Location: Back and forth between Arizona and Florida
Posts: 2,701
"Personalized" Radiator Testing on the Torture Rack
I'm struggling with how best to setup my current build in the Black Pearl, so decided to do a bit of testing first to decide which radiators and fans I want to use. I've used "personalized" in my title, because I'm not trying to do a universal test to the precision and detail of our pro testers - just wanting to answer some basic questions for my own use. Thought I'd share as others may find some of the info helpful, and also, I know I'll get some good suggestions.
I'm still working out my "testing protocol" . I plan to test with the CPU at stock and at least one OC, maybe two.
My stressing procedure will be based on crunching for the XS WCG team. CPU loads typically stay above 99%, and this way I'll also be doing something productive while testing. Using 64-bit Boinc Manager.
My room setup is not ideal - a garage with it's own A/C but still not super stable. I'll be measuring air in to the radiator and the water temperature in my loop. Currently measuring air into the radiator with a temp probe connected to the Scythe Kaze Master. Measuring my water temp using the port on top of the XSPC res, unless anyone can give me a better idea. I have a Crystalfontz 635 with scab on order so I can record temps and measure at more locations as need be.
Using Real Temp to log my CPU core data. Just some quick results from today's efforts:
(Oops - I see I need a bigger pic - temps hovering around 60 degrees C, running at 4 GHz and 1.35 Vcore).
EDIT - larger screeshots uploaded
I'm also pleased that the Vreg temps on the Micro X58 board do not seem to run as hot as the regular EVGA X58 board I have in my main rig:
I'll post a bit more info on my temps and delta, but gotta take a quick break to jump in the pool.
And, cause we all love pics, here are a few more of the setup as is:
__________________ skinnee labs test schedule: In Testing: Too much to list. Writing Reviews: finally getting my backlog cleaned up too. Planned Tests: Plan, what the heck is a plan.
Hay, you said in your " Tagan black Pearl work log" that your Torture Rack, as a test rack was not to be "pretty".
Well, I kinda think it,s pretty sweet myself.
Did you build the perspex case your self or buy it?
Damm! There's yet another work log to add to the list of "Not yet read"
My eye's are going square and strange word are comming out of my mouth during conversations lately.
I think I have found an inner geek in my subconscious thats slowly taking me over.
Location: Back and forth between Arizona and Florida
Posts: 2,701
hehe ... thanks, guys. I've resisted the testing virus, but now that I have more time there are some things I'd like to do in a more organized fashion. Some of you have seen my collection of radiators:
And, I have a few more not shown in the pic. I seem to buy radiators like most women buy shoes
I'll fix the screenshot pics and post a bit more temperature data tomorrow, , 'cause right now I've got a date with Tiger Woods. And, I forgot to mention my ambient was around 26-27 degrees C.
(The shots were taken with the Canon 5D Mark II, most at 3200 ISO). Future updates will focus more on data and less on photos, but it's fun to take pics of the baby Classified. And yes, I said it wasn't to be "pretty," but I start twitching every time I look at those unsleeved pump and fan wires...
Location: Back and forth between Arizona and Florida
Posts: 2,701
Edited first post to put larger screenshots in. Doing more testing today ... of course, my husband decides to open all the garage doors cause it's such a nice day outside. Ambient temps have dropped 4 degrees C ... which is fine, but they will fluctuate as the desert warms up ... arggghhhh...
Current Temps:
I'm operating at a 5 degree delta (air in - water temp). This is about right, using the data from skinnee's charts on radiator testing for my GT fans running at 1450 RPM (still need to verify actual fan RPMS).
I'm operating at a 5 degree delta (air in - water temp). This is about right, using the data from skinnee's charts on radiator testing for my GT fans running at 1450 RPM (still need to verify actual fan RPMS).
Nice... will be good to see variance from the radiato heat load charts once you get the RPM reading.
__________________ skinnee labs test schedule: In Testing: Too much to list. Writing Reviews: finally getting my backlog cleaned up too. Planned Tests: Plan, what the heck is a plan.
Nice clean looking test setup, good pictures! The red and black color theme is awesome.
I also have a Swiftech MCR320 radiator... For trying to get 4.2ghz+ its pushing it IMO... Good sized radiator though, fits in a lot of cases!
You are getting pretty nice temps... Are you planning on lapping the CPU/waterblock?
Location: Back and forth between Arizona and Florida
Posts: 2,701
Quote:
Originally Posted by skinnee
Nice... will be good to see variance from the radiato heat load charts once you get the RPM reading.
Thanks, skinnee - the CrystalFontz is to arrive tomorrow, so should be able to get some more readings. I may need some help with the temp probes.
Quote:
Originally Posted by Linuxfan
Nice clean looking test setup, good pictures! The red and black color theme is awesome.
I also have a Swiftech MCR320 radiator... For trying to get 4.2ghz+ its pushing it IMO... Good sized radiator though, fits in a lot of cases!
You are getting pretty nice temps... Are you planning on lapping the CPU/waterblock?
Thanks! I probably won't be lapping anything at this time. Just planning to compare radiators and fans.
Quote:
Originally Posted by Sadasius
Nice work Shazza! Good to see your putting your rack to good use (insert deviant smile here)!
That's never been a problem ................................................ the DD TORTURE Rack is very effective
Quote:
Originally Posted by NaeKuh
Shazza my avatar is speaking to you.
More rads ... I got more rads than any one person can use in a lifetime!
LOL ... while I was tearing apart my Black Pearl Case, the TR build was crunching away. Glancing at the temps, I couldn't figure out why my water temp was creeping up ... as much as two to three degrees. Even I could figure out what was happening when I took a closer look:
Location: Back and forth between Arizona and Florida
Posts: 2,701
Question for you experts out there. I received my Crystalfontz 635, SCAB, and cables today. I found this link for preparing the temperature probes for use in water. Anyone have any other suggestions or links? Also, for my modest testing purposes, am I okay measuring the water in the res, or would I be better off putting the temp probe somewhere else?
Question for you experts out there. I received my Crystalfontz 635, SCAB, and cables today. I found this link for preparing the temperature probes for use in water. Anyone have any other suggestions or links? Also, for my modest testing purposes, am I okay measuring the water in the res, or would I be better off putting the temp probe somewhere else?
Thanks!
Your linked method is by far the best method I have seen yet.
The guide you linked is by far the best guide out there, same guy had a build log for a multi-socket system in the LC worklogs section. I didn't use his exact methods, the guide wasn't around when I made my first set... but this next set will use some of his ideas. This next batch that I make will have two sensors per G1/4 plug, you can have 32, why not use them.
__________________ skinnee labs test schedule: In Testing: Too much to list. Writing Reviews: finally getting my backlog cleaned up too. Planned Tests: Plan, what the heck is a plan.
Location: Back and forth between Arizona and Florida
Posts: 2,701
Thanks guys. Sounds like this may be the current "best way" to do it (although not the easiest, I'd guess). I'm gonna put the man of the house on it ... this is something that's right up his alley! I do agree the idea to put two in isn't a bad one all.
What I do is a little different than that, but that guide is great...I use Bitspower 1/4" fatboy fittings (yeah, 1/4" fatboy, go figure, lol) and JB-Weld. I put two probes in per fitting (they don't both fit in the ID of the fitting, but the fitting holds the rubber end of the probes + the wires and has a cavity that the main body of the probe sits nicely in, leaving only a bit sticking up).
I can't tear down and show it in detail (system is being used right now), but here are some external shots of what I do...you can see the JB-Weld coming out the back of the fitting, it fills the entire space that the probes don't
EDIT: realized I could take pics from the front of my res and see the probes sticking up a little
Excellent info on the temp probe's there Shazza and Vapor.
I take it then skinnee, you also use the Dallas DS18B20 temperature sensors.
Is there any reason why you don’t just use the ready made BP temperature sensors fitting found here? http://www.performance-pcs.com/catal...ducts_id=24759. I am guessing that you find the Dallas DS18B20 to be more precise, is that right?
Location: Back and forth between Arizona and Florida
Posts: 2,701
I'll tell ya what I know (which isn't much, I admit). I have the BP temp sensor - they work fine if you have something to connect them to. For example, they work okay with the Aquaero. And some mobos have a place to connect the two pin temp sensor. Not sure on the specs, I don't know if they are as sensitive as the DS18B20s, but I do know it takes a while for them to stabilize. I wanted to be able to log temps, so that's why I went with the CF 635 and SCAB - I have experience with crystalfontz and figure I might be able to get the software working.
I don't think the temp range will be an issue with the display, as I wouldn't think most systems are going to actually get to 50 degrees C (the CPU core temps and Vregs, yes ... but not the whole system).
Yep - I saw that link as well. I have some of the bare sensors, but believe that mod is beyond my capabilities at the time.
And, thanks skinnee, BrokenArrow, and Vapor for the responses on my question. I've got the 3M Epoxy and some syringes on order. I did a quick test one using water-proof silicone, but it's curing now, so haven't had a chance to test it. Am I gonna blow up my controller if the temp sensor gets wet, or just short out the sensor???
You'll just lose that sensor, but at $4 per its a good learning experience about what truly needs to be sealed. Well, thats how I justified it to the wife when she asked why I was cussing like a sailor as I went walking to the garage.
Kgtiger, the 0ºC to 50ºC is the operating temperature, meaning the PCB is designed to run in a meat freezer to a really hot test lab. Here is the Spec sheet on the DS18B20.
The overall package of a CF PCB, LCD and Sensors is just perfect for monitoring and logging temps. If I where looking to run my daily driver, then I would go with something like the Aquero and TMS-200 where you can power and monitor your pump(s), monitor a couple temps and have safety features like those boast. I'm not sure what the ratings are on the Aquero/TMS-200/HeatMaster are, but I know I only get 1.5A max per fan header (4 fan headers), can run 32 sensors and can take data samples every second.
Just like with GPU's, RAM and so on... we have options. I just see the CrystalFontz fitting a different application than the Aquero/TMS-200/HeatMaster.
__________________ skinnee labs test schedule: In Testing: Too much to list. Writing Reviews: finally getting my backlog cleaned up too. Planned Tests: Plan, what the heck is a plan.
Thanks Shazza and Skinnee for your fast response to my questions, you have been very helpful and I do appreciate your time and effort you have put in to answering it.
I will look at the list of items “Aquero/TMS-200/HeatMaster “you have mentioned for me later Skinnee, as I need sleep.
Just one other question, the DDC 3.25 pump is only rated to 60 degrees C.
If I was to have my set-up running in a TJ07 like, res (home made large in size) going to-DDC 3.25 pump's (in parallel, hopefully watercool's dual res top)-rad XSPC 120x2-rad XSPC 120x4, then split into two, one line to the CPU i7, and the other one going to graphic’s cards, 2x GTX 285 SLI, then both lines back to res/ end of loop.
How hot might the water get when it gets back to the pump?
I know you can only guess, however, if water temps in the res do reaches 60 degrees. Do the pumps die fast at this temp?
I understand there are so many variable’s. Now I have written the question out in front of me. I think I may have lost the plot asking it. Perhaps I am just too tired to think straight.