Tricks and Secrets for Successful CD Burning CDR  Burning CDRW Burning, Burning Issues, CDR, CDRW, CD Burning forum, burningissues for Burning & copying copy protected CD's, CD burners, CD writing, CD CDR CDRW Burning

Intake-air temperature - not ambient [room] temperature - is one of the three basic components of a °C/W equation. With the kit specified here, it's easiest to use the "indoor" [encapsulated] thermistor of a hacked 63-1024.

The 63-1024 comes with a useful bent wire mounting-bracket; but it is a little short for our purposes, & when testing socket 7/370/462 CPU's/fansinks, it helps to make up a wire bracket to hold the thermistor at a suitable distance from the fan-centre - somewhere around 50mm seems to avoid the worst local-flow effects - it is obviously necessary to use plastic-coated wire &/or to mount the bracket between fibre washers & around a non-conductive sleeve to avoid direct conduction of heat from the fansink body.

A simple plywood widget I use to assist measuring the intake-air temperature on Slot 1 mobo's is this:

. . . it clips solidly into a SDRAM slot & holds the "indoor" [encapsulated] thermistor of a hacked 63-1024 at 3-4 convenient distances of your choice - I usually aim at around 50mm - to the core & fan of a Slot 1 processor/slocket.

Techniques: measuring "CPU-case" temperatures:

At best . . . now he tells us . . . you can only indirectly measure CPU-case temperatures using an external device in the way recommended here - see the notes below: one favoured technique possible using the 1.3mm thermistor from a hacked 63-1024 is to fix it to the underside of a socketed CPU then to thread the very thin insulated wires separately between the processor-pins. The wires are so thin the CPU will seat down near as flat as before onto its socket.

Fixing the thermistor so it contacts the backside of the CPU firmly may be done with any tape warranted to withstand the temperatures or with a tiny blob of a suitable silicon sealant - it helps to degrease the CPU with cotton-wool & isopropyl alcohol first. Another technique - shown on the right above - is to cut a small pillow of stiff foam material slightly thicker than, & fitting inside, the socket's cavity. When you push down the CPU & secure it, the bead is pressed up against the CPU by the foam's resilience. Remember - any material used here must be able to withstand the temperatures - up to 65°C or so.

You can just see the very thin wires entering below the lower-left corner of CPU #0 on this mobo. With a matched pair of thermistors to each CPU & fansink, it is possible to get °C/W's from a SMP system.

Obviously, doing this sort of thing to or near delicate parts of your system requires due care: it is the responsibility of anyone trying the procedures described in this article to take all proper precautions against static, mishandling, inappropriate materials etc - having said this, the only CPU I have yet damaged was chipping a corner of a FC-PGA when using the [dodgy] clip on a Molex FC-PGA - but note that threading wires, however thin, beneath a socketed CPU will increase & make asymmetric the necessarily strong clamping force of modern fansinks.

Notes: since the 1.3mm bead-thermistor is so small; it is practical for the suitably skilled to drill a clearance hole into the side & parallel to the base of an heatsink & to push in the thermistor as neighbourly to the CPU case-top as you dare. Less frighteningly, you can push the bead right down between the fins of a heatsink till it reaches the base directly over its contact-point with the CPU core, then hold the thermistor in place [& insulate it from the direct effects of the fan] with a carefully applied blob of a suitable silicon sealant - there are many similar techniques for getting this tiny bead as near the CPU core as is possible or practical with your system.

A nearly ideal solution - though complex & risky if you use fansinks held down by spring-clamps onto fragile FC-PGA & socket A CPU's - is to insert between the fansink & CPU-case-top a shim of low-spreading-resistance material [copper/silver] extending outwards & having the thermistor fixed to an end. This measures exactly what & where we want - & is the way a laboratory measurement would be taken. Experience so far suggests that backside measurements taken as suggested here are quite accurate; a public-spirited fansink designer/manufacturer has volunteered to conduct some detailed analysis of this question & we will report their findings.