(follow up) Would you be willing to cut into your CPU socket? - Page 3

jaydee · 06-22-2003, 06:50 PM

Alright i just did this socket mod on my ECS K7S5A. Being i fanally got around to downloading SetFSB so I can adjust the FSb in windows.

So here is the oddities. I got the new thermocoupler type K last week and thats what I used. I half assed calibrated it with another probe I have testing them against room temp and then in a hot glass of water while I stired it. The temps rose and fell pretty damn close together. I was satisfied enough. But then i got the probe installed and a HS mounted and booted the system and got a little freeked out. The thermocoupler shot up instantly and just kept going. Once in windows I turned on MBProbe checket the temps against the thermocoupler. MBProbe said 48C and the thermocoupler said 66C. I thought that was alittle odd. Then turned on SETI and ran it for a while. MBprobe went up to 52C and and the thermocoupler when up to 69-70C flipping back and forth. BAH! Guess I will have to start over.

joemac · 06-22-2003, 08:38 PM

It would be so much easier if AMD included a circuit in the CPU that read the diode and was auto calibrating- The CPU already has millions of transistor what a dozen more ? till then mod on ..

jaydee · 06-22-2003, 08:49 PM

Went over everything again. Reseated everything and added a better fan the HS. Was using a 15CFM 60mm above now have a 38CFM on there.

Here is what I got: (24C room temp)
XP1600+ 1400mhz 1.75V bone stock ECS K7S5A.

Bios Idle:
Onboard 39C
TC 49C

WinME Idle:
Onboard 39C
TC 49C

Running SETI:
Onboard 41C
TC 53C

XP1600+ 1540mhz 146FSB 1.75V Using SetFSB (http://www.vector.co.jp/soft/win95/h.../se239139.html) in Windows.
Onboard 43C
TC 57C

I have to lean towards the TC. As you can see there was only a 2C jump from idle to load from the onboard. Certainly not what I would have expected. But the TC shows a 4C gain which is more realistic with a SETI load. I think the onboard probe is calibrated wrong. Seems the higher the temps get on the TC the less they jump on the onboard by a consistant 2C. I am not to sure what to think about this right now.... Will be trying this on my Abit KD7 when it arrives later next week and will see what happens with it.

pHaestus · 06-22-2003, 11:02 PM

It freaks you out when you see how hot your CPU is actually running eh Jaydee?

hehe

jaydee · 06-22-2003, 11:13 PM

Quote:

Originally posted by pHaestus
It freaks you out when you see how hot your CPU is actually running eh Jaydee?

hehe

I was expecting high temps because of the cheap aircooler, but sure wasn't expecting a true 10C-15C over the onboard. And the onboard for some reason starts loosing consistancy after 40C. Seems for every 2C the onboard goes up the TC goes up 4C. Weird stuff. Will see how the Abit makes out, but if this is a consistant problem on all boards then you can't even attemp to use relative temps from the onboard. Certainly a bad thing. Who knows if their board has a off the wall probe reader or not unless they really do put a external probe under there. More to come on this later. Just preliminary stuff here.

Groth · 06-23-2003, 05:43 AM

In the 30-40 range, the socket thermistor in my Abit KX-7 reads almost exactly the same as the diode temperature. At least it does now, after I carefully bent it's legs to get solid contact with the CPU and insulated the socket with a chunk of old mouse pad.

During a test of how hot my Palomino would tolerate, I throttled the water flow down to ~1 liter/hour and watched the temperatures creep up. The difference between the two temps grew (and grew at a faster rate) the hotter it became. After a hair over an hour, the final reading before a spontaneous reboot was 73 thermistor, 101 diode.

Ketchak · 06-23-2003, 07:20 PM

Quote:

It freaks you out when you see how hot your CPU is actually running eh Jaydee?

I am not jaydee

but I also made some tests and the CPU this very hot

Hot Hot Hot

All my temperatures are in the table of Excel

I need a good thermometer

Ketchak.
www.devilmaster.cjb.net

pHaestus · 06-23-2003, 11:54 PM

I knew I had a graph somewhere...

The 6657 is the diode reader, the CF633 is a dallas one wire sensor epoxied under the CPU core, Asus CPU is my wonderful onboard CPU measurement from an A7V-133, and Asus JTPWR is a flat thermistor placed on the side of the chip touching the core. CPU was a Duron 1GHz with Morgan core.

jaydee · 07-01-2003, 08:21 PM

Quote:

Originally posted by jaydee116
Went over everything again. Reseated everything and added a better fan the HS. Was using a 15CFM 60mm above now have a 38CFM on there.

