hm.

this seems like it would be a great fan for someone wanting to cool a pentium D or such, as most of those solutions are very loud. 35 decibels is actually quiet for that type of thing, plus it isnt huge like a zalman.


I wonder how one would know whether or not their rheobus controller was giving them above 9volts vs 8 volts. i jsut never know myself on my Papst 120, but that even at like 6 volts provides adequate cooling during non gaming scenarios.

thanx for the review!

I can't imagine anyone actually setting the fan speeds in his system according to voltage. We always mention voltage in our reviews because we have to, it's part of our documentation regimen. But for the computers I run and use, I set everything by ear and by temp sensors: I set the fans so they're quiet enough while the components are cool enough.

BTW, all of our top-rated HSF can fit socket 775 -- this includes Zalman 9500, Ninja, XP/SI-120. They will all do it more quietly than the F7P, w/ lower temps.

Not sure why you say it's "not huge like a Zalman" -- it's 126mm tall, which is marginally taller than the 9500, and its other dimensions are about as big.

Page 5, last paragraph: "stomp all over the 7000s at the full 12V speed, but hey are close in price".

Should be a "they" instead, small typo. Otherwise interesting to see how the tightly packed fins affects the performance of the heatsink.

I wonder if it would benefit from putting another fan on the other side, so that the fans are in a push-pull configuration. Adding a second fan should greatly increase the air pressure though the heatsink, right? Perhaps by doing this it could be quieted without loosing so much performance.

Hi, in the spec sheet it says 4 pin PWM for the fan. Doesn't this suggest that the speed is intended to be controlled by the motherboard, in response to temp so you get as much cooling as you need?
The only 775 systems I've worked with are Dell's at work and they certainly has 4 pin PWM fan control, fire up CPU Burn and away they go! As a side note they (Optiplex GX280s) use a very similar heat sink design, 3 U shaped heat pipes and a bunch of Al fins. The fan (92mm) is on grommets on the back of the case and uses Dell's traditional bright green plastic duct.
Seb

" The individual heatpipes are located quite close together. This is bit surprising; distributing them farther apart would seem better for heat dissipation."

The processor cores are usually much smaller than the IHS. And i think even dual-core ones are smaller than the space covered by the three pipes in the AC7.


Also, benefit of cooling the VRM-section is longer life-time and maybe some increased stability/overclockability. Especially electolytic capacitators are very sensitive to temperature. Usually manufacturers say 10°C increase halves their expected life-time. Although MOSFETS are not so sensitive, they still perform better at lower temperatures.

Another thorough review, cheers MikeC.

One small typo on the 2nd page, in reference to the fan:
"This may be due to the absecne of a frame."

Al

You are right, that is the intent, and perhaps we should have explored this a bit more.

In our defense...

It doesn't change the cooling performance and noise we documented at various drive voltages, which is important to know if you are looking for quiet performance.

There's no guarantee that every BTX motherboard's fan controller will be set up the same way. In fact, I bet $$ that they are all a bit different. The test AOpen board has a Smart fan setting in the BIOS, but it was disengaged for the testing. Using this kind of auto-variable controller external to the HSF is no way to test a HSF.

It also means that the fan speed will go up and down in response to load/temp, which is not something most quiet seekers like. Variability is one of the most annoying aspects of noise for most people.

True...but once the heatpipes left the area of the slug on their journey to the top of the heatsink, the two outer pipes could have been angled outward from the center one.

my thinking as well. an extra fan in back would also direct air through the heatsink better, which might especially make a difference because of the frameless fan.

Any idea when the Freezer 64 Pro will be reviewed, seeing as it was mentioned in the article? I'm interested to see how the lower fan speeds work with the Athlon 64's which don't produce the same amount of heat. Maybe it won't suffer the same dramatic drop of in cooling performance.

I'm not working for SPCR so I can't tell you anything definite, but.....
SPCR doesn't review heatsinks on the A64 platform. IIRC there has been discussion about this before and the reason was that the exact energy dissipation of the multiple versions or cores on the A64 platform was less known than that on the P4 platform. This normally isn't a problem since most heatsinks are identical (except for some mounting clips) for the different platforms. AC's Freezer series is one of the notable exceptions. This unfortunately gives these nice coolers a disadvantage, since now the louder version is tested, while the majority of the readers interested in this thing will have a need for the slower version.

