View Full Version : My serial vs parallel experience.
TheWeaseL
11-02-2003, 12:15 PM
Just thought I would post my results of testing my new WW block and my dual heatcores, both in parallel, and serial. Using my home-made Y-fitting and running my heatcores in parallel, my temps would be around 36C~38C under full load. Now I switched to serial last night, and right now my temps are at 40C full load, but according to my high/low log in MBM5, my temps have hit as high as 45C, and the average is 42C. This is with my 2500 @ 217x11.5 using 2.03v in the bios, so I get about 1.95~1.97v in reality.
Needless to say, when my nice Y-fittings that I ordered from Cool Tech get here on monday or tuesday, I'll be switching back to parallel. I had originally thought that maybe serial would do a little better, cause even though I had been running the system for a couple days, when I turned off the pump, I saw air still in the lines up at the spouts of the heatcores. Not a lot of air, just a little.
naughty_guy
11-02-2003, 01:26 PM
thx for sharing, and now i'm totally convinced to go for a parallel setup, however, i'm thinking of a setup which is a combination of parallel and series. Which means 3 radiators.
pump-->blocks-->y-split to two rads in parallel-->y-split for merging-->3rd radiator-->res-->pump
KnightElite
11-02-2003, 01:57 PM
Yeah, I've never actually had a watercooling system with a single rad ;). And my WC has always been in parallel. From my point of view anyway, it just makes more sense.
TheWeaseL
11-02-2003, 02:16 PM
KE, do you still see a small amount of air in your circut if you shut off the pump up by the spouts of your rads? I think it verrrrrrrry slowly was bleeding out, but I didn't really give it enough time to do so if it was...
nikhsub1
11-02-2003, 03:10 PM
I always try to tell people that parallel rads generally works better than serial about 95% of the time. I also tell people to test it both ways to see which performs better in your specific setup. That was exactly the advice I gave to TheWeaseL and I am glad that he tested it both ways :D I find serial can be beneficial with a serious hipo pump like the MD-30rzt otherwise parallel will almost always win.
sandman
11-02-2003, 03:16 PM
I thought that you and cathar were talking about the Md-30rzt almost being too much pressure for most heatercoer walls?
KnightElite
11-02-2003, 05:01 PM
Originally posted by TheWeaseL
KE, do you still see a small amount of air in your circut if you shut off the pump up by the spouts of your rads? I think it verrrrrrrry slowly was bleeding out, but I didn't really give it enough time to do so if it was...
I know what you're talking about, but I've never noticed it on my heatercores, but I have noticed it on other people's systems, at the outlets of various barbs.
TheWeaseL
11-02-2003, 10:24 PM
After thinking about it, and looking at the short segments of tubing used to connect the Y's to the rads, I noticed they had some decent kinkage to them. The homemade Y fittings I made were rather large, and made it hard to get them to fit without the tubing kinking, so I think that might have been part of the problem...:o
Craig
11-02-2003, 11:03 PM
A nice comparison Weasel.
With Y's that don't cause the lines to want to kink you may get a small bit more gain yet.
What fan type, how many fans and how are they arranged?
TheWeaseL
11-02-2003, 11:58 PM
Originally posted by Craig
A nice comparison Weasel.
With Y's that don't cause the lines to want to kink you may get a small bit more gain yet.
What fan type, how many fans and how are they arranged?
When I get it all installed, I will post pics, but my friends have described it, as a bird house on top of my case. :)
What I did to house my two heatcores, built a wood box to sit on top, and it is 8"x17"x12". Now on each end, I mounted the exhaust fan on the outside of the case, and then used zip ties to hold the fan to the box, and the shroud to the box, creating a bit of a tunnel. Then I had drilled two holes in the shroud, one on each side, so I could wrap a zip tie around the rad to hold it against the shroud. I did this on each end with each rad.
Then on one side, I cut two spots for two intake fans, one mounted above the other. I used the same zip tie method, but no shroud here. All 4 of my fans are Panaflos, capable of 114cfms, but I run them at 7v, and they go quite nice there. It looks rather ghetto, and shows off my 1337 wood working skills. :) Any more questions, just ask.
TheWeasel, great job mate!
I am glad it worked for ya as we discussed in this massive thread :)
Yea, piccies will be a treat for us, make sure you post 'em asap :)
Could you try setting your fans to max and measuring temps?
Since you got this great setup ....
it'd be great to have air flow effect on temps measured.
About wind tunnel I found it working perfectly well in my case (it's taller than full size Chieftec tower on castors).
I am using sound attenuating design which is demountable according to needs. I noticed a diff in temps of 4C (CPU/GPU temps) between fully attenuated and jet-loud modes.
Craig
11-03-2003, 03:35 PM
Originally posted by sandman
I thought that you and cathar were talking about the Md-30rzt almost being too much pressure for most heatercoer walls?
From what I've seen Cathar post before, he says the MD30rlzt adds to much heat for most cooling loops to handle. So even though your flow increases your temps go up if the rad(s) can't handle the added heat.
Cathar has 2 large carefully chosen heater cores and some VERY powerfull blowers in his rig that has his MD30rlzt in it. And he says his gain from the MD30 is still modest due to it's heat output. Like adding another O/Ced CPU is another way I've seen him phrase it.
KE,
Thanks
sandman
11-03-2003, 03:45 PM
ahh, i see. Yea, that's a lot of heat.
Nookie420
11-03-2003, 06:20 PM
i was going to get another raddy and run dauls in my system in series, guess i'm gonna do parrallel.
Originally posted by Craig
From what I've seen Cathar post before, he says the MD30rlzt adds to much heat for most cooling loops to handle. So even though your flow increases your temps go up if the rad(s) can't handle the added heat.
Cathar has 2 large carefully chosen heater cores and some VERY powerfull blowers in his rig that has his MD30rlzt in it. And he says his gain from the MD30 is still modest due to it's heat output. Like adding another O/Ced CPU is another way I've seen him phrase it.
KE,
Thanks
Imho (I admit I don't know MD30) even if this pump was adding 10watts of heat into WC loop with double rads in parallel and massive airflow it wouldn't make any difference whatsoever. I believe that this pump is simply massively oversized and flow rate is so massive that radiators are not able to dissiate heat due to extremely short time period water spends in them (even thou it's a parallel setup and rads are massive...
But of course without knowing any details I may be totally wrong here :)
I think he'd be much better off with smaller pump :)
TheWeaseL
11-05-2003, 12:08 AM
You know, I might have been a bit premature about the serial setup being inferior to parallel...:eek:
After running it for a while, I settled around temps of 38C~40C...:/
And right now with my parallel setup using my Y fittings from Cool Tech, I'm getting temps of 44C...Soooooo maybe serial would be better for my config...Now I'm all confused...I'll let this one go until friday, and see what temps are on the weekend I think. Then decide for good.
TheWeaseL
11-05-2003, 07:40 PM
And now my load temps are 36C again!?!?! Although I tipped my system on its side, and re adjusted the cpu block, and I got this huge glop of air.
Kosmic
11-05-2003, 08:12 PM
Gotta be 110% sure all that air is out, or your temps will really suffer.
TheWeaseL
11-05-2003, 10:13 PM
Well, they are back at 42C...:( After I had that air bubble, I look at my heatcores...Only one was actually flowing water, the other was just sorta trickling...To get both with full tubes, I have to squeeze one to force the water up into the other...After the line gets full, I can let the first one back to normal, then they both flow fine.
CSOFT
11-06-2003, 05:03 AM
Anyone tryed coming out of a white water to two heatercores/rads then into a rez?
I'm no MSobe PAINToshop expert, but this is what I was thinking...
TheWeaseL
11-06-2003, 10:15 AM
I have a WW actually, but the only problem with that setup, is I have to cool my gpu and nb as well. Plus usually they recommend putting the heater cores between the pump and the cpu block to pull the heat out of the water that the pump puts out...
KnightElite
11-06-2003, 05:21 PM
Originally posted by TheWeaseL
I have a WW actually, but the only problem with that setup, is I have to cool my gpu and nb as well. Plus usually they recommend putting the heater cores between the pump and the cpu block to pull the heat out of the water that the pump puts out...
Interesting. I've never used that approach, I've always gone as follows:
pump-->block-->T for filling-->splitter-->rads-->splitter-->pump
TheWeaseL
11-07-2003, 01:18 AM
Really KE? From what I've been told, always from pump to rad to cpu for coolest water possible...
TheWeaseL->
How did you configure your parallel rads setup?
To work they must be 100% level with each other (inlets and outlets) otherise water will flow through the one which is verticaly closer to pump (that's probably why one was gathering air :) )
TheWeaseL
11-07-2003, 08:40 AM
They are level with each other. They are both mounted in a box I made for them on top of my computer.....I will say, with parallel rads, my temps have been all over the place. I've seen them as low as 34, as high as almost 40...
Take pictures and post them, top, front, side, angled views...
Are they both new rads?
TheWeaseL
11-07-2003, 10:58 PM
Well, new to me. I got them from dtek customs. One in August I think, and one just a week or two ago I think it was. Tomorrow I'll have my gf bring her digi cam, and I will take some pictures...
Or would you prefer I wait until I install my new pump and res first?
vBulletin® v3.7.0, Copyright ©2000-2008, Jelsoft Enterprises Ltd.