|
|
General Liquid/Water Cooling Discussion For discussion about Full Cooling System kits, or general cooling topics. Keep specific cooling items like pumps, radiators, etc... in their specific forums. |
Thread Tools |
03-16-2005, 10:56 PM | #126 | |
Cooling Neophyte
Join Date: Dec 2004
Location: Geelong, Australia
Posts: 16
|
Quote:
Patriots and Majors are nice, but they are loud, even when undervolted. Edit: Very nice rad though. Last edited by phextwin; 03-17-2005 at 12:43 AM. |
|
03-17-2005, 12:31 AM | #127 | |
Thermophile
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
|
Quote:
The fan that I suggested that be matched with was the Papst 4412FGL, while the ThermoChill 120.2 in comparison was using a pair of Papst 4312L fans. The shroud really only needs to be 15 degrees off the vertical, but 12.5 degrees is fine too. That figure, however, is an "optimal" shroud depth - meaning that is as good as it gets. Could happily get away with a 45 degree angle and only sacrifice maybe 5% of air-flow performance - heck - it may even be less than that. I mentioned the 15 degree angle thing purely because that is an industry accepted ideal. If the shroud is going to be obstructional, which it clearly appears to be in your small tower case, then by all means make it smaller. I had originally envisaged that such a radiator would be placed in the lower front of a case, and the shroud would just be deep enough to suit. |
|
03-17-2005, 12:39 AM | #128 |
Cooling Neophyte
Join Date: Nov 2002
Location: Hollister, CA
Posts: 44
|
Very nicely made radiator. Keep it up!
Can't wait to see some test data. *lurks on...* |
03-17-2005, 05:06 AM | #129 | |||||
Cooling Savant
Join Date: May 2003
Location: Wakefield, West Yorkshire, UK
Posts: 486
|
Quote:
HAL: Of course.... WACC completely slipped my mind!! Cheers! Quote:
Quote:
Quote:
Quote:
Based on O-CuK sales in the UK... Last edited by Marci; 03-17-2005 at 05:11 AM. |
|||||
03-17-2005, 08:07 AM | #130 |
Cooling Savant
Join Date: May 2003
Location: Wakefield, West Yorkshire, UK
Posts: 486
|
VERY loose idea of performance... FX51 setup
FX51 overvolted to 1.75v but at stock clocks (2.2Ghz), s940 1Gb ECC Reg 3200 Ram Asus SK8V OCZ PowerStream 520w Ambient measured as air into radiator, 2" from core surface. CPU Core measured via bios. ( ) ----------------------------------------------------------------------------- HE120.3 with 3x Xinruillian 0.40a fans at 7v. Specs are (at 12v): # Current - 0.4 # Speed - 2400RPM # Airflow CFM - 91.63 # Noise dBA - 40 # Depth - 1" Delta between ambient and cpu core: 4 deg C Fans are still audible in this scenario, but only just... ----------------------------------------------------------------------------- PA160.1 with 1x Xinruillian 0.18a @ 12v, specs are: # Current - 0.18 # Speed - 1700RPM # Airflow CFM - 65.12 # Noise dBA - 30 # Depth - 1" Delta between Ambient and CPU Temp: 6 deg C Fan is completely inaudible... for all intents and purposes you can't tell it's there unless you stick your ear RIGHT beside the fan... ----------------------------------------------------------------------------- That's literally remove one rad, hook up the other.... same tubing, same case, same everything else... ambient at a constant 28 deg C (fairly warm) due to central heating. But to be honest, as y'all know those results mean nothing conclusive... (Just for a VERY loose meaningless idea... can't provide more accurate test data until I send em over to JoeC; awaiting a reply from him at the mo) Last edited by Marci; 03-17-2005 at 08:14 AM. |
03-17-2005, 08:27 AM | #131 |
Cooling Savant
Join Date: Oct 2004
Location: Canada
Posts: 260
|
Was just reading this.
We make heat exchangers, there copper core with plastic tanks and folded fins for 80mm fans. There produced just like automotive units, and I believe are very inexpensive given how many are produced. We make them for an OEM customer (not for PC cooling), but maybe we can mod them for you guys. Think we will just need different tanks. I'll ask around. |
03-17-2005, 12:59 PM | #132 |
Thermophile
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
|
If were you after some semblance of meaningful data for yourself, put the CPU under full overclocked load, and measure the air-water delta for either scenario with at least a 0.1C accuracy calibrated thermometers for both the air and water temps.
It won't be super-spiffy accurate like the engineers are doing, but should give you a better rough ballpark of what's going on. Don't measure radiator performance by looking at the CPU temps, especially with the AMD on-die diodes - those things are pretty whacked from boot to boot, depending on the mobo of course. |
03-17-2005, 07:33 PM | #133 | |
Cooling Savant
Join Date: Jan 2004
Location: Horsham, UK
Posts: 140
|
Quote:
__________________
Member of the paramilitary wing of CAMRA
|
|
03-17-2005, 09:28 PM | #134 |
Cooling Savant
Join Date: Dec 2004
Location: Orlando, Florida
Posts: 383
|
My theory as to why Cath's idea works so well:
1. A fan will produce maximum flow when it is in open air (100% air available to it)
2. A radiator, such as a BIX, hinders air flow. For purposes of discussion we will say that air flow is reduced to %50 of the fans maximum. 3. Having a larger core, in this case 16x16cm, allows for more area through which air can enter the fan. At the same time it provides more frontal area. Through maintaining adequate airflow and increasing frontal area the larger 16x16cm core can be said to be "optimized". I would be curious to see how it performs against a BIX with a higher CFM fan. I wouldn't be surprised if the performance increase provided by the 16x16cm core didn't scale linearly over its BIX counterpart. In other words, I wouldn't be surprised if a 16x16cm core wasn't optimized (to some point) for any 12cm fan regardless of airflow. Anyone following me here? :shrug: :shrug: :shrug: If you think you are missing my point or misinterpreting it let me know. Assuming that 16x16cm are the optimized HxL demensions for a 12cm fan, one could simply vary depth to suit fans of different air pressure (some can translate this as air flow). Higher pressure fans would work better with deeper cores. Likewise lower pressure fans would work better with more shallow cores. Last edited by maxSaleen; 03-17-2005 at 09:35 PM. |
03-17-2005, 10:42 PM | #135 |
Thermophile
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
|
Could either vary depth, or FPI. FPI would allow for finer grain control, since depth is done in units of an entire core-row. Would speculate that it would take an extremely meaty 12cm fan such that a high-FPI (24-30) single-row core wasn't enough such that moving to a dual-row lower-FPI core was really necessary.
From my limited comprehension of the relationship between fans, performance, and pressure, it would seem to me that double-row cores (deeper cores) very, very rarely make sense for axial fans that aren't in danger of giving their user's tinitis. It would seem to be far better to have an "extreme" model that just uses a higher FPI, rather than a deeper core. Would cost less too I would imagine. |
03-18-2005, 01:23 AM | #136 | |
Cooling Neophyte
Join Date: Aug 2002
Location: London
Posts: 96
|
Quote:
4312L http://uk.farnell.com/jsp/endecaSear...U=719742&N=401 4312F/2GL (with rpm) http://www.alphacool.de/perl/shop.pl...6&art_id=24602 |
|
03-18-2005, 06:00 AM | #137 |
Cooling Savant
Join Date: May 2003
Location: Wakefield, West Yorkshire, UK
Posts: 486
|
Cheers for the fan linkage folks... shall be getting those ordered somepoint today...
Yep, would be cheaper and certainly easier to increase the FPI rather than increasing core depth. You only have a limited amount by which you can increase the core depth before it has a knockon effect on the sidewalls and tanks, ie: you'd have to increase those as well. At the moment as it stands, including the shroud as shown in the pix, costs would leave it retailing at the same price as the current ThermoChill HE120.2 - Any more expensive and I doubt the product would sell - it'll be hard enough to shift it to the US market due to Exchange Rate as it is... so anything that would increase it's current cost would probably be a bad move... As it goes, I think as long as quantities were maintained between the 14fpi model and higher FPI model then costs wouldn't be THAT much different. At the moment there is clearly a hole in the market for a rad that performs in silence... the hole for screamers is already filled by current offerings. I guess what I'm saying is, I personally have to be careful not to make the original ThermoChill HE120.x series rads redundant... or I have to make the choice to discontinue the current HE120.x series and replace it entirely with an optimised PA160.1 series where area is constant and only FPI changes... HOWEVER, I don't think the market could accomodate us only producing 160mm rads purely due to the size and space requirements.... that would be shooting ourselves in the foot... If we were positioned elsewhere so that economics were on our side to make it easier to sell to an international market, then perhaps we could start a revolution of moving folks over to 160mm rads, but with the limitation of only being able to realistically sell inside of Europe.... At the moment these would be aimed at a specific sector of the market - those wanting silence. Those who aren't bothered about the noise will likely just choose any other rad on the market for size-based convenience (? opinions ?) and cost-based convenience (go compare price of a HE120.2 vs a BIXII from DangerDen...) The scales of balance are bouncing from side to side at the mo... |
03-18-2005, 06:52 AM | #138 |
Thermophile
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
|
Marci, been plugging through the calculations with a flow calculator and the shroud angle.
