Where does Epox 8k3a+ read AthlonXP temps from?

[AluminumStudios]

I've read several conflicting posts on where the Epox 8k3a+ mobo reads the temp of an Athlon XP from. Does it use the thermal diode in the middle of the CPU socket, or the thermal diode that is built into the Athlon XP?

I have an XP2000+ (AGOIA) in an 8k3a+ with 512 megs Mushkin PC2700 and an Antec server case with 2 exhause and 1 intake fans and a Volcano 7 HSF with silver thermal compound. The inside of my sytem is cool with good airflow. The CPU heat sink is not very hot to the touch, yet my monitoring software from Epox reports the CPU at 62 c. when under a load in a comfortable room. This is pretty high compared to what other people have said their Athlons run at. I"ve reseated the heat sink and ensure it had good contact and a thin layer of silver comopound. I'm thinking that if my temp is coming from the CPU core and other's are coming from external thermal diodes that might account for the 15-20 degrees difference (I"ve seen people reporting their Athlon XPs at 43 c. with air cooling.)

Any thoughts? My system is rock solid. The CPU has never gone above 63 c. even under intense processing overnight with no A/C on in the room.

It's not overclocked, all voltages and frequencies are defaults.

 

[Mikewarrior2]

Int. Diode for the 8k3a's.


What kind of system temp do you have? Depending on that, your temps are generally fine.



Mike

 

[Markfw]

My XP2000 is in a desktop case with the cover off all the time. I didn't want to buy a new case since it had a 300 watt PSU, so I made a riser for my monitor and left the cover off. My cpu temps run 51c idle and 55-56c full load (seti), and that is with the retail HSF, and a fan blowing air into the area where the cpu is to help keep it cool. This bought me 5c (it was 60-64c). This is on an ASUS A7V333. I thought it read the internal diode, and this would explain my readings (high), but with the case cover off. I think your readings are right on. Mine is also rock solid. If your diode was external, the readings would probably be 15-20C lower. That would be 42-47c, which is what other people are showing with external diodes.

 

[AluminumStudios]

I haven't watched my system temp as closely as my cpu temp. It seems to hang out around 33 c. or so if I remember accurately.

I assume that the "system" temp is from a thermal diode near the north bridge?

If my 8K3A+ is using the XP's thermal diode, then chances are my 62 c. would probably register as something less on another board that uses an external diode ... so my temps. are probably pretty normal (that's what I'm seeking confirmation of.)

I need all of the hoarspower I can get, so I am thinking of overclocking just a tad. I know a 2000+ won't go much futher, but if I can get it to up a 2100+ that would be a little helpful. I just don't want to kill it in the process.

I know AMD says 90 c. is the chips max temp. so I should have a little room to work in (provided the chip is stable.)

 

[AluminumStudios]

Markfw900 - I had to leave the case off my previous PIII system. I didn't want to do that again, so I bought a pretty nice Antec case and loaded it up with fans. it has really good air flow. Prettymuch everything inside feels room temp when I open it up (unlike my overclocked PIII which would have flames shooting out of it when I opened the case

 

[PreDatoR]

The 8K3a+ read the internal thermistor on the XP's... Right now i'm running about 134f with a Dr Thermal EXTREME. Dr thermal HS with a 68CFM delta 80mm on top. In MBM 5191 you can also read the temp from the external thermistor in the socket. When i switch over to that i'm at 96F... If i put the retail HSF on i hit 160's and 170's for temps if i try to overclock. These temps are with a 1700+ @ 1.83 gig and 2.0v's too...

 

[Mikewarrior2]

External Socket-thermistors will read lower than your Int. Diode reading on the 8k3a+. Your temps are generally fine.

As far as the a7v333 readign teh int. diode, it does not. Fellow Forum member Mechbgon kindly did some tests that showed the readings to be from a surface-mount resistor (not uncommon in terms of socket-thermistors). Asus C.O.P. reads the internal diode, but it doesn't give that information to the end user.


