Pc V3.3

[Duffman]

i did make my own topic, but on the subject of clockers pc, how come no one has tried to rig up some air conditioner type thing to cool the computer

[lynx]

Originally posted by Duffman3@24 November 2003 - 22:35
i did make my own topic, but on the subject of clockers pc, how come no one has tried to rig up some air conditioner type thing to cool the computer

Because, and if you'd been around a little longer you would know this, it's really all about getting the NOISE level down without compromising the cooling. I hardly think an 80db hurricane quite fits the bill.

[Duffman]

2 birds with 1 stone in the summer

[clocker]

Originally posted by Duffman3@24 November 2003 - 15:35
i did make my own topic, but on the subject of clockers pc, how come no one has tried to rig up some air conditioner type thing to cool the computer

Duffman,

Vapochill makes this refrigerated unit for a mere $900. Intel or AMD. Several different models.
Belly up to the bar, boy.

@Kunal & Lynx,
This pic shows the u-shaped bracket that mounts this fan to the Zalman cooler.
It looked like it would just slip between the fins of the AMD HS...

And, son of a bitch!, it does. simple tension has held it in place so far.

As effective as this set up may be, I hate to keep it as it renders my Zalman HS useless for resale without this rather specialized fan. Till I come up with something better though, it stays.
Lynx, without a shroud, nothing is really directing the airflow specifically into the HS finning. That fan just floods the entire area, which is an effect that I rather like.
The first set up ( with the 80mm fan adaptor), really focused the airflow and a steady stream of heated air could be felt escaping from the sides of the HS. On the RH side unfortunately, this escaping air blew directly onto my first stick of RAM.
Hardly ideal.

[Duffman]

well my idea involed a ac in the wall, plastic wrap, papertowell tubing and massive amounts of duct tape, imagine.. if you will
plastic wrap coming off the ac in a triangle type thing ending with cardboard tubing that runs the the nearby computer and then into the computer. although the computer would have to be wrapped not to let too much air out. lol just picture it isn't it........ cool, get it cool

[clocker]

Duffman I have no doubt that there already exist several rigs set up as you propose.

I'm not really a duct tape, paper towel roll kinda guy, though.

[lynx]

Originally posted by clocker@24 November 2003 - 23:45
@Kunal & Lynx,
This pic shows the u-shaped bracket that mounts this fan to the Zalman cooler.

It looked like it would just slip between the fins of the AMD HS...
And, son of a bitch!, it does. simple tension has held it in place so far.

As effective as this set up may be, I hate to keep it as it renders my Zalman HS useless for resale without this rather specialized fan. Till I come up with something better though, it stays.
Lynx, without a shroud, nothing is really directing the airflow specifically into the HS finning. That fan just floods the entire area, which is an effect that I rather like.
The first set up ( with the 80mm fan adaptor), really focused the airflow and a steady stream of heated air could be felt escaping from the sides of the HS. On the RH side unfortunately, this escaping air blew directly onto my first stick of RAM.
Hardly ideal.

It's impossible to tell from that picture, but I assume that there's a hole in the middle of that u-shaped bracket. If not, you are shielding the most important part of the heatsink - the bit immediately on top of the cpu.
Edit: that's why I like copper based heatsinks, they make the area above the cpu less important.

I think the reason you are seeing an improvement in temps with the back panel fans blowing air inwards (compared to outwards) is that they are blowing across the cpu heatsink and consequently doing part of the job of a shrouded heatsink fan.

I hadn't considered the effect of hot air from the cpu blowing across the memory. Something I need to look at next time I start overclocking.

Speaking of overclocking, I was thinking about the trick of connecting pins on the back of the motherboard (for Thoroughbred and Barton processors) and suddenly realised that there might be a possibility of using some sort of jumper to try this trick out. The important bit was going to be the spacing of the pins. I have found out that the pin spacing horizontally and vertically is 0.1", the rows of pins are offset by 0.05" relative to the next row. The pins which need to be joined are on adjacent rows (shifted by half a row).

Converting this to metric, pins spacing is 2.5mm in any row, with 1.25mm between adjacent rows, offset by 1.25 mm. So the distance between a pin on one row and the nearest pin on an adjacent row is 1.414x1.25mm = 1.77mm. I'm pretty sure that there are some jumpers with pin spacing at 1.75mm, I'm also pretty sure I've got some, somewhere.

With this method, if it works a blob of hot-gun glue makes it semi-permanent, a craft knife to cut the glue restores the board to the original condition.

[clocker]

Lynx,

SpeedStrip looks like the fastest, easiest and most reversable method of unlocking the multipliers on your chip to me.

No, there is no hole on the center of the bracket. And the fan hub is fairly large. The fact that the fan assy. is raised over the HS probably helps unmask the finning somewhat, but still, it's hardly ideal. A work in progress, to be sure.

I'll be moving my temp probe onto the RAM stick in question this evening. See what shows up. At the very least, a simple deflector could easily redirect the escaping air upwards rather than directly at the RAM. Might be helpful.

[lynx]

I've seen that Speedstrip idea before, I'm not too keen on putting anything between the cpu and the socket, even if it only half a millimeter. I can well believe that the extra strain from the heatsink bracket could well crack the cpu die.

The method I propose joins the same pins on the back of the motherboard. If anything goes wrong the cpu has not been tampered with in any way, and until I'm happy with the results neither has the motherboard. In fact I don't need to even disturb the hsf etc.

Now I just need to work out what I've done with those jumbers.

I was thinking along the same lines as you with regard to having a deflector between the cpu and the memory. I suspect I've got exactly the same problem, and it probably explains why my mobo temp is so near to my cpu temp (the mobo temp chip is between the cpu and the memory).

Btw, my main heating has been down this weekend, so the ambient temp has been down at about 15C first thing in the morning. This dropped the cpu temp (with the fan at 100%) to 28C. Mobo temp was also 28C, which tends to confirm the thoughts about the heat from the hsf warming the mobo chip.

[clocker]

Right now my chip temp is 42C and the mobo is 32C.

I have always had a good differential in those two readings. Board temp tends to be 2-3degC over ambient case temp.
BTW, my first reads on the memory temp show 31.1C- very similar to the board temp. This doesn't really surprise me as that fan is blasting air all over. Maybe not the most efficient for the HS, but a boon for the rest of the neighboring components.