Page 3 of 4

Posted: Mon May 07, 2007 6:24 am
by IsaacKuo
butters wrote:I wonder if it would be easy enough to make one out of a 120mm fan.
Back when we last had a big thread about testing fans, that was essentially one of the ideas. The basic idea was to have two fans attached to a huge box, along with a "U" shaped pressure sensing tube (a bit of water at the bend--one side is inside the box and the other is outside). One fan is the test fan; the other is the "standard" fan.

The steps would be:

1. Zero the pressure sensor with both fans turned off.
2. Turn on the test fan.
3. Adjust the speed on the "standard" fan until the pressure sensor is equalized.
4. Record the speed of the "standard" fan.

You could replace the test fan with a CFM sensor to create a graph of CFM vs RPMs on the "standard" fan.

Posted: Mon May 07, 2007 3:47 pm
by lm
jaganath wrote:
I don't care about cfm. What I do care about is noise and temps.
temps are strongly correlated with cfm, so actually you do care.
You are not getting my point. I don't really care about how the level of quiet and the temps are achieved. Those 2 variables are the end results, and there's of course more variables if you dig deeper.

I think it would be great to find out which fan to attach to a ninja to achieve the lowest temperature on some predefined noise level, or lowest noise on some predefined temperature.

This could be measured directly instead of just measuring those other variables and then trying to figure it out myself based on the cfm and whatnot.

Posted: Mon May 07, 2007 3:52 pm
by Wirrunna
A thought provoking article.
Since I first replaced a Zalman Fanmate control with a CP Emma designed thermistor fan controller on a Zalman CNPS7000B-AlCu I have been grappling with the reported CPU temp and the lag and minor increase in fan speed. I suspect a lot of fans are run too fast.
Has SPCR ever considered thermal imaging of Heatsinks under load?

Posted: Wed May 09, 2007 2:55 am
by Jose Hidalgo
MikeC wrote:
Spanki wrote:Just as an example, I recently read another review, which included some medium speed (and even some high speed) fans. Airflow, actual cpu cooling, measured dbA levels and subjective noise/sound comments were listed for all. I found this an extremely useful review, because it showed fans like the Sharkoon Silent Eagle 2000 that I hadn't considered or even heard of before. In that particular case, I wouldn't run the fan at full speed (too loud), but it out-performed fans like the Yate Loon and Silverstone FN121 when cranked down to the same dbA levels as those (which likely means that it was still spinning faster).
The absence of any details about the tools used or pics / clear description of the setup makes it impossible to make a fair assessment of the roundup you linked to. The reviewer tells us, "After gathering the testing equipment, I formed a methodology that would provide extremely consistent and telling results".... but he doesn't show us any real evidence that this is true.
The question for SPCR is : when are you finally going to review the 120mm Sharkoon Silent Eagle 1000 ??? (please :mrgreen: )
lm wrote:I don't care about fan voltage. I don't care about cfm. What I do care about is noise and temps. What I'd like to see is comparison of fan curves on a temperature/noise plane. Hook them up to a ninja and see which one farest best at any noise level, ignore voltage and cfm.
+1000000000. That's it !!!
The folks at madshrimps have already understood that, and their results show... you guessed it : the clear superiority of both Sharkoons over the Noctuas and the GlobalWin in a PC enclosure. I hope SPCR will be able to confirm these results *soon* (it's been 6 months now since madshrimps said that, and SPCR hasn't been able to tell us anything about these fans yet :( ).

The link : http://www.madshrimps.be/?action=getarticle&articID=516 (see their graph on page 3 - that's exactly what SPCR should do IMHO)

Posted: Wed May 09, 2007 4:10 am
by jmke
That guy at XS had similar findings to mine concerning the Noctua and Sharkoon fans; so it confirmed the pressure issue with the Noctua fans.

I'm sure SPCR can come up with a solid testing method for evaluating fans; the real issue and challenge is set up one test method and bench and do the tests so that they can be 100% repeatable and able to recreate them; I got a selection of new 120mm fans here which I'm going to test again too, the anemometer approach is certainly useful, really like the case design with in-take fan, very close to real world design; would even consider stuffing the case with an old (even broke) motherboard/cpu/vga/hdd/cdrom to get even closer to real world results;

for CPU pressure tests why not combine that large tube with a heatsink?

