Sythe Zipang 14cm fan "blow-down" CPU cooler

Want to talk about one of the articles in SPCR? Here's the forum for you.
Post Reply
MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Sythe Zipang 14cm fan "blow-down" CPU cooler

Post by MikeC » Sun Jun 01, 2008 6:26 pm


colm
Posts: 409
Joined: Tue Jan 31, 2006 8:22 am
Location: maine

Post by colm » Sun Jun 01, 2008 8:45 pm

darn. it was almost real looking from a distance. ...

my qualms with it are
the base not going all the way up to the fantasy pipes. It is the fins cooling, pipes holding it together, just like the rest of them. Could make something easy enough out of it, but expect to see a model like it for duct design and reverse cooling...maybe..someday... the xp90 may have to get modded for my future.

the fan is way to big for the fin spacing.

Those bottom row of pipes are doing jack squat....and the base must be split in two somewhere for two different direction of pipes? or is it all one mass of piping....

I guess I shrug yet another one off. :roll:

Modo
Posts: 486
Joined: Wed Apr 16, 2008 3:32 am
Location: Poland

Post by Modo » Sun Jun 01, 2008 9:03 pm

colm wrote: Those bottom row of pipes are doing jack squat....
They support the fins, and they transport a little heat to the bottom part of the fins. It's not much, but it's more than "jack squat". ;)
colm wrote: and the base must be split in two somewhere for two different direction of pipes? or is it all one mass of piping....
It's clear from the pictures that the heatpipes wrap around. There is no separate set.

elpibe10
Posts: 114
Joined: Fri Apr 06, 2007 3:57 pm
Location: Out of this world !

Post by elpibe10 » Sun Jun 01, 2008 9:56 pm

6 to 9 deg C higher than a TRUE

and

2 deg C higher (& 2 dBA noisier) than an Andy

That's really poor for a heatsink of such size and price. :!:


Welcome to my WiNDy House of Soldam :

Image

http://www.freewebs.com/elpibe10/

lethul
Posts: 57
Joined: Thu Apr 10, 2008 1:35 pm
Location: Sweden

Post by lethul » Sun Jun 01, 2008 10:14 pm

You've forgotten the C in the title :)

mexell
Posts: 307
Joined: Sat Jun 24, 2006 11:52 am
Location: (near) Berlin, Germany

Post by mexell » Mon Jun 02, 2008 1:07 am

Typo on page one: "Covers the entire entire base..."

Bluefront
*Lifetime Patron*
Posts: 5316
Joined: Sat Jan 18, 2003 2:19 pm
Location: St Louis (county) Missouri USA

Post by Bluefront » Mon Jun 02, 2008 1:37 am

This has got to be the first heat-sink I ever owned and was using.......before an SPCR review. And my impressions are similar, but not exactly the same. My Zipang replaced a Thermalright XP-120, on an Intel D865-GLC board, using a P4-3.4EE (hot). It idles at 110w and goes to about 200w draw running CPUBurn.

I did many bench tests and measurements before installing the setup in my big Lian Li case. The fin pack of the Zipang is about 25mm further from the board than the fins of the XP-120. What I immediately noticed......many of the MB components were running much hotter than when the XP-120 was in there......no matter what fan I tried. I replaced the Zipang fan during the tests with a variety of fans, some 140s, and some 120s, including a 1600rpm S-Flex. The stock Zipang fan returned the worst board temps, although the CPU temp was ok.

It didn't look like the stock Zipang setup was going to work for me when it was in the case......so I did a few mods, and ended up with some "skirts" for the Zipang, and switched over to a 1300rpm 140x20 Aerocool fan. This gave much better board/CPU temps at only a slightly higher noise level.

Image

My impression was that the increased distance from the MB, combined with the lower pressure of the Zipang fan was to blame. The relatively low CFM, combined with the larger diameter of the fan, resulted in less airflow actually reaching the MB components. The new mod setup performs better....for me with my setup. YMMV.