Well, an interesting article. I've been using the Freezer 7 Pro for over a month and I must say that if I had read the article before I bought the heatsink I would've definetely went for the Big Typhoon. Anyway, the Freezer 7 Pro is a nice heatsink and good value for money.

I found it odd when you said in the article that the CPU throttles below 700 rpm. My motherboard has the Asus AI Quiet feature which I have set to silent mode and the fan is currently running at ~550 rpm with P4 520 @ 1,35V. The POST screen even keeps complaining about broken fan if I don't set it to ignore. The bios temp sensors page says it's idling at 45 C, which could indicate that the BIOS is bit too optimistic, but it's definetely not throttling, even at full load. I have a two rather high CFM case fans however (Akasa 120mm in the front @ 12V and Zalman 92mm @ 5V in the back), so they might help a bit on the matter.

The only negative side I can think of is the fastening mechanism, which is the same as in the default Intel heatsink. Altough the installation is easy and you don't have to remove the motherboard from the case, it just doesn't feel very secure with the cheap looping plastic tabs and all. The motherboard even bends slightly when the heatsink is in place. Backplate plus screws would've been a better choice in my opinion.

The performance characteristics of the two known "control" heatsinks (the Ninja and the XT-120) bring up an interesting question.

Is it reasonable to assume that the performance of these coolers will scale linearly? In otherwords, if one keeps adding watts will the characteristics of °C/W MP will be consistent (obviously until the processor throttles)?

Perhaps the answer to this is obvious -- each design will have a max amount of heat it can dissipate based on material, airflow and surface area. Clearly all of these devices have an upper bound independent of the processor's heat tolerance. Up until that point, however, is the curve linear?

Ultimately, what I'm asking is this: provided the tendency for chips to increase in output heat, do SPCR's standards of using the 2.8 Prescott (I believe) give realistic estimates for those of us with potentially hotter chips? Can we reliably extrapolate the numbers upwards for hotter chips?

I have no reason to believe that it won't -- but it seems like an important consideration (and was prompted by the notably drop in the Thermalright's performance by comparison to the Ninja's, whereby they were relatively comparable in the initial SPCR reviews). The wearing of the fan clip is noted, but is that the only variable?

Jeremy

If you're using a hotter chip than a 2.8 Prescott you haven't been reading the forum enough. Whatever your needs are, there is likely to be a cooler, quieter solution, most likely manufactured by AMD.

That's fair . I was assuming that the tendency of chips would be to get hotter and hotter (as has been the general trend the last decade) but at the same time I think it's arguable that we've hit a ceiling; people are demanding smaller and quieter and yet many heatsinks can barely fit in midtowers . It's clear that manufacturers recognize this as well, with the Pentium M and Semphron making it into desktop machines.

I, in fact, have an AMD (X2 3800; idles at 29*C and rarely goes above 40*C and that's with the stock cooler). I suspect even with two core the AMD is cooler than that chip, so your point seems to nail the issue. Thanks for the prompt response!

Short answer: Yes.

Longer answer: Chances are you'll hit non-linearity if you try extrapolate HS performance with C/W beyond, say, a 50% boost in CPU heat... but by then the airflow will have to increase substantially to provide effective cooling, so it becomes a non-issue.

I recently completed building a Pentium EE 840 (3.2G) system and was looking for a quieter solution to the stock jet-plane HSF. Many of my co-workers were recommending the Zalman 9500, but after researching both, the F7P seemed to be the best bang-for-the-buck, so I went with it.
After installing it, I noticed the temps weren't that much better than stock (high 50's @ idle). The fan was only running @ 350rpm, however. Turning the fan-control off, thereby having it run at 2400rpm, temps went down into the 40s. Of course, the noise level went up, to the point of being too loud. The sweet spot seems to be manually setting the rpm (via Intel's Desktop Control Center program) to 1400rpm. This keeps temps around 50deg while also not noticing the fan noise (my PCI-E video card's fan is louder).
My point here is that the review seems to be right on wrt to the cooling performance (or lack thereof, rather) at lower RPMs.
Two other friends built up the exact same system (everything the same), with one using the Zalman and the other using the Ninja, and both have idle temps in the 40s, so this is another testament to your review, Mike.
The F7P is an adequate cooler, especially given the cost, but there are higher performers out there.

