Experimental Water Pump?

[Haggis]

I just found this link at silentpcreview. A new waterpump experiment.
http://www.newscientist.com/news/news.jsp?id=ns99994086

[joemac]

Electro-osmosis wow what a concept. From the description of the pump “consists of a disc of glass two millimetres thick and five centimetres in diameter” This thing sounds cheap to make (materials any way). Cooligy's needs to price these the same or very close to current pump prices to make this mass market product (my opinion). Reading Paul Lee comment from QuietPC makes me wonder if he has ever compared a water-cooled rig to a fan heat sink one. Seems to me Paul thinks {water + PC = Fried computer user} wait We are still here..

[jaydee]

Thnis just doesn't make sence:

Goodson's experiments have produced a flow rate of 200 millilitres per minute. Keane says this would be enough to cool chips that radiate 120 watts of heat per square centimetre, with hotspots of up to 500 watts. In comparison, Intel's Centrino chip dissipates 35 watts.

200 milliliter = 0.0528344 gallon [US, liquid] That's 3.170064 gallons per hour!

And what is this Centrino BS? That is a mobile P4 is it not? 35watts? The regular P4 is like 80watts. I don't understand how they think they can cool a CPU with 3GPH. And there theory about how heat doubles every 18months with CPU's is very incorrect. The T-Bird 1400 (1400mhz) runs much hotter than a XP1700+ (1470mhz). If that was true the Barton 3200 would not be coolable without super water cooling.

Nothing adds up in that article...

[Haggis]

I agree totally Jaydee. I Just thought it would be an interesting post.
I don't know if area = flow rate with this design. If that was the case then a 5meter disc is needed for 300gph. Maybe it could also be used as a skylight

[myv65]

Hey JD,

Not everyone in the world overclocks or cares about getting low differentials from the chip to the water. Three gph is way more than enough to keep a stock chip stable. Given water's heat capacity and say, 75 watts, 3 gpm results in a water temperature rise of about 8°C. Take a block that has surface area more on par with a normal air-cooled heatsink and you would do just fine.

Take a gander at the HSFs packaged with CPUs. These typically let chips run much warmer than the DIY crowd would like, but obviously they do work. This is little different.

For the record, the Centrino is not the mobile P4. It is more akin to a mobile P3. There is indeed a mobile P4, but it is the Centrino that runs much like XPs, ie lots of work per clock cycle.

[jaydee]

Originally posted by myv65
Hey JD,

Not everyone in the world overclocks or cares about getting low differentials from the chip to the water. Three gph is way more than enough to keep a stock chip stable. Given water's heat capacity and say, 75 watts, 3 gpm results in a water temperature rise of about 8°C. Take a block that has surface area more on par with a normal air-cooled heatsink and you would do just fine.

Take a gander at the HSFs packaged with CPUs. These typically let chips run much warmer than the DIY crowd would like, but obviously they do work. This is little different.

For the record, the Centrino is not the mobile P4. It is more akin to a mobile P3. There is indeed a mobile P4, but it is the Centrino that runs much like XPs, ie lots of work per clock cycle.

Your missing one critical peice of the puzzle. How is 3gph (with no head) going to push water through a radiator and a water block? 3GPH is with NO restrictions. Thats like a 370GPH pump running with no hoses attached the the outlet. Hell my 500GPH HydroThruster gets bogged down 1-2 hundred GPH just by the radiator alone... I don't see any way this pump is going to provide any head pressure to be used effectivly. The water has to be cooled somehow no? Water has to be pushed through a cooling system? How does this work with 3gph with no head pressure?

[TerraMex]

http://www.intel.com/products/mobiletechnology/

Centrino. It's a mobile, low power processor. Like it's been said , it's not a p4 mobile. It has several built in features, like Wifi.

About flowrates.

You dont need huge amouts of flowrate to make a system work. Good example, Aquarius II. I wouldnt buy it, but i know who did. And the thing works within acceptable limits (ok, a bit high), as far as temperature goes. Just dont overclock .

Again, lots of users over here (Europe) use AquaComputer, Innovatek, and similar systems. Shipped mostly with 1046, and if you check Joe's test of the innovatek kit (as a reference point) you can see it doesnt fall behind larger (and with larger pump) systems.

http://www.overclockers.com/articles545/

http://www.overclockers.com/articles373/

However, i do agree with you on the point that 3GPH is exceptionally low. But we know nothing of how they are dissipating the heat. Or more details on the pump. Or even how is the waterblock, internal designs.

