Bowing the D-Tek Fuzion after lapping

[Exahertz]

So ever since i have had my rig i was doing direct die cooling to a dual core E2160, thus i needed to lap my Fuzion block parallel and shinny :p:

however i just upgraded to a q9650 with the soldered IHS. I am not going to remove this IHS and i must "re-bow" my fuzion block.

Im getting the quad core accelerator nozzel kit, so i already have an opportunity to take the block apart.

Is it possible to re-bow the fuzion?

and if so, what materials are required?

How much of a bow would be optimum for the q9650?

Thanks! ;)

[nikhsub1]

How did you lap the base? If you lapped it unassembled then the block was still bowed.. If you put an EK barb o-ring in between the top and mid plate, this will bow the Fuzion (V1 fuzion, not V2). Next, do NOT use the quad nozzle, proven to work worse than no nozzle. I would use the 5.5 or 6.5mm nozzle depending on the rest of your system.

[Tsaroth]

Not to be off topic but you said the quad nozzle works worse than no nozzle?

I am running a Q6600 (which is not his chipset I know) and at idle running 3600 on 1.272 i'm getting 26-27-30-30. Load is about 42-42-47-47.

Are you telling me without the nozzle I'd be getting better temps?

I know that the quad nozzle is restrictive so if you were to run a full cover GPU block along with it, it would be better to not have that in there and just run it with no nozzle.

What am I not seeing about the Quad Nozzle?


How did you lap the base? If you lapped it unassembled then the block was still bowed.. If you put an EK barb o-ring in between the top and mid plate, this will bow the Fuzion (V1 fuzion, not V2). Next, do NOT use the quad nozzle, proven to work worse than no nozzle. I would use the 5.5 or 6.5mm nozzle depending on the rest of your system.

[Exahertz]

How did you lap the base? If you lapped it unassembled then the block was still bowed.. If you put an EK barb o-ring in between the top and mid plate, this will bow the Fuzion (V1 fuzion, not V2). Next, do NOT use the quad nozzle, proven to work worse than no nozzle. I would use the 5.5 or 6.5mm nozzle depending on the rest of your system.

The whole block was lapped while assembled. and im almost sure its a V1.

So you're saying that a simple buna or viton o-ring will bow the fuzion?

[Eddie3dfx]

No reason to touch the bases, outside of cleaning them with alcohol. If you lap it, you will never get it back to where it originally was from the factory.
That is unless you have the equipment to do proper lapping.

[Exahertz]

darn... I knew i shouldn't have messed with it, i just had to DD cool my cpu...
anyways

Eddie, are you implying that the bases are not "bowed" but are more or less lapped so the material is thicker in the middle and thinner on the edges (like a plano convex geometry)???

[Eddie3dfx]

darn... I knew i shouldn't have messed with it, i just had to DD cool my cpu...
anyways

Eddie, are you implying that the bases are not "bowed" but are more like lapped so the material is thicker in the middle and thinner on the edges (like a plano convex geometry)???

What I'm saying is that should lap your cpu (but I'm not sure about direct die)
The only reason they bow the blocks is because of the convex/concave shape of the cpus. Once the cpu is lapped, there is nothing that can explain why a bowed block would perform better on a flat surface.

I read your initial question/info wrong. I am pretty sure you can rebow it, since it's the rubber that creates the bow. I doubt the sanding took off that much copper where it's impossible to get the same bowed effect.
So sure, try it!

[NaeKuh]

Not to be off topic but you said the quad nozzle works worse than no nozzle?

I am running a Q6600 (which is not his chipset I know) and at idle running 3600 on 1.272 i'm getting 26-27-30-30. Load is about 42-42-47-47.

Are you telling me without the nozzle I'd be getting better temps?

I know that the quad nozzle is restrictive so if you were to run a full cover GPU block along with it, it would be better to not have that in there and just run it with no nozzle.

What am I not seeing about the Quad Nozzle?

scotts comment only applies to a 45nm quad. And not the old kentsfield.

[Henry8601]

When the o-ring is inserted to bow the block, how come the copper block bows and not the delrin top??

[Sadasius]

When the o-ring is inserted to bow the block, how come the copper block bows and not the delrin top??

You know what? I asked myself the very same question and honestly don't know. I cannot tell which part is giving in here and I actually don't think any part is bowing and that the ring is just getting squished more then anything. But it does keep the water from entering that area again and that's why I used it because there used to be a gap.

[mozis]

When the o-ring is inserted to bow the block, how come the copper block bows and not the delrin top??
The copper is thin at the bottom and top is (relatively) massive and arch-shaped, even if its made of delrin. Delrin wins :up:

[Sadasius]

The copper is thin at the bottom and top is (relatively) massive and arch-shaped, even if its made of delrin. Delrin wins :up:

Well the funny thing is nothing is pushing down on the center of the copper bottom. If it did would the pins bend? Just seems that the insert in the middle of the block would bend first since it is pretty thin itself. I tried to see a difference and could not see it.

[Tsaroth]

Thanks for the clarification Nae,

Sorry Scott............ :yepp:


scotts comment only applies to a 45nm quad. And not the old kentsfield.

[mozis]

Well the funny thing is nothing is pushing down on the center of the copper bottom. If it did would the pins bend? Just seems that the insert in the middle of the block would bend first since it is pretty thin itself. I tried to see a difference and could not see it.
I must admit i dont own a fusion myself so cant really tell but i think you can make the base bow a few microns before the pins start bending.

[lucas81]

Sorry for bringing this old thread from the deaths, but hypothetically speaking, wouldn't lapping the IHS give the same results (or probably better since the water-block was lapped assembled) compared to re-bowing the water-block?

I'm trying to understand the whole lapping and bowing thing. I remember reading somewhere here on xs that;

Lapped CPU + Lapped water-block = Good
Lapped CPU + Unlapped water-block = Ok
Unlapped CPU + Lapped water-block = not as good performance

Is this tested and does it still holds?

[SNiiPE_DoGG]

the thing with bowing is that it gives higher center pressure over the cores whether the CPU is lapped or not, this not only gives better contact but also thins the thermal paste at the center better resulting in an all around superior mount.

the problem with a lapped cpu and lapped waterblock is that yes they are flat, but due to the thinness of the waterblock base and the larger size of the waterblock with respect to the CPU, when applying pressure to the joint of a lapped-lapped configuration the waterblock base itself becomes slightly concave resulting in worse contact.

sorry if that is too convoluted to understand....

[lucas81]

Thanks for clarifying.....I do understand the explanation perfectly. The thing is there's just a lot of info here on the forums, which is simply not true. Unless someone would delete all the posts with all that false info, it takes some time for someone to filter out what is true and not.