Bowmans Mod Shop

My two ocents guys.
I also deal professionaly with thermal resistance, heat transfer and cooling but in a bit of a different way;)

Copper in this case (copper evap block) acts as INSULATOR no matter the thickness and are. Why? because it has a thermal resistance value, Even if you used diamond to make a block off it'd still be a 'barrier' between heat load and coolant.
To work around it you'd have to make direct die contraption and only in this case all the 'cooling energy' could be 100% used.
One may add mass in all the other parts of the block to increase thermal inertia of the thingie and that's desirable, but imho the thinner the plate the better...

Your definition of insulator is misguided. A heat insulating material should have a large delta T between the hot surface and cold surface with a small thickness whereas a heat conducting material should have a small dT beween the hot surface and cold surface.

This image should demonstrate.
http://www.russellpetrie.plus.com/im...empprofile.bmp

It would be like saying copper is an electrical insulator because it has a resistance to electricity, now thats just nonsense. :D

When you improve the heat transfer, T1 and T2 come closer together. I'll leave you guys to figure out how to express it mathematically. I'll start:

T1 - T2 = ????????

Q = 150W

Your goal is to reduce T1 - T2.

If you have noticed that my contributions to this thread have been reduced to monosylavic grunts and repetitions of the much abused term "awesome", I am sorry. The technical jargon has me bewildered. Even then, I have been able to glean a few concepts out of all this.
More pictures please....me likes the pretty pictures! :D

You're not missing much, Dude. They are just figuring out the right answers to the wrong questions. :D

i see the 2 been battleing it out for a while it looks like.i had been busy today with no chance of getting on here to get into the fun.

you guys are getting into some interesting things while i was out...:toast:

by the way guys the base thickness i have found that works the best is .125".i have gone thinner down to .100 but it seams the cpu will heat up too fast.and .250 was too thick,not that it didnt work good.but it is a little slow on the catchup.
so i have stayed with .125 this last one i did is about .150 thick before the start of the fins.the walls of the outer shell is .100 that surounds the fins and the top that connects to the baker block is .125.but i milled a resess in the top so the baker block sit close to the fins.
i though i would giv e you guys the specs so round 2 of thermo dynamics can start...ring ring.....:toast:

buy the way most of my day was working on finishing up a promie or 2....:D

dale yours is on a vacuum right now.i like to pull a good vacuum.let it set overnight and watch for leaks.then i will triple evac it fully tommorow.then i pre charge with r134 for a day or 2.i like to see it is working fine and holding it own,before i move to 404 then its another vacuum and final charge.:)



had to clean up before my ex dropped my chridren off this afternoon.i had the cascade in my living room doing some tube bending and finishing up my roll around cabinet for it.

i have spent some time on the roll around.got remote pressure ports installed so i dont have to get all under there to check pressures.i have a massive electrical box made.relays swtiches and the works.i am not in a hurry so i like details.

:banana: :banana2: :banana4: :banana3: :eleph: :banana3: :banana2: :D

Quote:

---

Originally posted by TheDude
If you have noticed that my contributions to this thread have been reduced to monosylavic grunts and repetitions of the much abused term "awesome", I am sorry. The technical jargon has me bewildered. Even then, I have been able to glean a few concepts out of all this.
More pictures please....me likes the pretty pictures! :D

---

:toast: Yes more pretty pictures !!!!! lol

Quote:

---

Originally posted by Russell_hq
Your definition of insulator is misguided. A heat insulating material should have a large delta T between the hot surface and cold surface with a small thickness whereas a heat conducting material should have a small dT beween the hot surface and cold surface.

This image should demonstrate.
http://www.russellpetrie.plus.com/im...empprofile.bmp

It would be like saying copper is an electrical insulator because it has a resistance to electricity, now thats just nonsense. :D

---

Sorry for not being clear on this one.
I never said coper is an insulator, that's pure bollo$k$:) It is one f the best heat conductors.
Copper 386 W/mK
Silver 430 W/mK
Diamond 2300 W/mK
Thats thermal conductivity desrcibing behiaviour of same thickness and size samples of these materials when temp difference is 1Kelvin between sides.
Don't confuse with coefficient of heat ransfer W/m^2K - quite different thing :)

What I meant was that ANY material inserted between a heat source and a coolant acts like insulator in liquid cooling applications (in high amperage elec cables design there are equations describing certain thickness of isolating material at which it acts as a radiator helping to cool cable down, go any thicker and you toasted, go any thinner and at given amperage you're toasted too).

Making it simple, what's better for cooling a CPU
1. Phase-change cooled heatsink, or
2. Direct-die phase change cooling?

