipes. So, the KingWin version is the best deal at NewEgg ($4.01 less shipped than the Xigmatek):

http://www.newegg.com/Product/Product.a ... 6835124019
http://www.newegg.com/Product/Product.a ... 6835233003
Last edited by NeilBlanchard on Thu Apr 03, 2008 7:22 pm, edited 1 time in total.

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

Post by MikeC » Thu Apr 03, 2008 7:04 am

Tzupy wrote:I personally don't trust this "direct touch" approach, I firmly believe that heatpipes should be protected from the mounting pressure (and accidents).
Those who want to use these heatsinks however, should know that they perform better on larger (older process) CPU dies, as shown in a Xbitlabs review.
When using on a 45nm CPU the performance could be dissapointing. Of course, SPCR shouldn't have tested them *only* on an obsolete 65 nm CPU.
One more thought: the Xbitlabs review found out that the orientation of the 8 mm heatpipes matters a lot (but they tested the Xigmatek D1284).
I don't buy this -- (re- older CPUs vs newer ones) -- heat is heat underneath the heatspreader. It's true we don't have as much practical experience with the Xigmatek HS (as most of the other high end coolers we like). We also haven't played much with 4-cores. This is something we're willing to put to the test (with a 4-core CPU vs our old D850) -- when we can find some time.

I can't see any validity in your argument about protecting heatpipes -- do you have any idea how hard copper is? There's absolutely no way you could "damage" the heatpipes in the base of these "direct touch" heatpipe HS while installing it. You'd have to take a hammer and/or chisel to it.

I also have to point out that there is a major difference between the D1284 and other HS in that xbitlabs review (as well as the models we tested): Yes, the D1284 has 4 heatpipes, and they're configured in such a way that they act like 4 heatpipes. In all tower heatsinks, because each heatpipe is bent into a U, each heatpipe actually acts like 2 heatpipes. I'd expect the D1284 to lose out to other HS that effectively have a larger number of heatpipes.

Tzupy
*Lifetime Patron*
Posts: 1561
Joined: Wed Jan 12, 2005 10:47 am
Location: Bucharest, Romania

Post by Tzupy » Thu Apr 03, 2008 8:07 am

@MikeC: thank you for answering my rant. :wink:
I'm looking forward to read the SPCR review(s) that will clarify this old vs. new process (actually the area matters) cooling difference.
I believe that a smaller core, like the popular E8400, might only make direct (sort of) thermal contact with the middle 2 out of 4 heatpipes.
In such a case a Thermalright approach, with the 4 heatpipes soldered to the base on both sides, makes better use of the outer heatpipes.
BTW, any chance to review the SI-128 SE (the one with bolt-through) anytime soon?

From my experience with copper, it's rather "soft", scratches easily, and I don't know how thick the heatpipe walls are.
I'm used to the nickel plated heatpipes in the Thermalright products. :lol:

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Thu Apr 03, 2008 8:15 am

Tzupy wrote:I personally don't trust this "direct touch" approach, I firmly believe that heatpipes should be protected from the mounting pressure (and accidents).

Those who want to use these heatsinks however, should know that they perform better on larger (older process) CPU dies, as shown in a Xbitlabs review.

When using on a 45nm CPU the performance could be disappointing. Of course, SPCR shouldn't have tested them *only* on an obsolete 65 nm CPU.
Well that is certainly an elitist attitude you have there.

What about my 130nm and 90nm chips I'm still using? I have a 65nm system as well but I don't care as much about the nm as the TDP.

If INTC or AMD offers me a CPU with TDP, Price, and performance all in a range I'm willing to pay for I don't care what process node it comes from.

Quad cores might be getting most of the buzz but I'll be plenty happy with dual core chips for some time to come and I might buy a Tri core chip just to say I have one. I have no need for 4 or more cores in my current usage patterns.

I expect the primary driving force that will push me to having more than 2 cores will be price. At some point I won't be able to buy a dual core for a significantly cheaper price than a Tri Core so I'll get a Tri. At some point after that I won't be able to get a Tri Core significantly cheaper than a Quad core so I'll buy a Quad.

Right now the cheapest Single core is about $35, cheapest Dual core is about $50, and the cheapest Quad core is about $200.

