More good points and I happen to agree with most of them.
At some point (or in some setups) it might be an issue, but for the most part, I still don't think it would make any difference. For example, my system is:
MSI P6n Platinum mb (nVida 650i-based)
E6400 @ 2.66GHz - stock volts (1.325)
Cooler Master Hyper Tx heatsink w/fan usually running @ ~9v (actually, it's PWM, running at 60%, so it gets 12v, 60% of the time)
XFX Extreme 7900 GS vga card
1 x 250GB Seagate Barracuda 7200.10 SATA drive
2 x 120GB Maxtor IDE drives
1 x TDK Dvd/CD IDE drive
1 x Antec Tri-Cool 120mm fan
1 x Sharkoon Silent Eagle 1000 120mm fan
2 x Yate Loon 120mm fans
(the Antec runs at ~5v, the others run at 12v)
1 x 60mm Vantec Stealth (NB cooler) @ 12v
...powered by a Xclio 550W GreatPower psu.
I have a power consumption meter on my UPS and at idle, the system draws ~202W. When running Orthos blend, it draws ~222W. Of course if all the drives were seeking around reading/writing files, it might draw a few more watts, but that's still well less than half the available wattage (actually the important measure is amperage, but I'm not too handy with the conversions ). Adding a few (or a dozen) fans is not going to put any strain on my power situation.
Just a "for instance" number. My Antec P182 doesn't have a removable tray, but I know that they are one of the checklist items people often look for when buying a case. I hadn't done research to see exactly how much clearance any particular case has, but I've seen several people say something like "and I can finally remove my tray without removing the HSF first - yay!". Again, this would only be a benifit if it happened to work out that way in your case, so I was just using that as one possible checklist item some people might also be considering when they are looking for a HSF.
Far too often, I see someone start a thread with "Which heatsink should I get?" And the (equally bad) answer is usually "Get X, it rox!" (where X is whatever can be found at the top of some review chart at the time). Then a week later, heatsink arrives, user installs it, then returns to the forum... "hey, you guys said this was the one to get, but it (insert one or more of the following) -
* is way more expensive/heavy/sizeable/loud than I needed for my non-overclocked/mildly-overclocked system
* doesn't fit my motherboard
* doesn't fit my case
* is loud as hell
* my Orthos/TAT "load" results are much higher than I'm seeing on AT! (that's partly because they don't use those to measure with, but more on this later...)
* cools worse than the Y cooler I had
* won't fit with my side fan installed
* doesn't come with a fan!?
* has to be lapped???
* has to be installed with the motherboard removed?
* wouldn't slide back into my case, so I had to unmount it, slide my tray in and then mount it again - grrrr
* snapped my mb in half when I went 4-wheeling to the LAN party
* is not black!!
* looks like crap in my pimped-rig
...etc. . I guess people not stating thier criteria is a different subject, but the point is that people do have all sorts of criteria when shopping for a heatsink, so the more readily available info is, the easier it will be to find (if not by the people asking, at least by the people trying to help).
Yeah, my biggest concern is not that they're not testing in 2-3 different configurations (that would be a lot of work), but more that they're drawing conclusions based on the current setup, without accounting for other reviews that show contrasting (sometimes a complete reverse) results.
Thier current accounting is:
- they've turned on the case fans and while the top-blowers improved, so did the side-blowers, so they consider that a wash.
- they've tested with the side removed and (IIRC) saw similar results to the above - but haven't tried a closed case with side fan, vs closed case with side vent (fan removed for Tower sinks, due to clearance issues) as far as I know.
- they say it's because they are using a modern (Core 2 Duo) chip and other sites are using older P4 chips (but it's unclear to me how/why a top-blower cools a P4 chip at some heat level than a side-blower does (reverse of the results AT gets) on that same chip/heat in those other reviews).
I'm still a bit puzzled as to exactly 'what' the difference is, but AT's setup seems to be tougher on top-blowers than other tester's setups. Not that they are the only ones seeing this 'general' result (MADSHRIMPS talks somewhat about this in the latest review, but with a smaller sample of top-blowers being tested), I'm just not convinced that all the data is in yet, so I wouldn't personally be comfortable drawing those conclusions until I had a decent understanding of the contrasting info (for example here and here, just to list a few).
Yeah, at least it's a fairly objective means of testing cooling efficiency. I do find FrostyTech's reviews to be a helpfull data point, but since it removes the case entirely, it leaves some open questions (I'd want to understand the mounting pressure/mechanism better on that setup).
Speaking of "real world" testing... Let me start by saying that I'm not trying to bash AT, thier methods or any of thier staff - I'm only bringing up these various points (any/all previous and future posts) to help make people aware that they need to read and understand the testing environment, before they can decipher the results they see in the charts effectively. And this is true for all sites/reviews - read the fine print! .
