F2 Extreme Radiators for Extreme watercooling

[copyman]

I´ve been developing watercooling products for some years. In this time, I´ve tested most of the products available, about 90% of the high efficiency wc systems available in the market. In all of them, besides evaluating their performance, I also looked for mistakes, always trying to understand the tecnology used, and trying to find ways to improve their efficiency. In all of them, what caught my attention were the radiators.... All were very inefficient, and highly flow restricted. These factors limited the cooling systems I was evaluating in extreme Ocs...

Of course they can´t do the cooling work all by themself: a high performance water block and a strong water pump is also needed... I tested most of them, and will do a brief explanation of what I think is important concerning watercooling

About the tubing and hose barbs: I use modified hose barbs... I´ll explain: My hose barbs are modified, having the liquid flow coming from the tubing in mind. They have their inner width widened to the max and polished. The mods on the inlets of these hose barbs is to keep a straigh line from the tubing to them, avoiding turbulance in the incoming liquid that would decrease flow rate. I use a 50% silicone ½ inch tubing, which is extremely flexible and has a bigger inner width, when compared to other ½ inch tubings available in the market, and it fits my hose barbs perfectly.
These tubings are custom made for me by a manufacturer near my town, and are capable of doing a complete 4 inch radius circle without any obstruction or deformation. I use these tubings and hose barbs in all my water cooling kits, extreme ones or not.

About the F2 Extreme radiators: I would like to give some personal beliefs about the commom radiators used today for water cooling before explaining my radiators.
None of the radiators I tested, at least from what I know, are made exclusively for watercooling . For example, Dtek, one of the most used radiator, is simply a regular air heater radiator used in cars. The only difference is its size and headers. All the rest, including the Z form flat tubes, the tanks, the fins, are the same that are used in 90% of today´s cars air heaters. They are probably made by Delphi, since its identical, except for its size and headers, to the ones I tested. They are exactly the same heatcores made for the Chevette air heater, usually recommended in this Forums as a cheap solution. In my opinion, they´re highly restrictive and inefficient when used for watercooling.
The same thing happens with the old (refrigerator´s condenser) and new Innovatek radiators, as well as the Black Ice and Black Ice Extreme from Danger Den, which are also made by Delphi, and uses the same flat tubes, headers and fins used in other heatcores made by Delphi, most of them used in air heaters for cars. Again, these too are extremely restrictive, with low heat exchange and not innovative.

If you look for a air heater radiator in na auto´s parts shop, you will see that 50% are identical to the ones used by Dtek, and the other 50% are exactly the same used by Dangerden and Innovatek... Always the same thing, the only difference is the size, inlet type and, sometimes, the material used in it...

I wanted much more than the market could offer me... I knew how a watercooler radiator should be, and I would not do as the others did, using radiators based in trivial heatcores, inefficient and flow restrictive, usually used for air heaters... What I wanted was a perfect watercooling radiator, at least from my point of view.... Specially designed for watercooling use... And that´s how my F2 Extreme project started...



Some characteristics of the radiator F2 Extreme:

That radiator is made with 3 arrays of flat tubes opposed and 12 layers of laminate pure copper for each inch of height,
it has in the total of 54 flat tubes divided and opposed in 3 arrays

it works in an only fluid passage, he has TWO entrance of 1\2 in the superior camera and TWO exit of 1\2 in the inferior camera

the entrance has a special camera that feeds of fluid in a homogeneous way all the flats tubes .

all weld they are them done with 57% of silver and all the connections are re-worked for over diameter and soft angles of entrance and exit
























CPU: xp1700+
Stepping: 0310
Cooling:Watercooler F2 Extreme ( no petier , chiller or vapos)
Voltage: 2.1v
Motherboard:abit nf7-2 rev 2.0 Bios 1.3
System Memory: oczel3200 dual channel
Overclock: 2744mhz

Watercooler Setup

Pump : Quietone 1100 gph
Radiator : F2 Extreme
Block: LRWW


OBS : nf7-s bio1.4






OBS : nf7-s bio1.3








Extreme oc in wc sistem

[Blackeagle]

Welcome to ProCooling! !

Anyone who can machine and finish that rad should be interesting to have around.

You have a very strong o/c there, very nice.

You mentioned the inlet and outlet design to insure ease of flow. But in all the pics you posted they arn't shown.

The rad looks to be very large (those are 120mm fans are they not?) what are the L X W X H demensions of the rad?