There is a wide, WIDE range of products and users out there. 45nm quad core users will not be a large percentage of SPCR readership anytime soon, at least not for some number of months or years to come...

NeilBlanchard
Moderator
Posts: 7681
Joined: Mon Dec 09, 2002 7:11 pm
Location: Maynard, MA, Eaarth
Contact:

Post by NeilBlanchard » Thu Apr 03, 2008 8:30 am

Hello,

Please remember a couple of things: The heatspreader is (relatively) flat, and it is wider than ~2 of the heatpipes. Also, there are aluminum "ribs" between the hollow copper heatpipes -- so, the aluminum would have to also be crushed in order to bend the copper heatpipes.

It isn't called a "heatspreader" for nuthin' -- all the heatpipes are in contact with the heatspreader. Also, the grooved aluminum base is in contact with the heatspreader and with the heatpipes; so I think that the heatpipes that are a little off center from the actual CPU will still get heated.

In fact, they will get heat more "directly" than they do on the big Thermalright. The Ninja has a second level of heatpipes, and I think it is the contact (or lack thereof) with these that hurt the Rev. B's performance.
Last edited by NeilBlanchard on Thu Apr 03, 2008 8:38 am, edited 1 time in total.

Mats
Posts: 3044
Joined: Fri Dec 26, 2003 6:54 am
Location: Sweden

Post by Mats » Thu Apr 03, 2008 8:34 am

vick1000 wrote:Besides, the actual size of the cores is increasing as well, since they add transitors and cache with most updates.
Have a look here. The latest C2D (6 MB, 107 mm^2) have ten times more transistors and dual cores compared to the first P4 (256 kB, 217 mm^2), while still having half the die size.
It is close in size to the Dothan (2 MB 84 mm^2) and Yonah (2 MB, 90 mm^2).

Jipa
Posts: 167
Joined: Sat Dec 22, 2007 4:06 am
Location: Tampere, Finland
Contact:

Post by Jipa » Thu Apr 03, 2008 10:34 am

The only downside that I can see on these direct touch heatpipes is the fact that one solid block of copper is so much easier to make flat than three pipes and a milled block of aluminum. (here are some images to demonstrate the flatness of the base)

Despite the possibly uneven base there's no doubt that the Xigmatek HDT-S1283 is (atleast one of the-) best in price/performance ratio right now, dunno if it's cause this special base or not...

FlorisNielssen
Posts: 141
Joined: Fri Aug 10, 2007 1:33 pm
Location: The Netherlands
Contact:

Post by FlorisNielssen » Thu Apr 03, 2008 10:44 am

MikeC wrote:
FlorisNielssen wrote:Interesting review.

One point of objection. In the summary you say:
Review wrote: * LGA775 mounting could be improved
While earlier on you said:
Review wrote:Once the heatsink was secured, the bowing of the board around the CPU socket made it clear that there's quite a bit of tension applied between the base and the CPU. Inadequate pressure is not likely to be a problem here unless you mount the pins improperly.
I think you can leave it out the bad points of the summary. It cools very well, with the stock Intel pushpins and the mounting seems good. So why a bad point? I know you don't like the pushpins and I know why. But here it's sufficient.
It looks to me as if you have something personal to these pushpins, even metioning them as bad when they are not.
I don't see why we can't hold both opinions at the same time? They are not really self-contradictory. Yes, it works well enough but it could be better -- and we'd be happier with a better mounting system. And yes, we do have something personal against the pushpins -- umpteen cuts and many drops of blood trying to use them over the years with many different HS. This does tend to create a certain attitude. :? :x :lol:
Then I misunderstood. I thought it would (again) be about the pressure, but since it's enough this time, that seemed to be misplaced. But then now it isn't and I see your point.

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

Post by MikeC » Thu Apr 03, 2008 10:44 am

Jipa wrote:Despite the possibly uneven base there's no doubt that the Xigmatek HDT-S1283 is (atleast one of the-) best in price/performance ratio right now, dunno if it's cause this special base or not...
What's not to know? The base is the only thing that makes the xigmatek significantly different from other similar coolers.

