I understand that. I wanted to see how the performance carries over when the HD is set up in the parallel cooling configurations.Originally Posted by jayhall0315
I understand that. I wanted to see how the performance carries over when the HD is set up in the parallel cooling configurations.Originally Posted by jayhall0315
It's been a while since this thread's had any activity. Any one done anything with different inlet/outlet configurations or seen something somewhere where someone tried things?
upgrading...
Probably because know one cares.
It a cheap plastic block with High Restriction, the pin matrix is also poorly machined (like all swiftech blocks)
Nothing to see here really.
I don't 'no' that that's true. I own one and am pretty happy with the performance so far. There are some parts which show obvious poor machining/quality standards (not just the pin matrix). Like I said though, the performance is good and I want to know if anyone else has utilized the additional ports in ways not described by swiftech before I tear down my loop again.
upgrading...
I hope you are not referring to the HD as a "cheap" block. I have one and it does pretty good and I have had probably 20 some different blocks. As far as restriction goes, I just dont see that, I actually got better flow than my previous EK block. Nothing scientific, but I wouldn't call Swiftech cheap in any sense.
CPUID http://valid.canardpc.com/show_oc.php?id=484051
http://valid.canardpc.com/show_oc.php?id=484051
http://valid.canardpc.com/show_oc.php?id=554982
New DO Stepping http://valid.canardpc.com/show_oc.php?id=555012
4.8Ghz - http://valid.canardpc.com/show_oc.php?id=794165
Desk Build
FX8120 @ 4.6Ghz 24/7 / Asus Crosshair V /HD7970/ 8Gb (4x2Gb) Gskill 2133Mhz / Intel 320 160Gb OS Drive, WD 256GB Game Storage
W/C System
(CPU) Swiftech HD (GPU) EK HD7970 with backplate (RAM) MIPS Ram block (Rad/Pump) 3 x Thermochill 120.3 triple rads and Dual MCP355's with Heatkiller dual top and Cyberdruid Prism res / B*P/Koolance Compression Fittings and Quick Disconnects.
I agree, it's not cheap. I like mine. So do we know anything about additional configurations? Anyone done/seen a 2inlet, 2outlet configuration where the downstream blocks are split between the out lines - should it matter?
upgrading...
That's kind of what I'm asking. Is there a reason why the port labeled in is the in port and the other 3 are outlets only. I've not studied the internals. So if I have dual loops, can I tie them together at the cpu block, i.e. 2 inlets, 2 outlets?
upgrading...
If you used one of the outs as an inlet it would be like a Y fitting "AFTER" the CPU block, so you would NOT be running the CPU in series anymore. It is really designed as a 1 in 3 out to match their new radiators which have 3 input ports.
Martinm is correct. I mentioned that I would have tested the HD with the new Swiftech radiators if they had sent me a test sample, which they did not. I assume Stephen or Gabe are not so happy that I find the Raystorm to be even better (by a very slight degree). I can say though that we have over fifty different Swiftech parts in our lab in various machines and none of their products that my grad students have installed are what I would call "cheap". Swiftech and XSPC are my two favorite brands in the watercooling community for their bang to buck ratio. Gabe and Michelle have also sent me numerous small pieces over the years that they have never tried to extra bill me for or nickel and dime me over (uhhhmm... as some other firms often try). Big fan of Swiftech.
Maybe they will get out a sample to Martin with their new radiator series for full parallel testing. Would look forward to it.
No screen shots to post but I have ran this block with 1 in and 1 out and with 1 in and 2 out and it cools better with 2 outs. Also I am not using any of the other radiators that were designed to go with this block.
I think 1 in 2 or more out helps improve flow through or heat dissipation.
I have been thinking about adding another pump to my loop to run double pumps on my loop just to see if the extra flow power will help cool down the loop temps.
Right now my loop is like this. Pump > 360mm radiator > 120mm radiator > HD CPU Block > 2 out puts to video card for separate vrm block and GPU block > reservoir > Back to pump
I like the versatility that the HD block offers and it cools well.
Update.
That card in the photo is an MSI PE GTX 560 ti 448 that has a modified ported Apogee XT CPU block on the gpu and a Koolance VRM block on the video cards VRM areas. I think both of these block that I used on the video card were extremely non restrictive and this is why the short tube distances didn't really hurt my over all flow and temps were very nice. The big benefit of the set up was to help keep the video card very cool. It over clocked better than any other GTX 560 ti 448 that I have ever seen on water.
The card.
The port work on the Apogee XT CPU block.
This set up worked great with the Apogee HD and I wouldn't have been able to make it work as easily with the Raystorm. This is the big benefit of the Apogee HD. I think the Apogee does suffer some from being so restrictive.
Last edited by SonDa5; 08-17-2012 at 03:17 PM.
Another important update.
I know now for certain that some restrictive blocks being fed from any of the 3 output from the Apogee HD will suffer from low pressure flow. The fluids greatest flow rate is in the path of least resistance.
Took out the water cooled GTX 560 ti 448 and added a HD7950 with a
Watercool HEATKILLER® GPU-X³ 79X0 Ni-Bl full cover block.
I couldn't find the Apogee HD screw in cap to cover the outlet so I ran that tube to an input to my reservoir which is no more than 8 inches away. Very low resistance to the water flow. The other outlet of the Apogee HD went to the new video card. Here is my first test I did for temps on the water cooled HD7950.
28C ambient temps.
11+ minute test at 1205/1575
VDDC set to 1.16
Board Power Li at +15
Click on image below for HIGH RES.
I thought the Heatkiller block would do better on the VRM temps.
Also during my Furmark run there are some low spots on the temps and that happened when I pinched on my Swiftech Apogee HD CPU block tube that goes directly to the reservoir
For this configuration my loop would benefit from the more non restrictive Raystorm block.
I'm sure my video card temps will improve with some flow changes and once I get the Apogee HD outlet cap screwed in place.
It's nice to have the options with the Apogee HD design for different configurations but you really have to plan out your outlets and figure out which outlet will suffer or benefit from block restrictiveness.
Last edited by SonDa5; 08-17-2012 at 03:45 PM.
Yes, by connecting one outlet of the HD to the reservoir, you made a shortcut: most of the flow will want to go that route instead of going into the restrictive GPU block. Your GPU temps suffered from an extremely low flow rates.
Well what you are saying is a basic hydrodynamics concept. When you provide various ways for the fluid ( in this case water ) to flow it will chose the one with less restriction, you can do one out of three thing, run them all in series, run them in unbalanced parallel ( this will make the most restrictive blocks to have a small flow ) , run them in a balanced parallel. Basically what you did with that tube was plain wrong ( I'd call it a rookie mistake ) and you are striving your GPU of flow.
I just call it a learning experience.
The Apogee HD is designed to be ran in series or parallel with up to 3 different blocks so it seemed ok to do and I have done it befor with different gpu block. I did the same pinch test with the same video card and with a less restrictive Swiftech MCW82-7900 and because that block is much lesss restrictive the temperature changes were barely anything. Also my cpu temps were a little better the reservoir tube there. Heatkiller GPU block is just way to restrictive for that extra tube and I need to get rid of it.
Flow restriction on the Apogee HD hurts the over all flow as well.
Last edited by SonDa5; 08-17-2012 at 04:32 PM.
Yeah. Worked well for CPU cooling but not so well for GPU cooling.
Photo of the configuartion with some results from i5-2500k.
CPU-Z during test.
http://valid.canardpc.com/show_oc.php?id=2192781
Bookmarks