• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Installing a TT smartfan 2?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

SenorBeef

Member
Joined
Apr 23, 2002
Some dumb questions, probably...

Where's the best place for the thermal sensor to be installed?

Can someone give me a quick guide for what wires going to what? I'm sure I can figure it out, but if someone could explain it to me so I don't screw anything up, I'd appreciate it. I noticed there were quite a few wires on there, and I'm not sure what the exact install procedure is. TT's site doesn't have one. I want to run with the thermistat.

Anyone have the 'coolmod' fan? I got that because excaliberpc.com didn't have the regular version. It looks kinda cool, but it seems like the plastic 'arms' cover a decent bit of space on the fan and would reduce air flow. Does it, significantly? I think I'm just gonna toss the LED plate, or give it to my friend, because it's not worth impeded airflow.
 
Take the power for the fan from a 4 pin molex with the supplied cable. Connect the fan to this and the single yellow wire to your mobo header (cpu fan, with RPM mon.) Remove the little blue jumper from the wires at the corner of the fan and add your sensor.

Best place for the sensor would be on the bottom of the CPU. Tape it diagonally, corner to corner, so it clears the socket and test that it does not interfer with any under socket thermister.
Bring the two wiress out from the two adjacent corners either end of the edge labled Socket etc - this will help the pins to make good contact and the chip to remain steady.
 
Ive got that coolmod thing....Im not convinced that the LED panel really does limit airflow significantly...not sure :\

I didn't use the thermister thing, incase it fell off or something and my fan just died :D
 
start said:
Ive got that coolmod thing....Im not convinced that the LED panel really does limit airflow significantly...not sure :\

I didn't use the thermister thing, incase it fell off or something and my fan just died :D

Just use a small piece of tape,electrical works great,across the joint of the thermister plug to prevent it falling off. I use it on all my molex connectors to be safe.
 
The best place for the thermal sensor is in the trash. Use the Rheostat it came with. That allows you to be flexible all the time, not just when the sensor says so.
 
CSaddict said:
The best place for the thermal sensor is in the trash. Use the Rheostat it came with. That allows you to be flexible all the time, not just when the sensor says so.


agreed i hated that thermal stat, i dont think they work that well unless there on the processor. The Rheostat is a much better idea and can be controlled to your liking.
 
Yeah , I would have to agree , dump the thermal sensor , and if you're so inclined get rid of the led facia as well . I found that it made a large "hole" with no airflow. Anyways theyre soo cheap that its not a big deal to get rid of the extras , especially if you want pure performance.
 
The thermal probe is actually not that bad, but it is not very aggressive if you intend on overclocking. So in that case the rheostat is better for control. If you want it more aggressive its possible to fix. I wired a 47k pot in parrallel with the probe and am using it on an overclocked system so the probe is more aggressive. You can also use a plain 47K resistor which seems to make the probe work just about right.
 
I also would install the probe on top of the cpu, not under because it doesn't let the cpu lie flat in the socket. Which can effect how the heatsink sits on the cpu. clip the probe end with nail clippers as close as possible. Put a very small drop of thermal paste on the tip. Stick the probe using the double sided tape to the area around the cpu core, having the tip make contact with the side of the cpu core. Before you stick it there make sure the wires won't interfer with your heat sink or the cpu pads. Also make sure the probe or wires are no higher then the top of the cpu core. I stuck another piece of tape on tape for good measure. Pull off the backing of the tape and put some baby powder on it so if you pull off the heatsink it won't pull the probe up with it.
 
If you mount it under the cpu the thin wires tilt the cpu slightly in the socket. That tilt to me is not a good idea with a heatsink you want as flat as possible with the cpu core.
 
crull said:
If you mount it under the cpu the thin wires tilt the cpu slightly in the socket. That tilt to me is not a good idea with a heatsink you want as flat as possible with the cpu core.

Clips are bent/sprung to have one point of presure which occurs over the centre of the CPU and so the HS will find its own level (as long chip is stable).

Though I also prefer manual control.
 
Might be slightly off topic, but here goes...

What is the controlling mechanism for the SF2? I thought in manual mode that it was a simple rheostat that burned off watts when the fan was turning slowly. however, someone else told me recently that there was some kind of on-board circuitry that was adjusted by the rheostat. Exactly does one of these fans adjust speed, and how is the thermistor wired in? Thanks.

Ken
 
Robin Hood said:


Clips are bent/sprung to have one point of presure which occurs over the centre of the CPU and so the HS will find its own level (as long chip is stable).

Though I also prefer manual control.

Think about what your saying. Your tilting the whole core and in order for the heat sink to be flat against it, it also has to be tilted. Not to mention more stress on one side of the core then the other from lack of support on one side from not lieing flat in the socket. The spring clip does twist but one side of it is flat to hold against the heatsink, twisting the clip will put more pressure on one edge then the other, not to mention un-even force on the socket tabs if your using all 3. In order for the best possible themal thansfer all components have to be flat to each other.
 
crull said:
Think about what your saying. Your tilting the whole core and in order for the heat sink to be flat against it, it also has to be tilted. Not to mention more stress on one side of the core then the other from lack of support on one side from not lieing flat in the socket. The spring clip does twist but one side of it is flat to hold against the heatsink, twisting the clip will put more pressure on one edge then the other, not to mention un-even force on the socket tabs if your using all 3. In order for the best possible themal thansfer all components have to be flat to each other.

