More heat problems; new power supply showing *high* temperature

[joea64]

Still having problems controlling the system temperature. The latest incident is that I just today checked MBM5 and SiSoft Sandra and the power system/aux temperatures on my nearly-brand-new Antec TruePower 430W are much higher than I can remember them being. Right now, for example, SiSoft Sandra shows the power/aux temperature as being 141.8 degrees F (61 degrees C). This is a good deal higher than the temperature mark was this past week, and I haven't put any new hardware in since I reconstructed my system last weekend (at that time I installed a WD hard drive with 8 MB cache and 512 MB Crucial PC133 SDRAM), and temperatures weren't out of whack last week. I did have to do some work yesterday where I was pulling and reinstalling a couple of drive units, but I don't know if that's got anything to do with the current temperature issue. I have two fans in the case front blowing air in with another two fans plus an exhaust blower sending hot air out at the rear. A list of the devices that are being powered by the Antec can be seen after this message in my signature. (Actually, I wonder if the external temperature sensor on my Thermaltake Volcano 9 HSF might be detecting the hot air and giving an erroneous readout - at least, I think that little wire doohickey with a sensor on the end that extrudes from the Volcano 9 and goes out through the back of the case is an external temperature sensor. And right now, the power system temperature is at 145.4 F / 63.0 C.)

-Joe-

P.S. I _did_ try having a small fan blowing into the open case week before last when I was having my reboot troubles (the one that eventually turned out to be due to a defective motherboard) but it didn't really do that much to bring temperatures down.

[Whyzman]

at least, I think that little wire doohickey with a sensor on the end that extrudes from the Volcano 9 and goes out through the back of the case is an external temperature sensor.Where did you put the sensor from the Volcano 9 ? :confused:

http://www.thermaltake.com/products/heatsink/v9.htm

[joea64]

[QUOTE]Originally posted by Whyzman
[B]Where did you put the sensor from the Volcano 9 ? :confused:

I might have been mistaken after all - that's probably not the temperature sensor that's extending out from the back of the case, it's the controller. I do remember week before last when I took my computer in that the tech told me that that part had to be out of the case.

Frankly, I'm not sure what to make of all this. Here are the current voltage readouts, according to SiSoft Sandra:

CPU Core - 1.81V
Aux Voltage - 2.56V
+3.3V Voltage - 3.30V
+5V Voltage - 5.03V
+12V Voltage - 11.54V
-12V Voltage - -7.09V
-5V Voltage - -5.41V
+5V Standby Voltage - 2.20V
Battery Voltage - 2.03V

Chassis fan speed - 5273 rpm (this fan is a high-performance unit, and I actually have two of that model, one for intake and one for exhaust)
CPU fan speed - 5192 rpm

The +12V readout is the one that bothers me, is that rail supposed to be that low? (bearing in mind that I really don't know all that much about electrical requirements). It also bothers me that this is a top-of-the-line PS (Antec 430W TruePower) and it's not operating quite the way I'd expect it to operate. The strangest thing of all is that nothing bad has happened - at least, not yet.

-Joe-

P.S. I should add that neither the board nor the CPU seem to be operating at excessively hot temperatures, all things considered: the board has been running steady at 107 degrees F and the CPU is fluctuating slightly between 123 - 125 degrees F (bear in mind that this is an Athlon).

[mjc]

The 12V rail is fine.....most companies allow ± 10% for variations on the voltage....which for the 12V rail would be 1.2V (10.8 to 13.2 volts). It would probably show much closer to 12(or 12 flat (if it were unloaded (no drives or AGP card connected).

Also do the BIOS temps agree with Sandra temps?

I tend to trust the temps/voltages for
MBM 5 (http://mbm.livewiredev.com/) a little more than those from Sandra, but if they agree with the BIOS then yes, Houston, we have a problem.

One thing I would consider is lapping the HSF.

[joea64]

I'm not sure how much good lapping the HSF would do when the locus of the high temperature is in the power supply/auxiliary readout. MBM 5 says 107 for the board, 125 for the CPU, 147 to 149 for the PS (and a readout of 11.85V for the +12V rail). I would have to reboot to see what the BIOS says because I'll have to go into CMOS setup.

-Joe-

[Whyzman]

Since you've been working with a computer shop you might want to check on the situation with your warranty if you were to take this endeavor upon yourself...

http://overclockersclub.com/heatsinklappingguide.php

It's really not difficult and will drop your temps significantly...I think mjc usually says around 5C...

[joea64]

Well, the computer shop is supposed to try to get a replacement for the XP 2000+ I bought from them last month that got screwed up when the motherboard went bad. If they can do that, I'll see if they will lap the HSF for me when they replace the CPU.

Truthfully, though, while this will certainly reduce the CPU temperatures, I don't quite understand how this will affect the power supply / auxiliary temperature, which is the temp readount that's really gone out of whack over the last couple of days; if I recall right, back around Wednesday or so that readout was at about the same temperature as the board (105 to 110 degrees F). I don't understand what's changed to make it shoot up so significantly since then.