-- = tube
## = fan
@@ = scythe ninja/infinity


-----------------------------------------------------------------
##@@
-----------------------------------------------------------------

than measure the CFM that way, the airflow will be restricted through the fins of the heatsink and CFM can be measured, you can compare the CFM with/without the @@ heatsink to see if the tested fan is good for cooling areas where pressure is needed.

as a case fan the Noctua does okay, for CPU cooling it falls behind. with that tube testing method you would "catch" that flaw and be able to point it out. Then you can create XY charts with dBA/CFM at different voltages; you already have the data, just need to present it differently in a chart

Posted: Wed May 09, 2007 12:25 pm
by Devonavar
For the record, we did do some experiements using a wire fan grill to introduce impedance into the airflow path. The results showed virtually no drop in airflow — worst case was a measly 2 CFM drop with a grill on the intake of the Noctua at full speed. That's about a 5% difference, and it's lower than our confidence interval for our test rig.

The results of the constant airflow test on the Ninja also seems to show that the Noctua doesn't seem to be seriously affected by impedance.

I have no problem accepting that the Noctua has pressure issues with in high impedance situations, but instead of arguing about whether or not that is the case, why not take SPCR's results as demonstrating that there are situations where the Noctua's lower pressure is irrelevant? Obviously, that's not all situations.

I think if we let the results of this test guide our design approach, we should take home what Mike has been saying all along: Airflow design, specifically, low-impedance airflow design, is at least as important as the choice of fan. And, in a situation with good airflow, the choice of fan is more or less irrelevant.

Posted: Wed May 09, 2007 1:26 pm
by Spanki
Devonavar wrote:The results of the constant airflow test on the Ninja also seems to show that the Noctua doesn't seem to be seriously affected by impedance.
Cool, so we know how to deal with exactly one heatsink (with one particular input temperature) - "Buy a Ninja".
Devonavar wrote: I have no problem accepting that the Noctua has pressure issues with in high impedance situations, but instead of arguing about whether or not that is the case, why not take SPCR's results as demonstrating that there are situations where the Noctua's lower pressure is irrelevant? Obviously, that's not all situations.
So, the answer to your question just above there is obvious, and answered by the statement which I highlited.
Devonavar wrote: I think if we let the results of this test guide our design approach, we should take home what Mike has been saying all along: Airflow design, specifically, low-impedance airflow design, is at least as important as the choice of fan. And, in a situation with good airflow, the choice of fan is more or less irrelevant.
The problem is, not everyone owns/wants/has-the-room-for a Ninja. I have no problem with the 'recommendation' for some particular airflow design (low-impedence), but that's just not a all-encompassing answer.

What's the best (quietest, highest performing) fan I can put on a Ultra 120 that came in my computer (or that I already own for whatever reason) - without having to go out and buy a Ninja?

What's the best (quietest, highest performing) fan I can put on a Scythe Infinity that came in my computer (or that I already own for whatever reason) - without having to go out and buy a Ninja?

What's the best (quietest, highest performing) fan I can put on a Big Typhoon that came in my computer (or that I already own for whatever reason) - without having to go out and buy a Ninja?

...the list goes on.

I'm sure that there may be other heatsinks with similar airflow designs of the Ninja, but what about all the other heatsinks (or temperatures) where that's not the case?

Show me temperature performance results of a Noctua fan and a Sharkoon 2000 Golfball (down-volt the golfball to the same noise level as the Noctua) on a Big Typhoon/Ultra 120/Infinity. Let me hear comparitive recordings. If the Golfball performs better than the Noctua for the same amount of noise on some particular heatsink, isn't that valuable information that your readers might want to know? If I can then turn that Golfball up a bit to get better cooling (while gaming or whatever) without a huge increase in noise, isn't that valuable information?

Obviously there are hundreds of heatsinks and hundreds of fans, making thousands of combinations, so testing them ALL would be rediculous. Having said that, I think an interesting/enlightening article could be done with the following:

- pick 4-5 top/popular heatsinks with replaceable 120mm fans (Tuniq, Ultra 120, Ultra 120X, Ninja, Big Typhoon, Infinity, XP-120/128, Enzotech Ultra-X, etc).

- put a screaming delta fan on them and test at various voltages (1-2v steps, ideally) to determine the point of diminishing returns for airflow on both stock and hot/overclocked cpu...