Here is what I got: (24C room temp)
XP1600+ 1400mhz 1.75V bone stock ECS K7S5A.

Bios Idle:
Onboard 39C
TC 49C

WinME Idle:
Onboard 39C
TC 49C

Running SETI:
Onboard 41C
TC 53C

XP1600+ 1540mhz 146FSB 1.75V Using SetFSB (http://www.vector.co.jp/soft/win95/h.../se239139.html) in Windows.
Onboard 43C
TC 57C

Now I am real freaked out. I dropped on my Mini Channel block showed below. Now the temps have wildly changed and I don't mean overall temp wise, I mean consistancy wise.

Now under a load the TC reads 41C and MBprobe (not MBM as used above) reads 44C!!! The onboard is warmer then the TC. Only thing that changed is the water cooling setup replaced the air cooler! :shrug:

hydrogen18 · 07-02-2003, 12:43 PM

how does one modify the socket for the pentium 4 socket 478 to allow wires to be pased around the pins?

jaydee · 07-13-2003, 07:36 PM

I don't know about a P4 system. Wouldn't think it would be any different as the sockets are designed similar I think. The P4 socket is smaller though.

Anyway earlier today I did the mod on my Abit KD7. Instead of poping the top off I just dremeled it. I had it all setup already with oneof the meatal cutters. Took about a minute to make a nice little channel. The KD7 onboard is reading 3C cooler then the thermocoupler. AND I figured out what happened above. The battery in my multimeter was just about dead giving false readings. Changing the batter brought it back to normal.

hydrogen18 · 07-13-2003, 08:13 PM

the pins are so dense, and the spacing so small, considering expoying a probe onto the side of the IHS or taping to the back of the mobo...

jaydee · 07-13-2003, 08:38 PM

Quote:

Originally posted by hydrogen18
the pins are so dense, and the spacing so small, considering expoying a probe onto the side of the IHS or taping to the back of the mobo...

Probably be better off with the onboard than either of those options. Not sure though.

hydrogen18 · 07-14-2003, 12:47 AM

i have this suspicision the chip that reads my onboard temps if fu}{}{0r3|). Because no matter what cooler or what fan speed, it constantly reads 42.5 C. I even put the voltage up to 1.7 volts once and saw no change in temperature...

jaydee · 07-14-2003, 09:20 AM

Quote:

Originally posted by hydrogen18
i have this suspicision the chip that reads my onboard temps if fu}{}{0r3|). Because no matter what cooler or what fan speed, it constantly reads 42.5 C. I even put the voltage up to 1.7 volts once and saw no change in temperature...

Maybe the CPU is throtteling down to keep that temp? Or maybe the probe is just busted....

Boli · 07-14-2003, 10:54 AM

Trust me its busted... mine on an OC mission rocketed up to 99 (max it would go) before the system shut itself down.... (I must have knocked my WB off its mounting while reseting my BIOS.)surprisingly the only damage was... nothing. Only my CPU temp pernamently reads 35*.

hydrogen18 · 07-14-2003, 12:28 PM

my zalman may be mounted wrong, damn screw in system...

Nuson · 07-27-2003, 03:12 AM

what's wrong with temp probes onboard again? Besides the thrill of nearly ruining your motherboard for better temp readings, why would you do this? Seems like a pretty expensive risk if you ask me, esp if you have on die or motherboard sensors.... :\

jaydee · 07-27-2003, 10:48 AM

Quote:

Originally posted by Nuson
what's wrong with temp probes onboard again? Besides the thrill of nearly ruining your motherboard for better temp readings, why would you do this? Seems like a pretty expensive risk if you ask me, esp if you have on die or motherboard sensors.... :\

If your worried about runining your board your on the wrong site.

We are talking about people trying to test their blocks here. You can't test blocks with +-15C accuracy the onboards give. And as I am finding out the onboards on not consistant or liniar. Also we are tired of people spamming their usless, laws of physics breaking, impossible temps around that no one can use. Note the title of this site starts with "Pro". Things are done differently here than your average half assed tech site that is more interested in it's members clicking their adds and feeling good about what they post even if it's wrong.

