High spike in temps

[Snafu]

Anyone have this happen?

I am doing some testing on airflow and the first thing on the list is to clean up the insides of the case. Testing involved running Prime95 and F@H together to put the system under full load until the temps settle for about an hour.

Cleaning up the case involved some re-routing of wires, covering the psu's side vent, removed one of the hard drive frames (not used), moved PCI cards to the bottom slots. It really cleaned up well and should allow air to flow straight through the case (front to back).

Ran one air flow test, results were higher temps by 4-6C. This is my first puzzle as there should have been better airflow through the case. Room temp was about 2C higher than baseline test so this could cause some of the higher readings.

I started to run a second test (using side/top fans on hi) and after about 15-30 minutes the alarm went off with the proc @86C :yikes: (both Asus probe and MBM5). CPU fan speed reported at 1500rpm and PSU fans speed was 0. Not sure of the cause.

Some points:
- MBM5 and Asus Probe were reporting sporatic spikes and drops in temp, speeds and volts: CPU fan @1500, PSU fan at 0, spikes and drops in voltage and fan speed (PSU fan was running on hi)
- External probe reported proc @47C (probe lies under proc)
- Heat sink was not warm to touch (RAM in 3rd dimm was fairly warm but not hot - could keep hands on it)

Took a chance and rebooted into Bios, temp readings were warm but normal (proc @47C) but Proc fan speed was 1500rpm), re-ran Asus Probe and MBM5 and still showed a high proc temp and no readings for CPU and PSU fan (both were running). I shutdown the system and during this shutdown received 2 exception errors (divide by zero in module and access error in module - have details if interested).

I suspect that Asus Probe and MBM5 are not reporting correctly but not sure why. I used Asus probe to record the measurements. When there is a spike or drop reported by the graph the recorded value is NA rather than the spike or zero value.

Possibilities:
- I snafu'd in re-routing the wiring with some wires running in parallel causing false readings or reconnected incorrectly.
- When warm the board seems to report false values
- I am now seeing an instability with the oc. Puzzled by this as I encountered no problems before at this speed and I had dropped the vcore from 1.65v to 1.6v before testing
- One other possibility. While the case was open I was curious to find out the details of the proc. I partially lifted one of the levers but then my brain kicked in reminding me that I would need to reapply thermal grease and this would negate any baseline testing before adding the new CPU heat sink. I put the lever back in position. It is possible that this slightly lifted the heat sink off the proc and when reseated some contact was lost.

I rechecked the connections and all are okay (at least they are well seated). Rerunning the test with all fans on hi and dropped the FSB to 230. Asus Probe and MBM5 seem to be reporting correctly. There are a couple of zero readings reported but again the recordings show N/A. Proc and case temps still higher even with a lower FSB (proc higher by 1C, MB higher by 3C).

It's puzzling to say the least :scratch: :scratch: :scratch:

[Jim]

I've seen where running MBM and Asus Probe together at the same time can cause fan reporting issues. Have you seen the same thing when using just Asus Probe?
I also read that the creator of MBM had decided to no longer update his fine program, so we'll all need to use something different iin the future. :(

[Snafu]

Thanks Jim.

Interesting. It could be a reporting conflict. I have not tried Asus probe alone (something to try). They both did report the hi temp with the alram going off. Even after shutting Asus Probe down MBM5 continued to report a high temp. I ran both as a comparison to validate one against the other and to compare against the external temp probe (the external tends to report 4C low under hi loads). Since both are software I was thinking it could be related to instability - once the system becomes unstable then they did not run properly and gave false readings.

When I dropped the FSB to 230 and ran both they both report normally for the most part. There are the occasional readings where the fan speed drops but nothing like a huge spike in temp. I will up the FSB to 235 and only run Asus probe. Will see if I can repeat the same spike. If it fries then the up side is I have an excuse to look for a new proc ( :mischeif: :chit: :lol: ).

It's a shame that the MBM creator will no longer update this program. It is a great tool. About the only thing Asus Probe has over MBM5 is the ability to record the readings.

[flexkill]

i know for a fact that running MBM and asus probe(at same time) is a very bad idea. they do not play nicely together. :yup:

[Snafu]

Thanks guys. Good to know for future testing. Something seemed very odd. With alarms going off at 1am you'd have thunk I had oc'd a prescott :lol: .

Still puzzling why the proc/case temps were higher with better airflow through the case :scratch: . My guess is that the PSU vent was still pulling air out and closing it off kept more warm air in the case.

Will run some more tests and then open up the vent to see what works best for temps.

[flexkill]

well if temps are higher then the airflow must not be better. make sure you have all fans working with each other and not against each other. so many times i see people sticking a million fans on a pc when 2 or 3 well placed fans could have done a better job. :yup: :thumb:

[Snafu]

It's a puzzle. I just cleaned up the inside. No new fans added, no direction changes nothing. I was running the 2-80mm in front and 2-80mm in back on hi (side/top fans on low). I wanted to compare the proc/case temps before and after cleaning up the insides. If anything the temps should have stayed the same.

I suspect that the PSU vent was actually pulling some hot air out of the case. Or I somehow snafu'd on wiring the fans & controller. Gonna look at this too but I shudder everytime I look at TT Hardcano controller and it's a mass of wires and connections - plain ugly and there is no reason for this! Gotta clean this up.

[flexkill]

i have found less intake more exhaust to be cooler. such as one intake two exhaust. have you tried this? :thumbs:

[Snafu]

Haven't tried that exactly. I think I have a negative pressure with 2 in and 3 out on full, 2 in and 1 out at half speed (can't shut them off entirely without disconnecting them from the fan controller). Shudder to think about disconnecting these until I can create a better system for connecting the fans to the controller (the worst thing about this case is the hardcano controller 7 connections and each of the 7 fans have a male/female molex connector).
Time to visit the local Radioshack and some re-wiring/soldering I hope I can salvage this thing. Need to take some pics if I can make something presentable.

[flexkill]

cool :thumb: