Reserator with extra radiator for GPU
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
Reserator with extra radiator for GPU
I am considering a WC loop for my Phenom 965 and 4870.
I know the reserator will not cope with both in the same loop, and I don't want to buy 2.
However, would the reserator pump be able to handle a 2nd internal radiator with 2 slow speed fans (using 2 of my S-Flex Es so not adding any noise to the current setup)?
I could then have the loop something like this:
Reserator -> CPU block -> Radiator -> GPU/RAM -> Reserator
I know I could just buy a custom loop, but I like the idea of the reserator, and if my above idea works I can get the basic kit now and add the GPU side of things later.
So would a reserator cool a stock (undervolted to 1.375v) Phenom 965 C2, and then would the pump handle a second radiator?
I can find the following info so far:
Integrated Water Pump : 5 W, Qmax 300 l/h
Maximum Lift : 0.5 m
But don't know what the means!
I know the reserator will not cope with both in the same loop, and I don't want to buy 2.
However, would the reserator pump be able to handle a 2nd internal radiator with 2 slow speed fans (using 2 of my S-Flex Es so not adding any noise to the current setup)?
I could then have the loop something like this:
Reserator -> CPU block -> Radiator -> GPU/RAM -> Reserator
I know I could just buy a custom loop, but I like the idea of the reserator, and if my above idea works I can get the basic kit now and add the GPU side of things later.
So would a reserator cool a stock (undervolted to 1.375v) Phenom 965 C2, and then would the pump handle a second radiator?
I can find the following info so far:
Integrated Water Pump : 5 W, Qmax 300 l/h
Maximum Lift : 0.5 m
But don't know what the means!
-
- Posts: 324
- Joined: Wed May 07, 2003 10:35 pm
- Location: Surrey, B,C
Based on the info I can find, the reserator's pump is rated at 300l/h and 0.5m head. That means that it can move up to 300 liters every hour. It can also lift the water 18" up. The actual rate in your system will be different, based on how restrictive the components are, but it can get a bit complicated.
Essentially, the reserator's pump is weaker than the MCP350; rated at 3.3m head and 440l/h. More head means that the system can deal with more components.
The pump would at least be able to run for a while with the components in it. I personally think the system would be able to boot and run light Windows programs. It may be good for some mildly strenuous work. You can always purchase a secondary pump later.
The Reserator is capable of dealing with over 200w of heat. The CPU alone should be fine.
An alternative to adding a second radiator that you can try with what you have on hand is to put a fan on top of the Reserator, blowing up, and add a shroud around the top half. Forcing air over the fins will increase the heat dispersal. The issue with this is the ugliness that you'd be applying to your Reserator.
Essentially, the reserator's pump is weaker than the MCP350; rated at 3.3m head and 440l/h. More head means that the system can deal with more components.
The pump would at least be able to run for a while with the components in it. I personally think the system would be able to boot and run light Windows programs. It may be good for some mildly strenuous work. You can always purchase a secondary pump later.
The Reserator is capable of dealing with over 200w of heat. The CPU alone should be fine.
An alternative to adding a second radiator that you can try with what you have on hand is to put a fan on top of the Reserator, blowing up, and add a shroud around the top half. Forcing air over the fins will increase the heat dispersal. The issue with this is the ugliness that you'd be applying to your Reserator.
-
- Posts: 324
- Joined: Wed May 07, 2003 10:35 pm
- Location: Surrey, B,C
-
- Posts: 324
- Joined: Wed May 07, 2003 10:35 pm
- Location: Surrey, B,C
Yeah was gutted!
Still I have spent a few hours stripping down the PC removing all the now un-needed fans - sticking to 2 S-Flex Es - 1 intake, 1 exhaust to keep the NB/SB cool.
Two further questions:
1. If I go for a 2nd internal radiator, which is a good one to go for to work alongside the low power pump and 2 slow fans (S-Flexs @ 600 rpm).
2. Where can I get distilled water??? I bought 3L of de-Ionized water today before realizing it was the wrong stuff. I don't want to bugger up the loop before I even start.
Still I have spent a few hours stripping down the PC removing all the now un-needed fans - sticking to 2 S-Flex Es - 1 intake, 1 exhaust to keep the NB/SB cool.
Two further questions:
1. If I go for a 2nd internal radiator, which is a good one to go for to work alongside the low power pump and 2 slow fans (S-Flexs @ 600 rpm).
2. Where can I get distilled water??? I bought 3L of de-Ionized water today before realizing it was the wrong stuff. I don't want to bugger up the loop before I even start.
Connected the Reserator up with old M/B for leak testing etc:
Images don't show the Essence STX in the bottom PCI-E slot as I didn't want to risk it getting wet in case of a leak.
So far so good - took a while to get all the bubbles out, ended up removing the card and CPU block and standing the reserator up higher than them, and also tilting the reserator 40 degrees in each direction - seems to have cleared all the bubbles out.
Can't wait for motherboard to arrive tomorrow so I can get it all in properly and tested!
Images don't show the Essence STX in the bottom PCI-E slot as I didn't want to risk it getting wet in case of a leak.
So far so good - took a while to get all the bubbles out, ended up removing the card and CPU block and standing the reserator up higher than them, and also tilting the reserator 40 degrees in each direction - seems to have cleared all the bubbles out.
Can't wait for motherboard to arrive tomorrow so I can get it all in properly and tested!
New board arrived today - all fitted
First impressions - amazing - it idled at 30c CPU and 19c GPU with about 20c ambient.
But after a bit of gaming / OCCT testing the CPU idles around 40c and GPU about 30c.
Max temps have been around 60c for CPU and 40c for GPU.
I think I am going to go down the 2nd radiator option, but just need to decide on which one, and wether a single or double is best to go for.
Considering the XSPC RX series as they are designed for low-speed fans. Did consider Swiftech but their suggested fan speed is quite high.
How much of a temp drop can I expect from adding a single or a double rad?
First impressions - amazing - it idled at 30c CPU and 19c GPU with about 20c ambient.
But after a bit of gaming / OCCT testing the CPU idles around 40c and GPU about 30c.
Max temps have been around 60c for CPU and 40c for GPU.
I think I am going to go down the 2nd radiator option, but just need to decide on which one, and wether a single or double is best to go for.
Considering the XSPC RX series as they are designed for low-speed fans. Did consider Swiftech but their suggested fan speed is quite high.
How much of a temp drop can I expect from adding a single or a double rad?
-
- Posts: 324
- Joined: Wed May 07, 2003 10:35 pm
- Location: Surrey, B,C
Suggested fan speed doesn't really mean anything, XSPC's radiator is almost identical in performance to the MCR series with fans of the same speed and power. Buy whichever is cheaper.
Since you already have a reserator, performance will be sufficient with a single fan radiator.
I would expect a drop of 5ºC to 10ºC. YMMV.
Oh, your GPU temps aren't accurate, 19ºC with an ambient of 20ºC is impossible. CPU temps seem to be more accurate, what were your previous temps?
Since you already have a reserator, performance will be sufficient with a single fan radiator.
I would expect a drop of 5ºC to 10ºC. YMMV.
Oh, your GPU temps aren't accurate, 19ºC with an ambient of 20ºC is impossible. CPU temps seem to be more accurate, what were your previous temps?
Last edited by 1398342003 on Sat Apr 24, 2010 3:30 pm, edited 1 time in total.
+11398342003 wrote:Suggested fan speed doesn't really mean anything, XSPC's radiator is almost identical in performance to the MCR series with fans of the same speed and power. Buy whichever is cheaper.
.
The difference in performance at low (<1000 rpm) fan speeds is so small amongst all the big names its not worth worrying about.
-
- Posts: 324
- Joined: Wed May 07, 2003 10:35 pm
- Location: Surrey, B,C
-
- Posts: 324
- Joined: Wed May 07, 2003 10:35 pm
- Location: Surrey, B,C
so its the rx or swiftech, whichever is cheaper or available.
What barbs should i use? Cant seem to find any concrete info on if 8/10, 8/11 or 8/12 will fit or if all will?
Also is there any issue with the type of metal used for rad/barbs? I am using the standard zalman fluid plus deionised wate (quietpc said deionized is ok to use).
Is there any limitation to position of the rad in terms of flow rate? Is it ok to mount at the bottom of the case, below the other components,?
What barbs should i use? Cant seem to find any concrete info on if 8/10, 8/11 or 8/12 will fit or if all will?
Also is there any issue with the type of metal used for rad/barbs? I am using the standard zalman fluid plus deionised wate (quietpc said deionized is ok to use).
Is there any limitation to position of the rad in terms of flow rate? Is it ok to mount at the bottom of the case, below the other components,?
Use whatever barbs you like, as long as it has a G1/4 thread it will fit. I don't know what your tubing size is... but again, use what works, any G1/4 barb will fit in the rad (I think the swiftech might come with barbs from memory).
Mount it anywhere, no issues at all with where you mount it (I have one at the top and one at the bottom of my loop). Some positions make bleeding/draining slightly harder, but no big deal.
Mount it anywhere, no issues at all with where you mount it (I have one at the top and one at the bottom of my loop). Some positions make bleeding/draining slightly harder, but no big deal.
I am not sure of the exact ambient temp - our flat is quite cold usually unless we put the heating on, and I run these tests without the heating, so it could be as low as 17c - not sure.Oh, your GPU temps aren't accurate, 19ºC with an ambient of 20ºC is impossible. CPU temps seem to be more accurate, what were your previous temps?