There is plenty of articles and threads on this site covering this subject. Look for them.

Here is one that was responsible for this thread starting:
http://forums.procooling.com/vbb/sho...&threadid=6943

Nuson · 07-28-2003, 01:34 AM

Must I suffer the eternal damnation of losing all my posts!? I need to learn to copy before hitting submit or preview... (my second tries never seem to be as good as the first

)

I'm not AGAINST socket modding, but not everyone has the money to backup their board if something goes horribly wrong, and even if they do have the money, it's still a pretty big investment. I hope the "pro" denote that members are infalible or privileged, jd...

And it's not like socket mods are going to save us from bogus reports of cooler performance either, unless you can convince everyone that this is the only way to go which seems improbable since a good 40~ percent of those polled would not chop their socket...

My motherboard seems to report consistant temps and not even modded sockets can beat on die readings. Lots of smart hardware people, including hardocp.com and tomshardware.com, trust onboard reading devices; why shouldn't I?

jaydee · 07-28-2003, 09:23 AM

Quote:

Originally posted by Nuson
Must I suffer the eternal damnation of losing all my posts!? I need to learn to copy before hitting submit or preview... (my second tries never seem to be as good as the first )

I'm not AGAINST socket modding, but not everyone has the money to backup their board if something goes horribly wrong, and even if they do have the money, it's still a pretty big investment. I hope the "pro" denote that members are infalible or privileged, jd...

And it's not like socket mods are going to save us from bogus reports of cooler performance either, unless you can convince everyone that this is the only way to go which seems improbable since a good 40~ percent of those polled would not chop their socket...

My motherboard seems to report consistant temps and not even modded sockets can beat on die readings. Lots of smart hardware people, including hardocp.com and tomshardware.com, trust onboard reading devices; why shouldn't I?

Now I am pretty sure your just trolling for a flame war. Go back to [H] or toms where you belong.

bigben2k · 07-28-2003, 09:28 AM

Quote:

Originally posted by Nuson
My motherboard seems to report consistant temps and not even modded sockets can beat on die readings. Lots of smart hardware people, including hardocp.com and tomshardware.com, trust onboard reading devices; why shouldn't I?

You're going to have to go back to the first post, and track down the other threads that specifically discuss why the on-board probe is useless.

As for socket modding, it's an alternative, not a single solution: you would have picked that up if you had read the entire thread (I know, it's long!).

If you want to mention Tom and [H], you ought to run a search within these forums, to find out exactly how we feel about them...

jaydee · 07-28-2003, 10:27 AM

Quote:

Originally posted by bigben2k
You're going to have to go back to the first post, and track down the other threads that specifically discuss why the on-board probe is useless.

As for socket modding, it's an alternative, not a single solution: you would have picked that up if you had read the entire thread (I know, it's long!).

If you want to mention Tom and [H], you ought to run a search within these forums, to find out exactly how we feel about them...

He knows how we feel about [H] and toms and he knows how serious we are about block design. He's just trolling. Thats why he posts garbage like he has in this post: http://forums.procooling.com/vbb/sho...662#post82662.

Am I the only one that finds it odd he manages to touch upon all the issues that stirs things up around here?

Quote:

Originally posted by Nuson
SHIT! I lost my reply because my login time expired. That happens to me alot ... Here's my best recreation of what I wote:

I don't need to be insulted, and I don't need to brag about my physics knowledge to qualify what I say. Water block design is not exactly serious thermal engineering. The chaotic nature of the physical processes happening inside a water block practically rules out this type of indepth analysis completely. It's more a matter of trial and error and best guesses.

If I didn't take your word for it, jd, it should be of no suprise to you, since it is the whole nature of scientific investigation to not rely a small number of results. I would have more readily believed you if you had referred me to other reviews instead of simply yelling seniority. I never even disagreed with you that those results don't make sense. Why are you getting so excited about it?

Quote:

Originally posted by Nuson
I lost my post again, but I don't think I timed out Oh well...

For your information, jd, I did do my own research, but only after no one would help me out. In the past, I have had pretty bad experiences searching for water cooling info on the web, so it was pretty frustrating for me. My first reaction to this thread, therefore, was not to crawl the net searching for more info. I guess times have changed... (consider this: the last round up of water cooling products here was in jan 2002 so that sorta illustrates the still sparceness of reliable water cooling info on the web).

