"Radius" by BigBen2k - Page 14

jaydee · 12-30-2002, 06:10 PM

Quote:

Originally posted by LiquidRulez
I dont think ANY watercooling solution' performance is getting close to its limit for cooling with water.
I think it still has alot of tweaking to do.
After all, water is the best liquid thermal conductor known to man isnt it?? We still have a way to go before we hit the wall with water for cooling.

I do not see how this is possible as Cathars blocks are getting around 9-10C above ambient? If you cannot get ambient than that leaves what? 7-8C more room roughly depending on the load. Where is the point to where it hits the "wall" as you speak? You must know that answer or you wouldn't have made that statement?

Non the less thats not really what I was asking anyway. Basically I want to know if all this "heavy" equipment will do any better than more conventional systems. Faster flow means the water will heat up a little as it is creating frinction. So the faster the flow the more heat it will create. He is talking about an extream amount of flow and I just wonder if it is necassary or even worse if it may hamper it.

jaydee · 12-30-2002, 07:18 PM

Quote:

Originally posted by jaydee116
I do not see how this is possible as Cathars blocks are getting around 9-10C above ambient? If you cannot get ambient than that leaves what? 7-8C more room roughly depending on the load. Where is the point to where it hits the "wall" as you speak? You must know that answer or you wouldn't have made that statement?

Well my numbers above must not be even close to right.

I pulled this from another thread:

Quote:

add 3 or 4 for the coolant/air differential and the CPU would have to be below 20W to even be close to a real 10°C differential

That was said by unregistered in a thread I started here:
http://forums.procooling.com/vbb/sho...7&pagenumber=3

freeloadingbum · 12-30-2002, 07:54 PM

Quote:

by LiquidRulez:
I dont think ANY watercooling solution' performance is getting close to its limit for cooling with water.

I disagree. As far as waterblocks are concerned, I don't think you are going to see much better than the "whitewater". If you look at one of Bill Adams tests, you'll see that the baseplate of the liquidCC block is about 7C hotter than the water @100watt load (1.5gpm). Cathar's block should be about 3 or 4C better, so the baseplate should be about 3C above water temp. Since the effort needed to get the baseplate the same temp as the water would be unrealistic, I doubt we will ever see more than a 1 or 2C improvement. (Since cathar's block hasn't being properly tested against the LiquidCC, the actual numbers are debatable)

The biggest problem is the loss at the T.I.M. joint. No design, aside from direct die cooling, should have any affect on that loss.

bigben2k · 12-31-2002, 09:27 AM

As BillA pointed out, I've tried to incorporate the best of the best into one block, and finding out that it may not be possible to have everything.

In any rig, there is certainly a limit to the cooling ability, and it is dictated mostly by the "gradients" that BillA previously mentionned. In short, for as long as there is any kind of material between the heat source and the cooling fluid, there will be a difference in temperature, but more importantly there will be a hot side and a cold side to the baseplate. The baseplate is what really defines the limitations of the temp (along with the TIM joint, of course).

One thing that Radius does well is optimize the flow speed for the hot spot (over the core). You just can't get better than that (jet inpingement excluded). The baseplate is at a minimal thickness (2mm, which may be reduced to 1mm), which is just about optimal for the fin pattern, but not necessarily the thinnest of all design, and that's one of the limitations of Radius: White Water has continuous fins, where most of Radius' fins are "broken", for flow reasons.

I've specified that the baseplate will also have tiny holes over it, about the size of a 1 mm drill bit head, similar to the drill-press blocks and the Swiftech unit: it increases the surface area.

One thing that is unique, is the seperating res. Many people have thought about making something like it, but I've never seen one. It's advantaged by minimal restrictions, but disadvantaged by it's enormous size.

As for the pump heat, the motor is isolated (direct-drive of a magnet assembly), so that's not an issue (except that it will need a good vent). The pumping heat is still somewhat nebulous to me, as I'm finding out that there's heat from the propeller, as well as from the restrictions throughout the loop, which means that the nozzle at the entrance of the block will generate heat, but it should be relatively small compared to the heat that it will take away.

In any case, I am trying to achieve a Reynolds number (a dimensionless number indicating wether the flow is laminar or turbulent) in excess of 4000, which corresponds with turbulent flow, which is needed for the extra cooling ability. Of course I would need to optimize the flow and block so that the flow is turbulent only in the critical area: over the core. So far, it looks like my pump can't achieve that on its own, and so turbulators will have to compensate for what the pump can't do. Also, as has been pointed out earlier, it's far more efficient to add turbulators, than to use a very powerfull pump.

