Asus A7V333 & 2100+XP Temperature

ViperZ

Member
Jun 11, 2001
43
0
0
Anyone running with this particular combination?

What are your CPU temperatures read from the Asus Probe?

Currently mine is reading 57-60C @ 1898 mhz (13x146) This seems a bit high to me, however even at stock 2100+ XP speeds, the CPU reads somewhere around 55C.

Is this because of the diode in the XP CPU? I had heard the diode reading can be different that themistors on the mother board.

My case seems very cool to the touch and the case temperature is at 30C. There are no hot spots in the case as it seems.

I have enable Qlogic and noticed the CPU fan spin at 3200 when the system is just started and then most times it seems the A7V will run it at around 2700RPM.


Specs...

Asus A7V-333 RAID
AMD 2100+ XP @ 18.2 Ghz
Alpha PAL8045 / 3200 RPM 80mm Sunon Fan
512MB Samsung 2700 DDR RAM
Hercules GF3 @ 490/220 (Gainward Ti 4600 GF4 ... soon)
2 x 40 GB Maxtor DL740 ATA 133 in RAID 0
1 x 40 GB Quantum FireBall ATA 100 Backup Drive
Pioneer 106s DVD drive
Plextor 24x10x40 CDR-W
Intel Pro 100 NIC
SB Live Platinum 5.1
Enermax 431W ps
Lian-Li PC 65

Here is a picture of rig

 

IRJack

Member
Jun 6, 2002
101
0
0
That sounds about right. Before the board died from an attempted bios flash, my A7V333 XP1600+ with Thermalright AX-7 would idle at 49-51C, mobo temp was about 27C.
Currently I'm running the XP1600+ on an Abit KT7A which uses the socket thermal probe. With this setup, the reading is 38-40C idle, which is BS.Everything I've read (on this forum and other web sites) indicates that your ondie diode is much more accurate than a mobo's socket temp sensor.
Hope this helps
 

Mikewarrior2

Diamond Member
Oct 20, 1999
7,132
0
0
Actually, the reviews saying the A7V333 read the internal diode are only partially correct:

Yes, Asus C.O.P (cpu overheat protection), does read from the internal diode of an xp/mp chip, providiing security and thermal protection to a good degree.

Unfortunately, at the same time, user-derived temps in windows are NOT internal diode temps. Fellow Forum member MechBGon gladly donated his time to find this out. Windows readings on an A7V333, even though they are higher than other motherboards, are still the rotten, run of the mill socket-thermistor.

The only difference between the asus socket-thermistor and other socket-thermistors is that asus compensates the readings mroe than others do. But as soon by Phaestus' Procooling.com internal diode reads, even asus compensation cuts it short when compared to actual DIE temperature.



Mike
 

VSEKH

Member
Jun 10, 2002
151
0
0
My AMD 2100+ system, when idle, the temperature stays around 125F and when playing games like quake 3 arena, it would hit around 145F. It is normal for the temperature to go high because the AMD processors have always run hotter than the P4's.
 

ViperZ

Member
Jun 11, 2001
43
0
0
Thanks for your repsonses guys!



Originally posted by: Mikewarrior2
Actually, the reviews saying the A7V333 read the internal diode are only partially correct:

Yes, Asus C.O.P (cpu overheat protection), does read from the internal diode of an xp/mp chip, providiing security and thermal protection to a good degree.

Unfortunately, at the same time, user-derived temps in windows are NOT internal diode temps. Fellow Forum member MechBGon gladly donated his time to find this out. Windows readings on an A7V333, even though they are higher than other motherboards, are still the rotten, run of the mill socket-thermistor.

The only difference between the asus socket-thermistor and other socket-thermistors is that asus compensates the readings mroe than others do. But as soon by Phaestus' Procooling.com internal diode reads, even asus compensation cuts it short when compared to actual DIE temperature.


Mike

Mike, I'm confused by your comment... How does the A7V333 report thermistor CPU readings if it does not have a thermistor under the CPU? The A7V333 and another Board are supposedly the only 2 current boards to provide CPU temperature via the XP's internal diode. Suffice to say if you were to use a T-Bird you could not attain CPU temp for there is no thermistor for the CPU on board.

The temperature reading is the same in the Asus Probe (running in Windows) as it is in the Bios Hardware Monitor... so I asusme it's being read at the same point (XP CPU's internal Diode).