...nevermind. I realize that I probably just lost my audience in this venue - "hot/overclocked cpu" is irrelevent here, only quiet matters - "there are plenty of other sites that cater to the enthusiast crowd". :(

Jmke, thanks for your review work as well! I think it better suits my needs (my only gripe is that you left out recordings when you did the Golfball reviews :) ).

Posted: Wed May 09, 2007 1:47 pm
by IsaacKuo
The thing is, the Ninja has a good edge over all of the competition in anything but high airflow situations. And the Ninja isn't even expensive.

Imagine yourself spending the last two years in a monotonous routine of reviewing YET ANOTHER heatsink markedly inferior to the Ninja (even those made by Scythe). By now, there's really nothing exciting on the horizon, except for maybe the mini-Ninja for HTPC cases.

Is there really a point to testing the fans on non-Ninjas? It's just going to give worse results all around, within the regime we care about.

Posted: Wed May 09, 2007 2:22 pm
by Spanki
IsaacKuo wrote:Is there really a point to testing the fans on non-Ninjas? It's just going to give worse results all around, within the regime we care about.
Nope. Looks like you guys are all set until the Mini comes out.

Cheers.

Posted: Wed May 09, 2007 3:33 pm
by Spanki
Just to keep you amused while waiting for more Ninja models to hit the street, you might be mildly interested in SilentFlux Bubble Pump technology ;).

Posted: Wed May 09, 2007 5:18 pm
by kaange
I think Spanki has a point which is why I think the experiment/setup #2 results have some relevance since back pressure is not negligible in some situations. I posted this earlier but seem to have been ignored.

In fact what might be useful with setup #2 is a pressure reading of the enclosure so that the amount of pressure that the fan builds is recorded.

Posted: Thu May 10, 2007 7:48 am
by MikeC
kaange wrote:I think Spanki has a point which is why I think the experiment/setup #2 results have some relevance since back pressure is not negligible in some situations. I posted this earlier but seem to have been ignored.

In fact what might be useful with setup #2 is a pressure reading of the enclosure so that the amount of pressure that the fan builds is recorded.
The problem with the pressure in setup #2 is that it appears to be dynamic and dependent on more than just airflow -- it varies with fan size as well as the RPM/airflow of the fan.

Posted: Thu May 10, 2007 8:07 am
by Spanki
MikeC wrote:
kaange wrote:I think Spanki has a point which is why I think the experiment/setup #2 results have some relevance since back pressure is not negligible in some situations. I posted this earlier but seem to have been ignored.

In fact what might be useful with setup #2 is a pressure reading of the enclosure so that the amount of pressure that the fan builds is recorded.
The problem with the pressure in setup #2 is that it appears to be dynamic and dependent on more than just airflow -- it varies with fan size as well as the RPM/airflow of the fan.
But would it be useful for comparative purposes, between fans of the same size, spinning at the same rpm? Edit: Or, more importantly, at the same noise level.

Posted: Thu May 10, 2007 9:03 am
by Jose Hidalgo
Yeah, but I am asking again : will SPCR finally test the Sharkoon Silent Eagle 1000 fan against the other top fans soon ???
Thanks.

Posted: Thu May 10, 2007 1:20 pm
by MikeC
Jose Hidalgo wrote:Yeah, but I am asking again : will SPCR finally test the Sharkoon Silent Eagle 1000 fan against the other top fans soon ???
Thanks.
When and if we get samples. A request has been sent to Sharkoon. You (and anyone else interested) could probably help expedite matters by making the request to Sharkoon for review samples to SPCR.

Posted: Thu May 10, 2007 1:27 pm
by jmke
MikeC, send me the email you want to send to them, I'll forward it to my Sharkoon contact, and put you in cc:

Posted: Thu May 10, 2007 2:35 pm
by Jose Hidalgo
+1 : I'll do the same thing gladly if needed. Thanks in advance.

Posted: Thu May 10, 2007 3:59 pm
by kaange
Spanki wrote:
MikeC wrote:The problem with the pressure in setup #2 is that it appears to be dynamic and dependent on more than just airflow -- it varies with fan size as well as the RPM/airflow of the fan.
But would it be useful for comparative purposes, between fans of the same size, spinning at the same rpm? Edit: Or, more importantly, at the same noise level.
Yes. Unrestricted flow is pretty good for case exhaust (and probably inlet) simulation but for HSF and PSU fan replacement, the restricted flow rate (set for standard rpms as per CPU HS tests) of different sized fans would be very useful (plus their noise level/signatures, of course).

