PSUs that keep fans running after shutdown... balogna?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee, Devonavar
PSUs that keep fans running after shutdown... balogna?
I've read about some recent PSUs that keep their fans running for a period of time after the computer has shut down.
What benefit does this provide? Once you shut down the PC, the only heat left to remove is residual heat, which should disperse naturally anyway (albeit more slowly... but who cares?).
Gimmick? Or am I wrong?
What benefit does this provide? Once you shut down the PC, the only heat left to remove is residual heat, which should disperse naturally anyway (albeit more slowly... but who cares?).
Gimmick? Or am I wrong?
-
- *Lifetime Patron*
- Posts: 5316
- Joined: Sat Jan 18, 2003 2:19 pm
- Location: St Louis (county) Missouri USA
If you run your computer at 100%, then immediately shut down, the CPU temp will actually get hotter for a short time before cooling off. In theory, leaving the exhaust fans running for a time after shut-down, avoids this problem.
Is it necessary? Who knows? Probably it would be just as effective to let your computer return to idle temps before shutdown. I'm not buying.
Is it necessary? Who knows? Probably it would be just as effective to let your computer return to idle temps before shutdown. I'm not buying.
Where does this extra heat come from? When you turn the stove off but leave the kettle on the burner, does the kettle get hotter before it starts to cool?Bluefront wrote:If you run your computer at 100%, then immediately shut down, the CPU temp will actually get hotter for a short time before cooling off. In theory, leaving the exhaust fans running for a time after shut-down, avoids this problem.
What about the thermal energy stored in the heatsink? Once you shut the fans off, you suddenly and drastically degrade the ability of the heatsink to get rid of heat. Now the heat is going to flow back into the electronic component which is now cooler than the heatsink and which has a much lower thermal mass (i.e. for a given quantity of heat energy absorbed, its temperature will rise faster than a large hunk of metal). I don't enough thermodynamics to say if this can actually make the chip hotter than it was when it was powered, though. Seems hard to believe.
It's kind of like pushing a boulder up hill with the aid of a tow rope pulling from the top of the hill. Now somebody suddenly cuts the tow rope while you're standing behind it pushing. Where does all the potential energy stored in the boulder go?
It's kind of like pushing a boulder up hill with the aid of a tow rope pulling from the top of the hill. Now somebody suddenly cuts the tow rope while you're standing behind it pushing. Where does all the potential energy stored in the boulder go?
The heatsink is never, ever hotter than the device it is cooling. That's the whole point. So with no active airflow, the whole component/heatsink assembly won't cool down *quickly* after it is turned off. So... who cares?! Say you're fine running your CPU all day at 50C, why do you care how quickly it cools down after you turn the machine off?
Just thought of something:
While it's true that the hottest components in your system (or PSU) will not get any hotter after shutdown, the other components COULD heat up to a higher temperature than they normally run at due to the lack of any airflow. For example, lets say that the hottest part of your psu is, say, 50C. Now lets say there's another component in your PSU that doesn't generate much heat, and it stays at less than 30C while the PSU is running. When you turn off your PSU all the heat from that 50C component is dissipating into an enclosed box with no airflow, so the air inside your PSU (and thus all the other components) could heat up to, say 40C, whereas they would normally be much cooler.
Now, this probably makes zero difference for the lifespan of the PSU, but I just thought I'd point out how shutting down your computer COULD cause certain components to become hotter than they would be during operation.
Bryan
While it's true that the hottest components in your system (or PSU) will not get any hotter after shutdown, the other components COULD heat up to a higher temperature than they normally run at due to the lack of any airflow. For example, lets say that the hottest part of your psu is, say, 50C. Now lets say there's another component in your PSU that doesn't generate much heat, and it stays at less than 30C while the PSU is running. When you turn off your PSU all the heat from that 50C component is dissipating into an enclosed box with no airflow, so the air inside your PSU (and thus all the other components) could heat up to, say 40C, whereas they would normally be much cooler.
Now, this probably makes zero difference for the lifespan of the PSU, but I just thought I'd point out how shutting down your computer COULD cause certain components to become hotter than they would be during operation.
Bryan
-
- *Lifetime Patron*
- Posts: 5316
- Joined: Sat Jan 18, 2003 2:19 pm
- Location: St Louis (county) Missouri USA
putz....it has to do with sudden stoppage of airflow, or coolant-flow when the component/system is under high thermal load. To counter the adverse effects of this, a car's radiator fan may suddenly start when the car has been turned off. In a turbo car's manual, you are advised to let the engine idle for a few minutes before turning off the engine.
I think PSU makers are applying this to their new cooling feature. Whether a computer would benefit is debatable.