 

ViperZ

Member
Jun 11, 2001
43
0
0
VSEKH, is that 51.6C read from the diode or thermistor?

What Mother board are you using?

I guess that would place me in the ball park then if I'm Overclocked and my 2100+ is running at 2300+ levels. 60C doesn't seem like such a stretch using your values as a comparison.
 

Mikewarrior2

Diamond Member
Oct 20, 1999
7,132
0
0
ViperZ,

the A7v333 uses surface mount resistor to read temps. if you go back to look at the older Asus motherboards, none of htem have the "traditional" thermistor. They all use surface mounted resistors for temp reading. Certainly not accurate, but the hootenany about Asus boards reading too high was very off.

The COP temps are internal diode temps in the bios. However, there's no way to read COP temps within windows. Temperatures from within windows are read from teh surface mount resistor. See this thread for more details:

MechBgon's tests wtih the a7v333... scroll down for them

MechBgon's tests were accurate, since an internal diode would not show the drastic temperature drops when cooling the socket-base :D.


Mike

Of Course, one might ask why there are no internal diode readings from within windows.... Judging from the internal diode tests that Phaestus has done at procooling and voidyourwarranty.net, internal diode temperatures are even higher than the socket-thermistor temps reported by Asus motherboards in the majority of cases. Which motherboard manufacturer wants to be teh first to report internal diode temps of 80C? For example, the Soltek DRV5 supposedly reads the internal diode for its thermal protection scheme, but also does not give an end user the option fo reading that temp in windows.
 

ViperZ

Member
Jun 11, 2001
43
0
0
Thanks for that informative Link Mike, that certainly clears that up, but it does raise some concerns... Why the heck do they not use the diode for all temperture transactions? :disgust:

That link in your signature was a plethora of informaton and education, Thanks :)

Using the tools presented it shows that my theoretical CPU temperature should be 65.0C, I guess that I'm indeed in the ball park and that was my biggest concern. As well it seems MechBgon's temps were at the high 50s to low 60s with a 1700+ XP so it seems I have nothing to worry about.

I do wish that we all really knew what temperature our CPU's are really at.


Great work as well MechBgon, if you happen to read this thread.

__________
*** Added

Unless ambient is assumed to be 24C and not actual case temperature (30C) that would place me at 58.9C, pretty much bang on!

Nice Article!
 

Mikewarrior2

Diamond Member
Oct 20, 1999
7,132
0
0
remember that case temp under industry terms is measured at CPU fan inlet, one inch from the CPU fan...

Usually, its a tad lower than "system" temp on most motherboards, but a lot of factors play into this, including yoru case airflow, etc.

Asus temps have always been the best "guessers' in terms of socket-thermistors, and they aren't "too high" like a lot of people like to say.

As far as why asus didn't allow internal diode readings that end-users could read, who knows the real answer. If you ask me, its because there are a lot of OEM systems out there with sub-coolermaster level performance heatsinks, that would probably show diode temps of 70-80C with system temps around 30.


Mike
 

ViperZ

Member
Jun 11, 2001
43
0
0
Ahh, thanks again Mike!

I guess measuring temp of the air 1" from the Fan of an Alpha 8045 would difficult, especially as I'm expulsing the air away from the heat sink rather than blowing through it.

Close enough :)

It would seem that Asus's temperature guardband is indeed close.

The thing that I find funny is I was told my temps were way too high by people that state their temps are in the low 40's.
 

mechBgon

Super Moderator<br>Elite Member
Oct 31, 1999
30,699
1
0
Great work as well MechBgon, if you happen to read this thread.
Thanks ViperZ, glad it was helpful. I had to know the truth! :D And I will be dancing for joy if Asus ever releases some magical utility or BIOS update that does give us the readings from the on-die thermistor.
 

VSEKH

Member
Jun 10, 2002
151
0
0
I have a Tyan 2466M motherboard running two XP2100+ cpu's(running as MP2100+). It has the winbond W83627HF chip that monitors the temperature and it is via diode. This motherboard has two XP cpu's, so it gets really hot when the weather is warm outside and playing games like Quake III and Unreal tournament. Nice system overall. My friend built the same system with my help once he found out that I was able to get two XP2100+ cpus to work as MP2100+ after connecting the L5 bridge. NOTE: if I use the thermistor option, it shows up as 190F temperature.
 