The quiet one pump is a unusually powerfull one. How would the rad perform with a more usually seen pump such as a Hydor L30? Or even a pretty strong one such as the Iwaki MD-20rlzt or MD-15rlt?

Are the fans specs as ultra powerfull as the pump chosen?

Have you had any testing done on the rad outside of your O/Cing with it?

And once again welcome to ProCooling!

[TheDanMan]

That is a very nice radiator. I now wish I had one like it :-)


Excellent Work!!!1

[Blackeagle]

The more I look this over the better it looks. I've seen flat tubes like he's using at a site that is involved in new and advanced designs that intends to replace the now common alum. car heater cores & rads with these flat thin walled tubing along with copper fins.

Just the fact his tubes are copper instead of brass should offer a pretty pronounced boost in heat transfer. Then couple in the lower flow resistence of the flat tubes vs. a heater cores Z type flat tubes and it gains a second boost from increased flow rates.

I like your design work Copyman, looks to me like you've done some serious design work.

Would still be interested in the inlet/out let design and the L x W x H demensions.

[copyman]

Excuse, don't think that I don't answer for negligence, that week I am in trip :shrug: and I get to almost access the net time, tomorrow I will answer all the involved subjects and we can talk more about those radiators


now i only wanted to thank the praises

[]'s

Ivo Guilhon (Copyman)

[redleader]

Thats the only intelligently designed "custom" radiator I've ever seen. Nice work. The use of copper tubes (verses brass in heatercores) will probably help performance a little too.

[Blackeagle]

I looked for the article showing the new copper type rads I thought I'd seen.

I was in error. The only thing I found was ones with brass tubes. There are new brazing methods and thinner walled tubes in use that do save weight and improve heat transfer, but the ones I found are brass not copper.

Very sorry for the mistake.

[Joe]

Man... that thing is a freaking work of art... just beautiful!

[copyman]

Originally posted by Blackeagle
I looked for the article showing the new copper type rads I thought I'd seen.

I was in error. The only thing I found was ones with brass tubes. There are new brazing methods and thinner walled tubes in use that do save weight and improve heat transfer, but the ones I found are brass not copper.

Very sorry for the mistake.

Yes, the flat tuber are made also of copper but to 85%, it is a league, it is not gotten to work the pure copper in the format of flat tubes, he doesn't support that when it is modeled, that and been worth for any radiator with flat copper tubes, the only difference is that the manufacturers don't mention that, that was me informed for the it manufactures that produces the core for me.

And yes all flat tubes in every rad (BIX, BM..) are in brass because copper is unworkable with a such small thickness (0.05mm to 0.1mm) but it's not a problem for thermal exchange with this thickness

They consider leagues above 80% in an only material as pure, to the they think to be unnecessary you mention it
The own dteck announced your radiators with being 100% cooper, today her no longer itmentions more that, the chanbers are of brass and the flats they are mixed
My chanbers are 100% copper fastened with it welds to 57% silver, as well as the flats, they chambers are underneath of the cases in lateral inox

[copyman]

Originally posted by Blackeagle
Welcome to ProCooling! !

Anyone who can machine and finish that rad should be interesting to have around.

You have a very strong o/c there, very nice.

You mentioned the inlet and outlet design to insure ease of flow. But in all the pics you posted they arn't shown.

The rad looks to be very large (those are 120mm fans are they not?) what are the L X W X H demensions of the rad?

The quiet one pump is a unusually powerfull one. How would the rad perform with a more usually seen pump such as a Hydor L30? Or even a pretty strong one such as the Iwaki MD-20rlzt or MD-15rlt?

Are the fans specs as ultra powerfull as the pump chosen?

Have you had any testing done on the rad outside of your O/Cing with it?

And once again welcome to ProCooling!

Those pictures are of a removed rad of the the first lot of 10 pieces,it was not still with the mounted chanbers, but he has two entrances on a side and two exits of the other in 1\2 inch.
The chanbers with the connections are underneath of the case in lateral inox
it uses fans of 120mm yes

it is indicated for bombs of 500gph or larger, because your use is for extreme watercooling, and using good watercblocks as the cascade, tc4 or lrww, he works very well with a MAG 700 and a tc4, logical that as larger and less heat the bomb to review to liquid and more efficient the block, your results will be more positive, with the mag 700 and the tc4 the differences were in 1c of delta and 3c of temperature to 2733mhz 2,1v, compared with the quiet one and the lrww

thanks

[BillA]

so do you mean
"Red Brass, 85% (Copper Alloy No. 230) — An alloy containing nominally 85% copper and 15% zinc and generally available in flat products, rod, wire and tube."
thermal conductivity 159 W/(m°C)

as compared to
"Yellow Brass (Copper Alloy Nos. 268 and 270) — An alloy containing nominally 65% copper and 35% zinc and generally available in flat products, wire and rod."
thermal cnoductivity 67 Btu/(h.ft².°F/ft) ~ 116 W/(m².K/m)

the thermal conductivity of Red Brass is 37% better than Yellow Brass (and both are FAR below 'pure' copper)
but since the air side convection is the limiting factor, so what ?