10-15 degrees shroud angle from the vertical is optimal for the 16x16 -> 12x12cm shroud, with a pressure drop of 1.9Pa given a 50CFM air-flow, for which 35-50CFM is about the sorts of air-flows we're targetting with the low-noise optimised design. A 30 degree shroud angle (4cm shroud depth), yields just a 2.0 Pa pressure drop. A 45 degree shroud angle (2cm shroud depth), yields a 2.1 Pa Pressure drop Given that the sorts of low-noise fans we're talking about are generating 10Pa air pressures at the sorts of air-flow rates we're talking about, then the variation here is trivial at best. We'd only be talking about impacting air-flow rates by 1% by converting the shroud into a very convenient and compact 4cm depth device (fan mounted on outside), and by 2% from "optimal" if the shroud was an even more compact 2cm depth (fan on outside). In fact, even moving to a super-compact 1cm depth yields just a 3% loss of air-flow from the "optimal" depth. I would however suggest that 4cm (30 degree angle from vertical) would be a very nice shroud depth for both the blow and suck fan orientations, and still make for a very compact design. I feel somewhat responsible that you had that monstrosity of a shroud made up based upon my recommendations. Had I known that you were going to do so, I would have munched the math sooner to realise what the impact would be with shallower (and hence more compact and user-friendly) shroud sizes. I also had not realised that you were the manufacturer for the Thermochill radiators. Am happy to engage someone who is serious about seeing these things made. As for your market, I would say that it would extend to pretty much any nation with a healthy economy outside of the USA (i.e. including Australia and New Zealand, as well as parts of SE Asia). Down here we would have plenty of buyers for such a radiator. Would also suggest that for ease of mounting, since we're considering this as a serious commercial product, that the shroud be avaible for both side, so users could easily attach the radiatior to a 12cm fan mount point in one of many of the modern cases that now support such. |
03-18-2005, 06:55 AM | #139 |
Cooling Savant
Join Date: May 2003
Location: Wakefield, West Yorkshire, UK
Posts: 486
|
Heh... we're working on the same thing from opposite ends... I'll see ya in the middle in a mo - just got a pic to upload...
|
03-18-2005, 07:02 AM | #140 | ||
Cooling Savant
Join Date: May 2003
Location: Wakefield, West Yorkshire, UK
Posts: 486
|
Quote:
Quote:
Not a problem at all... rad will be supplied with one shroud by default, 2x Nickel Plated 3/8" BSP > 1/2" Hose barbs... individual shrouds will be available separately for those that want push > pull. As it stands at the mo, if you look in the pic above you can see the self-tapper that attaches the shroud to the rad. If you look just above that screw head on the sidewall there's the same hole on the opposite face so that shroud can be mounted on either face or 2x shrouds used simultaneously... I've always beleived in giving the customer the final choice on what side of the rad they wanna use if only using the one... Extrapolated the angles for 25 deg and 45 deg to see what space saving it would give with intention of posting the above and asking for airflow calcs etc to assess suitability... Will get 30 deg / 40mm shroud made up! Ta d00d! Last edited by Marci; 03-18-2005 at 07:20 AM. |
||
03-18-2005, 07:11 AM | #141 |
Cooling Savant
Join Date: May 2003
Location: Wakefield, West Yorkshire, UK
Posts: 486
|
Should be done somepoint beginning of next week... and now I've got a motorbike to repair after dropping it this morning doing 60 round a roundabout
|
03-18-2005, 07:14 AM | #142 |
Cooling Savant
Join Date: May 2003
Location: Wakefield, West Yorkshire, UK
Posts: 486
|
Could also be useful to produce a list of suitable fans for use with it from other manufacturers as well as PAPST as an alternative in case of market availability... These Xinruillian cheapy fans I muck about with are silent in the 0.18a format but have no info on generated air pressure... will poke about and see what I can find out...
|
03-18-2005, 07:18 AM | #143 |
Thermophile
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
|
Do you have access to the Yate-Loon/Tricod fans over there, which are slightly modified, rebadged and sold as the Nexus brand fans?
These are perhaps THE fan of choice when it comes to low-noise radiator fannage. Not quite as powerful as the Papst 4412FGL's (~80-85% at a guess), but certainly MUCH quieter. If you're having trouble finding them locally, I can send some over to you. |
03-18-2005, 07:28 AM | #144 | |
Cooling Savant
Join Date: May 2003
Location: Wakefield, West Yorkshire, UK
Posts: 486
|
Ah.... Nexus... yes, most likely.... will find a link for you to check...