F for Fragging
Posts: 46
Joined: Tue Dec 11, 2007 11:08 am
Location: The Netherlands

Post by F for Fragging » Thu Apr 03, 2008 12:50 pm

Very interesting heat sink, if it performs so well for a lower price than comparable heat sinks from Scythe and Thermalright. Unfortunately it isn't widely available it seems, I can't find it for sale over here in The Netherlands.
While $35 USD should be equal to €22,42 according to Google, the European webshops listed in the "Where to buy" section of the Xigmatek website sell it for prices around €30. And the Dutch webstore where I intend to buy my hardware sells the Thermalright HR-01 Plus for €35, which is a negligible price difference. And the HR-01 Plus doesn't have push pins.

Thanks a lot for the review. But as was already written in the comments on the first page of this topic, the review doesn't mention what the 92mm reference fan is, please add this information to the review. And on the first page of the review, the dimensions of the heatsinks are given like this: "120(W) x 50(H) x 159(D) mm". Even if this is exactly how it is written on Xigmatek's website, could this please be changed to: "120(L) x 50(W) x 159(H) mm"?

In the review it is written that the Kingwin Revolution and the OCZ Vendetta are respectively identical and very similar to the Xigmatek HDT-S1283, does that mean it wouldn't be worth the time to review these as well?

Hydrazine
Posts: 18
Joined: Sat Mar 29, 2008 12:03 pm
Location: betweent keyboard and chair (chair lives in the Netherlands)

Post by Hydrazine » Thu Apr 03, 2008 1:04 pm

F for Fragging wrote:Very interesting heat sink, if it performs so well for a lower price than comparable heat sinks from Scythe and Thermalright. Unfortunately it isn't widely available it seems, I can't find it for sale over here in The Netherlands.
While $35 USD should be equal to €22,42 according to Google, the European webshops listed in the "Where to buy" section of the Xigmatek website sell it for prices around €30. And the Dutch webstore where I intend to buy my hardware sells the Thermalright HR-01 Plus for €35, which is a negligible price difference. And the HR-01 Plus doesn't have push pins.

Thanks a lot for the review. But as was already written in the comments on the first page of this topic, the review doesn't mention what the 92mm reference fan is, please add this information to the review. And on the first page of the review, the dimensions of the heatsinks are given like this: "120(W) x 50(H) x 159(D) mm". Even if this is exactly how it is written on Xigmatek's website, could this please be changed to: "120(L) x 50(W) x 159(H) mm"?

In the review it is written that the Kingwin Revolution and the OCZ Vendetta are respectively identical and very similar to the Xigmatek HDT-S1283, does that mean it wouldn't be worth the time to review these as well?
this dutch site has it for sale :wink:
http://sallandautomatisering.nl/
unfortunately it costs 33 euro's :(

seraphyn
Posts: 322
Joined: Wed Nov 28, 2007 1:26 pm
Location: Netherlands

Post by seraphyn » Thu Apr 03, 2008 1:17 pm

So annoying when shops seem to do a straight $1 = 1 Euro.

Would be interesting to see Thermalright use this type of base on their TRUE.

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

Post by MikeC » Thu Apr 03, 2008 1:27 pm

92mm reference fan -- Nexus 92.
article wrote:Testing was done according to our unique heatsink testing methodology, and the reference fan was profiled using our standard fan testing methodology. A quick summary of the components, tools, and procedures follows below.
See the top of page 5 for the link. http://www.silentpcreview.com/article818-page5.html

Mats
Posts: 3044
Joined: Fri Dec 26, 2003 6:54 am
Location: Sweden

Post by Mats » Thu Apr 03, 2008 1:30 pm

seraphyn wrote:So annoying when shops seem to do a straight $1 = 1 Euro.
Taxes is usually the reason, compare the prices without it and you'll see what I mean.

cmthomson
Posts: 1266
Joined: Sun Oct 09, 2005 8:35 am
Location: Pleasanton, CA

Post by cmthomson » Thu Apr 03, 2008 2:56 pm

The heat path for all LGA775 Intel CPUs is the same. The die is mounted transistors-down (known as flip chip mounting) to give the best electrical path from the die pads to the socket lands. This means that the heat from the transistors is conducted through the die (silicon) to a TIM layer to the bottom of the IHS (integrated heat spreader). Then through the IHS to the external TIM, and finally to the external heat sink.