Back to "real world"...
- AT loops the Far Cry River demo for 30 mins, throws out any 'spikes' and averages the other temperatures they see to come up with a 'load' result. Interestingly, they do this in an attempt to use more "real world" conditions (rather than some artificial Orthos/TAT loading), but the problem it causes may make it more trouble than helpful... if you've spent any amount of time reading through forums and reviews about heatsinks/cooling, when you see someone mention 'load' temps... do you think "that must be thier Far Cry River demo looping for 30 mins" results? No. What you see are generally Orthos/TAT or some other artificial loading results. That doesn't make AT's method invalid, but it may cause confusion.
- they use a X6800 (14x multiplier). Raise your hand if you have one of those. Of course AT is an enthusiast-oriented site and they want to try to hit some really high overclocks, but it might make thier results harder to compare to your own. Just to be clear, at least they ARE using a Core 2 Duo, and if you know what stepping it is and what voltages are being used at each level, you can get some idea... other sites often use some hot-running P4 cpu, which is no better. The only meaningful conclusions you can draw from those might be the relationships between coolers tested under the same conditions (but even then you have to consider the rest of the test configuration).
- they have case fans turned off for testing. Raise your hand if you don't have any case fans (ok, so some minority of people don't have any case fans and some minority of people have X6800s too). I'm not sure why they disable the fans (I assume to do the dBA testing, but they could just disble them for that test), but it's a bit odd that they tune thier "load" conditions to try to simulate "real world" conditions, but then use a processor that noone owns and then turn off all thier case fans - puzzling .
...anyway, sorry for rambling and kudos to anyone who bothered to read all of this - some people collect stamps, I spent the last 4+ months reading heatsink reviews and forums... maybe it's time to switch to stamps .
Originally posted by: gorobei
I wasn't actually talking about the cost of electricity, but rather if you do have to install 2-3 additional fans you'll want to upgrade a (arbitrary number here) 450W PSU to a 550W. For high quality PSUs that extra 100W can be a $30 bump in price. I knew someone might misinterpret, I was too lazy to bother clarifying. My bad.
At some point (or in some setups) it might be an issue, but for the most part, I still don't think it would make any difference. For example, my system is:
MSI P6n Platinum mb (nVida 650i-based)
E6400 @ 2.66GHz - stock volts (1.325)
Cooler Master Hyper Tx heatsink w/fan usually running @ ~9v (actually, it's PWM, running at 60%, so it gets 12v, 60% of the time)
XFX Extreme 7900 GS vga card
1 x 250GB Seagate Barracuda 7200.10 SATA drive
2 x 120GB Maxtor IDE drives
1 x TDK Dvd/CD IDE drive
1 x Antec Tri-Cool 120mm fan
1 x Sharkoon Silent Eagle 1000 120mm fan
2 x Yate Loon 120mm fans
(the Antec runs at ~5v, the others run at 12v)
1 x 60mm Vantec Stealth (NB cooler) @ 12v
...powered by a Xclio 550W GreatPower psu.
I have a power consumption meter on my UPS and at idle, the system draws ~202W. When running Orthos blend, it draws ~222W. Of course if all the drives were seeking around reading/writing files, it might draw a few more watts, but that's still well less than half the available wattage (actually the important measure is amperage, but I'm not too handy with the conversions ). Adding a few (or a dozen) fans is not going to put any strain on my power situation.
Originally posted by: gorobei
Was there a specific case with a ~150mm opening you referred to or was that just a "for instance" number? I'm not trying to be nit-picky or challenge your position, just curious what other removable tray types are out there. (I have a perverse plan to stick a tower H/S inside a SFF cube. Most cubes have a 3in or 75mm tall opening for the removable tray, so the Enzo Ultra-X(118mm) wouldn't fit either.)
Just a "for instance" number. My Antec P182 doesn't have a removable tray, but I know that they are one of the checklist items people often look for when buying a case. I hadn't done research to see exactly how much clearance any particular case has, but I've seen several people say something like "and I can finally remove my tray without removing the HSF first - yay!". Again, this would only be a benifit if it happened to work out that way in your case, so I was just using that as one possible checklist item some people might also be considering when they are looking for a HSF.
Far too often, I see someone start a thread with "Which heatsink should I get?" And the (equally bad) answer is usually "Get X, it rox!" (where X is whatever can be found at the top of some review chart at the time). Then a week later, heatsink arrives, user installs it, then returns to the forum... "hey, you guys said this was the one to get, but it (insert one or more of the following) -
* is way more expensive/heavy/sizeable/loud than I needed for my non-overclocked/mildly-overclocked system
* doesn't fit my motherboard
* doesn't fit my case
* is loud as hell
* my Orthos/TAT "load" results are much higher than I'm seeing on AT! (that's partly because they don't use those to measure with, but more on this later...)