rads are ALL about airflow

[copyman]

Originally posted by unregistered
so do you mean
"Red Brass, 85% (Copper Alloy No. 230) — An alloy containing nominally 85% copper and 15% zinc and generally available in flat products, rod, wire and tube."
thermal conductivity 159 W/(m°C)

as compared to
"Yellow Brass (Copper Alloy Nos. 268 and 270) — An alloy containing nominally 65% copper and 35% zinc and generally available in flat products, wire and rod."
thermal cnoductivity 67 Btu/(h.ft².°F/ft) ~ 116 W/(m².K/m)

the thermal conductivity of Red Brass is 37% better than Yellow Brass (and both are FAR below 'pure' copper)
but since the air side convection is the limiting factor, so what ?

rads are ALL about airflow

yes, these data are right, but it excuses, I didn't understand what you ask me

you repeat being clearer?

[nikhsub1]

Originally posted by copyman
yes, these data are right, but it excuses, I didn't understand what you ask me

you repeat being clearer?

Copyman I believe what he is saying is that yes, your copper tubes in the rad are a bit more efficient in terms of thermal conductivity but, that fact matters little and that rad performance is mostly about airflow.

[j813]

Being in Brazil & all wat's your ambient temp in ur computer room?
Here in Asia in summer my room can get hot as 32celcius.

Thanks

[Cathar]

Going by those thermal conductivity numbers, and the total surface area of the radiators, and the thin-ness of the tubes and fins.

There are 54 tubes. Let's say that each is 1cm wide, 24cm long and with a wall thickness of 0.15mm

The thermal conductive resistance of the tubes works out to 0.26m^2 (cross-sectional surface area of the tubes) x 159 W/mK / 0.00015m = 274752 W/K, or basically 275000W/K

If the tubes were made out of pure copper, this increases to 665000W/K

For worst case 150W heat dissipation through the radiator, the walls of the tubes for copper vs red brass is causing a decrease of (150 / 275000) - (150 / 665000) = 0.0003C

i.e. not an awful lot. Heck, even for the ~4kW or so that a car would push through the heater-core we're still talking about a 0.01C difference between red-brass and pure copper.

Okay, so let's work in the fins as done with red-brass vs pure copper.

5 fins per cm, 54 tubes, medium fin length is 4mm (half of the angular distance between tubes). Fin thickness is 0.1mm.

Total cross-sectional surface area is 0.013m^2

Assuming an even heat spread to the halfway point along the fins between each pair of tubes, heat spreading resistance of the fins works out to around 156 * 0.013 / 0.002 = 1014W/K for red-brass, and to 2502 W/K

For a 150W heat source, this means ~0.15C for red-brass, and 0.06C for pure copper

All up, the difference for making the radiator using pure copper tubes and fins vs red-brass tubes and fins, for a ~150W heat source is less than 0.1C in terms of the added thermal resistance due to the metal used in a radiator of the size being used here.

Radiator's are about two things, air-flow and size.

The radiator's used here will work better than the DTek Pro style cores because they are larger for starters, and offer almost 2x the total fin->air surface area of the DTek Pro cores, and really, that's where the benefit is coming from, not from the use of copper over red-brass.

[BillA]

Cathar

its Red Brass vs. Yellow Brass for the tubes
both have copper fins
the diff is even less than your calcs suggest

[Cathar]

Originally posted by unregistered
Cathar

its Red Brass vs. Yellow Brass for the tubes
both have copper fins
the diff is even less than your calcs suggest

Oops, my mistake in the interpretation.

Yes, in that case, the difference is getting down to around a 0.0001C difference based on the tube wall material, since the bulk of the difference I was calculating was due the incorrect assumption of the fin material.

Point still stands though that if the radiators perform better, then it really comes down to the difference in size, rather than what materials are used, oh, and that two fans are being used rather than one over that larger area, so therefore more volumetric air-flow.