In most heat sink designs (including TRUE and Ninja), the path through the external heat sink is from the external TIM to the base block, through another TIM or solder or in some cases just pressure to the outside of the heat pipes, through the body of the heat pipes to the working fluid, then up the pipe via evaporation to cool spots where the heat pipes touch the fins, then to the fins via condensation and conduction through the walls of the heat pipes and then through solder (or sometimes just pressure) to the fins themselves. At last the fins shed the heat to the air flowing by. Whew!

Contrary to a post above, all Intel IHSs are copper (with a thin layer of nickel on top for hardness; if you lap your CPU, the nickel will be mostly or entirely removed). The IHS conducts better than the die or the two layers of TIM it sits between, and it does indeed spread the heat for a larger effective surface area to contact the external heat sink.

The history of the Ninja is instructive: the first version had the heat pipes soldered to the base, and had a high-pressure backplate-based mounting system (because at the time AMD 939 mounting was dominant, and the backplate and rail system was the simplest way to emulate it). In response to complaints about installation difficulty (need for tools, etc), Scythe adopted the push-pin mounting with the Rev B Ninja, but in the process reduced the mounting pressure resulting in lower performance. Then with the copper Ninja, they went to just press-fitting the heat pipes, so there was both low mounting pressure and poor heat conductivity between the base and the heat pipes, resulting is very surprisingly poor performance.

Now Xigmatek comes along with direct contact between the external TIM and the heat pipe bodies, eliminating two steps in the path, and with a push-pin mounting that is much snugger than Scythe's. Net result: finally an available heat sink competitive with the original Ninja. Yay!

andyb
Patron of SPCR
Posts: 3307
Joined: Wed Dec 15, 2004 12:00 pm
Location: Essex, England

Post by andyb » Thu Apr 03, 2008 4:26 pm

Well put "cmthomson"
The history of the Ninja is instructive: the first version had the heat pipes soldered to the base, and had a high-pressure backplate-based mounting system (because at the time AMD 939 mounting was dominant, and the backplate and rail system was the simplest way to emulate it). In response to complaints about installation difficulty (need for tools, etc), Scythe adopted the push-pin mounting with the Rev B Ninja, but in the process reduced the mounting pressure resulting in lower performance. Then with the copper Ninja, they went to just press-fitting the heat pipes, so there was both low mounting pressure and poor heat conductivity between the base and the heat pipes, resulting is very surprisingly poor performance.

Now Xigmatek comes along with direct contact between the external TIM and the heat pipe bodies, eliminating two steps in the path, and with a push-pin mounting that is much snugger than Scythe's. Net result: finally an available heat sink competitive with the original Ninja. Yay!
I'm glad I have a rev 1 Ninja - its got a scary amount of pressure, and that as SPCR have proven makes a big difference.

Xigmatek have got off to a good start, now to make it even better, make square heatpipes and forget the aluminium between the heatpipes totally, and sort out some decent mounting option for AM2, and bolt through options for 775.


Andy

Mats
Posts: 3044
Joined: Fri Dec 26, 2003 6:54 am
Location: Sweden

Post by Mats » Thu Apr 03, 2008 4:47 pm

The surface mounted heatpipe technology can only become better:
If you can remove the aluminium between the pipes in the picture and add another heatpipe, without loosing stability, you probably get even better cooling performance.
Combine this with heatpipes that are flattened vertically, whch will make them use less width, add one or two more pipes and machine the surface completely flat, and you have the next high end Thermalright
Image
Pretty much like the base of a Zalman CNPS7000 is made, but with heatpipes instead of sheet metal.

Edit: LOL, Andy was quicker.
Last edited by Mats on Thu Apr 03, 2008 4:51 pm, edited 1 time in total.

Mats
Posts: 3044
Joined: Fri Dec 26, 2003 6:54 am
Location: Sweden

Post by Mats » Thu Apr 03, 2008 4:51 pm

Is it possible to remove the push pins, and use bolts, nuts and springs instead?
I think I read how to do that with a Scythe infinity.