Simply designing and building water blocks does not constitute thermal engineering. Engineering is a design process which involves the ability to accurately predict how the design will work. I'm not saying that your experience has not helped you to design and build effective water blocks, but it's not an exact science as you've made it out to be.

Roady · 07-28-2003, 10:38 AM

i'd cut up my zif (is it zif or ziff?) socket, but being a modder i'm not afraid of plastic

i'm new and only read 4-5 posts through but shall be on a catchup mission tonight, excellent content so far guys!

bigben2k · 07-28-2003, 11:31 AM

Welcome to ProCooling!

This thread stands well on its own, but for background info (i.e. why I asked), you have to refer to the linked threads, in post#1.

BTW, it's ZIF (Zero Insertion Force).

I proposed cutting a lot more of it earlier, but it's really not that wise:
http://forums.procooling.com/vbb/sho...t=cpu+backside

View Poll Results: Would you chop your CPU socket?
Yes	63	46.32%
No	73	53.68%
Voters: 136. You may not vote on this poll

07-27-2003, 03:12 AM	#68
Nuson Cooling Neophyte Join Date: Jul 2003 Location: USA Posts: 41	what's wrong with temp probes onboard again? Besides the thrill of nearly ruining your motherboard for better temp readings, why would you do this? Seems like a pretty expensive risk if you ask me, esp if you have on die or motherboard sensors.... :\ __________________ Frag Out!

07-28-2003, 01:34 AM	#70
Nuson Cooling Neophyte Join Date: Jul 2003 Location: USA Posts: 41	Must I suffer the eternal damnation of losing all my posts!? I need to learn to copy before hitting submit or preview... (my second tries never seem to be as good as the first ) I'm not AGAINST socket modding, but not everyone has the money to backup their board if something goes horribly wrong, and even if they do have the money, it's still a pretty big investment. I hope the "pro" denote that members are infalible or privileged, jd... And it's not like socket mods are going to save us from bogus reports of cooler performance either, unless you can convince everyone that this is the only way to go which seems improbable since a good 40~ percent of those polled would not chop their socket... My motherboard seems to report consistant temps and not even modded sockets can beat on die readings. Lots of smart hardware people, including hardocp.com and tomshardware.com, trust onboard reading devices; why shouldn't I? __________________ Frag Out!

07-28-2003, 11:31 AM	#75
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Welcome to ProCooling! This thread stands well on its own, but for background info (i.e. why I asked), you have to refer to the linked threads, in post#1. BTW, it's ZIF (Zero Insertion Force). I proposed cutting a lot more of it earlier, but it's really not that wise: http://forums.procooling.com/vbb/sho...t=cpu+backside __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

06-22-2003, 06:50 PM	#51
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Alright i just did this socket mod on my ECS K7S5A. Being i fanally got around to downloading SetFSB so I can adjust the FSb in windows. So here is the oddities. I got the new thermocoupler type K last week and thats what I used. I half assed calibrated it with another probe I have testing them against room temp and then in a hot glass of water while I stired it. The temps rose and fell pretty damn close together. I was satisfied enough. But then i got the probe installed and a HS mounted and booted the system and got a little freeked out. The thermocoupler shot up instantly and just kept going. Once in windows I turned on MBProbe checket the temps against the thermocoupler. MBProbe said 48C and the thermocoupler said 66C. I thought that was alittle odd. Then turned on SETI and ran it for a while. MBprobe went up to 52C and and the thermocoupler when up to 69-70C flipping back and forth. BAH! Guess I will have to start over.

06-22-2003, 08:38 PM	#52
joemac Cooling Savant Join Date: Dec 2002 Location: Dallas Posts: 339	It would be so much easier if AMD included a circuit in the CPU that read the diode and was auto calibrating- The CPU already has millions of transistor what a dozen more ? till then mod on ..

