Bowmans Mod Shop

Quote:

---

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

---

I've come to that conclusion - I am not talking about the theoretical.

Quote:

---

... if dT equals zero ....

---

I am not at all concerned with the dT of any (single) substance in this matter. dT implies that a material (not two different sides of it!) has changed temperature.

Quote:

---

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

---

No, that is not a proper observation. What is important is the evaporating temperature that the refrigeration system is rated perform the 100 watts of work. Typically a compressor rated at 100 watts, R-404A @ -60F evaporating temperature, will perform perhaps 200 watts of work at -10F, and perhaps 300 watts of work at +20F. This is due to the increase in density of the return gas (resulting from the higher pressure) which increases the volumetric efficiency of the compression cycle and allows the compressor to compress more (pounds) of gas per unit time. This is what is referred to as "The rule of thirds" in my industry, pertining to guestimation of compressor performance.

If you exceed the design load of a system, the system just does not become "static". The system balances. As the load increases the capacity of the system increases accordingly, hopefully not in excess of the maximum rated capacity of the individual components.

Quote:

---

DT is still equals 0 since it’s max work but it’s not nearly enough to cool our 120watt core.

---

errrrrr.... wrong. Work is still being done. At even a greater rate than the design (lower) capacity.

Quote:

---

Now imagine the same 100 core but a new system which is capable of removing 300watts of heat.

---

Okay, I'm imagining.... Now, since the load is light, the suction pressure/evaporating temperature is going to drop (colder) which results in less dense gas returning to the compressor which in turn decreases the efficiency of the compression cycle. As such the system balances at a lower temperature to perform the 100 watts of work. Oh, and the TD will increase, being a factor in the system balance equation.

Quote:

---

If my line of thinking is incorrect, please do correct me.

---

I think I just did....

You are a breath of fresh air on this forum, Jabo.

:toast:

Quote:

---

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.

---

The "sides" of the plate are NOT changing temperature. The sides are "different sides, NOT one side changing temperature (dT). It is TD they are different sides. if one side went from 0degF to 10degF, the dT would be "10". If one side is 0, and theother side 10, the TD is 10 and there is no dT, because one side is 0, the other 10, and they still are for the sake of calculation.

Quote:

---

Originally posted by Gary Lloyd
If you lower the coolant temp, the core temp will follow, maintaining the TD....

---

Wrong.

If you lower the coolant temp assuming that the load at the presumed (higher temp) remains the same, you must conclude that the capacity of the refrigeration system has incresed, right? Let's say this is a result of a lower abient cooling our condenser, lower the head pressure and resulting in a higher vloumetric efficiency of the compression cycle.

Since we are (presuming) that we have a higher capacity, and are doing more work, we must realize that this means that there is a greater amount of heat transferring between the contact surfaces of the block and the package. Since we have not affected this contact area, we can anly assume that the TD has increased as a result of the observed greater amount of work.


(condenser capacity [btuh] @ SST) / (evap capacity @ 1degF TD) = TD

That's a massive amount of info guys, thx !:toast: :banana:

As you can sefrmo my sig I am more of tec blocke.
That's why i treated phase change as static heat pump
:D :slapass: - deserve one ;)

Jokes aside, WILL SOMEBODY ANSWER MY DIRECT DIE COOLING QUESTION???
You are useless!! ;) :D :D :D :D

And what about thermal inertia/capaity and base plate thickness equation theory?

Quote:

---

Since we are (presuming) that we have a higher capacity, and are doing more work, we must realize that this means that there is a greater amount of heat transferring between the contact surfaces of the block and the package. Since we have not affected this contact area, we can anly assume that the TD has increased as a result of the observed greater amount of work.

---

Wrong.

The CPU load remains the same, the SST drops reducing compressor capacity to match the load, and TD remains the same, because the CPU temp has dropped along with the SST.

You keep wanting to increase the heat output of the CPU. It ain't happening.

Quote:

---

Originally posted by Jabo
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.

---

I gotta disagree with this one, you cannot call it an insulator just because it resists heat. All materials resist heat, even diamond, using your logic this would make them all heat insulators. Now if we apply your logic to electronics, copper has an electrical resistance, using your logic this would make it an electrical insulator. As i said before this is nonsense.

