How is this for temps???

Mt_GoatMt_Goat Head Cheezy KnobPflugerville (north of Austin) Icrontian
edited March 2004 in Hardware
I finally got my W/C set-up all together, tested and running. I still need to get some new shorter SATA cables and a few odds and ends before I post more pics and give the full run down.

Here's my temps while folding after 2 hrs. Is this any good or am I missing something?

Comments

  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited March 2004
    Looks like you're missing about 10*C on that CPU temp... :nudge::D;D
  • DragstkDragstk Syracuse, N.Y.
    edited March 2004
    Temps might come down some, when the Thermal paste has a chance to "cook". I know my temps are a lttle higher, with fresh Artic Silver
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited March 2004
    DOWN!?!? DOWN!?!? The last thing he needs is for those temps to go DOWN!! 38*C under full load? Someone go put a kink in the tubing to mtgoat's radiator for me. :D
  • Mt_GoatMt_Goat Head Cheezy Knob Pflugerville (north of Austin) Icrontian
    edited March 2004
    Geeky1 wrote:
    DOWN!?!? DOWN!?!? The last thing he needs is for those temps to go DOWN!! 38*C under full load? Someone go put a kink in the tubing to mtgoat's radiator for me. :D
    I tkae it by your response that it's time to try to squeeze more out of this chip??? And I only have 1/3 the noise I used to. I guess all that work paid off! :bigggrin:
  • edited March 2004
    Very nice temps there mtgoat, sounds like a very worthwhile investment in both time and money.
  • csimoncsimon Acadiana Icrontian
    edited March 2004
    great work larry ...what is your vcore setting? 1.65v I guess? :rockon:
  • Mt_GoatMt_Goat Head Cheezy Knob Pflugerville (north of Austin) Icrontian
    edited March 2004
    csimon wrote:
    great work larry ...what is your vcore setting? 1.65v I guess? :rockon:
    Actually it's 1.750.
  • csimoncsimon Acadiana Icrontian
    edited March 2004
    mtgoat wrote:
    Actually it's 1.750.
    1.75 is good! :cheers:
  • Mt_GoatMt_Goat Head Cheezy Knob Pflugerville (north of Austin) Icrontian
    edited March 2004
    It looks like I have some more work to do. I got so excited to have this up and running again and it's so much quieter that I forgot I had the panel off when I took the screenie with the temps above. My CPU temp increased by 6C after about an hour after putting the panel on. I'm not sure yet why the system temp dropped though! I now know that I need to do something to get better airflow into the case. I am thinking since the panel has a window and I have some spare plexiglass that I can make a new window with an intake vent with an airbox on the radiator also made of plexiglass.
  • edited March 2004
    Just put a window with a 120mm fan in there down towards the bottom and run the fan at 7v to keep it quiet, you'll get more than enough airflow to drop your temps back to the non-side panel levels.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited March 2004
    I do not understand how, with an XP+, you are getting a core 1 voltage reading AT ALL (core 0 is right, in MBM 5 I have NO Core 1 voltage at all). Second, with a case of 27 and that FSB, I would be having a core temp of 47-50 C and CPU would be stable up to 49 C with zero issues. Airflow will not help a water cooled case scenario much unless you are making it directed directly at tank water or radiator.

    The 2.68 Core 1 reading should be mapped to RAM, methinks. Voltage is a tib high for CPU, because you are stretching FSB a bit far. I have a 13 multiplier, and BASE CPU rate is 167. Unlike a P4, the multiplier is different.

    My Barton is happy at 1.65 V average with stretched multipliers. The main reason I do this is during day today the case was at 31 C for a high, the Barton topped out at 44 C (I enabled high\low and interval loggin in MBM 5, the interval log is set for 15 min intervals and every time the MBM interval log fills up to 1000 lines, XP hangs, so I do a rename save every once in a while). As a wild idea, try another exhaust fan, side exit. You might have a case that is slightly positive pressure. Worst case, turn the side fan so it inputs. Oh, if it helps, Ceramique and Bartons get along fine if you use a very thin layer only on core (after about two weeks of run time it stabilizes at its max cooling ability and the decrease is almost a straight line to time for the first 200-250 hours of run time). It beats AS5 all hollow, sorry to say.

