Ever see a grown man cry?

[Bantre]

I swear nuclear subs have less cooling then I do for what I use and have better results.
This is a long read so bear with me.
Let me start by saying I have a good idea about airflow/cooling tactics as I have read about 50 zillion webpages and spent hundreds into making a cooling solution for this computer.( Alot of that is into products other then cooling as well though but overall adds to it).
I started off with this m7vib Biostar MB and a AMD 1.4 ghz Thbird.
The system came with a Fanner Tech 80mm intake fan( looks like a cheapy) and a standard PSU with only one rear fan.
The orginal heatsink was a Spree 24cfm type.
My orginal temperatures were 70c Idle 77C load.
Now I know, this seems BLAZING but the system never did crash in the 4 weeks I ran at this temp or seem to suffer slowdown.
I have long concluded my biostar board sucks at reporting temps.
My first changes to this system was to add a Vantec 6035d HSF with a Delta 60mm 37 CFM fan( stock on the 3065d) this shows a CW of .39.I also added a 17cfm triple fan HD cooler for additional intake.
Since my case only supports a 60mm exhaust fan( which it did not have ), I took the 60mm 24 cfm fan off the old heatsink and made it into my exhaust fan.
My results were impressive dropping to 52C Idle and 60-63C load.
At the time I did this last fall the Vantec was one of the higher rated Heatsinks. All copper is was quite nice.
Fast forward to this year:
Recently I added a Enermax 365 with 92mm intake and 80mm exhaust fan built in.
My +5v Stablized to 5.00 and 4.98 under load, and my C dropped a little over 1C in case/CPU.
I was happy!
I ordered round IDE cables as my interior case was a mess and moved all the wiring to bays not in use, so nothing is now restricting my airflow thru the case.
I also changed the intake fan to the enermax 80mm 25/40 CFM in take fan to replace that Fanner Tech I had. I can now feel the air being pulled thru it so it obviously is pushing more air in then before.
The last change I made was to install a brand new Mcxc370 copper Swiftech HSU which has a rating of .32, also with a delta 60mm 37 cfm fan.
My temps became 63C idle and around 70C load( yes much WORSE ).
I have taken the heatsink off multiple times changing the artic silver 2 every time. I tried near transparent and normal, to slightly thick to thick applications and let each burnin 24 hours all with only a small variation in temps.
I checked clearence around the heatsink and with the heatsink lugging all 6 tabs, know it is mounted stable and even.
This is a 40$ heatsink
I have tried different fan locations and positions to have positive and negitive airflow, nothing makes much difference.

Conclusion.---
This is a non overclocked machine!
It is reporting 20-30C higher then others with simular cooling.
Should I just give up and go buy a Nitendo,or apply to work at a heating company?

Also I would like to know of a good bay sensor kit that includes something to check the real as possible cpu temps.
I would love to avoid drilling holes if at all possible into the heatsinks. Anyone have any ideas here on this?

 

[Kendan]

I have long concluded my biostar board sucks at reporting temps.

you have done so much when you know this. I would get a compunurse or similar probe and place it next to the core to get a reliable read on your temps. You can also get a probe at radioshack but I do not know what kind to get from there. If you have never had a stability problem then I would not worry about it. Good luck

 

[Bantre]

Yah, you're right the actual reading is long overdue.
It is not so much stability as just wanting a cooly run system, I just added the orginal intake 80mm Fanner tech fan to the side directly over the CPU blowing in a few minutes ago it dropped it down to 60C so now I have a 80mm enermax frontal intake, a 17 CFM triple Hd cooler on the front blowing in, a 80mm Fanner Tech on the side blowing on the cpu, a mcxc370 with 60mm delta on cpu, a spree ( fanner tech) 24cfm exhaust and a enermax 92mm/80mm PSU sucking out on the back.
You would think for a unoverclocked 1.4ghz I should show something 45 or under, just sucks to see em so high
On the sensor thing, I looked over many websites/ebay but unable to find one that has the option for what I need. I know they make sensors that can go between the heatsink and waffer without rising above the cpu pads to insure good contact but not much luck finding one

This is all essentially a test for my new system I am slowing collection parts for, will be around a xp 1900 system I will start assembling in the next 2-3 weeks.
Cooling will remain a concern on this one as it will become my MP3/DVD/Mpeg machine.
While the other new one will become a gaming machine.

 

[Kendan]

A compunurse will work just fine it has a flat probe. Just do a search on pricewatch or google. here is the same thing under the name enermax. Scroll down and look for "Enermax Temperature Monitors. 2 LED display. 5.25" bay cooling device. " LINK Good luck.

 

[overdoze]

Bantre,
Clean your cpu die. Then look at your cpu die with magnify glass. Pay attention to the corner of your die. Are there any pieces of sillicon chipped off. Your cpu die temp will always be high if you have chipped sillicon die. If the chipped is severe you will get unstable system. If it is minor you can still use your cpu but it will be high temp.

