Northbridge doesn't like active cooling.

I don't think I've ever posted here before, but I've come here every now and then to check out the forums. I always learned new things here, and got some good ideas too.

But right now, I'm wondering if anyone else has ran into this kind of issue before. Currently I'm running my Epox 89KA-I motherboard in a 90F degree room, and I've noticed my northbridge chipset is a just below the ~40C mark. My CPU temps are hitting near or at 50C under load right now since it's so hot.

I decided to take off the side panel, since in the past it's always lowered my cpu temps back to more reasonable levels with a floor fan in front of it, however I noticed my chipset temps are climbing now! It's got a passive cooler on it, and it seems like adding air would only help it, right?

So I guess my question is, is it normal for a passive heatsinked northbridge chipset to climb in temps when you add airflow to it?

Have you overclocked the FSB more recently? That would make it jump some. And I also know that some ppl (including me), find that if you have good airflow in a case (say 2 intake fans and 2 exhaust fans), leaving the side panel on will produce better temps then taking it off.

If that doesnt work.....sell you comp and buy a window mounted air conditioning unit. Seriously...90F?!?! How do you live?

I am kinda disapointed that the winters comming because my window AC unit dosn't seem to work well when its cold outside. it Anyone know how to super charge a window AC unit to cool no matter the outside temps? Do the Epox boards really measure the north bridge temp?

Well, it could be that:

Before, the cooling of NB was provided by CPU fan, which expelled the "used" air ontop of NB.

Now there is a struggle between new fan battleing against CPU fan.

You could get a Northbridge cooler with a fan on it and see if that helps. They're not expensive.

And PaxMax: what's ethical overclocking when it's at home?

Ok, not sure if I am following all the way or not, so Im going to go with what I think your saying...

chipset temps were decent, removed sidepanel, temp on chipset went up. Right?
If so I would say that before, you had a solid flow of air going through the case, and in turn in a round about way, actively cooling the chipset. Now, when you remove that side panel, the air has alot more places to go, so the air that was going directly over the chipset heatsink, now isnt. So your now getting higher temps.

If I didnt understand what you said... then dunno... LoL

I would suspect that it is just a fault of the way your temperature is measured.

It simply seems impossible for you to replace your side panel with a floor fan (one of those 30cm diameter things right?), and for the cooling to become worse.

Unless the heatsink has come loose somehow.

shiyan said:

I would suspect that it is just a fault of the way your temperature is measured.

It simply seems impossible for you to replace your side panel with a floor fan (one of those 30cm diameter things right?), and for the cooling to become worse.

Click to expand...

Well it could be worse if the new airflow was with this

H m m m m . . .

Doing the math, if your ambient (surrounding air temp) is 90F then 50C for the CPU is not that bad nor the N'bridge at 40C.
My ambient is 77F and I see 105F (41C) on my CPU. My 77F is well below your 90F so I'd say your cooling system is doing it's job well considering the case is also 90F.

I have an EPOX 8RDA and 8RGA+ and I installed a 40mm fan on top of the N'Bridge HS using twist ties and it runs nice and cool like everything else. Go here and check out the blow hole in top. http://www.newegg.com/app/ViewProdu...ctory=1524&catalog=7&DEPA=0&sortby=14&order=1
I put one in the top of my boxes and it dropped my temps straight across the board by ten degrees. All it takes is a pair of aviation snips, a 30CFM fan and 80mm grille and you're in business. l

"So I guess my question is, is it normal for a passive heatsinked northbridge chipset to climb in temps when you add airflow to it?"

No, that is not normal but I feel it has to do with you using 90F air to cool the HS. Then to, how much air actually blows down inside the HS where it does the most good ? ? ? Take my advice, get a 40mm fan and some twist ties and mount it and cool things off.

Almost forgot: I took the N'Bridge HS off and found to my dismay only a small, square pink pad that covered maybe 60% of the chip. I removed the pink goo and replaced it with white HS compound covering the chip & HS as it should be, installed the 40mm fan and now we are rockin' nice and cool.

