Sceptr 2020 年 12 月 13 日 下午 12:00
Northbridge Chipset And Thermal Paste?
I'm in the process of upgrading an old LGA 775 motherboard with an Intel Q6600 PC into a new case and i've heard that it's a good idea to put some new thermal paste on the Northbridge and Southbridge chipset's especially if the parts are roughly 14 years old now.

Now the Northbridge chipset under the heatsink has a square window like padding around it but away from the actual chipset and the chipset itself has some clear looking sealant on it that makes the chip look like it has a glass like finish. I'm posting some pics of the chipset and under the heatsink of the chipset.

My question is: Can i add a small amount of thermal paste to the chipset as it is or should i remove the clear sealant first or just leave the heatsink and chip as it is?

https://i.ibb.co/JtgvtR1/IMG-3194.jpg

https://i.ibb.co/R6xRNL1/IMG-3190.jpg

https://i.ibb.co/yqPWVhX/IMG-3192.jpg
< >
目前顯示第 1-15 則留言,共 27
Illusion of Progress 2020 年 12 月 13 日 下午 12:48 
Now that you've even removed it, you need to redo it.

The chipset looks like a normally clean die to me.

The square pad on the heatsink can remain, as it sits around the die. The stuff in the middle of the heatisnk will need cleaned off. It looks like a thermal pad and maybe it can be re-used, but I'd remove it and replace it with thermal paste.

Apply a small rice grain sized of thermal paste.
Sceptr 2020 年 12 月 13 日 下午 2:00 
Thank you for your reply Illusion.

I think you are most probably correct in the chipset i have being a normally clean die, i've found a picture of another Northbridge chipset that is very similar to mine.

https://upload.wikimedia.org/wikipedia/commons/4/48/Pentium_E2220_with_Intel_i945GC_Chipset.jpg

What gets me is how many chipsets i've seen while researching them that had old thermal paste on them and so it was obvious to replace the old paste and then there's mine with just a thermal pad.
As for the thermal pad on the heatsink, it looks like it's in good condition and is firmly fixed in place, i've taken a couple of more pics of it.

https://i.ibb.co/Mf0xF6h/IMG-3199.jpg

https://i.ibb.co/s9yPWMH/IMG-3198.jpg

So do you think i should still remove the pad or apply some thermal paste in between the pad and the chipset? I'm sorry for double checking but it was only very recently that i realised that an old chipset could have new thermal paste added and so it was a shock to find that mine only had the thermal pad.
SeriousCCIE 2020 年 12 月 13 日 下午 2:32 
I have several LGA 775s in use. Two are modified to be LGA 771s; their chipsets got quite hot as a result of my admittedly abusive efforts at keeping them fast enough to run modern software comfortably.

I used thermal paste; typical CPU thermal paste.

I have used thermal pads and even recently just purchased sone for other purposes (I mention it so that I can try to appear to be unbiased--right tool for the job, etc).

I would suggest not going anything specifically special for your chipset cooling thermal regulation--a paste you trust or pad will do.

You also are likely to see acceptable results in just cleaning off as much dust from the pads as you can; I even ran mine under the kitchen faucet for a little bit and then wiped off the moisture. I was reluctant to use rubbing alcohol or anything that might dry the pads out--although the surface beneath the pad and on the heatsink I of course cleaned more thoroughly.


Illusion of Progress 2020 年 12 月 13 日 下午 2:51 
Yeah, that stuff is what attaches the die to the substrate as far as I know? I've definitely seen it on nearly everything I've used with an exposed die, including old CPUs.

https://upload.wikimedia.org/wikipedia/commons/0/00/KL_Intel_Pentium_III_Coppermine.jpg

Thermal pads aren't too uncommon, especially on chipsets.

Anyway, it may have been fine to use, but typically once you break contact, it's recommended to re-do to prevent air gaps within and ensure it has total coverage. But your newer pictures are quite interesting; the thermal pad (?) appears to have the same machining imperfections across it that the metal on the heatink itself does. That's... odd. It almost looks like a very thin tin-foil like material.

I'd hate to recommend removing it if you can't easily do so, but I'm not sure if mixing thermal paste AND leaving that on is a good idea either (normally, it wouldn't be, but that pad looks very thin, almost like it's an indent imprint on the metal itself rather than a pad?).
Sceptr 2020 年 12 月 13 日 下午 3:49 
Thank you for your input guys, it is very much appreciated.