I'll be trying out a CFD program, to see if I can graph the flow within Radius. It should at least help me optimize the outlet for a balanced flow.

freeloadingbum · 12-31-2002, 07:26 PM

Quote:

One thing that is unique, is the seperating res. Many people have thought about making something like it, but I've never seen one. It's advantaged by minimal restrictions, but disadvantaged by it's enormous size.

I was looking at that rez and I think I have a simpler way to do it. You cover the base with a large copper cap. You have a copper tube going through the center with a large round flange soldered to the end, covering the radius in the base. The water flows past the flange through the space between the flange and copper cap and up to 2 output barbs soldered to the copper cap. You can drill jet holes into the flange to speed up the flow. I drew a picture using ms paint.

jaydee · 12-31-2002, 09:05 PM

Thanks bigben for the quick summary. You still didn't directly answer my question but I think that is because you are not sure yet either. Glad you are making progress on this. It will be very interesting to see how it comes out.

bigben2k · 01-02-2003, 09:41 AM

Interesting.

I like the simplicity, and the lower height/width.

I'm still going to go ahead with mine: it has many clear parts, for a visual observation of the flow.

I'm re-doing the airtrap, out of 3/4 pieces and 1 inch OD tubing. Here's a pic of the original one, out of 1 1/2 fittings, and 2 inch OD tubing:

bigben2k · 01-02-2003, 09:59 AM

Quote:

Originally posted by jaydee116
Thanks bigben for the quick summary. You still didn't directly answer my question but I think that is because you are not sure yet either. Glad you are making progress on this. It will be very interesting to see how it comes out.

Ok, I must have missed your question... but I do that often!

I don't know exactly how much heat is added, but it can be calculated as a function of flow rate and pressure drop. In any case, it's not significant (I'm sure). I might take the time to try to calculate it, if it's of interest to anyone.

As for cooling below ambient (and a rad), that's part of phase 1. I do expect a drop in flow rate, with a rad, but that's ok. In phase 2 and 3, I'll add a Pelt, and a chiller (a la #rotor), which will eventually do away with the rad, restoring full flow.

So talking about phase 1 only... the temp will depend a lot on the air flow rate through the rad/heatercore (as you pointed out). The block will have to be tested with different baseplate thickness, to find the optimal one, but since I'm nowhere near having a proper test facility, I'll either "wing it" or send it out for testing. Either way, there will be a performance inprovement over "more conventional" setups, but as BillA pointed out himself, the benefits of running a 4 gpm rig enter an area where the cost/benefit ratio isn't wise. In other words, it might be 1 or 2 degree cooler than most blocks.

All in all, I don't expect to beat White Water, but I think that if any design could, it would have to be something awfully similar to this. If I could tweak out the water entrance point issues, it should beat Cathar's White Water, but I don't have enough expertise to do that.

hara · 01-02-2003, 12:44 PM

So, whitewater is unbeatable?

Cathar should patent.

bigben2k · 01-02-2003, 01:00 PM

He can't, it's already a patented concept, but he has copyright on the assembly, as a whole.

Out of all the waterblocks out there, they can generally be categorized: ~flat-plate (Swiftech), labyrinth (Maze, Spiral), etc...

We're at a point where some of these concepts have to be combined, in order to get more performance. In other words, you're not going to be able to tweak a Maze3 to get a couple more degrees, without adding something to it (fins, jet,...).

Cathar successfully combined a fast flow (not quite jet) and micro-fin.

bigben2k · 01-04-2003, 04:13 PM

Here's another pic (last on roll), which explains why I didn't go with flex tubing for the res. (in short: the 1 1/2 barb is WAY too long, so I cut it down for a rigid tube. It's cut even more than in this pic.)

Note: this thread now has the most number of views, with 12'842 !

LiquidRulez · 01-05-2003, 01:06 AM

SO when's it going to be done Ben? Cant wait to see it assembled and ready to mount.

jaydee · 01-05-2003, 08:37 PM

Quote:

Originally posted by LiquidRulez
SO when's it going to be done Ben? Cant wait to see it assembled and ready to mount.

What he said!!

:shrug:

bigben2k · 01-06-2003, 08:01 AM

Well, seeing that I'm building it for Barton, and that it isn't here... your guess is as good as mine! Maybe a month or two.

