I remember someone saying that changing the mesh made no difference to temps.
Printable View
I remember someone saying that changing the mesh made no difference to temps.
Better air flow is better air flow. Whether or not it made a difference in one person's loop is not a global indicator.
Check post #11 here, and read a little closer on the temp difference with the panels on/off.
if I take my panels out completely...I saw a 2-3C drop from stock..on my gutter guard temps are 1-2C better than stock mesh
no doubt more airflow helps
Fair enough then.
An Australian shortage of black TJ07's stunted my plans, but mine finally arrived yesterday.
Time to blow the cobwebs off the dremel :D
I took the opportunity today to do a little testing. I was curious what effect having both rads drawing outside air would have on temperature. When origionally setting this up, I kept in mind airflow...specifically going in one side and out the other taking the heat outside the case right away. It seemed to be the best way vs. having each rad drawing from the outside, then having to turn, then making its way upstairs to heat up the components residing there...etc.
But I wanted some definate answers if I would have better temp results doing the conventional thought of each and every radiator drawing outside air with no regard to airflow issues. I proceeded to flip my fans on the 120.3 so they would draw air in. I disconnected the center fans so they would not spin. I had thoughts of maybe removing them for more accurate testing as they might affect airflow, but with them not really sealed against anything and the built-in plenum space in the rads and gaps between the rads, it seemed whatever inaccuracies would occur would not be too drastic. That and I'm too lazy to tear my machine apart to be able to get at them. The fans were kept at 7 volts as is usual, and the case sides were on (note that I have 'modders mesh' for my grills so they flow a lot better than stock).
I turned on the AC and did some Orthos testing and awaited seeing some "OMG!!! :eek: " improvements in my temperatures. After everything warmed up, I noticed that the temperatures, well, sucked. At least compared to my memory of load test temperatures before. Time to do some 'scientific style' testing.
Using TAT's "100% load" function on both cores so I could have a really high and consistant heatload, I let that run until things stabilized and took a screenshot of what was happening...
http://i80.photobucket.com/albums/j1...TAT1007min.jpg
Rather high temps reported by TAT & Coretemp. Note the readings from the 8800/Rivatuner graph as far as ambient temp goes, as well as the MB temp from Asusprobe. Seems things get warm upstairs with radiator exhaust heating up the MB compartment.
I then put things back as they were with the airflow going in one side and out the other, same TAT load left on for almost an hour...much better temps.
http://i80.photobucket.com/albums/j1...AT10045min.jpg
No radiator exhaust blowing into the case keeps the 8800 ambient temp reading and MB readings cooler as well. The radiators are noticably cooler with the fans oriented this way, confirming the software temperature readings.
So, in the case of a TJ07 with wide thermochill radiators, it appears that it is better to have an 'in one side and out the other' fan setup. My results may have differed if I was using thinner rads, I removed the inoperative center fans during testing, I wasnt using a tower case where airflow would need to make a turn resulting in less airflow though the rads, etc etc etc.
I also did a bit of very rough experimentation with a kind of push-pull setup...
http://i80.photobucket.com/albums/j1...ki00/rough.jpg
While this is a really lame attempt improve flow with both rads drawing air from the outside as a lot of the intake area of these 'floor fans' are trying to suck though a sheet of metal, I wanted to give it a try anyway. It was so bad that it made no noticable changes so I wont bother persuing this idea further. I think the primary issue is airflow is so affected by having to turn reducing the amount making it though the rads at a given time really affects the water temperature. Seems the benefits of lots of air volume moved from 7 fans going all one way vs. 3 fans going one way, 2 fans going the other, are better in this application.
Hey man, it never hurts to test! Thanks for spending the time!
I couldn't agree more with NicePants42!
This worklog here has inspired me to improve my current setup...
Thats great info, cheers mate :)
Nice looking there, I have a quick question for you andy, Does that case use rivets to keep the frame together? If so...how did you get them off.
Nope, all screws, very easy to tear it down to a bare shell.
ah I see! Well thats good, dammit... Lian-li uses rivets and im trying to figure it out haha.
Thanks for the testing Andy, was considering doing that myself and now I know not to bother.
I suppose my question now is, does having the extra 120.2 rad help?
Would you get the same performance with just the 120.3 and great airflow? You could put another set of 3 fans to push the air out the other side if you ditched the second rad.
I think it helps. I'm coming from another triple rad setup, only using BIP's for that one. My max OC with that keeping the fans at 7 volts was 3.6Ghz. The thing would rapidly reach for uncomfortably high temps idle/load beyond there (I have a higher heat-producing "B" chip). Mind you BIP's are totally outclassed by the Thermochill for low-cfm setups, but that's what I'm building...high OC with little/no noise. Now that I have added mosfet coolers (P5K Deluxe gets real hot...and has a ton of mosfets all over the place), and ran the OC up to 3.8, this new setup really handles high ambient temps and low fan speeds with keeping temps where I like them. No way my old 'triple' setup could have handled that. Using a single Thermochill I think things would be a bit on the warm-ish side, I'd probably have to turn up the fan speed for the same cooling capacity.
andy i love the modders mesh thing so much how did you bend it to look like it came with the panel becuase i ordered the mesh but i am still tryin to figure out how i will get it to shape right
It might seem like some extra work...but if you go through the process of clamping it between two blocks and hammer the edge over to make a 90 degree angle before cutting it, the cut will be mostly hidden vs. just cutting to fit. I know the cells by chance happen to perfectly line up with the size of the vent hole, but doing the extra step makes for a better final result. I think that there is a guide on the mnpctech site on how to work with this stuff, and there is a worklog or two here using Home Depot gutter guard using this method (I'm forgetting who posted one). Here's a picture so you get a better idea of the end result...btw, a permanent marker makes a pretty good 'paint job' on this :D
http://i80.photobucket.com/albums/j1...ll_edge001.jpg
I really would like to know how did u connect those fan cables on the radiators together . It looks nice and neat. Is there a tutorial for this or maybe you can explain it somehow?
Check out this thread. Should be enough to get you started.
andy, thats amazing, well done. where did you get the 3/4 length standoffs to mount the hard drive cage up top?
the ones that have come with the 2 cases i've used so far were 5/16", less than half the length of what he used to secure the hard drive cage to the 5.25 bay.
just checked newegg pics and it does appear that the standoffs that come with the case are much longer that what i've gotten with other cases. if what comes with the case is 3/4 then that saves me a purchase which is very cool. i'm gettin a tj07 next month/x-mas, cant wait to have a REAL case
Nice work as usual Andy. Thanks for sharing the good photos.
Andy
Those wires on the fan spokes could do with a smear of black liquid electrical tape over them. Finishes them off nicely.
http://img.photobucket.com/albums/v4...Large_6365.jpg
64dragon: For my current project, I'm also using a TJ07 and my plan was to do the same thing as Andy. Unfortunately, I could not find those 3/4 inch standoffs at the hardware store. Probably because I didn't look hard enough. The standoffs that came with the tower are just about 1/2 inch and don't even come remotely close to the walls of the 5.25" bays:( In the end, I just bought hard drive adapters.
For peeps in the U.S.of A., there is only one source anyone should consider for all fastener/odd tidbit needs........Mcmaster