I see what you are saying though. If we take away the base and boil the refrigerant on the cpu core you think that we would get better performance. Beside the point that this wouldn't be feasible (pressures involved) you would only gain 5-6ºC lower temps on the core, the core temperature would of course be higher than this. Also you must note that the core has a small surface area, 101mm2, trying to remove 76watts would give an energy density of 752 kW/m2. Taking into account the poor film heat transfer coefficients( the amount of energy that can be transfered per unit area and unit temp diff between the surface of the core and the boiling refrigerant) of a boiling refrigerant then this may actually reduce performance. This is one reason that we increase the surface areas inside the evap because there is a big resistance between the evap surface and boiling liquid. Hope this makes some sense. I can draw some pics if you want :D :D :D

Quote:

---

Originally posted by Gary Lloyd
Wrong.

The CPU load remains the same, the SST drops reducing compressor capacity to match the load, and TD remains the same, because the CPU temp has dropped along with the SST.

---

Okay, I will regress, purely from a theoretical standpoint. :)

However, in reality I foresee an increase in the non-useful loads and transmissions that may be conducted through the heat exchange surface by way of heat transmission through the card and a foreseen increase in the TD between the card, ambient, mobo, etc.

Admittedly, these effects may not be that significant in a processor cooling application. but then again, the load isn't that significant... it's all relative.

Ahhhh... but we are discussing evaporator design. To compare evaps, we hold Q constant, we hold T2 constant, and we improve heat transfer (material, area, thickness), which drops T1 and thus TD. :D

Screw the formula, lets see how Bowmans performance on that evap turns out.

Quote:

---

Originally posted by Russell_hq
I gotta disagree with this one, you cannot call it an insulator just because it resists heat. All materials resist heat, even diamond, using your logic this would make them all heat insulators. Now if we apply your logic to electronics, copper has an electrical resistance, using your logic this would make it an electrical insulator. As i said before this is nonsense.

---

It all depends on from where are you looking at it. If you compare direct die (I am taking aboard your following comments and are inclined to agree with ya on that one) and block applications then copper ACTS (is not but behaves like one) like insulator.
With electricity is the same story. Drop teps to absolute zero and you have no rsistance to current pasage. Compared to this state every electrical conductor is electrical insulator - it's all relative :)

Quote:

---

I can draw some pics if you want :D :D :D

---

Nice fit ladies? Anytime m8! :slobber: :smileysex :D

About contact are, it's exactly the same between copper base plate and core as it is between boiling liquid and core ;)
The only thing is that copper is so much denser than boiling liquid that ieven thou there's a thermal resistance of copper as a middle man it still performs better than direct die. Agreed to that.

Idea:
Waht about using some sort of liquid coolant (like mercury for example, should be dense enough) to direct-die cool our cpu and making very short loop of it cooled by phase change 'double sided evap (finned area on both sides).
I should imagine that sth like mercury has incomparably bigger heat ransfer and storage capacity then any liquid coolant on the market. Are there any other nasty things like this to replace mercury, which quite unpleasant to work with?

Quote:

---

we hold Q constant, we hold T2 constant, and we improve heat transfer (material, area, thickness), which drops T1 and thus TD

---

Well, actually T1 would increase... because the gas in the evaporator is being heated more, which in effect would decrease the TD as you say. :D

The TD is the result of the efficiency of the (evaporator) to tranfer heat, relative to the capacity of the condensing unit.



(condenser capacity [btuh] @ SST) / (evap capacity @ 1degF TD) = TD



Let's say we have a condensing unit rated @ 300 btuh @ whatever SST, and an evaporator rated @ 30btuh/degF / degFTD


300btuh / (30 btuh / 1degF TD) = 10 degF TD

10x / x = 10


If we increase the efficiency of the evaporator, to say 40btuh/1degFTD :

300 btuh / (40btuh/degFTD) = 7.5degFTD

7.5x / x = 7.5

Of course, if the TD decreases, this results in a higher evaporating temperature, and as such a higher condensing unit capacity. As such, when doing a design/spec, I always start at the desired evaporating temperature to establish my condensing unit capacity. The I select an evaporator which balances the TD at a proper point for (desired results).