06-22-2003, 08:49 PM	#53
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Went over everything again. Reseated everything and added a better fan the HS. Was using a 15CFM 60mm above now have a 38CFM on there. Here is what I got: (24C room temp) XP1600+ 1400mhz 1.75V bone stock ECS K7S5A. Bios Idle: Onboard 39C TC 49C WinME Idle: Onboard 39C TC 49C Running SETI: Onboard 41C TC 53C XP1600+ 1540mhz 146FSB 1.75V Using SetFSB (http://www.vector.co.jp/soft/win95/h.../se239139.html) in Windows. Onboard 43C TC 57C I have to lean towards the TC. As you can see there was only a 2C jump from idle to load from the onboard. Certainly not what I would have expected. But the TC shows a 4C gain which is more realistic with a SETI load. I think the onboard probe is calibrated wrong. Seems the higher the temps get on the TC the less they jump on the onboard by a consistant 2C. I am not to sure what to think about this right now.... Will be trying this on my Abit KD7 when it arrives later next week and will see what happens with it.

06-22-2003, 11:02 PM	#54
pHaestus Big Player Making Big Money Join Date: Aug 2001 Location: irc.lostgeek.com #procooling.com Posts: 4,782	It freaks you out when you see how hot your CPU is actually running eh Jaydee? hehe

06-23-2003, 05:43 AM	#56
Groth Cooling Savant Join Date: Mar 2003 Location: MO Posts: 781	In the 30-40 range, the socket thermistor in my Abit KX-7 reads almost exactly the same as the diode temperature. At least it does now, after I carefully bent it's legs to get solid contact with the CPU and insulated the socket with a chunk of old mouse pad. During a test of how hot my Palomino would tolerate, I throttled the water flow down to ~1 liter/hour and watched the temperatures creep up. The difference between the two temps grew (and grew at a faster rate) the hotter it became. After a hair over an hour, the final reading before a spontaneous reboot was 73 thermistor, 101 diode.

06-23-2003, 11:54 PM	#58
pHaestus Big Player Making Big Money Join Date: Aug 2001 Location: irc.lostgeek.com #procooling.com Posts: 4,782	I knew I had a graph somewhere... The 6657 is the diode reader, the CF633 is a dallas one wire sensor epoxied under the CPU core, Asus CPU is my wonderful onboard CPU measurement from an A7V-133, and Asus JTPWR is a flat thermistor placed on the side of the chip touching the core. CPU was a Duron 1GHz with Morgan core.

07-02-2003, 12:43 PM	#60
hydrogen18 Cooling Savant Join Date: Feb 2003 Location: USA Posts: 248	how does one modify the socket for the pentium 4 socket 478 to allow wires to be pased around the pins?

07-13-2003, 07:36 PM	#61
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	I don't know about a P4 system. Wouldn't think it would be any different as the sockets are designed similar I think. The P4 socket is smaller though. Anyway earlier today I did the mod on my Abit KD7. Instead of poping the top off I just dremeled it. I had it all setup already with oneof the meatal cutters. Took about a minute to make a nice little channel. The KD7 onboard is reading 3C cooler then the thermocoupler. AND I figured out what happened above. The battery in my multimeter was just about dead giving false readings. Changing the batter brought it back to normal.

07-13-2003, 08:13 PM	#62
hydrogen18 Cooling Savant Join Date: Feb 2003 Location: USA Posts: 248	the pins are so dense, and the spacing so small, considering expoying a probe onto the side of the IHS or taping to the back of the mobo...

07-14-2003, 12:47 AM	#64
hydrogen18 Cooling Savant Join Date: Feb 2003 Location: USA Posts: 248	i have this suspicision the chip that reads my onboard temps if fu}{}{0r3\|). Because no matter what cooler or what fan speed, it constantly reads 42.5 C. I even put the voltage up to 1.7 volts once and saw no change in temperature...

07-14-2003, 10:54 AM	#66
Boli Cooling Savant Join Date: Apr 2003 Location: Blackburn / Dundee Posts: 451	Trust me its busted... mine on an OC mission rocketed up to 99 (max it would go) before the system shut itself down.... (I must have knocked my WB off its mounting while reseting my BIOS.)surprisingly the only damage was... nothing. Only my CPU temp pernamently reads 35*.

07-14-2003, 12:28 PM	#67
hydrogen18 Cooling Savant Join Date: Feb 2003 Location: USA Posts: 248	my zalman may be mounted wrong, damn screw in system...

07-28-2003, 10:38 AM	#74
Roady Cooling Neophyte Join Date: Jul 2003 Location: Mid-Wales, UK Posts: 3	i'd cut up my zif (is it zif or ziff?) socket, but being a modder i'm not afraid of plastic i'm new and only read 4-5 posts through but shall be on a catchup mission tonight, excellent content so far guys!

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)