ViperZ

Member
Jun 11, 2001
43
0
0
mechBgon, that would be nice if a Bios upgrade would do it. What's even funnier is most Hardware review sites have clearly stated the temperature is read via diode :Q

Nice set-up VSEKH!

What type of HSF are you using for those bad boys? :Q
 

ViperZ

Member
Jun 11, 2001
43
0
0
OK, now I'm really confused. There are published sites that indicate a 2100+XP outputs 72 watts under full load. Here

This makes we wonder why the Radiate program calculated the wattage output of a CPU @ 1733 Mhz (2100+) and 1.75 Volts to be 110 watts of output?

How realistic is this and what is the correct value?

Based on that table my CPU theoretical should be 50C, which is 10C lower than what my Asus is reading.
 

Mikewarrior2

Diamond Member
Oct 20, 1999
7,132
0
0
First Off, Was radiate ever updated with AthlonXp Values?

How are you calculating your theoretical temp? What C/W value are you using for the Pal8045?

I've been using AMD power Calc for amd wattages recently. Sorry, I don't have a link handy. E-mail me and I can send you a copy.

Based on AMD POwer Calc, at default clock, it matches the AMD published Max Wattage of your chip. Overclocked to 1898(What Voltage are you using to get to 1898? This is very important in determining heat), with let say a 1.85V vcore yields a cpu wattage of ~86W.

Now let's stake your "system" temp of 30C and a C/W of ~.32. At default clock/voltages, you'd be looking at a core temp under load around 53C.

Now let's take your overclocked wattage and apply it to the same situation. results are around 57C.

My recommendation wouldn't be to use radiate anymore. In part due to the C/W values they include being very faulty. The C/W values were derived from heatsink-thermocouple testing over at overclockers.com, and while accurate to a degree, doesn't give a godo representation of core-temp c/w versus the heatsink.

I still think your asus is pretty close to what your actual temps are. You may be one of the cases were the asus is reading a few C too high, but you're still looking at your motherboard giving you what your actual temps are better than the majority of other motherboards.


Mike
 

ViperZ

Member
Jun 11, 2001
43
0
0
I agree Mike, I think the Asus temp is pretty indicative of actual temperatures....

The C/W (.28) I used was from Alphas's site Here It does list it with a 46CFM fan @ 4000 RPM. So I realize this is not my actual C/W as my value will be higher.

The voltage used was 1.75, however I realized last night it was actually at 1.78

If anything I have learned that the thermal properties of what a CPU is actually at depens on a lot of unknown variables and really can't be modeled accurately in the small scope that I have attempted. It still is fu to gain an understanding of this regardless.

Mike, I'm very interested in that program you speak of, please e-mail it to me if it's not a lot of trouble. Mail To:
 

IRJack

Member
Jun 6, 2002
101
0
0
Mike, what kind of temp monitor setup do you use? I've seen various circuit diagrams that require you to solder leads to the back of the socket a to read the diode directly, and I just don't wanna do that. Also have read about people drill into their hs (just above the cpu core) and epoxying a thermal couple into the hole (again, not for me). These would give a nice mobo-independent reading, but both require extra hardware too.
Do you have any suggestions?
By the way, what temp is asus cop supposed to shutdown the system?
 

Mikewarrior2

Diamond Member
Oct 20, 1999
7,132
0
0
ViperZ,

yeah, which is why testing is a pain even with internal diode readings, but is downright awful when teh testing methods are incorrect ;) :p.


IRJack,

On my p4, I use MBM setup to read the internal diode.... On my athlon XP setup, I no longer use an external thermistor (ala digidoc), but just the socket-thermistor since i'm too lazy to mod it to read the internal diode, and since i've stopped doing personal testing a while back. The important thing when relying on the socket-thermistor is to try and find out what the max for your particular motherboard is.

A digital doc properly positioned can give good results within a few C of the actual core temp (usually will read ab it lower, though).


Mike
 

IRJack

Member
Jun 6, 2002
101
0
0
Thanks for the info, Mike. Nod to laziness.
All the methods to take more accurate temp readings on an XP seem to be more work than it's worth. Dig Doc would be ok, but the wife would probably kill me for buying more hardware.
Just got my A7V333 back from Asus last night, and everything is up and running now. Still reporting the same temps, so at least that's *somewhat* comforting. It's nice that asus went to so much trouble to add the COP feature, then not report the ondie diode. I guess when they say "Designed for XP", they don't mean "eXtreme Precision." :)