Put together my first loop starting with the XSPC Raystorm RS240 kit then added to it. First the system specs:
Intel 3770k CPU clocked at 4.5ghz with Raystorm wb
Asrock z77 Extreme4 MB
Crucial Ballistix 16gb DDR3 1600
EVGA GTX 680 core +100mhz mem +500mhz with XSPC Razor wb and EVGA backplate
Intel 520 120GB SSD
Hitachi 750GB 7200rpm HDD
CM 690 II Advanced case
XSPC RS240 rad - top mounted with 2 Gelid Wing 12 PWM in pull
XSPC RX240 rad - bottom mounted with 2 Gelid Wing 12 PWM and 2 25 mm shrouds in push
XSPC X20 750 pump/res combo - drain tube attached to 3rd port with Koolance inline drain valve on end
Gelid Wing 12 slim pwm in front
CM 120mm blue LED rear fan
GPU temps are fantastic and hardly reaches 50C even after hours of furmark. CPU temps still seems high. I have re-seated the Raystorm 5 times and it seems the MX-4 tim is spreading evenly and I have tightened it down as tight as humanly possible.
Max core temps with latest p95 64 bit clocked at 4.5 VID 1.224:
74 79 76 76
Min temps with Speedstep enabled\;
31 27 22 32
I am starting to think this is simply due to the IB IHS using thermal paste instead of solder. What do you all think and is there anything else I can do?
Core temps are essentially meaningless without water temps and water temps are meaningless without ambient temps.
Not recycling hot air from your bottom rad up to your top rad could make a difference to your ambient to water delta. It may be that your water temps are higher than they could be with a better air flow arrangement to the radiators and your core temps will lower accordingly. Its unlikely to be a huge difference with only the CPU loaded and decent speed fans running.
Ivy bridge is a hot design though ... there is no getting away from that without radical modification.
those temps are normal. even under decent water cooling the ivy runs hot.
its not that your water isn't pulling heat away from the chip, its that the ihs on the chip is not pulling heat away from the die.
best solution is to delid, like dopamin suggested. using the hammer and vice is incredibly easy and with coollabratory between the die and ihs, people have seen anywhere from 10-30 degree temp drops, if executed properly
being on air and being clocked at 4.4, my ivy struggles to leave the upper 50c range while prime95-ing.
and in all honesty your temps arent REALLY that bad for being at 4.5ghz no de-lidded but if you do de-lid you should see another 10c+ drop in temps across the board if you use CLU paste
So with my current airflow my fans drawn cool air in through the bottom of the case, up through the shrouds and then through the rx240. Front bottom fan bring cool air in as well. Top fans mounted above rs240 pull case air through the rad then exhaust out the top. Rear 120 exhausts air out. Waterflow is from res/pump to rs240 to cpu, to gpu, to rx240 and back to res/pump. Any thoughts on changes I should make? Thanks
Your temps are way better than mine. My max load (Prime 95) temps hit about the same as yours and min temps sit at about 35-40C. I am OC to 4.2 with my voltage set to auto. At 4.5 and 1.34v i have the same temps. Ive also deemed my i7 3770k, even after delidded, is a crappy chip. I think i got one from a bad batch, but im not to worried about, it still does the job.
Your temps are way better than mine. My max load (Prime 95) temps hit about the same as yours and min temps sit at about 35-40C. I am OC to 4.2 with my voltage set to auto. At 4.5 and 1.34v i have the same temps. Ive also deemed my i7 3770k, even after delidded, is a crappy chip. I think i got one from a bad batch, but im not to worried about, it still does the job.
why let it auto clock itself ..... it always over volts...... manually volt it down..... 4.5 @ 1.34v is WAY to high .... try to manually volt it at like 1.2v and see if its stable if not bump it by .1v at a time because you are way to high.
why let it auto clock itself ..... it always over volts...... manually volt it down..... 4.5 @ 1.34v is WAY to high .... try to manually volt it at like 1.2v and see if its stable if not bump it by .1v at a time because you are way to high.
I have a pretty good loop and I am at 1.22 volts 4.5GHZ and I also hit 79C max on one core almost 100% the same as you its just we cant pull anymore heat off the die with out delid
This was unexpected but I'll take it. Not as big of an impact as deliding but I re-worked my layout and these are my new temps:
3770k 4.5ghz 1.22 vid latest P95 Blend:
High: 69 74 71 71
Idle with Speedstep enabled:
28 24 20 27
5C drop!
Ditched my shrouds on the bottom RX 240 and mounted rad right to floor of case and fans to top of rad blowing down. They are fed air directly from the front 120mm intake. Top fans also blow down with outside air through RS240 and warm air axhausted by rear 120mm. Was a few hours well spent!
So I finally got the stones to delid the 3770k yesterday. Hammer and vice worked a treat but hitting that wood with a hammer was a little anxiety producing. My CLU won't arrive until Wed but I replaced the paste with MX-4. Temps are same as before I delidded but might have used too much MX-4 as I used a line instead of a grain sized dot. Just wanted to delid yesterday before I talked myself out of it. Will report back with CLU results as soon as I have them.
Yep, all the glue is gone. CLU is awesome. Max temp in prime down from 94 after I redid the MX-4 with a nice thin layer to 81 with CLU at 4.7 with 1.33v and ddr 2400 mem. Cores are much closer in temp. And now the wheels are turning and wonder whether I should ditch the IHS altogether and set up direct die cooling. Someone talk me off the ledge!
Don't think I have the greatest chip in the world. I suspect alot of leakage with it due to needing 1.33 to hit 4.7 stable. Your temps are beyond amazing for 4.9 at 1.4v . Any additional thoughts on improvements I can make would be greatly appreciated
im running caseless with a megahalems cpu cooler and 2x SP120 fans.
Although the best temps I got depended on how much CLU I used. I used a small blob on the die and spread it out with a q tip, then put a small blob on the underside of the IHS and spread it out again but in the shape of the cpu die. Doing it like this lowered my temps alot. Also there is zero black glue left on the cpu or on the IHS.
I just applied CLU to the die itself using the included brush. Didn't use much at all to coat it and didn't apply any to underside of IHS. I've read some that say putting it on both is too much but certainly doesn't seem like you are having any problems with it. I use MX-4 between IHS and Raystorm. What TIM do you use between IHS and Mega?
Medium mirror layer on die - ok
Minute mirror layer on die - awful temps over 100C
Medium mirror layer on die & thin layer under IHS - best temps mid 70s
Also to bear in mind, ive changed my CL LP a few times on the die which has resulted in some of it almost 'drying out' on the die and leaving a rough residue on it. I had to use the included metal scrub pab with the CL LP to remove it. You have to make sure your cpu die is perfectly clean and smooth before putting the CL LP on for best results.
Thanks for taking the time to share with me. I used arctic clean on the die and IHS witha lint free cloth to clean it before application. Gonna give it a try next with some CLU on both surfaces and see what happens
Im using H100 and 3770k.
At 4.5ghz with prime95 running my core temps are: 55, 59, 63, 57
At 4.8ghz: 90, 92, 84, 83.
Ambient temps is about 16 degrees.
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Related Threads
?
?
?
?
?
Ask a question
Ask a question
Overclock.net
27.8M posts
541.2K members
Since 2004
A forum community dedicated to overclocking enthusiasts and testing the limits of computing. Come join the discussion about computing, builds, collections, displays, models, styles, scales, specifications, reviews, accessories, classifieds, and more!