Fans, fans, fans

[time]

This is a thread for people to share their experiences with different cooling fans; a sort of group therapy where you can pillory brands or models that caused you pain and suffering.

There's a lot of hot air in fan specifications and claims. Air flow always seems to be overstated and noise understated. And there's no mention of v i b r a t i o n.

Here's a little puzzle to start things off.

One of the PCs here has a Thermoengine sitting atop a 1GHz Thunderbird. The original low-profile fan (which was reasonably quiet) started to make threatening noises after only a few months. It became quite antisocial, in fact, blowing one continuous raspberry.

Low-profile fans are hard to find here (another story), so I tried an 80-60mm adaptor. The towering cooler that resulted nearly touched the side of the case, didn't perform very well, and threatened to snap the lugs off the CPU socket with all that weight and leverage.

I've had a lug break off - it's not nice.

Then I read about Sunon's new "magnetic levitation system", with "the Magic of Magic Magnetic Plate". Obviously, these were the fans for me, so I bought a couple of 60mm ones.

After installation, the first thing I noticed was that the PC was trying to be a hovercraft. Unfortunately, this was not due to a stupendous flow of air, but rather by vibrating so much it no longer remained in the one place.

This particular Sunon model runs at 5500rpm, so I inserted a speed controller and dropped the revs to about 4000rpm. I could now hear myself think, and the PC stopped shimmering.

Interstate enquiries turned up a different Sunon model that churned around at just over 4000rpm, so I ordered a couple of these. Well, they were quieter alright, but I couldn't help but notice that they were still louder than the trusty 4300rpm YS-Techs that used to be everywhere.

As far as I can see, both Sunons are identical apart from motor current. But their design is nowhere near good enough to support the more powerful motor.

Anyway, I was still unhappy with the noise, so I slowed the fan down to 2750rpm. Now it was tolerable, but the CPU temperature ran several degrees hotter than with the original Thermoengine fan. Strangely, speeding up the fan didn't seem to impove this by more than a couple of degrees.

Recently, I ordered a Papst 60mm fan. It's rated at only about 16cfm and spins at just 3000rpm, so I didn't think it would be suitable for this application. But I tried it anyway.

Noise is about the same as the slowed Sunon, although I think this particular Papst may be slightly damaged (it was the last one in stock). However, CPU temperatures were now several degrees cooler!!!

I know that some fans are better than others under load. Papst and Panaflo quote pressure figures as well as unloaded CFM. But this has really rocked me. With two fans spinning at roughly the same speed, how can one be so devastatingly more effective?

 

[Jake the Dog]

did you take the HS off the board to change the fans?

 

[time]

No. :-?

 

[Jake the Dog]

well there goes that theory :roll:

 

[Mercutio]

I'd just like to contribute by saying that the piece of human filth who decided to put fans on northbridges needs to be removed from the no-doubt shallow end of the gene pool he's presently muddying.

Was that bitter enough?

No, probably not, but I somehow doubt anyone will disagree with the sentiment.

Either put a decent-sized heat sink on them or make reliable fan. Either way, I'm happy, but if I "fix" one more PC by unplugging a northbridge fan, you can all start calling me "The guy in the clock tower with the high-powered rifle" instead of Mercutio.

 

[blakerwry]

Personally, I've found Sunons to be short lived... they usually seem tpo die after a year if I'm lucky...


Panflo and Nidek have been good to me, I have several of these fans that have lasted over 10 years... Panflo is known for quality and quiet operation.

I've been happy with Panflo, Antec, and enermax fans... even though Enermax and Antec are most likely rebadging the fans.

 

[NRG = mc²]

Just to chime in - Sunon's 60mm Vapo-Bearing fans are pretty much free of any sort of bearing noise. There Zalman compact flower (5100 series) uses them, and they make very little noise when spinning, compared to sleeve and ball bearing fans I've seen.

I assume its some sort of fluid bearing system.