 

[Bantre]

There are 2 black specs on the top side that is near the L1/L2 cache.Dont have a magnifying glass but I have assumed these are small chip on the die.
By specs I mean near pin head sizedand are on the very edge.
So this is probably what is making the temps so high?
I noticed these back when I first got the system and put on the Vantec but back then knew very little of the AMD die.
I came over orginally from a Intel where you could work on them with a hammer

 

[overdoze]

Bantre,
That is the cause of it. I myself came from the celeron family too. Then I ran to the same problem as you. The more I try to fix the problem the worse it became. But I notice one thing the chip keep getting larger and larger until one day the computer refuse to boot. When I replace the cpu, temp get way down like it is suppose to be.

Any one want chipped AMD cpus. I have 6 of them here for sale

 

[Kaese]

Actually, Nuclear Subs have an exstensive cooling system, beyond anything you could imagine. Nevermind about the USS Enterprise, the number of tubes in the Main Condenser alone is mind boggling. Yet with the testing we do, we can detect a 10ml an hour seawater leak.

Anyway, back to your problem. If you have been running at those temps for the past year or more, the life expectancy of your cpu would be reduced to a few more minutes from now.

Get a temp probe, any temp probe to figure out what is actually going on. Compu Nurses or Digidoc's are fine. I don't know if you have updated your bios at all, but sometimes the temp reporting can be screwy. I know from experience with the Chaintech 7SID0, that the temps the board tells you and what they really are can vary greatly.

You should have an in socket thermistor on your board, right under the CPU. I don't know if your system gives you any other temps, but if they are screwy, then you can assume the board is to blame. If the other readouts are normal, then it may be the die.

But find out if the temps are right before doing anything else. You may just have a bad thermistor.

 

[Kendan]

. If you have been running at those temps for the past year or more, the life expectancy of your cpu would be reduced to a few more minutes from now.

Really! I would like to know how you can make that statement. If he has had no stability problems then I would say to hold off judgement on the problem until he gets a true reading of the temps. I would also like to know what kind of thermal compound batre is using.

 

[YMAN]

I would also like to know what kind of thermal compound batre is using.

I have taken the heatsink off multiple times changing the artic silver 2 every time. I tried near transparent and normal, to slightly thick to thick applications and let each burnin 24 hours all with only a small variation in temps.

Artic Silver 2

 

[Kaese]

I can easily make that statement based on the Thermal Guidelines from AMD themselves and that fact that higher temperatures and higher voltages than specified in turn cause a decrease in Processor lifespan and any other component for that matter.

Also, notice the key word "IF". IF the CPU has been running at those temps, then its lifespan will have been considerably shortened. Saying that it may die within the next few minutes may have been a little sarcastic, but that is how I am. But I actually have a lot of experience working in VERY hot places. You may have inferred that from my previous post where I expound upon the cooling systems of Nuclear Power Plants. The normal temps there were in the 120's (Fahrenheit) and the computers that operated there had a very short lifespan. You could expect a year out of them before the Blue Screen of death became a close friend of yours. We would swap CPUs and go on about our business.

Notice I also stated that obtain proper readings or at least a backup of the onboard thermistor readings before going further. I tend not to trust the Motherboard's readings as set in stone.

I am not being nasty or anything like that, I just hate being taken out of context or misunderstood. I am really much better at explain thermodynamics questions or the details of Nuclear Power Systems.

 

[Kendan]

Until he finds the true temp of his chip this whole discusion will not help him. And yes if it has been running at that temp it could die in the next minute or two or last for the next three years. I had not looked back to the top of the thread lately which is why I forgot about the AS2 thing. It just popped into my head. It could also be a chipped core but enough with speculation give him some time to get a true reading of his chip

 

[Bantre]

Heh no reason to start a war here!
I have ordered the enermax dual censor bay kit.
Thanks everyone for their suggestions.
Now comes the next issue.
If the chipped die is causing the heat, I will have to replace it, reguardless being I have used it for well over 6months no problems.
The motherboard I have is a biostar m7vib Ver 1.1, not supporting XP processors( but 1.2 does ).
So in this situation should I A) Buy a new 1.4 Thbird or B) buy a motherboard XP for this system?
I look over everyones suggestions on the MB's but heavy overclocking for me isn't needed, maybe in 2 years when X program comes out and the system isn't so new, but for now even a Tbird is nice.
So IF I replaced the motherboard I would want a cost effective one but not a crappy one( like I have now!) Raid is not needed, I am happy with my single 7200 rpm drive.

Last question:
Once I get the termal sensorand tape it to the waffer near the die( but not on it!) what temperatures SHOULD I expect?

 

[overdoze]

cheap and reliable is my middle name. Shuttle AK31A all the way for $57 refurbished at newegg. BTW, why replace a hot cpu's when it is still working? I had one and gave it to my wife, I underclock it to 1G from 1.33G and she could still read email after 6 months.

 

[Bantre]

Many reasons.
I am a online gamer holding a officer position in a guild.
It is vital I have access when needed.
My roomate runs some of her business thru this computer.
We both have vital emails time to time.
I hate things that are broken
To come home one day and have a meltdown at the wrong day could be very bad, especially if it took over a week to order and recieve a new CPU.
I will wait till I get the sensor and check the temps that way before any actions, but if it is indeed overheating cause of the chip, stability or not , it will have to go.
I place this with driving a car with bald tires as a only means to work that is 40 miles away. It is just not wise to wait till the disaster to clean up the mess.
Or as microsoft does, if it is not broken, fix it.

 

[ac3wooddog]

lol

lol