Your computer, your choices. Let me provide some data that may be of interest to your decision-making:

A CPU will run beyond the life of the equipment at 90C die temperature. IBM's mainframe department, which wants their equipment to run a week past the end of eternity, use 85C as their design standard. Your mobo fan/temperature control appears to be designed to hold the die temperature below 60C. At that temperature your CPU will run 7 eons past the end of eternity.

Choosing to run at lower temperatures is legal, moral, ethical and non-fattening. It does raise the noise level, however. I have no motivation to stop you, even if I could. I just want to point out that you aren't making your CPU any more reliable; it's already as reliable as it's possible to get at 60C. And if you load the CPU, the mobo control loop speeds up the fan so that's a non-issue.

Just trying to clarify the issue.

I'm not trying to improve the reliability, longevity, etc...
Having a space heater down by my legs is what I'm trying to avoid.

The amount of heat generated by your PC does not change with the speed of your CPU fan -- there is no change in the amount of heat coming out if it. If you had high enough airflow (way noisier) through the whole case, this might feel cooler simply because there is more displacement of air but the same amount of heat.

FWIW, as a former architect at Amdahl, I can tell you that most ICs are designed to last five years at 105C. Of course that's the mean...

Man y'all like to split hairs...
My goal wrt the cpu cooler, case, fans, etc. was to have a "quiet" system while also having adequate cooling. All the talk about parts being designed to last forever under high thermal loads, they're certainly true statements, but that doesn't mean that having better cooling isn't inherently better.
BTW, my quib about the space heater by my feet was because I actually feel the heat as compared to my old P4 system.
Overall, my orig. post was merely to corroborate Mike's orig. review of the F7P...

As someone who runs their CPU below 30C at all times (quietly!) the motivation for running it that cool is just personal preference; I don't like having anything in my case which is too hot to touch without burning myself. Also, if you are running it at 85C and using a PSU in the traditional ATX role the intake air temps will be higher, shortening the life of the capacitors in the PSU.

When was the last time you touched the CPU die while the computer was running? Heck, you can't even touch the much-cooler IHS because there's a HS or HSF bolted to its top!

As Mike just pointed out, the CPU die temperature is not related in any way the heat in the chassis, which is fixed. You apparently believe that an 85C die temperature (and we were talking in this topic about high 50's, remember) will result in an 85C stream of air entering the PSU. Uh uh.

Of course the amount of heat in the chassis is not fixed. Heat emission by the CPU varies with frequency, CPU load and supply voltage. Also heat in the chassis will vary with fan speed and concomitantly CFM through the case.



I never said that at all; you are misrepresenting what I wrote. I said the intake air temps for the PSU will be higher, thus shortening cap life in the PSU. It is a concrete fact that capacitor longevity is shortened by higher operating temps.

This CPU cooler is now just $20 from New Egg, definately the best thing in that price range.

I also got interested of this product because of it's low price. It's far from the best cooler around, but I'm considering updating my Prescott system with a slightly better cooler than the stock one.

http://www.arctic-cooling.com/cpu2.php?idx=79&disc

Do you think it's a bad idea to cut the back edge of the fins to make them shorter and less restrictive? Such long and tight passages cause quite a bit airflow resistance and with undervolted fan, the air travelling between fins soon warm up to temperature of the fins making further traveling useless but still harmful due to resistance.

Cutting the fins right after the heatsinks would reduce surface area by ~30%. Reduction in flow resistance is probably less than that because the front edge of fins probably creates turbulence. I would leave the curved few fins intact because they cool VRMs and they also have a wider fin spacing, thus don't contribute to resistance that much.

What's your guess, better undervoltability for the fan or a complete disaster?

Definitely not better undervoltability. Reducing the cooling fin area is never a cooling win. The fin spacing on the Freezer is greater than any Thermalright product save the HR-01, and the same as the well-reviewed Apack towers. Leave the fins alone.

However, if you're experimentally minded, wrecking a Freezer is an inexpensive and highly educational project. I've done worse...