Mike

 

[RaynorWolfcastle]

Originally posted by: Mikewarrior2
Int. Diode for the 8k3a's.

Are you positive? Anand's review indicates otherwise
LINK

-Ice

 

[AluminumStudios]

Hmmm conflicting info...

Based on my high CPU temp readings (~62 c.) in a rather well ventilated case, with a decent HSF and circulation I'm inclined to believe that my readings are coming from the CPU's thermal diode. I'm using Epox's USM utility which is in beta form on WinXP.

If my external diod is reading 62 c. then I think I have a heat issue ^_^ but seeing as how my heatsink doesn't feel crazy hot I don't think that's the case.

If anyone else can point me to any documenation that says what the 83ka+ does I'd appreciate it. I seem to remember some articles that mentioned some boards could read the thermal diode, but didn't implement thermal shutdown based on it ... maybe that is what Anand meant in his article?

 

[PreDatoR]

go to amdmb.com forums and ask there you'll have more than one person tell you it does. When the board first came out it wasn't supported but after bios 2312 it is... 62c on the external your internal temps would be in the 80C range if not higher. 60 C on the heatsink if your hand was there long enough would start to burn. I put a compunurse probe on the back of my HS right next t othe core and it was reading 36 which was 1 higher than mbm5 was telling me the external was... If your so worried about it and email Epox about the issue they'll tell you the same thing they told me. Its reading the internal temp...

 

[Mikewarrior2]

Originally posted by: icecool83

Originally posted by: Mikewarrior2
Int. Diode for the 8k3a's.

Are you positive? Anand's review indicates otherwise
LINK

-Ice

Yes, I'm positive.

THere are two settings that work via MBM on this motherboard... the Winbond 2 reading, and the WInbond2 - p2 Diode. When properly set to p2-diode, the 8k3a's (and +) read the internal diode.

Anand's review is incorrect. Anand (gasp), isn't always right. In fact, MechBgon proved anand wrong on the A7v333 intenral diode. It reads the diode, but the user doens't get to see this temp at all.

Further evidence that the reading isn't socket-thermsitor, but rather INt. Diode:

8k3a+ heatsink review

Look at the spread of temps. Compare to any run-of the mill socket-thermistor test. It shows 30+ C spread between teh top and bottom heatsink. Very clear indication that the readings aren't of teh compressed socket-thermsitor variety.

Compare these temps to these review

14C spread.

All indications are that the 8k3a+ is the first single processor consumer level (the fujitsu siemens board was high priced, not intended for end-users) motherboard to give itnernal diode readings to the end-user. The Soltek DRV5 and A7v333, while providing int. diode readings for the thermal safety equipment but not to the end user. The Aberration on the list is the MSI KT3-Ultra. According to MBM and user reports, even though it is supposed to support internal diode reading, it does not. MBM cannot be set to read a diode, and teh temp spread with the kt3-ultra is a mere 4-6C on most systems, which is consistent with compressed socket-thermistor readings.

And yet more evidence the 8k3a+ is a true diode reading motherboard: The reported user temps range from mid 30 to mid 60C. Give me another motherboard with as wide a reported temp range. There aren't any, since the socket-thermistor compression limits temp swings (adjusted for system temp) to 10-12C rather than 30C. Heck, even the a7v333 only shifts temps about 3-6C from idle to load.

AluminumStudios

Use Motherboard Monitor. Set a temp reading to read "winbond2 - Diode". That will then read the internal diode. As seen via this page (ignore the 8k3a non plus, it can be set to diode as well). MBM Epox page


Mike

P.S. Before anyone tries to say that the review linked above with 30+C spread was because the heatsinks were poor, actually read it first. Among the heatsinks 30+ C behind the SLK800 was an SK6 wtih a Low-speed fan.

The large delta was a result of int. diode readings, not because of bad heatsinks or other factors.