 

[time]

That would be a lot of trouble to go to just to shoot poxy 40mm fans ...

 

[Howell]

I understnad the difference to be in the way the blades of the fan are shaped possibly as well as the way the outer cylinder of the air path is shaped. Supposed to cause less turbulance and thus quieter operation.

 

[Buck]

Personally, there should never be the need for a fan to be placed on the Northbridge - as Mercutio points out, a quality heatsink should work very well. In addition (as I've lamented before) I wish fans could be abolished from higher-end graphics cards. Too many smalls fans result in too many problems (and generates unnecessary noise)!

Mercutio, your words regarding the "piece of human filth who decided to put fans on northbridges" were indeed kind.

 

[LiamC]

I can't claim this - saw it over on Ace's.

Volcano 7+ - take the wacky ThermalTake plate off of the top of the fan - it's noticeably quieter. Marketing gone amuck methinks.

Also, for the last four or five cases I've had, I've been adding 80mm exhaust and 80 or 92mm inlet fans. I perused AMD's thermal design PDF and let it sit in the back of my head for a while. Last week I unplugged the inlet fans in my wifes PC (80mm inlet in bottom front 80mm exhaust + p/s fan) and my server (92mm inlet in bottom of case - cooling hdd's, 80mm exhaust and p/s fan). End result - wifes CPU temps went down by a couple of degrees (C) - I checked and re-checked this, according to mobo monitor whilst my server CPU temps went up by a staggering 1 degree C. It might pay to do a little checking after modding.[/b]

 

[Handruin]

Personally, I've found Sunons to be short lived... they usually seem tpo die after a year if I'm lucky...


Panflo and Nidek have been good to me, I have several of these fans that have lasted over 10 years... Panflo is known for quality and quiet operation.

I've been happy with Panflo, Antec, and enermax fans... even though Enermax and Antec are most likely rebadging the fans.

All 6 of my sunons have been running well over two years now 24x7. All 6 continue to run great.

 

[blakerwry]

maybe I've just gotten the bad ones... MY case came with two, both died within 1 year... I also got 2 with a athlon slot A heatsink... they both died as well...

They are a trusted brand name, but from my personal experience they haven't performed very well at all.

 

[time]

Unfortunately, I'd have to agree with Blakerwry.

From memory, the first 80mm fan that ever failed on me was a Sunon.

I didn't mention my complaints about their 80mm fans, but they exist alright. Again, excessive vibration (even without Magnetic Levitation) was the killer. In particular, I was looking for a fan to replace the heap of doggie-doo-doo that Thermaltake shipped with the Volcano 7.

I had an assortment of Sunon fans, and I was determined to use them. But the only place I could use even the slowest was firmly bolted to the case - otherwise it turned the Volcano 7 into something favored by single women.

I eventually used a pair of 1900rpm Panaflos, one on the Volcano and one as the case exhaust. With the CPU socket orientated North-South, the cooler blew virtually straight into the exhaust fan, and that was one quiet PC.

To be fair, I don't think much of the 3000rpm Panaflo either. At least it's higher noise rating is believable. But IMO yet another fan that isn't good enough to support a bigger motor.

And finally, everyone knows how unbelievably bad Sunon 120mm fans are. Other 120mm fans are quiet, but not Sunon's.

 

[jtr1962]

I'll echo the sentiments about putting fans on the Northbridge or video cards. A large passive heat sink will do just fine in either case, or if you really must have a fan, don't have one of those miserable 25mm jobs that move less air than a moth and sound like mosquitoes on steriods. When my saw my friend's new PC last year with both types of fans I honestly asked him how the hell he could live with the noise they make.

