• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Hoot, I did your Digital Temp Mod, are my eyes playing tricks!?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

AMDGuy

Member
Joined
Apr 10, 2001
Hoot,

I followed your recent article (although now I can't find it on oc.com?) about putting in a digital themometer from crazypc.com. Mine came today (smaller than I thought) and I cut the rubber coating back to expose the wires, then trimmed the coating around the thermistor. I attached it to my Duron 650 @ 952 1.85V with a drop of superglue making sure the thermistor touched the edge of the die. I put my water block back on and fired it up. I watched as the temp quickly rose to 26 C. Room temp according to my cheap radio shack digital is 26.3? Then it goes into windows. At the logon screen I wait to see if the temp continues to rise.....28.4 C. I log in and when MBM5 shows up it reads 40 C (via insocket thermistor reading).

As I type this MBM shows 33 C and the crazy PC digital theometer shows 22.9 C. My RadioShack digital indoor outdoor shows a room temp of 26.3 and an ambient (measured 1" from processor) as 32.0 C. Are these crazyPC.com digital themometer temps reasonable. I guess I'm so used to MBM I'm shocked at how low the temp actually is. I ran prime95 torture for about 30 minutes and never saw temps go over 29 C (making my calculated C/W = .11 using radiate 2? Yet when I stopped the test the temp quickly dropped back to around 20 C. So I know the thermistor is touching the CPU as the temps rise and fall very quickly. I'm just suprised the insocket thermistor is this far off, and seems to get farther off as the temp rises.

What are your thoughts?
 
I would like to see that link that explained how to do that if you can find it.Sounds like the ideal way to monitor temp.
 
I got a senfu probe today and I'm really pleased with it. I'm going to try this mod to it which should work on yours as well really.

I must say the accuracy is very good and it is very sensative so I'm lovin finally knowing my cpu temp :D
 
Um, unless you're using a peltier or chilled water, I don't think your cpu temperature could be *below* ambient. Am I missing something?
 
Phil (Jun 21, 2001 11:17 p.m.):
I got a senfu probe today and I'm really pleased with it. I'm going to try this mod to it which should work on yours as well really.

I must say the accuracy is very good and it is very sensative so I'm lovin finally knowing my cpu temp :D

If you look he has one probe attached at connector position #1. In order to get reading from positions 2-5 he needs 4 more probes to connect.

The only problem I have with it is you either have to buy 4 more probes (to hook to the other 4 sensor locations on the switch) or you have to move the single probe from connector to connector. To me this would not be practical nor cost efficient unless you can buy just the probe for this device.

It is a clean looking install though. And dig that case...lookes like the cockpit of a plane!
 
Flash (Jun 21, 2001 11:48 p.m.):
Um, unless you're using a peltier or chilled water, I don't think your cpu temperature could be *below* ambient. Am I missing something?

That was my concern as well. I know this themometer is accurate as I tested it prior to putting it in. It's also highly responsive. When I held it in my hand it would jump to 34 C or so. When I let it go it would drop back to around 20 C. I also know it's installed properly as I made extra sure it's touching the die. I trimmed the thermistor coating close enough that a pin head size of the thermistor is actually exposed and that is what's touching the die. I then put a drop of superglue on it to attach it to the processor. I don't know what to think about this. That's why I'm hoping Hoot can post similar results with his to confirm this for me.
 
Congrats on your installation of the strip thermistor. Sounds like you did a great job. Yes, you can trust that reading. I am not surprised by the temps with water cooling. The in-socket thermistor error factor is the worst with water cooling. Less so with air cooling. This is due to not having the outflow of air around the CPU, which is normal with an HSF, where the fan blows into the HS.

If anyone wants me to send them a copy of the article, as an attachment to an email (appx 600k w/pictures), let me know.

Hoot
 
Hoot (Jun 22, 2001 08:08 a.m.):
Congrats on your installation of the strip thermistor. Sounds like you did a great job. Yes, you can trust that reading. I am not surprised by the temps with water cooling. The in-socket thermistor error factor is the worst with water cooling. Less so with air cooling. This is due to not having the outflow of air around the CPU, which is normal with an HSF, where the fan blows into the HS.

If anyone wants me to send them a copy of the article, as an attachment to an email (appx 600k w/pictures), let me know.

Hoot

You don't think it's odd that the strip thermistor reports a temp that's actually lower than ambient? This was what concerned me. I guess the radio shack themometer measuring the ambient 1" from the processor could be off due to that little plastic housing around it's thermistor (it reads 32 C) so my concern is that the processor is running below ambient?

Please advise.....I'm having a hard time accepting a Duron 650 @ 952 never goes over 29 C at full load.
 
I would have placed all my various sensors together and compared them before I began the experiment, just to get their relative relationship. I would not be so concerned about what the air temperature is near the CPU, but rather the air temperature measured at the intake side of the radiator. I would also try measuring the water temperature. With my RS Digital thermometer, I "shucked" off the plastic capsule to expose the thermistor inside it. I don't think that capsule upsets the accuracy of the thermometer though. Given the spec on the RS thermometer is something like +/- 1C and the Enermax strip thermometer is +/- 1C, there could feasably be a difference of 2C between them. It's too bad you didn't have a known good sendor and readout to calibrate both your RS and Enermax sensors against, before you started. I am not surprised by your loaded Duron temps with your water cooling system. I set my 600 Duron up the same and my Cohesive Air Cooling (Hoot Chute) got 31C. I would expect a properly implemented water system to do as well, if not better. I guess that is a compliment to your water system. Keep in mind that the c/w of a cooling system, whether air or water, typically goes down a little the lower the wattage device you are cooling. It certainly is not a linear function.
Don't Worry, Be Happy. Besides, it's Friday!

Hoot
 
Ok with a suggestion from Hoot (I shoulda thought of it), I determined the problem. My Enermax digital themometer appears to be defective. I set it next to other digital thermometers and it was always 6 C off. I'm going to contact crazypc.com to see if I can get a replacement. This makes my idle temp 28 C and full load temp 35 - 36 C. This seems more a more reasonable full load temp than the 29 C I was getting.

Now I just gotta see if crazypc will send me a new one. I'll post an update when it arrives.
 
I can gaurantte , unless your running a bong , your water is not gonn get below ambient , musch less your actual core temp, it just aint gonna happen :p
 
Hoot (Jun 22, 2001 08:08 a.m.):

If anyone wants me to send them a copy of the article, as an attachment to an email (appx 800k w/pictures), let me know.

Hoot

I just got two more emails off, but I'm afraid I missed somebody who asked for a copy. Please re-email me if you didn't get a copy of the article.

Hoot
 
Back