* cools worse than the Y cooler I had
* won't fit with my side fan installed
* doesn't come with a fan!?
* has to be lapped???
* has to be installed with the motherboard removed?
* wouldn't slide back into my case, so I had to unmount it, slide my tray in and then mount it again - grrrr
* snapped my mb in half when I went 4-wheeling to the LAN party
* is not black!!
* looks like crap in my pimped-rig
...etc. . I guess people not stating thier criteria is a different subject, but the point is that people do have all sorts of criteria when shopping for a heatsink, so the more readily available info is, the easier it will be to find (if not by the people asking, at least by the people trying to help).
Originally posted by: gorobei
AT is probably a little remiss for not being more specific about the actual case layout/setup. But testing each cooler in a standard atx and a case better suited to top blowers would probably be a whole other can of worms.
Yeah, my biggest concern is not that they're not testing in 2-3 different configurations (that would be a lot of work), but more that they're drawing conclusions based on the current setup, without accounting for other reviews that show contrasting (sometimes a complete reverse) results.
Thier current accounting is:
- they've turned on the case fans and while the top-blowers improved, so did the side-blowers, so they consider that a wash.
- they've tested with the side removed and (IIRC) saw similar results to the above - but haven't tried a closed case with side fan, vs closed case with side vent (fan removed for Tower sinks, due to clearance issues) as far as I know.
- they say it's because they are using a modern (Core 2 Duo) chip and other sites are using older P4 chips (but it's unclear to me how/why a top-blower cools a P4 chip at some heat level than a side-blower does (reverse of the results AT gets) on that same chip/heat in those other reviews).
I'm still a bit puzzled as to exactly 'what' the difference is, but AT's setup seems to be tougher on top-blowers than other tester's setups. Not that they are the only ones seeing this 'general' result (MADSHRIMPS talks somewhat about this in the latest review, but with a smaller sample of top-blowers being tested), I'm just not convinced that all the data is in yet, so I wouldn't personally be comfortable drawing those conclusions until I had a decent understanding of the contrasting info (for example here and here, just to list a few).
Originally posted by: gorobei
Interestingly Frostytech uses some sort of simulation test to determine overall heat increase over ambient temp at 85W and 150W loads. (This would seem to be a better standard for testing general potential ability to cool separate from case layout issues. Though far less helpful in giving the reader a practical use, "real world" performance rating.)
Yeah, at least it's a fairly objective means of testing cooling efficiency. I do find FrostyTech's reviews to be a helpfull data point, but since it removes the case entirely, it leaves some open questions (I'd want to understand the mounting pressure/mechanism better on that setup).
Speaking of "real world" testing... Let me start by saying that I'm not trying to bash AT, thier methods or any of thier staff - I'm only bringing up these various points (any/all previous and future posts) to help make people aware that they need to read and understand the testing environment, before they can decipher the results they see in the charts effectively. And this is true for all sites/reviews - read the fine print! .
Back to "real world"...
- AT loops the Far Cry River demo for 30 mins, throws out any 'spikes' and averages the other temperatures they see to come up with a 'load' result. Interestingly, they do this in an attempt to use more "real world" conditions (rather than some artificial Orthos/TAT loading), but the problem it causes may make it more trouble than helpful... if you've spent any amount of time reading through forums and reviews about heatsinks/cooling, when you see someone mention 'load' temps... do you think "that must be thier Far Cry River demo looping for 30 mins" results? No. What you see are generally Orthos/TAT or some other artificial loading results. That doesn't make AT's method invalid, but it may cause confusion.
- they use a X6800 (14x multiplier). Raise your hand if you have one of those. Of course AT is an enthusiast-oriented site and they want to try to hit some really high overclocks, but it might make thier results harder to compare to your own. Just to be clear, at least they ARE using a Core 2 Duo, and if you know what stepping it is and what voltages are being used at each level, you can get some idea... other sites often use some hot-running P4 cpu, which is no better. The only meaningful conclusions you can draw from those might be the relationships between coolers tested under the same conditions (but even then you have to consider the rest of the test configuration).
- they have case fans turned off for testing. Raise your hand if you don't have any case fans (ok, so some minority of people don't have any case fans and some minority of people have X6800s too). I'm not sure why they disable the fans (I assume to do the dBA testing, but they could just disble them for that test), but it's a bit odd that they tune thier "load" conditions to try to simulate "real world" conditions, but then use a processor that noone owns and then turn off all thier case fans - puzzling .
...anyway, sorry for rambling and kudos to anyone who bothered to read all of this - some people collect stamps, I spent the last 4+ months reading heatsink reviews and forums... maybe it's time to switch to stamps .