Further research and i've found out that pads are a one shot deal and mixing pads and paste is a no deal scenario so the current pad will be removed, i've also cleaned it with Isopropanol Alcohol like with the whole motherboard so there's another reason to remove the pad.

That's very observant of you Illusion with noticing the same imperfections on the pad as is on the heatsink, it does look like some matching wallpaper looking closer.

I have some MX-4 paste unopened that i bought back in January 2015 and some 2mm thermal pads and with the size of the chipset i'm now thinking of going like for like and using a pad instead of paste possibly, i think i would have to use a 0.5mm pad though i'm not sure.


xSOSxHawkens 2020 年 12 月 13 日 下午 4:22 
Just re-paste it bud,

The foam bit is just padding for the heatsink for it doesnt get off center or damage the silicon.
最後修改者:xSOSxHawkens; 2020 年 12 月 13 日 下午 4:22
Bad 💀 Motha 2020 年 12 月 13 日 下午 6:56 
I always did that, even on brand new boards that had both a north and south bridge chip. As the stock paste was always terrible, and on boards that old its more like a gum style glue then anything. Use rubbing alcohol and a thin plastic card or flat plastic tool to remove the old gummy paste. Using a touch of AS5 or ACMX4 should help. Same goes for the CPU.
Sceptr 2020 年 12 月 17 日 上午 10:37 
Thank you for all of your comments they were interesting and helpful.

I used thermal paste on the Northbridge and Southbridge chipsets and i think i used just a little bit too much but i've taken the paste off again to see how they had spread so i can reapply it again.

I'm going to show you some how i applied the paste and the spread of the paste pics.

Northbridge before and after

https://i.ibb.co/hd3W1C0/IMG-3209.jpg

https://i.ibb.co/VVxT0cW/IMG-3228.jpg

Southbridge just the after.

https://i.ibb.co/8zMkmHj/IMG-3225.jpg

CPU before and after

https://i.ibb.co/Hgkm70S/IMG-3216.jpg

https://i.ibb.co/MMxsLHs/IMG-3221.jpg

I did some test boots out of the case and used HWMonitor for the temps, the CPU tended to idle at the temps shown here and hovered around the 85c mark while running Crysis on high settings. Do you think the idle temps are normal or ok for the Intel Q6600 CPU while using the supplied heatsink fan (seen in the pics) and do you think the 85c is too hot while gaming. I'm pretty sure the CPU did briefly rise to 100c on a reset but came back down fairly fast eventually hovering round the idle temps mark.

Are the Northbridge and Southbridge chipset temps shown here, i can't make them out?

Idle Temps.

Hardware monitor Fintek F71882F
Voltage 0 3.34 Volts [0xD1] (+3.3V)
Voltage 1 1.20 Volts [0x96] (CPU VCORE)
Voltage 2 2.10 Volts [0x83] (VIN2)
Voltage 3 1.15 Volts [0x62] (VIN3)
Voltage 4 4.83 Volts [0x73] (+5V)
Voltage 5 11.88 Volts [0x87] (+12V)
Voltage 6 1.74 Volts [0x6D] (VIN6)
Voltage 7 3.34 Volts [0xD1] (VSB3V)
Voltage 8 3.22 Volts [0xC9] (VBAT)
Temperature 0 50 degC (122 degF) [0x32] (TMPIN0)
Temperature 1 39 degC (102 degF) [0x27] (TMPIN1)
Fan 0 1262 RPM [0x4A5] (FANIN0)
Fan 1 1388 RPM [0x439] (FANIN1)
Fan PWM 0 51 pc [0x82] (CPU)
Fan PWM 1 83 pc [0xD4] (System Fan 1)
Fan PWM 2 83 pc [0xD4] (System Fan 2)
Fan PWM 3 83 pc [0xD4] (System Fan 3)
Register space LPC, base address = 0x0290