Without more info, how can we take conclusions properly?

Anyway, the article is talking about working prototypes, IMO. Probably proof of concept. Not real, finished, comercial product.

[jaydee]

Originally posted by TerraMex
http://www.intel.com/products/mobiletechnology/

Centrino. It's a mobile, low power processor. Like it's been said , it's not a p4 mobile. It has several built in features, like Wifi.

About flowrates.

You dont need huge amouts of flowrate to make a system work. Good example, Aquarius II. I wouldnt buy it, but i know who did. And the thing works within acceptable limits (ok, a bit high), as far as temperature goes. Just dont overclock .

Again, lots of users over here (Europe) use AquaComputer, Innovatek, and similar systems. Shipped mostly with 1046, and if you check Joe's test of the innovatek kit (as a reference point) you can see it doesnt fall behind larger (and with larger pump) systems.

http://www.overclockers.com/articles545/

http://www.overclockers.com/articles373/

However, i do agree with you on the point that 3GPH is exceptionally low. But we know nothing of how they are dissipating the heat. Or more details on the pump. Or even how is the waterblock, internal designs.

Without more info, how can we take conclusions properly?

Anyway, the article is talking about working prototypes, IMO. Probably proof of concept. Not real, finished, comercial product.

I know all about flow rates. Been experimenting with it for 3 years. I been down to 10GPH with decent results. Problem is it was a 170GPH pump in that system that is capable of creating head pressure. I have no idea what your point is about dissipating heat. I thought it was pretty clear they are going to use a water block of sometype (being it is a water pump) and a radiator of somekind (no other way to "dissipate" heat in a water cooling loop without some extra help). You could try and stuff this in a one cooler setup like that failed all-in-one unit BUT if CPU's are going to keep getting HOTTER then this whole no overclocking argument is simply irrellevant. The heat we creat by overclocking today is the same amount of heat going to be on the future processors at stock. It is going to take MORE cooling capability in the future (from what they are saying anyway) not the same cooling capability as today. I see no way this current 3gph pump is going to effectivly cool a CPU running at 75+ watts. ANYTHING you add to the loop is going to stop the flow. A water block of ANY type and a radiator of ANY type. Even if the rad had one round copper tube and the water block was completely wide open with no channels. Just the connections would kill off the flow.

Sure 3GPH can in theory cool 75+watts, but how can you do it in reality. The water needs to be cooled somehow and there has to be a heat exchanger on the CPU of somekind. Both of which will need more than a 3gph pump IMO. Maybe a pump that after everyting is already hooked up got 3GPH, but not a 3GPH pump BEFORE the rest was hooked up.

These guys are scientists though. What do I know. :shrug:

[TerraMex]

I know you do. I've keeping track of your exploits in the forum, and good work too.

Just trying to make a point. Flow is not everything.
Not directly related to the subject, the osmosis pump (my bad).

> I have no idea what your point is about dissipating heat (...)
> Sure 3GPH can in theory cool 75+watts, but how can you do it in
> reality

Simple. We have no real info. We assume alot, but no real data to make those assumptions.

It can be obvious to us, they WILL be using this and that. But we have no ideia on which and how. We just have to wait and see .

> failed all-in-one unit

I've seen a device of such type for a laptop. I dont recall where, maybe even in Procooling. I'll see if i can dig up the thing.

> It is going to take MORE cooling capability in the future (from
>what they are saying anyway) (...) Just the connections would
> kill off the flow.

No arguments from me.

---- Edits (got it bit more organized)


http://www.lostcircuits.com/tradeshow/idf_2002/5.shtml

From Hitachi . I knew i saw it somewhere . Dunno what happened to it thou.

Looking more closely. Not actually a "radiator". Iit's using the backplate to dissipate the heat.

It's stainless steel. Acording to this article:
http://www.newsfactor.com/perl/story/16579.html

http://news.designtechnica.com/article704.html
And a bit more info from NEC.

[TerraMex]

Extra.

The laptop using that plate reminds me of this :

http://www.oc-card.de/assets/big/konvekt.jpg

It's a german passive radiator. I know someone who has one.

[myv65]

JD,

You may have been down to 10 gph, but you were not using a block designed to work at that flow rate. As to flow resistance, bear in mind that pressure loss is a function of flow and the resulting pressure required to move 3 gph is virtually zero.

At 3 gph, convection coefficients obviously would be quite low. This is why you would require a block with a whole lotta surface area in contact with the water, ala a typical air cooled heat sink.

