| Forum Home | ||||
| Press F1 | ||||
| Thread ID: 50784 | 2004-11-01 10:50:00 | Fans: Blast air on to CPU? Or suck it away from CPU? | george12 (7) | Press F1 |
| Post ID | Timestamp | Content | User | ||
| 286799 | 2004-11-02 00:11:00 | Oh, is that right? There was I all this time thinking it may have been Newton, but now I'm not really sure. The reasoning goes something like this: The amount of air displaced by the fan depends upon the resistance to flow. By reciprocity the flow resistance through the heatsink fins does not depend upon the direction of flow. So the quantity of air and its average velocity displaced by the fan to at least a first order of estimation is the same whether it sucks or blows. Thus the average heat transfer is the same in each direction. The only variable in this system (heatsink and fan) is the average air temperature, either being drawn into the heatsink, or being drawn into the fan and out through the heat sink. That is why the air temperature in the case is so important. The very tiny temperature rise through compression and work done on the air by the fan from inlet to outlet I'm ignoring. |
Terry Porritt (14) | ||
| 286800 | 2004-11-02 00:49:00 | Theoretically, then, a fan sucking air away from the heatsink shouldn't have a 'dead zone' in the middle? | POTUS (5276) | ||
| 286801 | 2004-11-02 00:51:00 | The temperature of the heatsink (heat exchanger ;-)) and the CPU to which it is reasonably well connected depends on these things: 1) the watts of heat dissipated. 2) the mass of air flowing through the fan/sink unit. 3) the input temperature of the air entering the fan/sink module. It doesn't matter whether the air enters through the fan or through the fins. If you get cooler air to the fan/sink , your CPU will be cooler. If the heated air can escape from the box, your CPU will be cooler. If you just stir the air in the box, it will reach a higher temperature, because (3) goes up. If you obstruct the fan, the CPU will get hotter because you are reducing the airflow. The figures I remember from a Marconi book are 8500 cubic ft/min with 20°C air input for 20°C temperature rise. (That's for a 100KW transmitter ;-)). |
Graham L (2) | ||
| 286802 | 2004-11-02 01:52:00 | > > As Peter has said, one of the most effective cooling > methods is a 120mm fan in the side of the case > blowing directly at the cpu fan. In this case the > cpu fan must also blow into the heat sink, so that > the flow directions dont oppose one another. Ducting: smaller fans, less noise, less engery consumed, greater efficency. D. |
drb1 (4492) | ||
| 286803 | 2004-11-02 02:22:00 | > > > > > As Peter has said, one of the most effective > cooling > > methods is a 120mm fan in the side of the case > > blowing directly at the cpu fan. In this case the > > cpu fan must also blow into the heat sink, so that > > the flow directions dont oppose one another. > > Ducting: smaller fans, less noise, less engery > consumed, greater efficency. > > D. May I be permitted a small correction. Larger fans = slower rotation for same (or greater airflow) = less noise. Cheers Murray P |
Murray P (44) | ||
| 286804 | 2004-11-02 02:25:00 | > > > > > > > > As Peter has said, one of the most > effective > > cooling > > > methods is a 120mm fan in the side of the > case > > > blowing directly at the cpu fan. In this > case the > > > cpu fan must also blow into the heat sink, > so that > > > the flow directions dont oppose one > another. > > > > Ducting: smaller fans, less noise, less engery > > consumed, greater efficency. > > > > D. > > > May I be permitted a small correction. Larger fans = > slower rotation for same (or greater airflow) = less > noise. > > Cheers Murray P Or larger fans running as fast as possible = heaps of air flow = PC sounding like a jet plane |
Pete O'Neil (250) | ||
| 286805 | 2004-11-02 02:28:00 | If one starts to go into fine detail, then there are differences between sucking and blowing. These are primarily at the intake of the fan. When the fan is taking air from the case and blowing into the heatsink, the intake conditions could be considered 'free air'. There will be swirl though, or vortices in the direction of fan rotation giving the air rotational velocity before it enters the fan. Like water going down a plug hole. When the air is sucked out of the heat sink, the fins will tend to act as a flow straightener. In either case the velocity at the eye of the fan will be lower than at the periphery, a dead zone if you like, as Potus mentioned, but there will still be some flow towards the eye. Smoke is a good visualiser, trouble is not many smoke these days, so a lot of good simple flow science is never seen :) The fine detail analysis would be very complex, that's why I said 'to a first order estimation'. It's very similar to a centifugal pump acting either as a pump, or being driven like a turbine, depending on flow direction. The simplistic equations are the same, just sign reversals. |
Terry Porritt (14) | ||
| 286806 | 2004-11-02 02:40:00 | > Or larger fans running as fast as possible = heaps of air flow = PC sounding > like a jet plane Yep, if you can crank them up that much without doing damage, how much could/would you over volt a 92 or 120mm fan? Think I prefer V8s to RX7's anyway ;) Cheers Murray P |
Murray P (44) | ||
| 286807 | 2004-11-02 02:44:00 | BTW Terry, I smoke, but it's a moot point to test on my gear as the fan is separated from the heatsink by about 50mm, it's gotta blow or it would suck :) Cheers Murray P |
Murray P (44) | ||
| 286808 | 2004-11-02 02:50:00 | > > Or larger fans running as fast as possible = heaps > of air flow = PC sounding > like a jet plane > > Yep, if you can crank them up that much without doing > damage, how much could/would you over volt a 92 or > 120mm fan? Think I prefer V8s to RX7's anyway ;) Ive seen fans running at 20+ volts before, doesnt take long to burn out though . I would imagine that 14-15v would be perfectly safe . Although you would in theory start to experience diminishing results, no matter how much air you blow over the HS there will be a certain temp you cant go below . Delta make some freaky fans, they draw some much current that its reckonmended running them directly off the PSU, you run the risk of burning out motherboard traces if you plug them into the 3pin headers . |
Pete O'Neil (250) | ||
| 1 2 3 4 5 6 7 | |||||