Imho direct-die is a clear winner here since coolant decompreses directly around a core without and 'middle man' to take any cuts (copper/silver/diamond blocks)

I hope this makes it a tad clearer :)

Great work Bowman, ya da man, as they say ;)


P.S. Russell_hq, all you said is 100% correct m8:)

I just realized, that what I wrote above requires one preemptive condition.
Cooling circuit has to be able to remove whatever a core is capable off :)

Copper (or whatever you use) in blocks simply introduces thermal capacity or heat ransfer lag between heat source (core) and heat evacuating media (boiling coolant).

If a system has not sufficient 'pick-up' or response time then more copper is better between core nad coolant.

Gary -> is this the equation you were looking for? Q=k*A*(t1 - t2)/L
where:
Q - the amount of heat flow in Watts
k - conductivity in W/(m.K)
A - area of material in m^2
(t1 - t2) - temp diff between either side of thickness of a material concerned aka delta-t
L - thickness of the material under consideration

It's the most fundamental law of thermodynamics.

Q is the biggest when delta-t is the biggest i.e

Q=const. * (t2 - t1)

where const. =k*A/L - sample's properties.

I just don't get what does it have to do with base plate thickness ;)

It's more suitable to illustrate the need to get your coolant to as low temps as possible :)

Now shift that formula around to where it starts with:

(T1 - T2) =

Then assume that Q is 150W.

In designing the block you are screwing around with k, A, and L, in an effort to reduce (T1 - T2).

To make it more of a challenge A is about 1/4 inch x 1/4 inch.

Dont worry I wasnt confusing the 2, ive been doing this at uni for 5 years. Thermal conductivity can be converted to a heat transfer coefficient simply by dividing the thickness of the material with it, this will give you the heat transfer coefficient for the object in question.

In this application we want the lowest core temperatures and not biggest heat transfer, as Q is constant (as good as) for this application so we must look at other variables i.e. x k and A.

If we do some quick calculation we get the following;

http://www.russellpetrie.plus.com/im...r/thermalK.gif
A pic for all that love pics :D

Q=76 Watts (Barton 3200+)
A=101mm2 (Barton Core)
x=3.175mm (0.125")

http://www.russellpetrie.plus.com/im...sfer/table.gif

Here we can see the effect of different materials and material thickness on the temp difference between the face of the evaporator and the face inside the evaporator. As we can see not much of a gain is made going from copper to silver, and using the 0.125" thickness we only loose 1.16ºC compared to 0.100" thickness but get a better performing block (from Bowmans tests).

Russell_hq,

I actually was able to follow and understand that! (must have been the pic that did it!) :D
Thanks

Quote:

---

In this application we want the lowest core temperatures and not biggest heat transfer, as Q is constant (as good as) for this application so we must look at other variables i.e. x k and A.

---

Exactly my point. The goal is to reduce (T1 - T2) aka dT.

As a practical matter, the way to compare evaporator performance for various block designs is to compare the difference between CPU temperature and SST (saturated suction temperature). This is the evaporator TD. (CPU temp minus SST = Evap TD).

Lower evaporator TD = better evaporator performance.

Quote:

---

Exactly my point. The goal is to reduce (T1 - T2) aka dT.

---

Wait a minute, T1 - T2 is NOT dT. It is TD. The difference in temperature of the two different masses on the opposite sides of a (seperator, thermal conductor, thermal insulator, whichever way you want to look at it).


Gary,

How much work is being done if the TD = 0 ?

Let's say the package is at 275degF, and the block is at 275degF. Is this the situation we are looking for?

Or would the package at -22F and the block at -40F, be more desireable.

The load on the evaporator is:

(capacity of the evaporator [btuh] @ 1degF TD) * (TD).

If the TD = 0, no work is being done.

In order for "heat" to move, there must be a temperature difference. The greater the difference in temperature, the greater the rate of heat transfer.

(condensing unit capacity @ SST) / (evaporator capacity @ 1degF TD) = TD


The greater the amount of work that we want done, we must increase the TD, because we cannot increase the surface area that we are contacting the package.

The only way to decrease the TD is to increase the contact area of the evaporator (we can't do that), or decrease the capacity of our condensing unit (or use and air-conditioning compressor - same thing :p ). Because of the limitation in contact area, the result of decreasing the TD is doing less work which results in the undesireable effect of a warmer core.

You are confusing yourself with the formula. Want to increase TD? Just put a layer of insulation between the evap and the CPU.

Good grief!

I don't want to give up on you, now.

The TD relates to the two surfaces in contact with each other (that is what is represented by T1 and T2), that are exchanging the heat. If you put insulation between the evap and the package, the TD would now be between the block and the insulation.

The net effect of doing this would decrease the TD, because the insulation would lessen the load on the refrigeration system. the surface of the insulation that contacts the evap would be at near the evap temp (presumably). the system would balance due to a drop in the suction pressure/temperature decreasing the volumetric efficiency of the compressor and as such the refrigerating capacity of the system.