Startup fan problems..... similar for me to the SPCR review. My controller gives the Zipang fan a 12v burst to get moving, then drops back to 5V till the sensor warms up. When the controller drops back to 5V, the Zipang fan stops completely, and doesn't start spinning till about 8-9V. At that time the CPU has reached a fairly high temp, as well as everything else on the MB. So you can see why I stopped using the Zipang fan.

IMHO.... a Yate Loon 140x25mm medium speed fan would be a better choice. I'm using the 140x20mm Aerocool mostly for clearance purposes. An Evercool Red Scorpion is another possibility....although noisier.

Remember I'm trying to cool a hot-running MB/CPU..... using a different setup (cooler) and the Zipang stock setup might be perfect. Using the stock Scythe mounting pieces, this cooler can mount in four different positions, so compatibility with many boards should be no problem.

I like the way mine performs in the case... after the modding.

Nice review....
:)

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

Post by MikeC » Mon Jun 02, 2008 5:45 am

elpibe10 wrote:6 to 9 deg C higher than a TRUE

and

2 deg C higher (& 2 dBA noisier) than an Andy

That's really poor for a heatsink of such size and price. :!:
Well, you can look at it that way if you like, but....

1) the zipang comes with a v. quiet fan, the TRUE does not.
2) nothing else we've tested comes close to the TRUE anyway.
3) the SPL is often misleading -- the 2 dBA difference in this case is barely audible. Just listen to the sound recordings.
4) 2C, imo, is too small a difference to consider a big deal. Take a slightly different pair of samples, and it might even go the other way -- note that at the more SPCR-likely 9V fan feed, the cooling result of zipang & andy was the same.
5) like the Andy, the price of the zipang will drop over time.

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

Post by NeilBlanchard » Mon Jun 02, 2008 8:05 am

Hello Mike,

I'm slightly confused by the insertion of the Nexus fan data -- I'm assuming it is there for reference only (in relation to the 140mm Scythe fan), and that the Scythe fan was used with the HS.

It is good to have the Nexus data down to the lower ambient level. But for this article, I think it would be clearer to put that paragraph in a box, to make it's purpose more apparent.

Erssa
Posts: 1421
Joined: Sat Mar 12, 2005 9:26 pm
Location: Finland

Post by Erssa » Mon Jun 02, 2008 8:08 am

Good review as always. I was more excited to see 15dBA ambient test results, then the actual performance of Zipang. Did you already do something to your test lab, or was the test just made in the middle of the night?

fri2219
Posts: 222
Joined: Mon Feb 05, 2007 4:14 pm
Location: Forkbomb, New South Wales

Push pin mounts. Just say no.

Post by fri2219 » Mon Jun 02, 2008 8:41 am

One thing I didn't pick up from the review- would it be physically possible to use a Thermalright bolt-through kit with this? It looks like the overhang would prevent it.

I'm glad non-tower, non-flower, heatpipe heatsinks are finally getting some attention from vendors- hopefully, the Arctic Cooling Freezer 7 LP actually shows up in retail channels some day.

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

Post by MikeC » Mon Jun 02, 2008 9:04 am

NeilBlanchard wrote:I think it would be clearer to put that paragraph in a box, to make it's purpose more apparent.
Good point; done.
Erssa wrote:Good review as always. I was more excited to see 15dBA ambient test results, then the actual performance of Zipang. Did you already do something to your test lab, or was the test just made in the middle of the night?
It was done at 5am -- but you're right that there was another change. The new SpectraRTA + ACO Pacific mic were used for the first time. We're finding differences in readings between the new PC-based SLM and the B&K 2203. They have to be expected anyway; no way for them to track identically across all frequencies and levels. The dynamic tracking is different -- the new SLM reads lower <20 dBA and higher at >30 dBA. Differences in the low end, where it counts the most for SPCR, are 1-2 dBA. Where the B&K might read 16 dBA, the new SLM reads 14-15 dBA. Between the old and new, we have to trust the new: The mic is brand new, its internal noise way lower, and the 24bit/96kHz soundcard also has lowe noise electronics and should have much better dynamic tracking all the way up and down the dB range.