00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F
00 FF 03 00 00 FF FF FF FF FF FF 00 51 55 4C 00 00
10 00 00 00 00 FF FF FF FF FF FF FF FF FF FF FF FF
20 D1 9C 81 61 72 87 6D D0 C8 FF FF FF FF FF FF FF
30 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF
40 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF
50 FF FF FF FF FF FF FF FF FF FF 03 04 10 19 34 FF
60 00 00 00 00 FF FF 02 79 00 00 FF 0E 40 24 FF 08
70 FF FF 32 FF 27 FF 80 FF FF FF FF FF FF FF FF FF
80 FF FF 64 4B 64 4B 55 4B FF FF FF FF FF FF A8 FF
90 00 0C 0C 00 26 FF 05 FF 44 02 FF 15 55 55 FF 1A
A0 04 A5 00 82 0F FF 41 3C 37 32 FF C0 A1 82 63 0D
B0 04 57 00 D4 03 FF 3C 32 28 1E FF EA D4 C0 AA 0E
C0 0F FF 0A FE 02 A0 3C 32 28 1E FF D9 A6 80 66 0F
D0 0F FF 0F FF 03 FF 3C 32 28 1E FF D9 B2 99 80 0F
E0 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF
F0 00 00 00 00 00 00 3B FF 00 00 FF 00 FF FF FF FF

Hardware monitor D3D

Hardware monitor NVIDIA NVAPI
Voltage 0 0.91 Volts [0x391] (GPU)
Temperature 0 41 degC (105 degF) [0x29] (GPU)
Fan 0 1530 RPM [0x5FA] (GPU #0)
Fan PWM 0 41 pc [0x29] (FANPWMIN0)
Clock Speed 0 405.00 MHz [0x195] (Graphics)
Clock Speed 1 324.00 MHz [0x144] (Memory)
Clock Speed 2 810.00 MHz [0x32A] (Processor)


Processors
-------------------------------------------------------------------------

Number of sockets 1
Number of threads 4

APICs
-------------------------------------------------------------------------

Socket 0
-- Core 0 (ID 0)
-- Thread 0 0
-- Core 1 (ID 1)
-- Thread 1 1
-- Core 2 (ID 2)
-- Thread 2 2
-- Core 3 (ID 3)
-- Thread 3 3

Timers
-------------------------------------------------------------------------

ACPI timer 3.580 MHz
Perf timer 10.000 MHz
Sys timer 1.000 KHz


Processors Information
-------------------------------------------------------------------------

Socket 1 ID = 0
Number of cores 4 (max 4)
Number of threads 4 (max 4)
Manufacturer GenuineIntel
Name Intel Core 2 Quad Q6600
Codename Kentsfield
Specification Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
Package (platform ID) Socket 775 LGA (0x4)
CPUID 6.F.B
Extended CPUID 6.F
Core Stepping G0
Technology 65 nm
Core Speed 1594.9 MHz
Multiplier x Bus Speed 6.0 x 265.8 MHz
Base frequency (cores) 265.8 MHz
Base frequency (ext.) 265.8 MHz
Rated Bus speed 1063.3 MHz
Stock frequency 2400 MHz
Instructions sets MMX, SSE, SSE2, SSE3, SSSE3, EM64T, VT-x
Microcode Revision 0xBA
L1 Data cache 4 x 32 KBytes, 8-way set associative, 64-byte line size
L1 Instruction cache 4 x 32 KBytes, 8-way set associative, 64-byte line size
L2 cache 2 x 4096 KBytes, 16-way set associative, 64-byte line size
Max CPUID level 0000000Ah
Max CPUID ext. level 80000008h
FID/VID Control yes
FID range 6.0x - 9.0x
Max VID 1.288 V



Temperature 0 63 degC (145 degF) (Core #0)
Temperature 1 62 degC (143 degF) (Core #1)
Temperature 2 52 degC (125 degF) (Core #2)
Temperature 3 55 degC (131 degF) (Core #3)
Clock Speed 0 1594.90 MHz (Core #0)
Clock Speed 1 2392.35 MHz (Core #1)
Clock Speed 2 2392.35 MHz (Core #2)
Clock Speed 3 1594.90 MHz (Core #3)
Core 0 max ratio (effective) 9.0
Core 1 max ratio (effective) 9.0
Core 2 max ratio (effective) 9.0
Core 3 max ratio (effective) 9.0




Bad 💀 Motha 2020 年 12 月 17 日 上午 10:45 
the die surface on any secondary chipset is very tiny, you only need about half a grain of rice worth of paste on there.
Sceptr 2020 年 12 月 17 日 上午 11:27 
引用自 Bad 💀 Motha
the die surface on any secondary chipset is very tiny, you only need about half a grain of rice worth of paste on there.
Yeah it was the first time that i've ever tried applying paste, the Northbridge chipset had a few attemps before the pic you've seen lol. I know now though.
Illusion of Progress 2020 年 12 月 17 日 下午 12:17 
That's probably a little on the high side for the amount of thermal paste, but emphasis on "a little". It's not enough to really hurt anything. It's usually overstated how much of a difference an amount like that and a little less will make (often, next to none). As long as the chipset heatsinks are on and you're not overclocking anything that would impact it (namely, the FSB), you probably don't need to worry about temperatures of those.