I'm done.

Quote:

---

Originally posted by herefishy
I'm done.

---

Another one achieved Nirvana ;) :D

First of all, I'd like to say, on behalf of all the not-so-technical ppl on this forum, I think my head just exploded. :frag:

Quote:

---

Originally posted by Jabo
About contact are, it's exactly the same between copper base plate and core as it is between boiling liquid and core ;)
The only thing is that copper is so much denser than boiling liquid that ieven thou there's a thermal resistance of copper as a middle man it still performs better than direct die. Agreed to that.

---

Except that the block either contains fins, ala the prommy, or forces the refridgerant thru many channels, ala Baker's blocks. All increase the surface area. As Russel already explained much more scientifically than I could, Direct Die just doesnt have the surface area to compete with a block design.

Quote:

---

Originally posted by Jabo
Idea:
Waht about using some sort of liquid coolant (like mercury for example, should be dense enough) to direct-die cool our cpu and making very short loop of it cooled by phase change 'double sided evap (finned area on both sides).
I should imagine that sth like mercury has incomparably bigger heat ransfer and storage capacity then any liquid coolant on the market. Are there any other nasty things like this to replace mercury, which quite unpleasant to work with?

---

That, sir, gives me a hard on.

While mercury's been discussed for liquid cooling applications (and summarily abandoned as impractical), that may be possible with phase change. Then again, it may just be too complicated and/or difficult to implement. But I'll leave that to those smarter than me.

Quote:

---

Except that the block either contains fins, ala the prommy, or forces the refridgerant thru many channels, ala Baker's blocks. All increase the surface area. As Russel already explained much more scientifically than I could, Direct Die just doesnt have the surface area to compete with a block design.

---

Yep, it's not just a simple matter of material and thickness. We can increase area by spreading the heat. We ideally end up with just the right trade-off between extended surface area and thermal resistance of the added material. I still like my idea of a spiked, drilled, or channeled hemisphere, spreading heat in all directions. I'm waiting for someone to try it. :D

Well, since that is over... peculiar that Bowman didn't get involved in the discussion, since it was his post regarding his theory of increasing block mass that started the whole dern thing! :confused:

Maybe he's just standing by, waiting for the soapbox to be vacated... :D

Quote:

---

Originally posted by Gary Lloyd
Yep, it's not just a simple matter of material and thickness. We can increase area by spreading the heat. We ideally end up with just the right trade-off between extended surface area and thermal resistance of the added material. I still like my idea of a spiked, drilled, or channeled hemisphere, spreading heat in all directions. I'm waiting for someone to try it. :D

---

Wow, you have just stolen my idea, mate ;) :)

:toast:

I wanted to try sth like that but didn't have tools... even both special copper block, but... al my drill bits broke after a while, even with machine oil used as coolant.
I am shaping up to use running water as a coolant for my drills (both some special titanium ones, next is only diamond tipped extravaganza...) and I haven't got anuthing to cut through 2.5"x2.5" of copper... :(

DarKcyde -> I am glad I could be of service, dude :cord: :cool:

Quote:

---

Originally posted by Gary Lloyd
Yep, it's not just a simple matter of material and thickness. We can increase area by spreading the heat. We ideally end up with just the right trade-off between extended surface area and thermal resistance of the added material. I still like my idea of a spiked, drilled, or channeled hemisphere, spreading heat in all directions. I'm waiting for someone to try it. :D

---

give me time..your design has interested me for a while.i will get one done to test one day.

Quote:

---

Well, since that is over... peculiar that Bowman didn't get involved in the discussion, since it was his post regarding his theory of increasing block mass that started the whole dern thing!

---

herefishy

not to say i would have loved to jump in,but i do have to work for a living...first rule of thumb....family first,job second,toys and hobbies last.
and i am very busy at home and work right now,i just stepped back in from visting some of my customers today,one a milliion dollar account of mine...(ryder truck rental in case you have heard of them)sorta got my attention.

but everyone was thowing ideas around.and i learned along time ago....if you see a crowd of people in a discussion or argument...just step back ,listen,and wait for it to be over.
i am old enough to know that .
but in this instance i have been away.and i am going back out until late tonight right now.