EDIT: http://www.kustompcs.co.uk/acatalog/...t_Fans_72.html But 120mm out of stock - have just e-mailed Graham for an ETA but usually a month or two, so if you could get me some any quicker that'd be most appreciated... Quote:
Last edited by Marci; 03-18-2005 at 07:55 AM. |
|
03-18-2005, 07:36 AM | #145 |
Cooling Neophyte
Join Date: Dec 2002
Location: Twin Cities
Posts: 67
|
I don't understand the obsession with 120mm fans. There are several OEMs of 180mmx50mm fans like the Patriot. While they are a tad big for the "Chevette" 150mm cores, they are a perfect match for the "Camaro" 180mm cores. The 24V Patriot, run at 12V is very quiet. With the fan and radiator having the same OD, you don't have to make a sloped shroud. Straight wall works perfectly. Given the tolerance of the Patriot for resistance, only one is needed in a pull configuration. I've tried two in push-pull with no improvement on heat exchange though they are louder. I have my shroud set at 25mm spacing from the intake side of the blades to the fins of the radiator. The Camaro radiator cost the same and sometimes is cheaper than the Chevette one, depending upon where you get it and it has 12mm ID tubes on it already. Despite the Camaros greater dimensions, it still fits inside a tower, though not offering much "wiggle room". You can pick up the 24V Patriots from several surplus houses here in the states for around $12US new or $10US used. Surely other countries have surplus houses? Food for thought...
Hoot
__________________
DFI LP UT NF4 Ultra-D 7.04-3 Bios A64 3700+ San Diego 11x257 (2827 Mhz) @ 1.57Vcore 2x1Gb Crucial Ballistix 7:6 (300 Mhz) 3-3-3-8-1T @ 2.77V Connect 3D x850xt 620/625 2x WD 36GB Raptors in Raid0 DIY Jet Impingement WB on CPU / Pipecap WB on GPU Shrouded Camaro Heater Core / Swiftech MCP655 Pump Antec P-160 Tower / Seasonic S12 600W PSU MAXIM 6658 Thermal Diode Reader Moderator at OCForums |
03-18-2005, 07:43 AM | #146 |
Thermophile
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
|
How many were you after?
They cost £4 each here (incl. tax), with shipping to the UK being around £2 each, unless you're talking about dozens of them. Alternately I'll ship you a quantity in exchange for a PA160.1 core. Unsure about shipping arrangements to Australia for you - typically these things are done by weight here in Oz, unless you're talking about "buying" a portion of a shipping container. |
03-18-2005, 07:45 AM | #147 | |
Thermophile
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
|
Quote:
Need to remember Hoot, we're talking about the near-silence obsessed here. Fans that you can barely hear (or not hear at all) in a quiet room, with everything else turned off, in the middle of the night. |
|
03-18-2005, 07:58 AM | #148 |
Cooling Savant
Join Date: May 2003
Location: Wakefield, West Yorkshire, UK
Posts: 486
|
Expect an e-mail from me imminently Cathar....
EDIT: YGM Last edited by Marci; 03-18-2005 at 08:09 AM. |
03-18-2005, 08:01 AM | #149 |
Thermophile
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
|
On that low-noise topic, the "goal" here is to approach something of a low-noise utopic setup. I would suggest:
Laing DDC (with optional inlet modification) The PA160.1 core as defined above A single or double-push-pull Yate-Loon/Tricod/Nexus/Papst 4412FGL fanned setup attached to a Rheobus 3/8" or 7/16" ID tubing A CPU waterblock of moderate pressure drop that performs well at ~5LPM which is the sort of flow rate we'd be expecting with said pump/radiator/tubing Such a setup would be marrying the multi-faceted goals of near top-end water-cooling performance in a compact(ish) and near silent (read as: have trouble distinguishing if system is turned on at all in a quiet room) format. Sure, can knock off a further 3C or so off a 100W CPU by boosting pump + radiator size, but could still achieve close to that in near silence as well by adding a second PA160.1 radiator. i.e. make water-cooling be all that it can be. |
03-18-2005, 08:14 AM | #150 | |
Cooling Savant
Join Date: May 2003
Location: Wakefield, West Yorkshire, UK
Posts: 486
|
Hmmm.... can anyone remember where the ThermoChill CPU Blocks fell in the testing tables?? I know they were more suited to low flowrate 1048ish pumps than 1250 pumps... but would they fulfil this criteria? We have loads of em that never sold for S462 / s478 that I was contemplating getting tops made for to conver tup to A64 / LGA775 to try and move the last of the stocks and then discontinue, but if the block fulfils this criteria then it may be a case of making the changes then putting them back into production and remarketing...
Obvious aim that I'd been looking towards here is ThermoTube res, PA160 rad, DDC or CSP-Mag Pump, ThermoChill Block so that we finally have a 1/2" (or 7/16" if I can find all the right barbs) based British Manufactured kit that has been made to the specifications of those who know, which was one of the original reasons for ThermoChill existing in the first place... Quote:
|
|
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
|
|