[copyman]

Originally posted by Cathar
.

The radiator's used here will work better than the DTek Pro style cores because they are larger for starters, and offer almost 2x the total fin->air surface area of the DTek Pro cores, and really, that's where the benefit is coming from, not from the use of copper over red-brass.

Hello noble friend stew, is a pleasure to have your post in my topic

But...

you disrespected some points of relevance in the efficiency of radiators.

1: Feeding of the flats tubes, the more homogeneous, better the use of the total area of the core

2: restriction to the flow, even if it reduced my rad to the size of a dteck, your resistance to the flow would be 3x smaller, first due to total area of drainage of the flats, second for the type (format) of flat tubes

3: Correct density of the fins X pressure of air, I tested it varies conbinations of densities leaving from 5 to 15 fins for inch,also varying your thickness, also varying the pressure of the fans in the tests, to arrive to the ideal group, I didn't simply remove a radiator of hot air of car and I placed a fan in him.

4: Behavior of the flow of air inside of the pits, my flow is not to laminate and yes turbulent, causing a more efficient thermal change, that due to having 3 arrays of flat opposed tubes.
To the I thwart of the detk that works with flow to laminate

The several other although I will post later, but those 4 items for the time being are enough

a series of differences that make that contruction type that I developed still exist much more efficient than a dtek, my use of size of the core x transfer of heat is very superior, and much larger than one used in a dtek.

my rad F2 is a king kong, don't have doubts, but he is not only more efficient for having the double of size of a dteck, thing that you will by itself be able to see, when it compares him with the two dteks that you use, logical that for that will have to generate enough heat for the differences to appear, as I did

[Cathar]

Yes, all those things are factors.

We're just dissecting the red vs yellow brass side of things and figuring out how much impact that would have.

Optimal fins per inch really is a bit open as well, as like pumps, different fans respond differently to different levels of air-pressure resistance.

Axial fans would indeed prefer a more open fin spacing.

Blower type fans ( like this one that I picked up) can provide greater pressure and will work better with higher FPI radiators. The radiators I use are 24FPI and the blower I use is barely affected.

Bill (unregistered) is more familiar with air pressure-drop relationship with radiators than I am though.

The heater-cores I use are constructed by Nippon Denso (Toyota subsiduary) and most (all?) of the qualities you describe are present in their design, including homogenous flow entry, low flow restriction, and concertinaed slits on the fins themselves disrupting laminar flow, and despite using 24FPI, still have a fairly low air-flow restriction because they are thinner.

I like the radiator's you are making, but it is also wrong to assume that all heater-cores are not made to the same level of optimisation that you are talking about. The Nippon Denso cores certainly are.

[copyman]

Originally posted by Cathar

The heater-cores I use are constructed by Nippon Denso (Toyota subsiduary) and most (all?) of the qualities you describe are present in their design, including homogenous flow entry, low flow restriction, and concertinaed slits on the fins themselves disrupting laminar flow, and despite using 24FPI, still have a fairly low air-flow restriction because they are thinner.

I like the radiator's you are making, but it is also wrong to assume that all heater-cores are not made to the same level of optimisation that you are talking about. The Nippon Denso cores certainly are.

did I already use radiators of 24fpi with plenty success in my rad F1, but are they impossible for me in you adapt them the any size that I want, and is the level of heat that I generate superior to the size that I had available, but did they need a lot of pressure in the fans, did they really have a very low restriction level of flow, could you show me one of those rads that you mentioned?

With relationship to the blowers are a show, it has a level of noise x pressure amazing

[Blackeagle]

The estimates of performance would be helped a good deal if Copyman could post the L x W x H demensions of his rads finned area.

By the crude means of just measureing the shourd (on screen)and comparing those measurements to the known size of the fans I'd have to guess his rad is at minimum 13" x 6.5" x 1.75" for just the finned area, and perhaps as much as 15" x 7" x 2.2" for a maximum. I may be wrong, but I think the rad is larger than Cathar's estimates.

The CFM of the fans used would be a second very helpfull stat. 120mm fans are on the market with CFM's of from 53 CFM up to 152 CFM.

Along with a quad fan set up and good shrouds (Copyman's shroud's look flawless), a rad somewhere within the above size range should have little problem dissipating all the heat from just about any non TEC system, even if medium noise level fans were used such as fans giveing 80-90CFM.

There are so many variables here.

Copyman, how about some hard data on the rads size and the specs of the 4 fans used?

[Cathar]

Originally posted by copyman
did they really have a very low restriction level of flow, could you show me one of those rads that you mentioned?