Posted: Thu May 10, 2007 11:14 pm
by Bemaitea2
Haven't been on SPCR for some time but I just checked out this article.

All I have to say: Bravo.

I was most impressed with the passion that was shown and the fact that you listened to your readers in your forums.

The ability to admit a fault and try to look for a solution shows a great deal of sagacity, much respect SPCR crew.

Posted: Sun May 13, 2007 7:19 am
by MikeC
jmke wrote:MikeC, send me the email you want to send to them, I'll forward it to my Sharkoon contact, and put you in cc:
Here's the reply from Sharkoon:
Dear Mike,

thank you for your email and your interest in Sharkoon.
Unfortunately we can't send you samples of Sharkoon Silent Eagle.
We only dispose of few review units and basically provide high-circulation german print magazines.
Who knows whether emails from SPCR fans (no pun intended) will change his mind....

Posted: Sun May 13, 2007 8:29 am
by jmke
don't do that:) mailbombing is never the answer imho, only thing it will cause is possible backlist?

anyway check your mail;)

Posted: Sun May 13, 2007 8:34 am
by MikeC
jmke wrote:don't do that:) mailbombing is never the answer imho, only thing it will cause is possible backlist?

anyway check your mail;)
You mean a blacklist? :wink:

Anyway, I saw the cc of the email you wrote Sharkoon. That's very civil & perhaps it will lead to some samples. Thanks very much for your effort!

Posted: Sun May 13, 2007 9:12 am
by jmke
ya blacklist:) or back of the list, backlist;)

Posted: Mon May 14, 2007 3:40 am
by jmke
check your mail Mike, good news:)

Fan noise reduction

Posted: Wed May 23, 2007 9:26 am
by phconnell
Can you do another roundup of the fans you tested but this time cut the motor and
fan or each from the four ribs connecting it to the frame. Try remounting them on new
ribs so that the distance between the ribs and fan tips is 5mm at least.

I have tried this and found that one of the main sources of noise in so-called
quiet fans is the interaction of the fan blade eddy flow with the four ribs, and
that you only have to make more distance to get near silent operation.
If you dangle the fan and run it you will hear almost nothing!

The problem is that the fan housing seems to to have a standard depth of about
25 mm and this doesn't leave much room for "exit space" beyond the trailing
edge of the fan blades. Extend the depth to 30 mm and I would bet all your fans
move into the "almost silent zone".

P.H.Connell

Posted: Wed May 30, 2007 3:38 pm
by cpemma
MikeC wrote:
jmke wrote:don't do that:) mailbombing is never the answer imho, only thing it will cause is possible backlist?
You mean a blacklist? :wink:
!
Purely in the spirit of helping a non-native to the English language, I think 'backlash' is the word. :D

The new test method looks better than earlier attempts, but I worry that it's still considerably less in impedance than even a well laid-out "real" system. Checking flow rates against manufacturers' claims doesn't bother me - it's how the fan is performing for real users. The Noctua 120 is a case in point; "comparison with both the Nexus and Scythe SFF21E 120mm fans, the Noctua still seems to have an edge in airflow at lower speed." But given the narrowing of the margin with the new test, would the margin disappear altogether in the real world?

I've wasted rather a large sum on ultimately-disappointing fans over the last few years; it's going to take a lot more persuasion from you guys before I swap out my Yate Loons. ;)

Posted: Wed May 30, 2007 3:46 pm
by jmke
no no MikeC got it right, bLacklist was what I was going for:)

Posted: Tue Jun 26, 2007 2:23 pm
by peTeMelster
Hi, this is my first post here and I have a few questions. They are summarized at the very bottom of this post for easy reference, but to understand the questions, I need to explain my thinking first.

To stray from the current discussion a bit, and back into what this thread was initially about, I have a question regarding which fan to use for the Scythe Ninja Plus B Revision in order to achieve the maximum cooling performance for my CPU (Intel Core 2 Duo E4300 Allendale 1.8ghz 800 FSB 2mb L2 cache, NOT overclocked).