Hope all this helped. OldBird

OMG SSS thats the same hairdryer my moms got!!! lol

maybe the new chipset cooler is badly seated?

You got it nailed down Shiyan.

Incomprehensible to me how moving more air over any cooling device would make the temps on object it is cooling go up. It just don't add up. That applies not only to chip & CPU HS's but a car engine radiator as well.

I feel the problem is with the HS cruddy pink compound and see my prior post for the quick fix. You gotta' do this mod before you install the MB or remove it to do the mod because the nylon retainer clips are released from the back side of the board.

I think I'm going to have to agree with the two suggestions that the chipset temp. sensor isn't reading right, and that it's got that crappy pink tape under it. I decided to just replace this Epox board with a SOLTEK SL-75FRN2-L that I ordered last week from Newegg. I was going to use it to build another AMD rig, but instead I'm jumping ship and building a P4

Ironically, the Soltek board seems to like the floor fan even though it's has a passive northbridge heatsink like the Epox board.

And no, I hadn't increased the FSB on the Epox board recently, it didn't like going over 180FSB very well, but the Soltek does. I put my 2100+ at 11x200 and it's working nicely.

jamespetts said:

And PaxMax: what's ethical overclocking when it's at home?

Click to expand...

I doesn't matter where you are.. Ethical overclocking can be practiced wherever there is overclockable equipment.

Ethical overclocking is essentially accepting that "If (for example) my cpu dies while overclocking it, I go out and buy another one, I don't RMA it since I broke it.

-OR-

If I overclock a chip and notice that I don't get as high overclock with it as I expected... then I go out and buy another one. I don't RMA the chip that didn't meet my expectations.

Temp sensors.

All of the MB makers adhere to "calibration standards" when they set the heat sensors and exactly what those standards are I do not know, but I venture to say I seriously they're off more than two degrees + or -. Two degrees is not that bad at all considering the mass production lines.

I have three boards, 8RDA, 8RGA+ and SL-75MRN-L. When I boot I see the temps at the bottom displayed by the BIOS and rarely do I see that much difference, if any, between that and what SpeedFan shows me or the monitor utility on the Soltek.

I guess we can cop out and doubt what the sensors display but somewhere, sometime we have to believe one is right, or close to being right. If I saw that much difference betwen the BIOS displays and the utility displays, then yes, I would also doubt them. Howsumever, such is not the case and I shall continue to believe what I see as long as both read within two degrees + or -.

Feature if you will the thousands of cars that roll off the assembley line in Detroit with engine water temp indicators. QC has the responsibility to check and make sure they are reading accurately and when the engine gets to hot you get a flashing light or some other warning signal just like on the MB's.

I set my MB's BIOS to give a warning when the CPU temp reaches 50C and shut down automatically at 60C in the event I am not here to shut it down myself when I hear the first warning. Other than that, ain't much else I can do. Another test is have the 'puter running and then take the side off the case and see if the temps drop five or ten degrees. If so you know the sensors are working so don't worry, be happy, use your computer and rock on and enjoy it.

IMHO, to some degree I think you're worrying about fixin' somethin' that ain't broke . . . . . OldBird

Heh, sorry to bump the thread, but I had to agree. It's not really broke per-say, but I just seemed odd to me and I wanted to know if anyone else had noticed what I considered to be a phenemenon<sp?> that counter-acted what made the most sense.(passive heatsink with more air providing cooler temps.)

And on ethical overclocking, I agree, you shouldn't whine if you fry you chip messing with it, it's on you, as you're the one who provided the circumstances to the cpu's demise. And never, never, RMA something unless it fails to run at the rated speed given by the manufactor.

Its possible that your passive NB cooler was getting a good deal of airflow over it from your the fan on your CPU's heatsink (that is assuming you are air cooling). Depending on the type of heatsink you have an the direction the fins face (if you even have fins on your HS) then its definatly possible you were getting a good deal of airflow over your NB HS. Blowing the fan directly into the case might be disrupting that airflow.