ASUS A7V PLL Chip Cooler

SUMMARY: A How-To Cool the PLL (clock Generator) chip on the ASUS A7V – knocked temps down to 35 C.

In my review of the ASUS A7V, I noticed that the KT133 and PLL chips run hot. ABIT includes active cooling for the KT133 chip and I used a TennMax Lasagna to cool it on the ASUS. After some puzzling over how to do it and what to use, I figured out an approach using (again) a Lasagna heatsink. It’s not hard to do but does involve some aluminum cutting and shaping.

PLL Cooler

Pictured above is the finished product. I had to use some neoprene tubing between the Lasagna tab and the nut because the nut doesn’t clear the Lasagna’s side; I also used nylon washers just to be safe. What follows is a detailed How-To:

PLL

The playing field.

Pictured above is the area where the PLL chip lives. Note that there is a tin-can (this is the reference crystal for the PLL chip – thanks Brainchasm!) next to it; the problem here is that it is taller than the PLL chip (less than 1mm), so any cooling must clear it. I thought about it and decided I had to make a spacer to fit any cooling solution on the PLL chip.

Cutout

Make a template…

First step was to make a template so I could fashion the spacer. I am using a TennMax Lasagna for this, so the I made the template by tracing the Lasagna’s outline on a piece of thin cardboard. I lined up the mounting holes and figured about where the tin can was, then cut a hole in the template for it.

Al Spacer

Use it to fashion an aluminum spacer…

I then laid the template on a thin (2mm) piece of aluminum and carefully cut out a square using a radial arm saw. I then used a stationary belt sander to cut the corners and chamfer the edges. I have a Dremel Drill Press and I used that to cut out the opening for the tin can. Finally, I used some 600 grit paper to lap the spacer flat.

I then used thermal grease and stuck the spacer on the Lasagna. Press it firmly together so that the grease has enough suction so it does not move. The heatsink is ready to go.

Bolts

Epoxy bolts to mount the heatsink…

Mounting was the thing that perplexed me the most until I remembered how I epoxied bolts to slotkets. Eureka! I was thinking of epoxying the heatsink directly to the PLL chip, but it’s too permanent. With the bolts, you can mount just about anything, anytime.

I used 5 minute epoxy (I don’t like to wait). First I marked with a felt tip pen where I wanted the bolts on the board. The I mixed a gob of epoxy and placed it on the board’s marked spots. I used steel bolts (4 x 32); in order to avoid any possible shorting of components, I waited until the epoxy was tacky.

I then placed the bolt heads (I sanded these down so they were flat) lightly on the nearly-set epoxy so it would not touch the board and let it set. When it’s almost set, the bolts will not fall down; I placed the Lasagna’s tabs over the bolts and set it in place to cure, about 3-4 minutes more. Once this is done, the bolts are set.

Fiberglass

Reinforce the bolts with fiberglass.

I then reinforced the bolt heads using fiberglass cloth. I cut a small square of fiberglass, carefully placed a dollop of epoxy around the bolt’s base, and slid the fiberglass over the bolt onto the epoxy. As you can see above, the glass virtually disappears when saturated.

WORDS OF CAUTION: Fiberglass is very sticky and permanent. I masked the RAM and Drive slots to avoid any possibility of shorting out pins and suggest you do the same. Also, remember that the Lasagna’s tabs are on its base, so do not over-epoxy or you will not be able to mount the heatsink to contact the PLL chip. You can use nail polish remover as an epoxy solvent – it’s acetone with perfume.

Wait 10 or 15 minutes and you’re all set to mount the heatsink.

I placed a dab of grease on the PLL chip, slipped the Lasagna with its spacer over the bolts, cut two small lengths of neoprene tubing, slipped them over the bolts, placed washers on top and finger tightened the bolts.

Voila! A PLL chip cooler.

RESULTS

So is all this worth it? I cut the PLL temp from 65+ C to 35 C at 900 MHz. I am now able to run at 115 MHz FSB whereas before I could only get to 113 MHz. Finally, I can run Prime95 @ 950 MHz for about 10 minutes before it freezes; previously I could not post into W98 @ 950 MHz. With some additional tweaking, I may get 950 MHz stable yet.

Email Joe


Be the first to comment

Leave a Reply