    You might try a pushme-pullyou fan forced air flow through RAD (one dual ball bearing fan blowing into rad from inside (pushme), one dual ball bearing fan pulling air out on outside (Delta high volume fans, say 80 to 100 CFM each to allow for back pressure caused by tubes, especially on PUSHME side??)), see if CPU temps drop 2-3 C. The totally forced air flow would probably help cool the water better adn since you have the rad and it can be used as an exhaust point "why not try it?" bacomes relevant to my mind. Yes, a duct of Plastic around rad perimeter, if you have totally forced air flow, might be good.

    John D.
  • edited March 2004
    I'd be willing to bet that what MBM5 is showing as core 1 is the vDIMM in reality.
    And the reason I suggested a 120 in the window is that mtgoat's system is a negative pressure setup in that the airflow out is greater than the airflow in so in putting the case side back on the rad is probably getting a bit starved for airflow which would explain the higher temps.
  • Mt_GoatMt_Goat Head Cheezy Knob Pflugerville (north of Austin) Icrontian
    edited March 2004
    Matt,
    you hit it on the head. I don't think John remembers the front intake issues I had going into this project from trying to mount my radiator in the front of the case in this thread.

    I have:

    Intake
    1x 120mm Enermax adjustable (64-105 CFM)
    1x 80mm (34 CFM)
    *** There is a lot of restriction to the intake of both of these fans***

    Exhaust
    1x 120mm rear (suction side of radiator) (85 CFM)
    2x 80mm Vantec Stealth (2x 22 CFM)

    *** Adjusting the Enermax 120mm on the front has no effect and I connected the sensor lead to my board and it doesn't increase much in RPM so I am certain the restriction is keeping it from spinning up as it is running out of air to pull. So I am back to either doing my above mod to get radiator air directly through the side of the case or modding the intake capabilities on the front. I still do not want to do anything to visibly change the front bezel of the case.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited March 2004
    madmat wrote:
    I'd be willing to bet that what MBM5 is showing as core 1 is the vDIMM in reality.
    And the reason I suggested a 120 in the window is that mtgoat's system is a negative pressure setup in that the airflow out is greater than the airflow in so in putting the case side back on the rad is probably getting a bit starved for airflow which would explain the higher temps.

    The front air flow being blocked or limited as MtGoat said would limit air flow in, and starve radiator. As Geeky has preached, you WANT a negative pressure inside case, but you need the potential for flow from front.

    A side fan makes sense, and given the stats MtGoat stuck in and specs, the rad has to be limiting air flow through it some. Tubes and narrow fin spacings limit flow volume. So, a side fan will help some, but the rad will also act as an air dam in part with just a sucking fan. Put a fan on inside, push sir at another fan on outside of RAD. The flow from the fan pushing air into Radiator (possibly with a short 1-1.5" duct btween pushing fan and rad)combined with the flow from fan on outside should take air flow to about 1.5 times the present value. Then water or coolant in the cooling system will get cooled better by the wind sucking calories (kilocalories) of heat out of fins and tubes. It becomes a flow induced WINDCHILL effect almost.

    Note that in a car radiator, the sucking fan has forced air flow behind it-- caused by car motion through air. So, create that effect with a fan in an unmoving case, and airflow volume improves.

    I would say start with just a hole where you would put a fan on side, see if starvation is mitigated by just hole and if temps iomprove. If not, add a pushing fan. Front flow area limits in case causes negative pressure I bet, not the equal fans.

    Opening up the front for fanning in front makes good sense also.

    As to voltage at CPU, I HAVE a Barton 2500+, Vcc is 1.62 typical per BIOS (half my actual 3.3 actual), and RAM is at 2.68-2.73 variable as is MtGoat's Core 1 showing in display. My Core 1 is ZERO. Opterons are multicore, and true multipipe, but the Barton is a single pipe and does not accept almost three volts in, nor does the P4. The core is likely to gradually heat crack at the builtup core material temps that would result.