I think PSU makers are applying this to their new cooling feature. Whether a computer would benefit is debatable.
a couple of these should give you all the after-cooling you need:
http://www.directron.com/rd7ca.html
http://www.directron.com/rd7ca.html
-
- *Lifetime Patron*
- Posts: 1740
- Joined: Sun Dec 21, 2003 11:24 am
- Location: 'Sunny' Cornwall U.K.
Alternatively, fold 24/7.
No need to worry about thermal dynamics, residual aheat, convection currents or thermal shock when you switch off at all then.
You also eliminate expansion/contraction issues as well, as everything should be at a reasonably stable temp.
Could help prevent aging of electronic components.
There you go... problem solved.
Your Welcome
Pete
No need to worry about thermal dynamics, residual aheat, convection currents or thermal shock when you switch off at all then.
You also eliminate expansion/contraction issues as well, as everything should be at a reasonably stable temp.
Could help prevent aging of electronic components.
There you go... problem solved.
Your Welcome
Pete
"putz....it has to do with sudden stoppage of airflow, or coolant-flow when the component/system is under high thermal load. To counter the adverse effects of this, a car's radiator fan may suddenly start when the car has been turned off. In a turbo car's manual, you are advised to let the engine idle for a few minutes before turning off the engine. "
Turbo is really different. You drive around hard and the turbos are sitting there red hot, oil is constantly pumping though them... if you don't let the turbo cool down a little before shutting the engine off, soon as power is cut oil stops pumping, but the turbos are still red hot. The oil heats up and gets burned on. Doing this alot builds up layers of burnt oil, turbos run even hotter and their life is killed.
Computers don't have anything similar, theres no huge store of energy anywhere. Cut put to a computer and find something that gets hotter and you could make some arguement, in the case of the car when power is cut the oil left in the red hot tubro heats up. Its also not that something gets a little warmer, but it gets significantly hotter, hotter than it would ever get in normal usage... not even that but that it got hot enough make a difference in the life of the part. I don't believe something is heating up to 90C when you cut the power, yet runs cool when computer is on.. probably the best reasoning for not needing the fans to stay on... how often do you see power supplies die? If there was some need they wouldn't last years and years, if there was some need you wouldn't see a VERY small % of power supplies that incorporated a timed faned.
Really when you do shutdown your comptuer isn't 100% loaded as its shutting down, its got a few seconds of light usage which is probably infinitely longer than it needs.
Turbo is really different. You drive around hard and the turbos are sitting there red hot, oil is constantly pumping though them... if you don't let the turbo cool down a little before shutting the engine off, soon as power is cut oil stops pumping, but the turbos are still red hot. The oil heats up and gets burned on. Doing this alot builds up layers of burnt oil, turbos run even hotter and their life is killed.
Computers don't have anything similar, theres no huge store of energy anywhere. Cut put to a computer and find something that gets hotter and you could make some arguement, in the case of the car when power is cut the oil left in the red hot tubro heats up. Its also not that something gets a little warmer, but it gets significantly hotter, hotter than it would ever get in normal usage... not even that but that it got hot enough make a difference in the life of the part. I don't believe something is heating up to 90C when you cut the power, yet runs cool when computer is on.. probably the best reasoning for not needing the fans to stay on... how often do you see power supplies die? If there was some need they wouldn't last years and years, if there was some need you wouldn't see a VERY small % of power supplies that incorporated a timed faned.
Really when you do shutdown your comptuer isn't 100% loaded as its shutting down, its got a few seconds of light usage which is probably infinitely longer than it needs.
-
- Patron of SPCR
- Posts: 700
- Joined: Thu Mar 13, 2003 2:38 pm
- Location: California, US
- Contact:
Let's say the chip is running at 55C, and the base of the heatsink is 50C, so we have a 5-degree gradient over the thermal interface. Now turn off the computer. The chip's temperature drops, but only to 45C, because the heatsink is still at 50C and slowly cooling. Then, as the heatsink slowly cools, the CPU and heatsink approach room temperature.Metaluna wrote:What about the thermal energy stored in the heatsink? Once you shut the fans off, you suddenly and drastically degrade the ability of the heatsink to get rid of heat. Now the heat is going to flow back into the electronic component which is now cooler than the heatsink and which has a much lower thermal mass (i.e. for a given quantity of heat energy absorbed, its temperature will rise faster than a large hunk of metal). I don't enough thermodynamics to say if this can actually make the chip hotter than it was when it was powered, though. Seems hard to believe.
It's kind of like pushing a boulder up hill with the aid of a tow rope pulling from the top of the hill. Now somebody suddenly cuts the tow rope while you're standing behind it pushing. Where does all the potential energy stored in the boulder go?