As I've said many other times, I would like to see smaller size fans done away with. A nice 120 mm fan running on 6 or 7 volts moves more air than an 80 mm one operating on 12V, and is practically inaudible to boot. The cooling requirements of a modern PC are not so extreme that entirely passive means can't be used, either. The problem is that for many of the people who design PCs(and most other electronics), heat removal is an afterthought rather than being considered at the first stage of the process. The best way to deal with heat is not to produce it in the first place. The second best way is to integrate the cooling into the case. A PC case certainly has enough surface area to passively dissipate 200 watts or less without getting too hot to touch.

One other pet peeve of mine about fans-all the dust they suck into a machine. I actually need to clean out my machines after a month of running time. Why somebody didn't think to use one large fan, covered by a washable filter, sucking in air rather than blowing it out, is beyond me. This would solve the dust problem at least.

 

[Jan Kivar]

One other pet peeve of mine about fans-all the dust they suck into a machine. I actually need to clean out my machines after a month of running time. Why somebody didn't think to use one large fan, covered by a washable filter, sucking in air rather than blowing it out, is beyond me. This would solve the dust problem at least.

The problem with fans sucking air in is that the case must then be air proof. If it isn't, less hot air comes out the box. This is why "all" fans suck the air out. You can buy some filters, put they make more noise and You have to wash them frequently or they will clog.

What most of You haven't thought yet is the environment of the fans. The retail HSF gets pretty hot, so the fan's lifespan is shortened. IIRC, Papst gives MTBFs in two temperatures.

Cheers,

Jan

 

[Yash]

heh, if those normal videocard fans drive u guys nuts, then that dustbuster FXFlow on the GFFX would be the killer. :roll:

 

[mubs]

I've never understood why cards in PCI and AGP slots face downwards in ATX cases. That's dumb; heat goes up, and this way, the heat has nowhere to go. In a WTX case, the PCI/AGP cards do face up like they should. They probably run a lot cooler that way, and the number of hotspots in the case are reduced too.

 

[Buck]

I've never understood why cards in PCI and AGP slots face downwards in ATX cases. That's dumb; heat goes up, and this way, the heat has nowhere to go. In a WTX case, the PCI/AGP cards do face up like they should. They probably run a lot cooler that way, and the number of hotspots in the case are reduced too.

Addtronics W8500 cases have the motherboard turned around so you work from the other side of the case, which results in the heat leaving adapter cards correctly. The downside is that they are very large cases and a bit expensive.

 

[Howell]

I've never understood why cards in PCI and AGP slots face downwards in ATX cases. That's dumb; heat goes up, and this way, the heat has nowhere to go. In a WTX case, the PCI/AGP cards do face up like they should. They probably run a lot cooler that way, and the number of hotspots in the case are reduced too.

I bet it has to do with differentiating them from ISA cards which do face up. Originally, at least.

 

[mubs]

Buck, that's because the Addtronics W8500 is a WTX case.

You're probably right, howell, but it's still a dumb idea. Ever try to stick a PCI card in an ISA slot? Even a newbie would be able to tell that it won't fit.

 

[LiamC]

Mubs

Even a newbie would be able to tell that it won't fit.

I dunno about that. I've seen some dumbass (as our northern hemisphere friends might say) things done involving fitting cards, motherboards and hammers (literally). <shudders>

 

[Buck]

Buck, that's because the Addtronics W8500 is a WTX case.

Yes, I know. That is why I bought one. But once I found out how nice the case worked with regular ATX boards, I went ahead and bought another.

 

[mubs]

I came very very close to buying a W8500 when they first came out. Tons of room, cleverly designed. I chickened out because they a) are big and b) cost too much (for me). Bought a couple of Antec SX-1030s instead. After some surgery with a Dremel, am quite happy with them.

 

[Howell]

Mubs

Even a newbie would be able to tell that it won't fit.

I dunno about that. I've seen some dumbass (as our northern hemisphere friends might say) things done involving fitting cards, motherboards and hammers (literally). <shudders>

Aye, they must grow 'em extra perceptive on the left coast.

PS. Who was the Aussie from Perth?