Edit: Here it is, with pictures of a properly lapped IHS.
Edit: How to lap an IHS.

lucas82
Posts: 88
Joined: Mon Jun 04, 2007 12:24 pm
Location: Santa Maria - Brasil

Post by lucas82 » Thu Apr 03, 2008 5:34 pm

There is another company with coolers that have the same technology.
http://www.3rsys.com/english/products/aircooler.asp
I don't know if it's the same product under a different brand, it seems so.

Mats
Posts: 3044
Joined: Fri Dec 26, 2003 6:54 am
Location: Sweden

Post by Mats » Thu Apr 03, 2008 5:45 pm

lucas82: Looks a lot like it, but it got four heatpipes instead of three .

lucas82
Posts: 88
Joined: Mon Jun 04, 2007 12:24 pm
Location: Santa Maria - Brasil

Post by lucas82 » Thu Apr 03, 2008 5:47 pm

Yes, and the shape of the aluminium fins is different, too.
And the LGA775 mounting mechanism seems to be better.

Mats
Posts: 3044
Joined: Fri Dec 26, 2003 6:54 am
Location: Sweden

Post by Mats » Thu Apr 03, 2008 5:52 pm

Now I saw the bigger one, the Prima 120. Bigger fins and AM2 mounting in the other direction.

Here you can see where to buy them, many countries included, like Brazil and Sweden! :D

lucas82
Posts: 88
Joined: Mon Jun 04, 2007 12:24 pm
Location: Santa Maria - Brasil

Post by lucas82 » Thu Apr 03, 2008 6:04 pm

http://www.anandtech.com/casecoolingpsu ... i=3010&p=1
According to anandtech this iCEAGE cooler is very good (but far away from TRUE), but they tested it with the stock fan (2000rpm).

Howard
Posts: 275
Joined: Wed Feb 08, 2006 5:33 pm

Post by Howard » Thu Apr 03, 2008 6:14 pm

Too bad they didn't plate the pipes for corrosion resistance.

FartingBob
Patron of SPCR
Posts: 744
Joined: Tue Mar 04, 2008 4:05 am
Location: London
Contact:

Post by FartingBob » Fri Apr 04, 2008 6:06 am

Howard wrote:Too bad they didn't plate the pipes for corrosion resistance.
Unless you plan to use the thing for a decade your probably ok.

nageltass
Posts: 10
Joined: Sat Mar 08, 2008 6:21 am
Location: Sweden

Post by nageltass » Fri Apr 04, 2008 2:47 pm

This will most certainly set the new standard, as every other new heatsink model released from now on will utilize this direct-contact thingy.

cmthomson
Posts: 1266
Joined: Sun Oct 09, 2005 8:35 am
Location: Pleasanton, CA

Post by cmthomson » Fri Apr 04, 2008 5:53 pm

Howard wrote:Too bad they didn't plate the pipes for corrosion resistance.
All TIMs will prevent corrosion, since they are basically grease/metal combinations, which exclude air (ie, oxygen). That said, it may be worthwhile to touch up the heat sink before mounting if it isn't shiny, with some acidic solution (eg, vinegar).

niels007
Posts: 451
Joined: Mon Nov 22, 2004 2:18 am

Post by niels007 » Sat Apr 05, 2008 2:41 am

SPCR, or anyone else, does the ThermalRight 775 bolt though kit work on this cooler, and more importantly, can it then be mounted in each 90deg orientation?

ta..;)

rpc180
*Lifetime Patron*
Posts: 309
Joined: Sun Feb 09, 2003 8:01 pm
Location: Washington, DC

Post by rpc180 » Sat Apr 05, 2008 1:19 pm

Apparantly, the cooler's been making rounds in the US for a bit of time! OCForums has a thread on it with the TR-775 bolt thru kit, check out this photo:

Image

About the orientation, on page 3 someone eluded that it could go in any direction... I should be getting one in a week or two with a new build.

Sooty
Posts: 625
Joined: Mon Sep 22, 2003 5:15 am
Location: UK

Post by Sooty » Sat Apr 05, 2008 2:37 pm

rpc180 wrote:About the orientation, on page 3 someone eluded that it could go in any direction... I should be getting one in a week or two with a new build.
Can I use that kit with my A8V Deluxe (K8 939), to get the HDT-S1283 facing rearwards, towards the exhaust case fan, rather than the PSU?

Image

Post Reply