It will work, but it'll never be used by the hardcore DIY crowd, who incidently make up far less than 1% of computer users.

[jaydee]

Hereis the companies home page: http://www.cooligy.com/

[jaydee]

Originally posted by myv65
JD,

You may have been down to 10 gph, but you were not using a block designed to work at that flow rate. As to flow resistance, bear in mind that pressure loss is a function of flow and the resulting pressure required to move 3 gph is virtually zero.

At 3 gph, convection coefficients obviously would be quite low. This is why you would require a block with a whole lotta surface area in contact with the water, ala a typical air cooled heat sink.

It will work, but it'll never be used by the hardcore DIY crowd, who incidently make up far less than 1% of computer users.

I don't see it working for any crowd. Not if heat is going to double and die size is going to shrink. Would have to be one hell of a block to be able to spread the heat out into a large WB with only 3gph pump. Not to mention the holes in this pump are 1 micron thick. How do you keep it from plugging up?

Tech that is over my head clearly.

[sevisehda]

First of all processor temps don't gradually rise. I'm sure if someone plotted it out, it would be more sawtooth in nature. When a new processor is released its normally relatively cool, as they start to tweak the design for all its worth is when temps start to rise. Early P4s were very cool but now the latest tweak will have the output at over 100 watts. When the P5 is realeased chances are the temps will be back down again.

I'm sure this would be a OEM solution. The pump/rad/wb would be a preassembled unit. As such the coolant could be distalled water with antifreeze or some other additive. You wouldn't have to worry about clogging the pump because the loop would be sealed.

Mainly though keep in mind this is basically a press release to promote a product. By no means is this the 'written in stone' future. Personally I think the next step will be heatpipes on CPU coolers.

[bigben2k]

Originally posted by jaydee116
I don't see it working for any crowd. Not if heat is going to double and die size is going to shrink. Would have to be one hell of a block to be able to spread the heat out into a large WB with only 3gph pump. Not to mention the holes in this pump are 1 micron thick. How do you keep it from plugging up?

Tech that is over my head clearly.

It's not terribly hard to grasp.

Imagine a spiral block, where the channels are just a couple of millimeters tall. We wouldn't normally use a block like that, because it would be too restrictive, but for a 3 gph pump, it's ideal. Who needs 1/4" tall channels anyways?!?

[jaydee]

Originally posted by bigben2k
It's not terribly hard to grasp.

Imagine a spiral block, where the channels are just a couple of millimeters tall. We wouldn't normally use a block like that, because it would be too restrictive, but for a 3 gph pump, it's ideal. Who needs 1/4" tall channels anyways?!?

:shrug: Your going to need a massive amount of surface area to account for the lack of flow. Much like an air cooler. How is a block with a couple mm tall channels going to get that kind of surface area?

[bigben2k]

Oh you'll still get a massive performance loss, but you have to optimize the block design the same way that we do with our relatively huge pumps. In this case, you'd also have to compensate for the lack of flow with a thicker baseplate, and that's where most of the performance loss will be.

Now if we had a 3 gph pump with a huge amount of pressure, we could really have some fun!

[Althornin]

um, whats the deal here folks: Just make it bigger!

Assuming flow rates are directly proportional to area, this post is wrong:

If that was the case then a 5meter disc is needed for 300gph.

Area of circle is PI*(.5*diameter)^2
area of 2.5^2*3.14= about 20cm^2
So for 100 times the flow, we need 2000cm^2, which is a diameter of 50CM.
while this is kinda large, even making it the size of a 120mm case fan gives it flow of 21 GPH (about seven times the area), and of course, it is so thin multiple could be stacked to negate head problems.

[sevisehda]

I still dont see this becoming the next 'thing'. I'm sure it may have some limited uses but for computers I highly doubt this will ever happen. There doesn't seem to be any benefit of using this over a normal pump. It may be 100% silent but there will still be HDs and Fans in the system that would drown out noise from a normal pump. Over an extended amount time those pores would eventually become clogged and it would lose performance over time(just like a normal pump). I think the article was more an advertisement than an ocjective report.

[zoson]

I think the real issue at hand here should be, if you need a waterblock the size of a standard heatsink, why wouldn't you just use a standard heatsink?

This seems like a terrible idea to me. At least with a standard heatsink the ONLY thing that can fail is the fan. With this you're needlessly adding system components to get the SAME results as a standard heatsink.