The following is a OCCT GPU test result from yesterday:
This is a OCCT PSU test (GPU plus CPU Linpack test) with my modified Zalman Cooler with both 92mm fans on 100%:
And the same cooler with fans on their minimum (silent) speed, think it was 1150rpm:
I did run a full OCCT PSU test yesterday, but the CPU topped out at 72c!!!! Thats 10c over the max temp of the chip
Ordered:
XSPC Dual Radiator Black - RX Series : RX240 with 3/8" Barbs
Is it possible to mount this at the bottom of my 690-II case, with the fan on top to draw air through the radiator, or do I need to have the fan/s mounted under the radiator? If so I will mount both GTs under the rad and have an S-Flex as my rear exhaust fan.
XSPC Dual Radiator Black - RX Series : RX240 with 3/8" Barbs
Is it possible to mount this at the bottom of my 690-II case, with the fan on top to draw air through the radiator, or do I need to have the fan/s mounted under the radiator? If so I will mount both GTs under the rad and have an S-Flex as my rear exhaust fan.
Going for the Rad mounted at the base of the case with 2 GTs running at 930rpm (6v) giving 17 CFM each and should be silent. Will then use an S-Flex E at 660 rom (7v) giving 15.5 CFM as an exhaust fan.
Been running some tests with the reserator with two sflexs zipstrapped together and then stood on their end blowing cold air at the tower. This lowered temps quite well, although OCCT linpack still came up with errors. I updated the BIOS on the motherboard and the errors went away, so thats that one solved.
OK, this is the OCCT Linpack graph for no fans, just the big ol Reserator:
This is the S-flex contraption, both fans set at 780rpm (19CFM each):
And the OCCT Linpack chart:
Just running a final test with the two S-flexs on full blast (1200rpm = not quiet!)...
Been running some tests with the reserator with two sflexs zipstrapped together and then stood on their end blowing cold air at the tower. This lowered temps quite well, although OCCT linpack still came up with errors. I updated the BIOS on the motherboard and the errors went away, so thats that one solved.
OK, this is the OCCT Linpack graph for no fans, just the big ol Reserator:
This is the S-flex contraption, both fans set at 780rpm (19CFM each):
And the OCCT Linpack chart:
Just running a final test with the two S-flexs on full blast (1200rpm = not quiet!)...
OCCT Linpack - External S-Flex fans on 100% -1200rpm
No external fans - max temp - 63c
Silent external fans - max temp - 56.5c
Loud external fans - max temp - 56.5c
Well that shows quite a difference from just having some slow speed fans blowing cold air on the resorator, and that increasing the fans above silent levels made no difference to max load temps (well at least for a 30 min test, didn't have enough time to let the tests run a little longer).
Hoping by adding the rad and having the GTs running silent, those temps will drop a little more. Would like that to be 50c max, but am a little doubtfull.
No external fans - max temp - 63c
Silent external fans - max temp - 56.5c
Loud external fans - max temp - 56.5c
Well that shows quite a difference from just having some slow speed fans blowing cold air on the resorator, and that increasing the fans above silent levels made no difference to max load temps (well at least for a 30 min test, didn't have enough time to let the tests run a little longer).
Hoping by adding the rad and having the GTs running silent, those temps will drop a little more. Would like that to be 50c max, but am a little doubtfull.
-
- Posts: 324
- Joined: Wed May 07, 2003 10:35 pm
- Location: Surrey, B,C
Well the reserator does hold a lot more water than most WC systems, so it'll take longer to get to the maximum temperature. Without the fans you may have a thermal runaway problem. The max dissipation of the reserator may not be enough to keep the processor cool over a several hour test.
You should try Prime95 for your testing if you're pressed for time. It doesn't give the CPU a 2 minute break every 8 minutes.
Also, if you look carefully at the high/low RPM tests, the high RPM test climbed slightly more slowly than the low RPM test. I would say that the reserator doesn't have enough contact between the water and the aluminium sides to transfer the heat to the fins effectively.
You should try Prime95 for your testing if you're pressed for time. It doesn't give the CPU a 2 minute break every 8 minutes.
Also, if you look carefully at the high/low RPM tests, the high RPM test climbed slightly more slowly than the low RPM test. I would say that the reserator doesn't have enough contact between the water and the aluminium sides to transfer the heat to the fins effectively.
Swapped out the Gentle Typhoons for S-Flexs as they GTs had an annoying ticking noise. Set the SFlexs to run between 50% to 65% with speedfan, with desrired temp at 50c to leave the speed increase as late as possible. They now run practically silent. The rear fan is also an S-Flex, set to run at 60% to 70% dependent on the northbridge temp with a desired temp of 60c.
Ran Prime95 Small FFT Test for 3 hours, temps seemed to have plateaued easily after an hour at 56/57c:
Sorted
Ran Prime95 Small FFT Test for 3 hours, temps seemed to have plateaued easily after an hour at 56/57c:
Sorted