-Joe-

[Budfred]

I didn't get from your earlier posts that the PS was the problem. You won't be able to get a temp read on that in the BIOS, it doesn't track that. The first thing I would do is make sure that the fans in the PS are running properly. Also, it is possible that you have changed the air flow pattern in the case and that this has reduced cooling to the PS. It sounds like you have a lot of air blowing in and not so much pulling out... Obvously the air will exit eventually, but you may have set up some areas of turbulence that are leading to heat build-up.

[Whyzman]

I forgot to mention earlier (Budfred reminded me with his observation about monitoring the PSU) regarding the HSF rpms...The Volcano 9 supposedly maxs-out at 4800 rpm...are you sure that you're reading the correct sensors with MBM5?

It could be that the HSF is within tolerances, but I think I'd have a look-see just to make sure...

[joea64]

Originally posted by Budfred
I didn't get from your earlier posts that the PS was the problem. You won't be able to get a temp read on that in the BIOS, it doesn't track that.

I see...well, then, I'm not quite sure what the "Power/Aux" readout in the SiSoft Sandra Mainboard Information window (which corresponds to the temperature readout #3 on my installation of MBM5) is supposed to display the temperature for. As I said, the board and CPU temperatures aren't excessive and have been holding steady. In point of fact, the Power/Aux temperature readout has dropped some since I started the thread, it's now at 134.6 F / 57 C.

The first thing I would do is make sure that the fans in the PS are running properly.[QUOTE]

I did that when I opened up the case earlier this morning, the fans were spinning properly. I have also verified that the HSF is locked in place and the fan is spinning properly.

[QUOTE]Also, it is possible that you have changed the air flow pattern in the case and that this has reduced cooling to the PS. It sounds like you have a lot of air blowing in and not so much pulling out... Obvously the air will exit eventually, but you may have set up some areas of turbulence that are leading to heat build-up.

As a matter of fact, I have the following:

1) Two 80MM fans in the front of the case, situated in front of the 3.5" drive cage, blowing air into the case
2) One 80MM fan and one 70mm fan in the back of the case, above the expansion slot area, exhausting hot air (and I had to look hard to find a vendor who had a 70mm fan!)
3) One turbine-type exhaust blower unit in the bottom expansion slot, exhausting hot air
4) One two-fan blower unit in the top 5.25" drive bay, exhausting hot air

I will check on the turbulence issue. When the computer shop put in the new mobo, they cleaned up the cable tangle some, but I've moved things around a bit since when I put in new RAM and the new HD. Problem is, it's hard to identify where the "hot spot" areas are because obviously they don't show up with nice neat red spots like they do in the diagrams; but even so, I'll have a look.

It might even be that the sensor that's reporting the power/aux temperature is giving a false readout because it's getting temps from an area that's actually hotter than the rest of the system. Also, the air that's blowing out from the system blower in the front doesn't feel excessively warm or hot.

For Whyzman: the TI-V77 HSF I had on my CPU before I put the Volcano 9 on also ran at 5000+ RPM as a matter of course. I would emphasize your word "supposedly" with respect to the ostensible max fan speed: if a high-quality power supply like Antec's can deliver, reliably and regularly, more power than it's technically rated for, why shouldn't a high-quality HSF be able to put out more RPM's than it's technically rated for?

-Joe-

[Budfred]

I am not aware of any mobos that can sense temperature of the power supply, but there certainly are temp monitors that you can buy that allow you to place sensors anywhere you would like. If you have such a sensor, it accounts for that third temp. If not, I wonder if it is even real?? I have tried to use the 3rd readout on my P4 and the 2nd on my HP with MBM5 and always get bizarre temp readings, usually something like -144C. That would be very cool indeed!!! I would double check that to make sure it isn't reading negative and to make sure it is actually measuring something. Also, just to be sure, reboot and go into BIOS and see if there is a 3rd reading there....

[mjc]

IS this Antec model one with adjust speed controls for the fans?

If so can you up the speed?

Another thought......make sure that one fan is pulling air into th PSU (should be the inside/bottom one) and that the other one is actually blowing air out....if one of them got installed backwards....