Think about the whole nature of water cooling. Most of us do it to move heat away from the processor faster than a standard heatsink can. And if you take into account the argument above that as processors are developed they become cooler (which I do NOT agree with at ALL), you would be able to use a slow and quiet fan, negating any argument about noise. Why risk having water in your system for no performance gain at all?

.02
-Zoson

Edit: I also dread the thought of running electricity through my coolant, which runs through (in most cases) more than one type of metal. Yay for corrosion! Not to mention one of those metal components is your waterblock sitting on a VERY sensitive and tiny block of several million transistors. I'd hate to be a transistor owned by someone who uses this pump in his cooling system.

[joemac]

I do not understand why people bash the reliability of the current brush less pumps we use. They work in the same principle as today’s PC fan. For those that don’t know the motor are of the same concept but instead of turning fan blades the motor turns an impeller. If the MTBF (Mean time between failure) is acceptable for the fans today then why is it not acceptable for the pumps (brush less). What called my attention to this pump is that in a way it reminded me of the peltier effect very cool but highly inefficient.

[jaydee]

I don't see water cooling being the real answer in the end. I think (as mentioned before) heatpipes will advance and other phase change things like the Vapochill. I am sure this tech isn't nearly as advanced as it could be. I think we may see vapochill style phase change that will fit in a stand computer case. I really like the heat pipe concept though. My buddy designs these where he works and we have been talking about it off and on. I really should get off my ass and develope a protoype....

[nicozeg]

That pump is a great product for a specific niche: Super silent notebooks.

First it need to be part of a complete system assembled in a carefully controlled environment to avoid contamination of the coolant.

We still have very little info on the specs, like how much power is needed to reach that 200cc mark.

We also dont know the head capability of this pump, but I'm prety sure that it behaves completely different than the pumps we know. This is an absolute guess on my own, but I believe that it have an incredible amount of pressure. Just think what happens when power is not applied in the pump: A centrifugal pump makes almost no force against backflow, but how much water can be pushed across a 2mm glass?

The potentials of this are amazing: stacking several discs in parallel can make a decent flow, and if the pressure is like I think it can make some truly incredible speed microjets.

[bigben2k]

Originally posted by jaydee116
I don't see water cooling being the real answer in the end. I think (as mentioned before) heatpipes will advance and other phase change things like the Vapochill.

I disagree.

I think that heat dissipation is going to be an ever increasing problem, and that water-cooling is the most affordable solution.

A refrigeration unit would only draw a lot of additional power, and that's going to have the EPA ("Energy Star") people up in arms: Can you imagine 100'000 people using a PC with a compressor to keep it cool?!? Yikes! If you thought that that last power blackout was bad...

So the future really, would probably be somewhere in the middle, where the water cooling solution is tuned to the CPU temp. As a guideline, Energy Star requires that a device on stand-by consumes no more than 5% of its max rate.

(Man, I ought to just shut up and market this stuff!!!)

[jaydee]

Originally posted by bigben2k
I disagree.

I think that heat dissipation is going to be an ever increasing problem, and that water-cooling is the most affordable solution.

A refrigeration unit would only draw a lot of additional power, and that's going to have the EPA ("Energy Star") people up in arms: Can you imagine 100'000 people using a PC with a compressor to keep it cool?!? Yikes! If you thought that that last power blackout was bad...

So the future really, would probably be somewhere in the middle, where the water cooling solution is tuned to the CPU temp. As a guideline, Energy Star requires that a device on stand-by consumes no more than 5% of its max rate.

(Man, I ought to just shut up and market this stuff!!!)

Well, we havn't even begun to break the ice in CPU development. They will have to start coming up with new designs that do not run as hot and still scale well. Something has to change as the dies can only get so small before nothing you attach will cool it. I think that is half the reason for these heat spreaders. Just a quick fix for the real problem they have yet to solve. In 20 years pocket PC's will have more power than our fastest computer today. They are not going to have water cooling systems attached to those. Tech will advance past to point to were any of this will be required. Just a matter of time. It is what happens between now and that time that is going to be interesting. I still think heat pipes are going to become a little more popular. hell I seen one for the new vid cards and for hard drives already. Not to mention the one's used in some laptops for a couple years now.

I predict PC are going to make some drastic changes in appearance and size in the not to distant future. Going to be interesting next 20 years.

Also I agree with one of the posts above. Why not just use a good 3GPH mechanical pump that doesn't have 1 micron tubes to clogg up. hell the corrosion off the blocks and rad alone I would think can get bigger then 1 micron.

I don't know though. This is still over my head I guess.