I have read my post(s) again and I don't see anything wrong in what I have said, so I shall try to rephrase it.

A heatsink levels itself to the CPU core, not the other way around. The core is the HS's only support (save the 4 round spunge pads that are in the same plain as the core) and it will therefore sit flat against it unless pressure is applied outside of the core - which it isn't.
The clip is not twisted, the socket tabs are not affected. I agree there isn't as much support to the CPU with the wires below but at least with a wire at each corner, the chip is stable.
 
Robin Hood said:


I have read my post(s) again and I don't see anything wrong in what I have said, so I shall try to rephrase it.

A heatsink levels itself to the CPU core, not the other way around. The core is the HS's only support (save the 4 round spunge pads that are in the same plain as the core) and it will therefore sit flat against it unless pressure is applied outside of the core - which it isn't.
The clip is not twisted, the socket tabs are not affected. I agree there isn't as much support to the CPU with the wires below but at least with a wire at each corner, the chip is stable.

I have read my post(s) again and I don't see anything wrong in what I have said, so I shall try to rephrase it.

A heatsink levels itself to the CPU core, not the other way around. The core is the HS's only support (save the 4 round spunge pads that are in the same plain as the core) and it will therefore sit flat against it unless pressure is applied outside of the core - which it isn't.
The clip is not twisted, the socket tabs are not affected. I agree there isn't as much support to the CPU with the wires below but at least with a wire at each corner, the chip is stable. [/B][/QUOTE]

I just pulled out an SLK-800 to check the clip, so I could be sure. OK think about this a moment. The clip that is putting actual pressure on the heatsink is around 1/4 wide. In order for that 1/4 to supply even pressure something has to twist such as the clip or it won't be pushing down with even pressure against the heatsink. There won't be even pulling force on all three tabs of the socket using the clip when twisted however slight of a twist there is. There also might be a risk of cracking your core because the heatsink as pressure is applied will make contact with not all 4 pads at the same time. Then the heatsink will make contact with only one edge of the core until full pressure is applied on the clip. The heatsink at that point might be flat against the core, but the pressure on the heatsink is only being applied by one corner of that 1/4 or if the clip is flat then it has been bent from being applied at an angle. Having the wires under the cpu also keeps the cpu pins from making the best possible electrical conection. If all 4 things are flat, such as socket, cpu base & core, heatsink base, and the bottom of the clip. Then if any one of those is at an angle then something has to be twisted. I am sure the clip was designed with a slight tilt in mind from very very slight inperfections, but having those wires underneath tilts the core more then a 1/16" which I don't think is good. Something has to be bent for everything to be sitting flat and if not then the bottom of the clip that is making contact with the heatsink is at an angle and not applying even pressure.
 
I'm sorry Crull I still don't follow you, nothing twists and how can a clip apply uneaven pressure to a HS with only one point of contact that is over the centre of the core?

I don't know what else to say so if some one else could step in here and put us on the right track, it would be appreciated.

I don't like the idea of wires under my cpu, but the 2 single green strands only measure 1/32 side by side, so individually they would raise the one edge even less - not more than 1/16 as you describe.

I too am worried about pin contacts which is why I suggested inclining the chip towards the "socket" label since this should put the end of the pins against the contacts first, in the action of the tray CPU tray closing.

Edit - V9 reference removed.
 
Last edited:
Robin Hood said:
I'm sorry Crull I still don't follow you, nothing twists and how can a clip apply uneaven pressure to a HS with only one point of contact that is over the centre of the core?

I don't know what else to say so if some one else could step in here and put us on the right track, it would be appreciated.

I don't like the idea of wires under my cpu, but the 2 single green strands only measure 1/32 side by side, so individually they would raise the one edge even less - not more than 1/16 as you describe.

I too am worried about pin contacts which is why I suggested inclining the chip towards the "socket" label since this should put the end of the pins against the contacts first, in the action of the tray CPU tray closing.

Edit - V9 reference removed.

I have installed my Palamino with the wires underneath, and looking at it from the side the cpu was tilted around 1\16". If the wires are underneath I would put them running parrallel to the length of the clip that attaches to the tabs on the socket. The shortest sides of the core, such as cpu info labelside or the opposite end. I don't believe that will interfer with the heatsink as much that way, because then the contact of the clip to the heatsink will still be even. What your saying would be true if the pressure from the clip was applied at a single center point to the heatsink, but its not. The bottom of the clip that applies pressure is around a 1/4" wide shaped like a V, looking at it from the side. If you put the wires under it then this might be OK "===>V<===", but not this "===>||<=== (socket tab end view)

The reason I think it gets tilted a 1/16" is because the socket is recessed and the wires don't bend at perfect angles to fit inside the socket. The only way to get that 1/32 you mention is to have those wires at perfect bends to fit the socket, but to get that you most likely will end of breaking the wires. and have extra wires under the opposite side to get rid of the tilt. If you only put something under one edge of something flat then the height of the outer edge will be higher because of the angle.
 
Back