The old B&K has dual switches to move its narrow 20-dB range analog meter from 0-20dB to 10-30db to 20-40db and so on. We find discontinuities as the ranges are switched. For example, when set to 10-30db, we might get a reading of 24dB, but when set to 20-40dB, we get 22 dB. Which is correct? Difficult to sort out, although it can and has been done. All the testing done on the new system indicates that it is the more accurate system. We will do more calibration testing on it before introducing it officially and laying out all the details.

The current challenge is to find a signal generator with a calibrated output. This would be used to simulate mic head output of very low signals to check the dB accuracy of the whole system.

bgavin
Posts: 160
Joined: Sat Apr 12, 2008 5:05 pm
Location: Orangevale, CA

Post by bgavin » Mon Jun 02, 2008 2:52 pm

Thanks for the excellent and thorough review.

As noted in your Summary, the mounting for this cooler is the weak point. I own three Zipang, and have one of them mounted with the stock push-pin arrangement. I view this to be flimsy at best, but this machine is used with adult supervision.

I have experimented with the LGA-775 kits. It can be used... with great difficulty. The problem is the bolts are Phillips only, no hex heads for wrench access. Access to the Phillips head is impossible unless you have or fabricate a special tool. There is no top-down access for a Phillips driver, due to tight fin clearance. Scythe tells me they are preparing to release a new mounting kit "shortly". No other details were given to me.

I like the symmetric support of the Zipang with heat pipes on both sides, plus the center base. This virtually eliminates any flexing and/or fracture of the heat pipes. As you noteded, one level of heat pipe has only partial contact with the fins. The 140mm wing span appearts to be an after thought, as the heat pipes are grouped under the 120mm fan section.

Installing a 120mm fan is really not an option unless one resorts to hacks such as zip ties. The springs do not retain the smaller 120mm fan. My retailer tells me Evercool is releasing their 140mm line of fluid bearing fans in June. Their speed models are similar to Panaflo, so there is an RPM to suit everybody.

I favor down-draft cooling, and am waiting to see how the AXP-140 fares.

leifeinar
Posts: 108
Joined: Thu Oct 04, 2007 1:24 am
Location: tromsø in norway

Post by leifeinar » Tue Jun 03, 2008 3:06 am

is it possible to buy the fan ?

bgavin
Posts: 160
Joined: Sat Apr 12, 2008 5:05 pm
Location: Orangevale, CA

Post by bgavin » Tue Jun 03, 2008 3:44 am

Bluefront wrote:What I immediately noticed......many of the MB components were running much hotter...
I noticed the same when testing the stock Zipang and the Intel cooler with a hot running Prescott P4 530.

The difference was cool vs. scalding. I could not keep my fingers on the VRM components with the stock Zipang fan... the temp difference was enormous. I used a 120x38 Panaflo 12H to get VRM temps down to the Intel level, but the noise level was unacceptable.

Bluefront
*Lifetime Patron*
Posts: 5316
Joined: Sat Jan 18, 2003 2:19 pm
Location: St Louis (county) Missouri USA

Post by Bluefront » Tue Jun 03, 2008 4:35 am

The relative temperature differences between the various MB components on various boards, differs greatly. Some boards just don't run extremely hot. The D865-GLC has some hot spots. The combination of the skirts and the different fan on my setup, lowered some of these board components >15C. (I used a laser temperature gun to check). Nothing burns be anymore.... :lol:
Last edited by Bluefront on Tue Jun 03, 2008 8:04 am, edited 1 time in total.

bgavin
Posts: 160
Joined: Sat Apr 12, 2008 5:05 pm
Location: Orangevale, CA

Post by bgavin » Tue Jun 03, 2008 7:06 am

Bluefront, do you see skirts as a viable solution for tower coolers and VRM?

It seems possible to skirt the lower 20% of a tower cooler and direct it toward the bulk of the VRM components on the I/O shield side of the socket.

Bluefront
*Lifetime Patron*
Posts: 5316
Joined: Sat Jan 18, 2003 2:19 pm
Location: St Louis (county) Missouri USA

Post by Bluefront » Tue Jun 03, 2008 12:00 pm

Heh.... start a new thread on the subject and I'll try to find the links to an invention of last year. It's a Ninja turned into a partial blow-down heat-sink. This thread isn't the place for that subject. :)

Post Reply