According to the review over at AnandTech of several heatsinks, including the top performers: Scythe Ninja, Thermalright Ultra 120, and Thermalright Ultra 120 Extreme, it seems that (as is the case on SPCR) the Ultra 120 Extreme is the best heatsink in terms of cooling.

Since this is my first post, I can't post URL/links, though I wish I could.

The AnandTech reviewer used a Scythe S-Flex fan (fast version 1600 RPM) on the Ultra 120 Extreme and received excellent cooling. They also tested the Scythe Ninja first using the stock Scythe fan that came with the kit (not S-Flex) and found that it was okay, but dismally wanting much more airflow to increase cooling performance. Therefore, they switched the fan to a SilenX Ixtrema fan 14 dBa fan (not the 11 dBa version tested at SPCR). They found this greatly enhanced the cooling performance to nearly match - only a few degrees hotter - that of the Ultra 120 Extreme with S-Flex 1600 RPM fan by using the SilenX supposedly high-flow, low-noise fan with the Ninja (manufacturer rated at 72 CFM/14 dBa...although if the tests for the 11 dBa version are any indication, the noise on this even louder 14 dBa version would be tremendous, even more than the purported 26 dBa at 1 meter SPCR measured for the 11 dBa version fan).

Since the methodology was changed for testing airflow and the Noctua 800 RPM fan that I bought to attach to my Scythe Ninja Plus B is actually pushing less air than I initially thought, believing the initial SPCR Noctua fan review (pre-this new airflow testing method), I'm worried about the cooling performance of this HSF unit with the Noctua 800 RPM fan.

If what AnandTech claims is true, that increasing airflow (which they did by switching to the SilenX 14 dBa fan) makes the Ninja heatsink perform MUCH better, then I'd like to try increasing my airflow as well while keeping noise down. It seems the SilenX fans are FAR too loud, even the 11 dBa version is too loud at a measured 26 dBa at 1 meter.

But I wonder what the airflow is with the stock Scythe fan? It's probably too loud as well, so I don't want to use it. This is why I purchased the Noctua 800 RPM fan, but now I'm hearing (according to SPCR) that it needs a 120mm fan at 1100-1200 RPM for maximum cooling performance for the Ninja. Well, the S-Flex fan used by AnandTech for the Ultra 120 Extreme was 1600 RPM, far faster than 1200 RPM. Furthermore, AnandTech used a hi-flow SilenX fan spinning much faster than 1100-1200 RPM to achieve best cooling with the Ninja.

So, if what SPCR says is true, that increasing airflow beyond what 120mm 1100-1200 RPM fans can push does not increase cooling abilities, then wouldn't it be true that AnandTech should be able to achieve the same cooling abilities for the Ninja and the Ultra 120 Extreme with slower, less airflow fans (as long as they don't drop below 1100-1200 RPM, yet they were using 1600 RPM fans to achieve the best cooling).

I guess what I really want to know is this: Which fan should I get to achieve maximum cooling with my Ninja heatsink? Now that the airflow of the Noctua fans has been found to be much lower than before, I believe the Ninja is not receiving enough airflow to do the best cooling job it can. Therefore, I want to get another fan to push the Ninja to its limit, but what fan should I get?

I'm not sure what CFM or what RPM the stock Scythe fan that came with my Ninja measures at, but if AnandTech tests show that the stock fan doesn't push the Ninja to its peak performance, that would imply the Ninja needs higher airflow. But do I really need to couple my Ninja with a really high airflow fan like the Scythe S-Flex 1600 RPM version or the SilenX 14 dBa version? My thinking is that if cooling doesn't improve beyond 1100-1200 RPM, then wouldn't the Noctua 1200 RPM fan push the Ninja to its limit and cool the best? That additional airflow beyond what the Noctua 1200 RPM fan can push would not benefit cooling, and would make things louder?

I worry about all this because at idle, my CPU is around 38-44C according to Core Temp, Thermal Analysis Tool, and SceneFan. Ambient temperature is around 70-80F (I think that's around 22-26C). Yet, the AnandTech reviews show that at idle, the Ultra 120 Extreme with S-Flex 1600 RPM fan and the Ninja with SilenX 14 dBa fan achieve 26C and 27C at idle. This is a huge discrepancy...so, I'm worried.