CPU is definitely running warm though. I never had one (only had Core 2 Duo E8x00s) so I can't comment on if that's usual for stock cooling, but it's warm regardless, especially for stock clocks. Whether it's going to be a problem is another question though. My current CPU does get near or sometimes slightly above 80C too, but Ryzen is a totally different matter (and highest I've seen is about 83C not 100C). If you're reliably seeing 85C and even above that up to 100C it's probably too warm. You'd probably want to look into better cooling for it.
Bad 💀 Motha 2020 年 12 月 17 日 下午 12:51 
Disable Intel SpeedStep in the BIOS.
Set Windows Power Options to High Performance.
Get a better CPU cooler such as HyperT4 Hyper212 EVO
Illusion of Progress 2020 年 12 月 17 日 下午 2:53 
Disabling Speed Step is going to potentially prevent it from clocking down at idle and may raise idle temperatures (not that it matters much since load temperatures are what matter and the issue here), but it isn't going to help lower load temperatures either. If not overclocking, there's little reason to disable it IMO (and I personally never disabled it even when I was overclocking and had no issues from leaving it on).

Better cooling is probably the only answer here short of undervolting and/or underclocking.
Sceptr 2020 年 12 月 17 日 下午 3:21 
I'm more confident about the chipsets now, i'll add a smaller amount of paste to the both of them and leave them at that. I pointed an infrared laser thermometer on the top of the chipsets heatsinks while i was test booting and the Northbridge read around 39c and the Southbridge around 27c, how accurate that would be i'm not sure but it's something to go off.

I will look for the intel SpeedStep in the BIOS, i don't recall seeing it so far it but i'll have a look, the BIOS is the old MBR type. I was looking at the Hyper212 last night as an alternative cooler but what i can do in the meantime is attach a Corsair H80 that i have in my old backup pc and see what the temps are with that for the Q6600. I will also have to have a word with my sister and see if she can afford another part for this pc,

The pc is for her granddaughter who's 8 for Christmas and the pc parts are from an old pre built Medion pc that i used to have but gave it to my sister back in 2011 i think it was. It wouldn't boot at first so i changed the CMOS battery, cleaned out the PCIE slots and the 3.5gb of RAM slots and it booted ok, then i couldn't find a graphics driver for its GPU a medion x740xl.

The old OS was Vista 32 bit and i wanted to install W10 64 bit into it for the updates so i had to install W7 64 bit first then upgrade to W10 64 bit, luckily the Vista and W7 are legit OEM licences so now it has a legit W10 64bit OS digital licence.

I'm putting the mobo and new 8gb of ram plus a Samsung EVO 500GB SSD and the old HDD into a Corsair Carbide Series 100R case and loaning my HX1000W PSU until she can get a good PSU for herself, i know the SSD is twice as fast as the DDR 2 mobo can handle but i got it and the case for a future build, for now though i don't have a lot to work with.

The case has an outtake fan at the back and i have a spare Corsair fan for a front intake fan, i've had to buy a fan splitter for the only mobo SYSFAN1 three pin fan header which i think is mobo controlled rather stuck on full power. I didn't want to use the CPUFAN1 for an extra fan

The mobo i'm working with is a MSI 7502, here's the manual http://msi-ftp.de/Manuals/MD8828_MS7502.pdf

I'm hoping the Q6600 and maybe a GTX 280 that i'm going to try and cool down or a GTX 570 that i have spare can handle two 24" monitors at 1080p 60Hz
SeriousCCIE 2020 年 12 月 17 日 下午 4:14 
Have you considered adding a small fan to or near the chipset?

In nearly all of my machines, I add a small fan. Sometimes I need a Y splitter to get power to it, other times there is an unused fan header that might work.

Now that you have found the best way to get the heat off the chip and into the heatsink, a fan would work more effectively than before you did any of the hard parts to get to this point!

< >
目前顯示第 1-15 則留言,共 27
每頁顯示: 1530 50

張貼日期: 2020 年 12 月 13 日 下午 12:00
回覆: 27