[joea64]

Budfred, I'm beginning to think you could be right, and that the reading might be a "phantom" one. I just opened MBM 5 (I have the latest version installed, of course) and the mainboard and CPU temps are still holding steady at 107 F and 125 F respectively, but the third temp readout (the power/aux temp on Sandra) has dropped all the way to 80 degrees F!! I've been watching it for the last couple of minutes and it hasn't changed any, either. At this point I should add something I forgot to report earlier; when I booted up again after the first time I noticed the "high" temperature reading, the power/aux temperature (Sensor 3 on MBM 5) climbed rapidly to 150+ degrees F, then dropped abruptly like a rock to 74 degrees F, then started climbing again much more slowly to ca. 140 degrees F, and now it's back down at 80 degrees F and staying there (at least for now), while the other two temperature readouts (Sensor 1 and Sensor 2 on MBM 5, measuring the mainboard and CPU), have been staying rock-steady within a degree or two of variation all day. Next time I reboot, I'll go to CMOS and take a look at the temperature readouts in the "PC Health" section (my mobo has a Phoenix/Award BIOS).

mjc, I don't believe that the Antec TruePower series allows fan speeds to be adjusted, but I could be wrong - I'll look and see. I did an old-fashioned "hand test" just now, sticking my hand in back of the case to see if air was blowing out of the fan in the rear of the PS, and it was, so I think the fans are installed properly.

-Joe-

P.S. Sensor #3 is still holding steady at 80 degrees F. :confused: Maybe I *was* making much ado of nothing, after all...

[joea64]

I just did another quick check. The power/aux temperature reading on SiSoft Sandra was up to 87 degrees F after some heavy browsing activity, but it went back down to 80.6 degrees F when I first checked it about five minutes ago. I shut the program down, then opened it up again a couple of minutes later, and that readout has skyrocketed to 163.4 degrees (!). Since the mainboard and CPU temperatures are still stable at 107 and 125 degrees F respectively - if the PS were that hot, I think those temps would be a lot higher than they actually are, and in fact the system probably should have shut down or frozen up - I am now beginning to think that there is some kind of glitch in the circuitry for that particular temperature sensor, or else it's reading some particular "hot spot" in the case and not accurately reflecting the actual temperature. I might also add that the back of the case where the PS is located is warm to the touch but not excessively so; if the temperature _were_ 164 degrees, I would expect the metal case to be be so hot as to literally be painful to touch for any length of time and the rest of the case to be very hot as well, whereas in reality, most of the rest of the case is cool to mildly warm, and even directly over the power supply, I can rest my hand on the case for an extended period of time without any discomfort.

-Joe-

[Whyzman]

Does your MB manual tell you how many sensors are onboard?

[BigBlue66]

Personally, I wouldn't worry about that aux temp. My P4 system at work does the same thing, wild flucuations, but everything is stable.

I did notice one thing of concern though, and that's the battery voltage. The CMOS battery should be 3.0 volts. I would replace that battery for a couple of bucks, just to eliminate any problems down the road.

If you're still using that 1.4 TBird, the default vcore for that processor is 1.75. I would check the voltage in BIOS and if it is indeed running over 1.8, then manually set the vcore voltage to 1.725 for the first go round. If that brings the vcore back down where it should be around 1.75, then good, and your temps will decrease because of it. If it's still too high, back it down another notch. You can even run it below default voltages as long as you retain stability with the system. That in and of itself will greatly affect your CPU temps.

Again though, disable that aux temp readout if you can. You're driving yourself nuts for no apparent reason.

[joea64]

Whyzman - my motherboard manual is in my car at the moment, but I;ll look it up and see where the sensors, if any, are placed.

BigBlue - I wasn't previously aware that the battery voltage needed to be 3v. I haven't noticed a problem so far, I have to say; the system clock, which should be one of the most obvious indicators of a battery problem, is in perfect sync with the time servers (I use Atomic Clock Sync to keep the clock properly timed). Even so, I'll see about getting a new/spare battery.

I just checked the CPU voltage and it fluctuates slightly between 1.79V and 1.81V, so it's probably operating at close to spec or just a little over. I'll check to see whether the CPU core voltage can be manually reset to the values you cite.

I might add that for some reason the system actually seems to be running a little hotter when idle than when in use; this may be because the various fans spin slower at idle, I don't know. I left it running idle overnight and when I got up, I found warm air blowing out of the case-front exhaust unit. After using the computer for a few minutes, the exhausted air is cooling down. Also, for what little it's worth (given your assurances!), the aux temp was at 131 F when I turned on my monitor again, but it's dropped to 120 F. The board and CPU were at 104 and 120, respectively, after an idle night. I hope I get that XP 2000+ back soon, though, because it hardly ever ran hotter than 120 F and usually operated at 114 - 116.

-Joe-

-Joe-

[joea64]

The BIOS shows the battery voltage as 3.28V, so I think that part is all right after all, and the various monitors (Sandra and MBM) are returning an incorrect value. BIOS doesn't show power/aux temp, but system and CPU temperatures are in line with what Sandra and MBM record.

I looked for the CPU voltage adjustment facility and must confess I'm not quite sure whether I should mess with it after all. The voltage is set to "normal" by default, and the available values to change are +25 mv, +50mv, and +75mv. As I have no idea what any of those would do to my system, I think the better part of valor is discretion, as they say, so I'll leave that alone for now.

Oh yes, and the system was perfectly fine today while I was gone, no problems. I guess I let myself be alarmed by a wild goose, after all...

-Joe-