I'm building this thing on the weekends, and since life is pretty busy here, I get maybe one half day, every other weekend. I do however have a lot more time to think about it!

Yesterday, I realized I made a measurement fault, so I've got to cut another piece of tube, for the part between the screw/barb cap, and the clear flexible elbow.

Lucky for me, I've got 4 feet of the stuff...

I've also started to look into the hardware needed for the mount. I'll go with the idea Utabintarbo had, which is a set of outriggers. Screws have to be 6-32, since the (new) hole size is now 0.150 inch. Once I've figured out which springs to use, I can spec the screw, and place my order with McMaster.

Right now, I still need to figure out which epoxy to use for polypropylene (the 1 1/2 barb and tube connection), since the plastic welder is useless.

bigben2k · 01-06-2003, 10:35 AM

Been googling... and I may have found a product that might allow me to bond polypropylene:

http://www.nbond.com/techdata.html

The problem I had came from this fact: polypropylene is a "low energy plastic", aka a non-stick plastic, much like Teflon.

http://www.nbond.com/faq.html#lowen

On the plus side, cyanoacrylate (superglue) might work too.

Still searching...

edit:
e-mailed 3M and Lepage.

bigben2k · 01-07-2003, 09:05 AM

Ok...

3M says they don't have anything that'll withstand methanol, but otherwise recommend DP-8005 or DP-8010, Structural Plastic adhesive. They sent me a PDF for each one. One note:

"However, these adhesives have .008" microspheres (glass bubbles) so you would have to design the parts; ID and OD to accept the microspheres."

I e-mailed Loctite. No reply from Lepage.

bigben2k · 01-07-2003, 10:16 AM

Here's the final render (thanks to Utabintarbo!):

Puzzdre · 01-07-2003, 11:10 AM

Looks beautiful!!!

Do you have the clamping mechanism ready or in progress?

bigben2k · 01-07-2003, 11:21 AM

As I just mentionned to Utabintarbo, I'll be going with his outrigger idea.

I think I'll make them stick out from the top: it'll avoid having to mill copper anymore than it has to be.

bigben2k · 01-07-2003, 03:55 PM

Devcon says: "We don't have a solution, but consider abs or nylon, instead of polypropylene, in which case you can use the Plastic Welder".

It's actually a good idea...

bigben2k · 01-08-2003, 09:34 AM

From Loctite:

"PP is the more difficult of your two substrates to bond. Our best
solution is to apply Primer 770 to the PP, then try one of our cyanoacrylates, such as 401. This should give you a good bond, but you would need to test for the fluid exposure, as cyanoacrylates do not typically have the best long-term resistance
to water. Another possibility with better fluid resistance, but lower adhesion to PP, would be a UV acrylic, such as 3105. The datasheets can be downloaded from our website."

bigben2k · 01-11-2003, 02:24 PM

For those still in the dark...(sorry!) I'm opting to replace the polypropylene barb, with a nylon one.

In the mean time, OC forums killed an idea to sponsor DIY blockmakers, for a round of tests by Bill Adams, but in the middle of it all, there was some info about Radius that I thought I'd repost here:

1-My pump may only deliver a pressure equivalent to 9 feet of head, at 250 gph. There seems to be many versions of the 2-MDQ-SC, I guess I hoped that mine was one of the best, at 10.5 feet.

2-I estimated the flow/pressure at 4 gpm, and 11 feet of head (through the block): if you've been following this thread, you know that this only referred to the inlet (still incorrectly), and not the flow inside the block, past the nozzle, so the block is actually a lot more restrictive. Les estimates a head of 35 to 50 feet, at 4 gpm.

Total heads loss (at 4 gpm):
3/4 ID tubing: 0.62" of head, per foot of tubing
rad: n/a (when it goes to a chiller)
res: about 2" of head

3-That nozzle (above) was to be 3/16. I've since decided/recalculated to go with 3/8. Either way, I'll be ordering a polycarbonate rod, 5/8 ID (to fit the inner res tube) so that I can try different nozzles.

4-More importantly, Bill pointed out that viscosity changes significantly at a lower temp, to which I replied that using the Windshield Wiper fluid will make little to no difference, at least according to Scott Gamble (corrections notwithstanding).

5-Bill has pointed out (correclty) that I have no modelling software. I do plan to use one, when I can find it (free CFD anywhere?) and have the time to mess with it.

Now all this started when I called Radius a high flow/high pressure block, which was of course objected to.

Since87 · 01-11-2003, 02:40 PM

Quote:

Originally posted by bigben2k

In the mean time, OC forums killed an idea to sponsor DIY blockmakers, for a round of tests by Bill Adams, but in the middle of it all, there was some info about Radius that I thought I'd repost here:

No, the sticky for requesting sponsorship is still there.

The idea of trying to collect donations from members for a prize is pretty well dead though.

bigben2k · 01-11-2003, 02:55 PM

Well, I almost lost that, in all the confusion!

Sponsor thread

As posted, I'm not solliciting any pledges at this time, because the thing ain't built yet!

Les · 01-11-2003, 03:17 PM

Quote:

Originally posted by bigben2k
.

Les estimates a head of 35 to 50 feet, at 4 gpm.

No.
"35 to 50ft" at 40fps and non-consequentially 4gpm ( 0.031 sq in X-section).

12-31-2002, 09:27 AM	#329
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	As BillA pointed out, I've tried to incorporate the best of the best into one block, and finding out that it may not be possible to have everything. In any rig, there is certainly a limit to the cooling ability, and it is dictated mostly by the "gradients" that BillA previously mentionned. In short, for as long as there is any kind of material between the heat source and the cooling fluid, there will be a difference in temperature, but more importantly there will be a hot side and a cold side to the baseplate. The baseplate is what really defines the limitations of the temp (along with the TIM joint, of course). One thing that Radius does well is optimize the flow speed for the hot spot (over the core). You just can't get better than that (jet inpingement excluded). The baseplate is at a minimal thickness (2mm, which may be reduced to 1mm), which is just about optimal for the fin pattern, but not necessarily the thinnest of all design, and that's one of the limitations of Radius: White Water has continuous fins, where most of Radius' fins are "broken", for flow reasons. I've specified that the baseplate will also have tiny holes over it, about the size of a 1 mm drill bit head, similar to the drill-press blocks and the Swiftech unit: it increases the surface area. One thing that is unique, is the seperating res. Many people have thought about making something like it, but I've never seen one. It's advantaged by minimal restrictions, but disadvantaged by it's enormous size. As for the pump heat, the motor is isolated (direct-drive of a magnet assembly), so that's not an issue (except that it will need a good vent). The pumping heat is still somewhat nebulous to me, as I'm finding out that there's heat from the propeller, as well as from the restrictions throughout the loop, which means that the nozzle at the entrance of the block will generate heat, but it should be relatively small compared to the heat that it will take away. In any case, I am trying to achieve a Reynolds number (a dimensionless number indicating wether the flow is laminar or turbulent) in excess of 4000, which corresponds with turbulent flow, which is needed for the extra cooling ability. Of course I would need to optimize the flow and block so that the flow is turbulent only in the critical area: over the core. So far, it looks like my pump can't achieve that on its own, and so turbulators will have to compensate for what the pump can't do. Also, as has been pointed out earlier, it's far more efficient to add turbulators, than to use a very powerfull pump. I'll be trying out a CFD program, to see if I can graph the flow within Radius. It should at least help me optimize the outlet for a balanced flow. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

01-02-2003, 01:00 PM	#335
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	He can't, it's already a patented concept, but he has copyright on the assembly, as a whole. Out of all the waterblocks out there, they can generally be categorized: ~flat-plate (Swiftech), labyrinth (Maze, Spiral), etc... We're at a point where some of these concepts have to be combined, in order to get more performance. In other words, you're not going to be able to tweak a Maze3 to get a couple more degrees, without adding something to it (fins, jet,...). Cathar successfully combined a fast flow (not quite jet) and micro-fin. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

01-06-2003, 08:01 AM	#339
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Well, seeing that I'm building it for Barton, and that it isn't here... your guess is as good as mine! Maybe a month or two. I'm building this thing on the weekends, and since life is pretty busy here, I get maybe one half day, every other weekend. I do however have a lot more time to think about it! Yesterday, I realized I made a measurement fault, so I've got to cut another piece of tube, for the part between the screw/barb cap, and the clear flexible elbow. Lucky for me, I've got 4 feet of the stuff... I've also started to look into the hardware needed for the mount. I'll go with the idea Utabintarbo had, which is a set of outriggers. Screws have to be 6-32, since the (new) hole size is now 0.150 inch. Once I've figured out which springs to use, I can spec the screw, and place my order with McMaster. Right now, I still need to figure out which epoxy to use for polypropylene (the 1 1/2 barb and tube connection), since the plastic welder is useless. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

01-06-2003, 10:35 AM	#340
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	polypropylene adhesive Been googling... and I may have found a product that might allow me to bond polypropylene: http://www.nbond.com/techdata.html The problem I had came from this fact: polypropylene is a "low energy plastic", aka a non-stick plastic, much like Teflon. http://www.nbond.com/faq.html#lowen On the plus side, cyanoacrylate (superglue) might work too. Still searching... edit: e-mailed 3M and Lepage. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101 Last edited by bigben2k; 01-06-2003 at 11:10 AM.