If anything, running the fans after the CPU is off is worse for the chip, because the chip's temperature will drop more quickly (like greeef said). Rapid temperature changes can be bad: see what happens when you put a hot glass in cold water. Of course, nothing this drastic will happen inside a PC, so running a fan or not after shutdown makes no difference whatsoever.
Your boulder analogy makes no sense because there's no potential energy stored in the chip. I mean, when you play a game for 2 hours and then shut off the computer, does the machine have to play the game in reverse for 2 hours to relieve all the stored energy? Running a PC is more like pushing against an immovable wall, because all the energy going into the system is coming out as heat.
Last edited by SometimesWarrior on Sun May 30, 2004 2:51 am, edited 2 times in total.
-
- Patron of SPCR
- Posts: 700
- Joined: Thu Mar 13, 2003 2:38 pm
- Location: California, US
- Contact:
That temperature graph is idiotic. Some guy just squiggled a couple lines and applied some arbitrary axis labels. It looks like the graphs I made up for my chemistry lab.mond wrote:a couple of these should give you all the after-cooling you need:
http://www.directron.com/rd7ca.html
Honestly, if there is no energy being drawn by the PC after shutdown, then where's that heat coming from? If these guys are right, then does that mean the thermal protection circuitry that shuts down an overheating PC is doing more harm to the system?
@Becks: interesting, thanks for the info.
Hmm...I reckon the whole thing is just snake oil, for the following reasons - please pick apart my logic, such as it is, where necessary....
The comparison with a car engine is completely spurious. A car's cooling system sometimes continues after switching the engine off to protect the *coolant*, not the engine. If you hammer a car down the motorway, suddenly turn the engine off and the coolant stops flowing, the heat from the block could boil the now-static coolant in the waterways (undesirable), whereas keeping it flowing ensures that it continues to be replaced by...umm...cooled coolant from the radiator. The same principle applies with oil flowing through a turbocharger, as someone mentioned above.
The engine itself *cannot* get hotter once the heat source (burning fuel) has been removed. Heat cannot go from a cold place to a hot place, it's against the laws of thermodynamics (and before anyone mentions heat pumps, the same principles apply, it's just so not immediately obvious). The only result of stopping the coolant flow would be that the block and head would cool at a lower rate, unless it blew a gasket or something as a result of boiling its coolant.
As far as computer components go, once the power is killed, they also cannot get hotter, and in fact they will immediately start cooling until they match the temperature of their surroundings. I suppose it's just possible that a small amount of heat from hot components could be convected to cooler ones once the airflow stops, as bcassell said, but really the temperature differentials would be so microscopic that they would make no practical difference.
What kills electronics even faster than heat is rapid changes in temperature...an ordinary tungsten filament lightbulb will last for years if left on continuously, it's the changes in temperature caused by switching on and off that shorten its lifespan (of course, the cost of the electricity leaving a lightbulb on all the time would be much more than replacing it once in a while). So really, for maximum life-expectancy computers should be allowed to cool at as slow a rate as possible.
Conclusion: it can't do any good, might conceivably do some harm, probably makes absolutely no difference whatsoever...
The comparison with a car engine is completely spurious. A car's cooling system sometimes continues after switching the engine off to protect the *coolant*, not the engine. If you hammer a car down the motorway, suddenly turn the engine off and the coolant stops flowing, the heat from the block could boil the now-static coolant in the waterways (undesirable), whereas keeping it flowing ensures that it continues to be replaced by...umm...cooled coolant from the radiator. The same principle applies with oil flowing through a turbocharger, as someone mentioned above.
The engine itself *cannot* get hotter once the heat source (burning fuel) has been removed. Heat cannot go from a cold place to a hot place, it's against the laws of thermodynamics (and before anyone mentions heat pumps, the same principles apply, it's just so not immediately obvious). The only result of stopping the coolant flow would be that the block and head would cool at a lower rate, unless it blew a gasket or something as a result of boiling its coolant.
As far as computer components go, once the power is killed, they also cannot get hotter, and in fact they will immediately start cooling until they match the temperature of their surroundings. I suppose it's just possible that a small amount of heat from hot components could be convected to cooler ones once the airflow stops, as bcassell said, but really the temperature differentials would be so microscopic that they would make no practical difference.
What kills electronics even faster than heat is rapid changes in temperature...an ordinary tungsten filament lightbulb will last for years if left on continuously, it's the changes in temperature caused by switching on and off that shorten its lifespan (of course, the cost of the electricity leaving a lightbulb on all the time would be much more than replacing it once in a while). So really, for maximum life-expectancy computers should be allowed to cool at as slow a rate as possible.
Conclusion: it can't do any good, might conceivably do some harm, probably makes absolutely no difference whatsoever...