[dfox]

With my Opty 165 1.47v at 2.8ghz and Orthos errors. So it looks like I need more volts, but Core Temp shows Core0 at load about 45-48c Core1 53c-56c

Sheesh doesn't seem like my Tuniq Tower 120 is cooling all that great, I mean if I'm going to push more volts and go for 3.0ghz and I'm hitting already 56c.

I lapped this darn thing. When I was dragging it across the sandpaper on the counter top I noticed one section that was off a bit, and for the life of me I couldn't get it even dragging it, so I held it in my hand and sanded it lightly to try and get that out.

HELL I bet I got the lapp messed up, CRAP. How can I tell if it's completly flat?

I don't have a micrometer to check the bottom for flatness would this be the only way for sure, or pull the HSF and CPU and clean them off and lay the CPU on the HSF and see how flat it lays?

How Can I Check If HS Lapp Is Flat/OK?

P.S. I didn't sand it that hard or with a course grit, I did it very light in circles while holding it in my hand trying to even out this section, and I can't imagine as hard as the metal is, I made it uneven at all.

[thideras]

Originally Posted by DasFox

With my Opty 165 1.47v at 2.8ghz and Orthos errors. So it looks like I need more volts, but Core Temp shows Core0 at load about 45-48c Core1 53c-56c

Sheesh doesn't seem like my Tuniq Tower 120 is cooling all that great, I mean if I'm going to push more volts and go for 3.0ghz and I'm hitting already 56c.

I lapped this darn thing. When I was dragging it across the sandpaper on the counter top I noticed one section that was off a bit, and for the life of me I couldn't get it even dragging it, so I held it in my hand and sanded it lightly to try and get that out.

HELL I bet I got the lapp messed up, CRAP. How can I tell if it's completly flat?

I don't have a micrometer to check the bottom for flatness would this be the only way for sure, or pull the HSF and CPU and clean them off and lay the CPU on the HSF and see how flat it lays?

How Can I Check If HS Lapp Is Flat/OK?

P.S. I didn't sand it that hard or with a course grit, I did it very light in circles while holding it in my hand trying to even out this section, and I can't imagine as hard as the metal is, I made it uneven at all.

Use a razor blade.

[dfox]

Now I tried putting the vcore up to 1.49 and Orthos still gives me an error, BUT what I didn't like was Core1 was running 58c-61c

That doesn't seem like Tuniq Tower 120 is cooling all that great OR maybe I tightened it down to much and mashed the grease out, LOL

[El<(')>Maxi]

By looking at it, thats the easiest way. If you want more a more accurate method mask the substrate and spray a little black paint on the IHS, then sand it very lightly for a moment and look for high/low spots. Hopefully it will be nice & flat. Also remember that the IHS has two sides we care about, the side you can see and the side you can't

[darkcow]

ketchup and a peice of flat see through glass. dark spots = bad.

[Mycobacteria]

SEM (scanning electron microscopy) a scratch on your HSF looks like the grand canyon.

An uneven surface starts to look like the surface of the moon.

I have not yet seen a hand lapped HSF that looks nice under SEM.

When we want our metals to be cut flush to the micron we usually get a argon/krypton lasor.

[twitchee2]

before you lapp it, draw on the bottom with sharpy, make sure your sanding block (w/e that may be, i use a piece of granite). Then when your sand if the middle stays colored then its concave, it the outside stays colored its convex, either way you need to sand it so that it is flat. It it looks fairly even, then its flat. hope this helps.

[nd4spdbh2]

Originally Posted by twitchee2

before you lapp it, draw on the bottom with sharpy, make sure your sanding block (w/e that may be, i use a piece of granite). Then when your sand if the middle stays colored then its concave, it the outside stays colored its convex, either way you need to sand it so that it is flat. It it looks fairly even, then its flat. hope this helps.

GOSH who did you learn this technique from... o wait i taught you.... N e ways this is what i do when i lapp my hs's... Draw a grid or color the whole damn thing... heck even a smilie face works just make sure you have ink over the whole surface of the base of the HS... make sure the piece of material that your sanding against is FLAT... granite works, a marble counter top is damn good, and i have heard that glass is quite good too... then go at it...

ALSO if you have a nice mirror finish on it you can always use the line test by reflecting a line off a paper on the hs and see if its straight or not.

[enz660]

if you even tried a lap job, im thinking it would be very difficult to make a finish thats worse than the stock one (if you put SOME effort in). If you lap both the heatsink base and the IHS and STILL get crappy temps, its probably not the lap job. Use less thermal paste and apply more pressure when you mount it. You dont want a layer of TIM between the IHS and HSF. Ambient temps? Case airflow? What TYPE if TIM? etc etc... many more things to consider. Good luck.

[Immortal_Hero]

I would be willing to make a bet that it is more the CPU than the cooler. My Opty 165 was very concave. It was hitting low 50's at stock speeds. I went at the IHS with 220 grit and I bet I removed .0010 to .0015 of an inch from the edges before it was flat. Check your CPU with a razor blade or ketchup as I think AMD has some bad quality control. This may be on purpose to keep less that savvy OCers from buying cheaper chips.