Further note, that according to AnandTech's review of the Noctua heatsink and fan combo, the Noctua fan does a great job of cooling (though the heatsink itself isn't on par with the Ultra 120 Extreme or Ninja). They tried switching out the Noctua fan for a higher airflow fan and the cooling difference was nothing. This seems to imply that SPCR is correct in saying that beyond a certain airflow point, increasing airflow doesn't improve cooling. This further implies that the Noctua 1200 fan is at the peak performance that the heatsinks (or at least the Noctua heatsink, but hopefully the same with the Ninja and Ultra 120 Extreme too) can perform at.

Basically my questions in a nutshell:

1. Is my Noctua 800 RPM fan not pushing enough air to push the Ninja to its peak cooling performance?
2. Is the Noctua 800 RPM fan worse (in terms of cooling only, not noise level) than the included stock Scythe fan?
3. Would getting a Noctua 1200 RPM fan push the Ninja to its peak cooling performance? Or do I really need to get a hi-flow Scythe S-Flex 1600 RPM fan or the SilenX Ixtrema 14 dBa fan to achieve this peak cooling as AnandTech suggests.
4. Finally, do the Noctua 800/1200 RPM fans really still push slightly more air than the Nexus 120mm and Scythe S-Flex fans at the same RPM? Because if they do, then it should be a no-brainer to get the Noctua fans over the Nexus and S-Flex fans, right? Or am I missing something completely?

Thanks! I'll post the links in the next post for the Ninja review, comparisons with other heatsinks, and the Noctua heatsink/fan review at AnandTech.

Posted: Tue Jun 26, 2007 2:25 pm
by peTeMelster
Sorry that this is a double post, but I couldn't post URLs/Links in my last post (since it was my 1st post here) which I needed to for reference.

Here is the Ninja review: http://www.anandtech.com/showdoc.aspx?i=2965&p=1

On this page: http://www.anandtech.com/showdoc.aspx?i=2965&p=5
you can see how the Ninja with various fans compares to the other heatsinks.

Here is the review on the Noctua heatsink: http://anandtech.com/casecooling/showdoc.aspx?i=2976

Posted: Tue Jun 26, 2007 3:36 pm
by tehfire
peTeMelster wrote:1. Is my Noctua 800 RPM fan not pushing enough air to push the Ninja to its peak cooling performance?
Because of the way the Ninja is set up (most air is free to escape from the sides and doesn't have to flow through the whole heatsink) it's very possible that the Noctua isn't giving you the best temps possible. A fan strong enough to blow air fully through the heatsink, however, may be excessive (and loud).
peTeMelster wrote:2. Is the Noctua 800 RPM fan worse (in terms of cooling only, not noise level) than the included stock Scythe fan?
SPCR is coming to the consensus that the Noctuas are very good at pushing air through free space, but faulter a bit when trying to push air against resistance. I would say the stock S-flex is better.
peTeMelster wrote:3. Would getting a Noctua 1200 RPM fan push the Ninja to its peak cooling performance? Or do I really need to get a hi-flow Scythe S-Flex 1600 RPM fan or the SilenX Ixtrema 14 dBa fan to achieve this peak cooling as AnandTech suggests.
The Noctua would do better than your current setup, but it's a war of diminishing returns...the faster the fan, the better the cooling. You have to personally draw the line somewhere. I'd stick with a 1200rpm S-flex max.
peTeMelster wrote:4. Finally, do the Noctua 800/1200 RPM fans really still push slightly more air than the Nexus 120mm and Scythe S-Flex fans at the same RPM? Because if they do, then it should be a no-brainer to get the Noctua fans over the Nexus and S-Flex fans, right? Or am I missing something completely?
After some exhaustive testing, SPCR found that the Noctua's really don't push any more air than conventional fans, so it hurts their appeal. Also, many people have been complaining about whining noises when their Noctuas have to go against any significant pressure. The Ninja lets air flow pretty easily, however, so I'm not sure how much of a factor this is. Personally, I'd stay with a conventional FDB fan.

I think the main question you need to ask yourself is whether your cooling is good enough (whether it requires any modification at all). 38-44C idle isn't that bad as processors go - mine's sitting at 47C right now. Granted, I'm using a Prescott, but silicon's silicon. I would load your processor with TAT. If it doesn't exceed 70C, then you don't really have to worry at all.[/u]