 

[Mercutio]

PCI cards face upside down so that a PCI card can share a slot with an ISA card that faces up.

... and we've done it that way ever since.

And I have a WORSE story about an ISA graphics card, a PCI slot and some machine tools.

Hammers? Pffft!

 

[Clocker]

the design of that addtronics case really intrigues me! I think I must replace my Fong Kai with one! How much $? Mmmm more cheap wine! Yimmy!

C

 

[Howell]

PCI cards face upside down so that a PCI card can share a slot with an ISA card that faces up.

... and we've done it that way ever since.

And I have a WORSE story about an ISA graphics card, a PCI slot and some machine tools.

Hammers? Pffft!

Yeah the PCI connector sharing a slot with the ISA connector. That makes sense.

 

[time]

I can't help but notice that the Athlon XP1700s I bought have different coolers from the XP2400s.

Firstly, the cooler is bigger: 77mm vs 68mm fin length, or +13%.

The fan has a shroud, mainly to adapt the 60mm fan to the 77mm heatsink, but IMO shrouds are a good thing.

The fan blades have 'wing tips', at least that's my layman's take on things. I imagine this is to cut down turbulence, although I also wonder if it doesn't stiffen the blades and thereby increase the pressure the fan can deal with.

Unlike the 5400rpm 11-blade rubbish found on most Athlon coolers, this is a 9-blade 3300rpm fan that could be classified as "quiet". It seems to push plenty of air through the heatsink, though.

I wouldn't be at all surprised if this cooler performs nearly as well as the 'high-performance' cooler on the XP2400s. I'd guess the 5400rpm fan is rated at 35cfm and manages about half that when coupled with the heatsink (and no shroud). Whereas the 3300rpm wing-tipped fan is maybe 25cfm, and achieves as much as two thirds of that with the shroud (pure conjecture on my part).

Coupled with 13% more surface area, you can see why visually at least, this cooler impresses.

Now, you only have to wait for me to test one ...

Has anyone else come across them? The brand on the shroud is AVC, coincidentally the brand of the best 80mm fan in my collection (powerful, quiet, minimal vibration).

 

[time]

Here are my results from testing the different AMD retail coolers.

Apart from the interesting AVC design, I also tested the Tai Sol model that seems to come with the XP2400+ and other faster Athlons. Topped with a Delta 4800rpm 11-blade low-profile fan, its most prominent feature is a nickel-plated copper plate bolted to the base. The Tai Sol's height is modest, which can be useful in some situations.

The CPU was a 1466MHz Athlon (XP1700+) running at 2192MHz (12.5x175.4). Motherboard was an Epox 8RDA+ (nForce 2). I ran Sandra's Burn-in module (arithmetic & multimedia only) for at least 20 passes. Readings are merely from the socket sensor, so take all this with a grain of salt. However:

Cooler          Peak  Ambient   Net
Falcon Rock II   47      21      26
AVC 3100rpm      53      22      31
AVC 4800rpm      50      22      28   (Delta fan)
Tai Sol 4800rpm  53+     23      30+  (Crashed)

The reading for the Tai Sol has to be an estimate - the computer crashed consistently once it reached about 52C, so presumably the true die temperature was rocketing up faster than the socket sensor could report it. :x

I removed the cooler and reapplied Arctic Silver, but it made no difference. It's hard not to conclude that AMD's layoffs included at least one engineer too many: the guy who selected coolers for retail bundling ... surely this POS was selected by a bean counter?

I see that THG has done a recent cooler comparo that included this Tai Sol OEM cooler (except that they didn't identify it as such) and the XP2000+ skived fin cooler (which is noisy as hell and not all that good - but small). As usual, they failed to notice the bleeding obvious: the XP2000+ cooler is better than the one supplied with XP2400 and higher!

It's also worth noting that the AVC coolers were easily the best all-round coolers in the line-up - which THG missed as well, of course. :roll: