Finally! Help me with my first w/c setup!

I've been cleaning everything out with 25% vinegar, 75% water mixture. Just finished the radiator. Everything drained clear, so if there was something in there, it wasn't much.

I dropped the silver coil in a glass of the concoction above, and it's clearing right up. It's been in for maybe 15 minutes and it's almost entirely shiny silver again. Just a bit of yellow/redish around the edges.
 
After ~30 mins in the vinegar mixture.

z1xq.jpg
 
Just a couple thoughts on this. I'd say either some kind of strange chemical reaction as your mentioned that happened very quickly. Other thought is it's not truly a "silver" coil. I seen mention of this on other forums of fake coils being sold. In my own experience, I've used the same killcoil for nearly 2.5 years and upon inspection from a recent teardown about a month ago it's only slightly darker/tarnished verses some brand new ones I have.
 
Let it run with just the pump overnight and the coil was the same color this morning, thankfully.

ds4m.jpg


Finally back up and running. Excuse the phone pic.

7kzk.jpg


Temps are *okay*, maybe you can tell me otherwise. 100% load, no OC, I'm getting about 47C. I was hoping for lower. Things that may be contributing:
1. crappy fans on the rad
2. stock TIM on the block
 
Your probably correct about your rad fans and stock TIM contributing, however even with great fans/TIM I wouldn't expect a huge drop. What programs are your using for testing 100% load and temp monitoring? How does your temps now being watercooled compare to your aircooled setup for these tests? There should be a considerably smaller difference between idle/load temps for water vs air.
 
Load temps are about 48C (with one core at about 40C). I run FAH 24/7 when not being used for something else.
 
My 3820 at 4.3 ghz tops out around 62 C, running auto voltage.

Overclock your 2500k and see what temps you get. With CPU temperature sensors, the farther you are away from TJMax, the less accurate they tend to be.

Martin's Liquid Lab has the 2600k topping out at 63 C with the Raystorm at 4.8 ghz with 1.4v, and 25 C air.
 
Ok, 4.5GHz @ 1.38V all day, room was about 20C.

Core1: 53C
Core2: 50C
Core3: 34C
Core4: 53C

What's up with Core 3? Can it really be 16-19C different than the rest? I don't remember that big of a difference when I was on air.

On Air I was at about 65-70C. That's about as hot as I'm comfortable with 24/7.
 
Hmm, are you sure a core wasn't dropped? That seems like the most likely scenario to me, it shouldn't vary that much.
 
I guess I'm not sure what that means. Does it shut off a core if things get too hot or overworked?

Speedfan shows all 4 cores loaded 100%. Temps now are:

Core1: 50C
Core2: 48C
Core3: 32C
Core4: 52C
 
What I meant was did a thread stop working for some reason, or is one core not working at full load?

Use OCCT Linpack to stress test.
 
Ok, well I don't do anything more stressful than folding, so if it folds on my OC I'm okay with it. However, I ran OCCT anyway just to try it out. I found out right quick that my OC was, in fact, NOT stable. It crashed in the first 5 minutes or so.

After playing with the settings a bit, I'm stable at 4.4 GHz with 1.35V. What's interesting is that my CPU voltage in the BIOS does not match any analytics program I've run. The BIOS setting is always higher. In this case, I'm setting BIOS at 1.35V, but CPU-Z, Speedfan, and OCCT are reporting 1.28V. What gives?

Regarding temps, I think Speedfan is just doing something funky. OCCT is showing all 4 cores within a few degrees, all hovering at about 54C at 100% load. So that's good.
 
It's probably vdroop. Voltage is always going to vary a bit.
 
I know this is a big debate, but I've heard for 24/7 use, it's best to stay at or below 1.4V for sandy bridge. If these programs are returning voltages of around 1.3V when I set my bios for 1.4, can I still safely increase voltage by 0.1V?
 
I would mess around with any LLC/vdroop options you have first.
 
Depends how much you trust the programs and how much you care. I'm no SB expert, I do not have one. It's always best to measure yourself if possible though. I don't know that it is.
 
Yeah, I've actually got headers on my mobo to measure. Unfortunately I don't have a DMM or budget to get one :(.
 
Something definitely wrong if you have one core that is that much lower. Either a bad sensor or a bad mount..
 
sok0 said:
Something definitely wrong if you have one core that is that much lower. Either a bad sensor or a bad mount..
Click to expand...

Yeah, I tried a different temp monitoring program and they're all within a few degrees. I'm chalking this up to SpeedFan.
 
jamsomito said:
Yeah, I've actually got headers on my mobo to measure. Unfortunately I don't have a DMM or budget to get one :(.
Click to expand...

Harbor Freight has coupons for free DMMs all the time. Use Google, you may have to buy something else though (get cheap batteries or something).
 
A cheap DMM will give you a good ballpark figure. It should be good enough for your purposes.
 
Ok, I actually went and got that cheapo DMM. Turns out my actual voltages are slightly higher than my reported voltages in OCCT by 0.01V. Guess I just have a mediocre chip. Oh well. 4.4 GHz isn't bad by any means.

To keep this water cool related... My Tygon tubing is getting cloudy after a couple months of 24/7 use. The water is clear in my res. Is this the tubing breaking down? Should I be concerned? Is there any way to keep your tubes clear?
 
You must log in or register to reply here.
Back
Top