01-07-2003, 09:05 AM	#341
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Ok... 3M says they don't have anything that'll withstand methanol, but otherwise recommend DP-8005 or DP-8010, Structural Plastic adhesive. They sent me a PDF for each one. One note: "However, these adhesives have .008" microspheres (glass bubbles) so you would have to design the parts; ID and OD to accept the microspheres." I e-mailed Loctite. No reply from Lepage. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

12-31-2002, 09:05 PM	#331
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Thanks bigben for the quick summary. You still didn't directly answer my question but I think that is because you are not sure yet either. Glad you are making progress on this. It will be very interesting to see how it comes out.

01-02-2003, 12:44 PM	#334
hara Cooling Savant Join Date: Oct 2002 Location: Malta, Mediterranean Posts: 662	So, whitewater is unbeatable? Cathar should patent.

01-05-2003, 01:06 AM	#337
LiquidRulez Cooling Savant Join Date: Sep 2002 Location: In Hell Posts: 322	SO when's it going to be done Ben? Cant wait to see it assembled and ready to mount.

01-07-2003, 11:10 AM	#343
Puzzdre Cooling Savant Join Date: Aug 2002 Location: Croatia Posts: 969	Looks beautiful!!! Do you have the clamping mechanism ready or in progress? __________________ 'Out of cheese error... ...please reboot the universe (press the GBL to continue)'

01-07-2003, 03:55 PM	#345
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Devcon says: "We don't have a solution, but consider abs or nylon, instead of polypropylene, in which case you can use the Plastic Welder". It's actually a good idea... __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

01-08-2003, 09:34 AM	#346
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	From Loctite: "PP is the more difficult of your two substrates to bond. Our best solution is to apply Primer 770 to the PP, then try one of our cyanoacrylates, such as 401. This should give you a good bond, but you would need to test for the fluid exposure, as cyanoacrylates do not typically have the best long-term resistance to water. Another possibility with better fluid resistance, but lower adhesion to PP, would be a UV acrylic, such as 3105. The datasheets can be downloaded from our website." __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

01-11-2003, 02:24 PM	#347
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	For those still in the dark...(sorry!) I'm opting to replace the polypropylene barb, with a nylon one. In the mean time, OC forums killed an idea to sponsor DIY blockmakers, for a round of tests by Bill Adams, but in the middle of it all, there was some info about Radius that I thought I'd repost here: 1-My pump may only deliver a pressure equivalent to 9 feet of head, at 250 gph. There seems to be many versions of the 2-MDQ-SC, I guess I hoped that mine was one of the best, at 10.5 feet. 2-I estimated the flow/pressure at 4 gpm, and 11 feet of head (through the block): if you've been following this thread, you know that this only referred to the inlet (still incorrectly), and not the flow inside the block, past the nozzle, so the block is actually a lot more restrictive. Les estimates a head of 35 to 50 feet, at 4 gpm. Total heads loss (at 4 gpm): 3/4 ID tubing: 0.62" of head, per foot of tubing rad: n/a (when it goes to a chiller) res: about 2" of head 3-That nozzle (above) was to be 3/16. I've since decided/recalculated to go with 3/8. Either way, I'll be ordering a polycarbonate rod, 5/8 ID (to fit the inner res tube) so that I can try different nozzles. 4-More importantly, Bill pointed out that viscosity changes significantly at a lower temp, to which I replied that using the Windshield Wiper fluid will make little to no difference, at least according to Scott Gamble (corrections notwithstanding). 5-Bill has pointed out (correclty) that I have no modelling software. I do plan to use one, when I can find it (free CFD anywhere?) and have the time to mess with it. Now all this started when I called Radius a high flow/high pressure block, which was of course objected to. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

01-11-2003, 02:55 PM	#349
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Well, I almost lost that, in all the confusion! Sponsor thread As posted, I'm not solliciting any pledges at this time, because the thing ain't built yet! __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)