Quote:

---

Wait a minute, T1 - T2 is NOT dT. It is TD. The difference in temperature of the two different masses on the opposite sides of a (seperator, thermal conductor, thermal insulator, whichever way you want to look at it).

---

If we are talking about the surface temperatures on opposite sides of a single mass (evap base) it is a delta-T. If we are talking about comparing the temperatures of two different masses (CPU and SST) then it is a TD.

Quote:

---

Originally posted by Gary Lloyd
If we are talking about the surface temperatures on opposite sides of a single mass (evap base) it is a delta-T. If we are talking about comparing the temperatures of two different masses (CPU and SST) then it is a TD.

---

Then, TD it is.

Okay, the TD we are concerned with is difference between CPU temp and SST. The best block design will give us the lowest TD.

I hope this is making sense to you. The formula still applies. When we are designing an evaporator, we are working with the type of material, its thickness, and its area. Q is held constant. When we change the material, thickness, or area, what changes in order to balance the equation is (T1 - T2).

You are all right here but as usual we are talking about different things thinking that we do about one.

Herefishy -> if dT equals zero we have a situation where our heat pump (air cooling, thermoelectric and phase change are all heat pump principle applications) does one of two things
1. It’s off :) :)
2. It’s working at it’s maximum capacity and removes max amount of heat possible (that’s the case with pelts at least). Imagine a phase change system being capable of removing only 100watts of heat. Our cpu produces 100watts of heat and asa result we have a cpu running at ambient temps. Now imagine a cpu giving away 120watts of heat with the same system hooked up. As you probably catching my drift by now it will be hotter than ambient by these 20 watts of heat which our pathetic phase change is not able to remove. DT is still equals 0 since it’s max work but it’s not nearly enough to cool our 120watt core.
Now imagine the same 100 core but a new system which is capable of removing 300watts of heat.
As a heat pump is not doing it’s max rated work since it’s not pumping 300 watts of heat but only 120w.
The rest is being converted (since 180watts of coolers capacity cannot just disappear) into energy deficit which equals below ambient temps.
3. If my line of thinking is incorrect, please do correct me. I’m always hungry for knowledge and since I am not a thermodynamics guru like you guys are, and only quite good at physics and maths in general I will accept any corrections given them being logical ;)

Russell_hq and Gary -> Yes, you are right, I thought about more global dT, not ‘local’ and what you are saying is 100% right. Let me just entertain Russell’s equations a bit further and hopefully you’ll see my point.
As you proved yourself by lowering the thickness of a base plate we ARE getting what we want, lower ‘local’ dT I am using terminology of local and global as not to confuse temp difference between coolant and core which the bigger the better and local for temp diff b’tween ‘hot’ and ‘cold’ sides of base plate.
Now, imagine reducing thickness to zero.. :) you’ve got yourself the smallest ‘local’ dT which is 0 :)
My point was from the very beginning that no matter how good heat conductor the material of base plate is going to be it WILL always have a thermal resistance value and this makes it an insulator.
Now, thermal inertia or thermal capacity which is ( or is it not) desirable here is what we are discussing.
I purposefully omit structural problems of to thin a base plate here.
I always thought that phase change systems are the most efficient cooling solutions available to overclocking community (I am not talking LN2 but 24/7 app) and the response time to increased load is so fantastic, that there’s no need for introduction of thermal inertia into the system (again, no structural/design/practicality issues are being considered here)

What do you think?

:)

Quote:

---

Russell_hq and Gary -> Yes, you are right, I thought about more global dT, not ‘local’ and what you are saying is 100% right. Let me just entertain Russell’s equations a bit further and hopefully you’ll see my point.
As you proved yourself by lowering the thickness of a base plate we ARE getting what we want, lower ‘local’ dT I am using terminology of local and global as not to confuse temp difference between coolant and core which the bigger the better and local for temp diff b’tween ‘hot’ and ‘cold’ sides of base plate.
Now, imagine reducing thickness to zero.. you’ve got yourself the smallest ‘local’ dT which is 0

---

Bigger is not better between coolant and core. The TD is just an extension of that "local" dT. If you lower the coolant temp, the core temp will follow, maintaining the TD, because the TD is a function of the evap design. If you lower the thermal resistance between the core and the coolant, the two temperatures move closer to each other (lower TD).

Quote:

---

Originally posted by Gary Lloyd
Bigger is not better between coolant and core. The TD is just an extension of that "local" dT. If you lower the coolant temp, the core temp will follow, maintaining the TD, because the TD is a function of the evap design.

---

He he he , that's what I meant.
I was just thinking about static state - pure theory.
In a running, dynamic system the lower TD the better.

What about dd cooling tou?