    Continuing with voltage, take a look at the fan power load draw, any chance the fans are not running at full rated RPM and thus running under the volume throughput they should have???? Bigger PSU might be needed for that.

    John D.
  • csimoncsimon Acadiana Icrontian
    edited March 2004
    I would at least start by replacing the "2x 80mm Vantec Stealth (2x 22 CFM)" with 2x 80mm SmartFan II's ...they are quiet and will kick up way more cfms ...exhausting will probably give you the most benefit. Also ...you have an awful lot of stuff connected in there ...I consider you quite fortunate to be getting these good temps IMHO.
    Also ...remove the integrated fan guards with dremel off of the case to improve flow and replace with wire guards...
    Basically ...you're generating a buttload of heat my man!!! :necro:
    Are these temps under load (ie folding)?
  • leishi85leishi85 Grand Rapids, MI Icrontian
    edited March 2004
    one thing u could do is move the heatercore outside of the case( but i assume you don't want to ) that is what i will do when i get a watercooling rig together.
  • csimoncsimon Acadiana Icrontian
    edited March 2004
    leishi85 wrote:
    one thing u could do is move the heatercore outside of the case( but i assume you don't want to ) that is what i will do when i get a watercooling rig together.
    ideally yes ...that would work best I'm afraid.
  • edited March 2004
    Personally I've never bought into negative pressure systems, I have run that way and I've had dust getting pulled into my optical drives and around every little gap in the case and I figure that the smoothest airflow will be where the air being evacuated from the case is equal to the air being supplied by the intake so I try to keep my airflow neutral and I've always had good luck with my temps.
    John, his temps with the case side off were fine as there was no restriction to airflow going into the rad so if the airflow through the case were equalized the temps should drop back down although I'll agree that to optimize the flow through the rad and to make up for the restriction created by the core a second fan on the rad would be a good idea especially with a shroud.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited March 2004
    The other reason for the shroud is that the fan is not right close to the rad. When fan is right close to RAD, the area of rad where hub and "motor" is will not get cooled worth beans. When you move fan 1.5-2.5 inches AWAY from the rad, all the rad gets airflow through it. THAT is the real reason for shrouds or short ducts.

    We understand each other. I run my Barton, now, on a negative to equal flow basis, and my P4 on a negative pressure basis. Both are stable, that is what counts.

    A passive side hole will give volume of air, that is what in part we are fighting to achieve, agree there. My Barton in an Antec SOHO case cools fine with just the vent in side-- been playing with fans and positions and volumes of fans in different places, so old descrips do not match reality.

    The dust thing-- my cases are not under the table, they are next to it, and raised from floor about 3.75" now from feet bottom to floor, thinking of raising them more (about 6-8" more). Raise whole case up, less dust gets in, and the bottom front slot for air is not needed if you have front venting like the Antec SOHO and 1000 series performance cases come with. Duct tape the front panel bottom slot vent, and dust input halves. The CPU HS gets cleaned every 8 weeks or so with cases raised, before raising it was an every 3-4 week thing.

    John D.
  • Mt_GoatMt_Goat Head Cheezy Knob Pflugerville (north of Austin) Icrontian
    edited March 2004
    The 2 Vantecs are doing their job of keeping the case cool but it is truely a matter of starvation on the front intake. Today I removed the front bezel that covers the front of the case and is only 1/4" from the case itself and my system cooled down immediately! All my fans also spun up faster with my system temp dropping to the lowest yet! This was probably a result of what John referred to but in the case itself.

    Chris,
    I did remove all the metal from the fan areas. Those temps are full load, folding Gromacs with all switches. :D

    My original plan was to mount the rad externally but decided to try this first as it looked much neater!

    Here is the problem with the front.

    attachment.php?attachmentid=5761&stc=1


    Edit:
    New PSU will be here on Thursday! :D
  • edited March 2004
    Well man, I hate to say it but you'll either have to modify the face of your case or put a fan in the window. If you decide to mod the face here's a mod article by a friend of mine that might give you some inspiration.
Sign In or Register to comment.