building a CPU die simulator
 

*edit* i JUST realized i posted this in the wrong forum. Can a mod please move it to the proper one? thanks.

Hey guys, I'm looking into building one of these blocks. Basically its for testing out waterblocks. I need some help. I have no engineering skill whatsoever, and i don't know what kind of material to use. I've seen many simulator designs, but i was looking for a simpler one? I know billA and bigben2k have talked alot about this, but i find it very confusing >.>'

Anyways, as most of you already know, i have a CNC machine at my disposal, so i don't mind making my own block. One thing that would help are cpu IHS dimensions. This has been such a pain to find. They're listed practically nowhere. I recall the AMD IHS being 38x38mm, but that's all i found out. I'm not even sure about the accuracy of that value.

The heaters are a problem to me, as i do not know which to buy and where to buy them from. However, i already have temperature sensors. I just need suggestions as where to place them.

Is it possible for someone to do a quick drawing of a cpu die simulator? Just a simple expanded sandwich diagram would be more than enough, and greatly appreciated.

Is there a special method for calculating heat dissipation? or is it simply temperature input compared with temperature measured at the waterblock?

Its all for the good of the forum, i'm testing out every design someone sends me, and all results will be published. Hopefully we'll be able to learn some new things from this.

*edit*anyways.. did a bit of reading after i discovered the waterblock testing forums -_-'. I see how alot of people are calculating the thermal capacity of the waterblock and whatnot. I was actually wondering if i could just calculate what the resultant temperature is. Wouldn't that be much cheaper/simpler? I'm trying to go for pretty much minimum cost... and since i don't have a system to mount this all onto, i'm forced to use a die simulator. The things i'm looking for are basically just... 1. measures final temperature (only need quantitative measurements, i'm going to directly compare other waterblock designs on the same rig.) 2. simplicity. (wanna get it up and going fast as possible. need to be ready by the time the waterblock is.) 3. cost effectiveness. (really need to save on costs. I mean the only thing the water cooling system is going to cost me is a water pump and tubing, and copper blocks. I'm machining everything to save on costs >.>') Accuracy really doesn't matter that much to me. a 2-3*C accuracy would be ok already. I can always upgrade it later =p.

On a seperate note, i've gotten a new heatblock and i have no idea where it came from. I know its a car, but that's it. I can't even tell if its single or double pass. It has no visible tubes. Its all copper, and i'm currently cleaning it with vinegar, then i'm going to get barbs welded on. Pumps and tubes will have to be purchased after.

Re: building a CPU die simulator
 

Moved here.


We have been over and over this and pretty much anything you do will be wrong. :hammer:

However I think Inco was on the best track. http://forums.procooling.com/vbb/showthread.php?t=12247


Good luck.

Re: building a CPU die simulator
 

Thanks Jaydee,

what do you mean anything i do will be wrong? 0.o there are no accurate ways to measure the effectiveness of a waterblock?

Re: building a CPU die simulator
 

Read
http://www.procooling.com/index.php?...icles&disp=140

and then

http://overclockers.com/articles1310/

Deep subject around here. Also search Intel TTV on the forums here.

Re: building a CPU die simulator
 

thanks for those two links, very informative and thought provoking.

It seems the most important question to ask is what are you trying to measure and WHY?

From that point some answers will start to flow.

If you want a test rig for experimenting with various WB designs for your own CPU without having to rebuild the PC each time , a piece of 10mm Cu plate with a raised area similar to the IHS and a metal clad resistor bolted on the back fed by a constant voltage PSU should give you a quantitative test rig.

Choose resistor values that will give you something near the max power you are hoping to poke into your CPU once it is water-cooled.

This will allow you to preselect the best of your WB prototype designs.

Though this will be a very crude sim it will tell you quantitativly if you are in the right ball-park with your cooling system or whether it's not even worth pulling your PC apart because you're wide of the mark.

Of cource if that was not your "WHY" you will need to think through your own what and why logic.

HTH 8)