With relationship to the blowers are a show, it has a level of noise x pressure amazing

Actually the blower is pretty quiet for the air-flow/pressure it's generating. At the second lowest speed setting it's pushing about as much air through my radiators as 2 x 104CFM Panaflo H1A fans did at full speed, and I can barely hear the blower at all. The 8cm medium speed case fan I have blowing air over the CPU area is the dominating noise in my setup, at least until the CD drive spins up and that is far louder than anything else.

Blower noise is far more tolerable than axial fan noise when pushed up higher. Blower's don't whine like axial fans. Forgive me for going on about the blower, it's just that I'm pretty happy with it after years of never being happy with axial fans. It's the best fan investement I've ever made.

As for the heater-core pictures. The fin area is 22cm x 14cm x 3.2cm

[copyman]

Originally posted by Blackeagle
The estimates of performance would be helped a good deal if Copyman could post the L x W x H demensions of his rads finned area.

By the crude means of just measureing the shourd (on screen)and comparing those measurements to the known size of the fans I'd have to guess his rad is at minimum 13" x 6.5" x 1.75" for just the finned area, and perhaps as much as 15" x 7" x 2.2" for a maximum. I may be wrong, but I think the rad is larger than Cathar's estimates.

The CFM of the fans used would be a second very helpfull stat. 120mm fans are on the market with CFM's of from 53 CFM up to 152 CFM.

Along with a quad fan set up and good shrouds (Copyman's shroud's look flawless), a rad somewhere within the above size range should have little problem dissipating all the heat from just about any non TEC system, even if medium noise level fans were used such as fans giveing 80-90CFM.

There are so many variables here.

Copyman, how about some hard data on the rads size and the specs of the 4 fans used?

my rad f2 has exact 12,5 x 6 x 2 of it blushes, me fis him to work with fans of low noise as the tT of 31db 79cfm, not having temperature difference when used fans of 89 cfm 105 cfm and 130cfm when tested to 2744mhz 2,125 vcore compared the tT 79cfm , to an enormous eficency surpluses here, I have cores in final phase of 6x6 48 flats and 5x5 40 flats that they owe these ready ones inside of one week using the same technology,

many did vary, very difficult to explain, I know because I have two dteks and I already used them in parallel

another complex subject is on types of rads construction, as I don't get with 2744mhz and 2,125 of vcore and two tecs of 70w (9800pro) to heat up liquid of the system in a proportion that gets to measure, in other words, it is always smaller than 1c above the room temperature, I don't have as telling the all it is more efficient

[copyman]

Originally posted by Cathar
Actually the blower is pretty quiet for the air-flow/pressure it's generating. At the second lowest speed setting it's pushing about as much air through my radiators as 2 x 104CFM Panaflo H1A fans did at full speed, and I can barely hear the blower at all. The 8cm medium speed case fan I have blowing air over the CPU area is the dominating noise in my setup, at least until the CD drive spins up and that is far louder than anything else.

Blower noise is far more tolerable than axial fan noise when pushed up higher. Blower's don't whine like axial fans. Forgive me for going on about the blower, it's just that I'm pretty happy with it after years of never being happy with axial fans. It's the best fan investement I've ever made.

As for the heater-core pictures. The fin area is 22cm x 14cm x 3.2cm

hehehe, to work with blowers is really something exciting when fans is had by the whole part, I already tested some but I didn't have the luck to find one that if it adapted my nessecidades and size x potency, your post it motivated me to continue seeking

from where they had left your blowers, it will be very difficult achal-them in the brasil, because I already sought plenty, but....

with relationship to your rads are a lot similar to the one of F1 (seemed this far away from me to affirm that are similar in efficiency, me that I swimed in the middle of rads I know what small differences do in acting)

I only am not right of the material used in that rad, could it clear me?

[Blackeagle]

Cathar,

I note that Copyman said he would be sending you a rad to test.

What will your testing set up & methods be?

Copyman,

Thank you for providing some hard info, and I'm glad to see that the rad is smaller than I thought it might be.

I'm afraid I didn't understand which of the fans you've spec'd were used when you obtained the results you posted early in this thread. Could you please restate that info?

Did you do testing on the ideal distance for the shrouds stand off from the fins? What stand off distance did you decide on?

Are the flat tubes you used in your rad dimpled on the sides to increase turbulence in the tubes, or did you find the smooth walled ones to be better performers due to higher flow rates?



A interesting thread & rad.