Yeah.... go ahead.

Wine and Dine.

Just can't seem to find good help anymore.

well yes i try and treat them like a customer should be.
and actually i am planning a deep sea fishing trip to take some of there men out for a good time.


but back on track:toast:

as this thread was started over everyone wanting to know what i am working on or creating.
it is time for some more pics that everyone likes.

shortly tonight i will be back with a prommie that belongs to a dear friend,.:D

and something i put together for a custom job for a man.i think it will interest some of the guys.it is small so it will go into a custom case but i hope large enought to keep his a64 AMD chip cold enough,only one i am doing since i wont be lucky enough to find another condensor like it.

Quote:

---

Yep, it's not just a simple matter of material and thickness. We can increase area by spreading the heat. We ideally end up with just the right trade-off between extended surface area and thermal resistance of the added material. I still like my idea of a spiked, drilled, or channeled hemisphere, spreading heat in all directions. I'm waiting for someone to try it.

---

I think you pretty much nailed it there.
I think my evaps getting close to your hemisphere idea... It uses tall pins over the die area decending into the suction line (only to the top of the block though the top plate) which is right over the die. The freon is injected as close to evenly as possible but will probly end up going mainly through the middle fins... either way all freon passes over the die going through outer areas first.
From the pic... the blocks sorta leveled in how i'm making it, bottom is the base + fins, second left goes over the fins to distribute the freon from the cap tube to all sides
top goes around the bottom two to close off the sides, a jacket sorta thing
right side piece is the top.
base plate thickness of 2mm , it doesn't need to be huge since the pins are right on top of it.
(sorry for the highjack but it might be good for discussion :explode: )

well here is the pic....
prommie on bottom(one with the first barker block installed.
and on top is the custom job.sitting there now in a vacuum.this unit was made to me in the upper case like a vapo.but much more powerfull.

well i know alot keep asking me for pics,so here you go.i should have a pile by the weekend.i will have 4 prommies in here at once.
anyways tommorow guys

Wow...Bowman, you really have made some changes since the very first prom you did. Can't wait to find out more about that top one. :D

ok i gata question for u bowman. say i bought a prommi and sent it to u.
1st question....would it be possible to get better temps with out changing the freon? It comes stock with r134 right. Now i keep hearing u talking about cap tubing and the lengths and such...
if its possible to just change the cap tubbing length to get better temps then how come chip-con doesnt realize this? i might be way off but hey its worth a try right:D i hope im not makin a fool of myself!
also if the ? is unclear i would clear it up need be

Quote:

---

Originally posted by Pimpsho
ok i gata question for u bowman. say i bought a prommi and sent it to u.
1st question....would it be possible to get better temps with out changing the freon? It comes stock with r134 right. Now i keep hearing u talking about cap tubing and the lengths and such...
if its possible to just change the cap tubbing length to get better temps then how come chip-con doesnt realize this? i might be way off but hey its worth a try right:D i hope im not makin a fool of myself!
also if the ? is unclear i would clear it up need be

---

ok well first..i never discuss who or what i am doing service for.only ask me personal questions in a pm.and i will try to anser them as quickly as i can.nothing personal,it is i like to keep the thread as just about what i am modding..and not so much who or if i charge anything or not.this is about me showing what i have done or doing,some wrong some right.but it is to help everyone who has every wanted to try or show pics of differant ways i do things sometimes.to help the most people i can at one time is the main goal.and you have a pm


second..i have gotten a couple proms about 3c colder with just a cap tube adjustment.but it isnt worth the time to mod for just that.but the only reason i can do that is....chip con or what ever the new name is...has to build theres to handle a mulitple selection of cpu mounts and heat loads,so you lose a small amount of effeciency by doing so.as i work it to a specific heat load.if the show was on the other foot..i proberly wuld do the same as they have on that.

gotcha:D well keep up the good work! and keep the pics comin;) what about that cascade big guy? hows that comin allong? this weekend if i remeber right ur supposed to be finished?