XP-2600-M VCore 2.1 heat generation?

[DDogg]

Has anybody actually measured the amount of heat put into the loop by the 2600 XP-m @ ~2.0-2.1 actual VCore? I'm having a devil of a time trying to find out what it actually is.

[bigben2k]

That's not an easy question to answer.

Because you're pumping more power into the chip, it's going to run outside specs, which means that the power output (heat) is going to be higher that the specifications.

How much more? Wouldn't we all like to know!

It's actually very difficult to measure. Right now, only our own pHaestus has the capability of making such a measurement, but he doesn't use that processor type...

[DDogg]

Originally Posted by bigben2k

That's not an easy question to answer.

Because you're pumping more power into the chip, it's going to run outside specs, which means that the power output (heat) is going to be higher that the specifications.

Kinda figured that one out already :-)

Originally Posted by bigben2k

How much more? Wouldn't we all like to know!

That actually makes me feel a little better if you guys don't readily know. I was feeling dumb for not being able to figure it out.

Originally Posted by bigben2k

It's actually very difficult to measure. Right now, only our own pHaestus has the capability of making such a measurement, but he doesn't use that processor type...

That surprises me as the heat load seems to me to be a basic need for the stuff the advanced members do here. Doesn't Cathar run a mobile on an nf7s, or is that my imagination?

Ben, what general part of the Lonestar are you from? Maybe I could buy you coffee sometime if around Big D.

[greenman100]

it actually varies from chip to chip

but think 130w would be a safe bet

[Cathar]

I very roughly estimated ~110-115W while running BurnK7 on an XP-M @ 2.7GHz/1.95v (actual).

At 2.05v actual, would probably guess around the 125W mark, but only when running BurnK7. Expect substantially less heat output when at idle or with "lesser" heat generating programs.

[bigben2k]

Sorry, I'm in Houston, but I still have a visit to the Big D on my list of things to do (you heard me thykingdomcome!); I'll have to take you up on that then

I'd guess 100 to 130 Watts, or roughly 30 Watts above spec, depending on the overclock.

[pHaestus]

give me a day or so and I'll answer this k?

[DDogg]

Originally Posted by pHaestus

give me a day or so and I'll answer this k?

K? Heck yeah!

[Cathar]

Originally Posted by pHaestus

give me a day or so and I'll answer this k?

/me has images of Phaestus's Epox board's MOSFET's overheating and desoldering themselves from the board.

Happened to more than one friend who liked to overclock heavily.

Make sure you stick a fan blowing over the mosfet area Phaestus. Those Epox boards demand that area be actively cooled when pushing hard unless you want instant board death. This is especially a problem with water and/or phase-change cooling where people lose the cooling effect of air coming off a heatsink. Of the two boards I know of that died due to heavy overclocking, one was phase-change, one was water-cooling.

[psychofunk]

Originally Posted by pHaestus

give me a day or so and I'll answer this k?

Great! Thanks Phaes. I am interested in this too as it is me well 2.7@2.125 but 2.6@2.1 is close enough.

[DDogg]

Yep, fan is critical on that area. Caps and mosfets get spit sizzling hot without it. A 7 volted 60m handles it ok on mine. Now, as to whether it actually getting hot enough to unsolder ...

[ferdb]

It used to be in the good old days that power dissipation in CMOS chips increased linearly with clock rate and by the square of the supply voltage. So double the clock rate, double the heat, Double the supply voltage 4 times the heat, Double both and it's 8 times. These relationships still hold but these days with .13 -.09 micron dimensions the leakage currents are very significant and so add another dissipation factor on top of the previous 2 so you can't so easily extrapolate from known dissipations.

Another major factor is that with increased Vcore the gate drive voltage is higher and the on resistance of the devices goes down (which is the whole idea of increasing Vcore to make it run faster) and the totem poles spiking current increases. This is the current where a gate is switching and both N and P devices are briefly on at the same time as they transition and essentially shorting the supply. The totem pole current increases with Vcore. It also increases with the reduction of switch resistance from increasing Vcore which means it goes up even faster than just 1:1 with voltage. And as the devices heat up the turn on thresholds of the transistors decreases so the totem poles now short out for a longer periods of time on each transistion.

What all this means is that increasing Vcore increases dissipation at a very high rate.
Increasing Vcore reduces switch resistance which makes things switch faster which is why we overvolt, but it produces a lot more heat. More heat produces higher temperatures which increase switch resistance which slows things down, which is why we try to cool as much as possible. At some point increasing Vcore produces less speed gain than the increased temperatures reduce the speed.

So use the minimum Vcore possible that will do the job.

[ferdb]

Originally Posted by DDogg

Yep, fan is critical on that area. Caps and mosfets get spit sizzling hot without it. A 7 volted 60m handles it ok on mine. Now, as to whether it actually getting hot enough to unsolder ...

I've seen parts unsolder before. If it's a semiconductor though it usually destroys the device on it's way to those temperatures.

[Ruiner]

FWIW, I sat a 92mm fan vertically on my vid card so that it blows air on both the mosfets and NB passive HS. It's dead quiet at 7v, and keeps things cool.

[DDogg]

Originally Posted by ferdb

It used to be in the good old days that power dissipation in CMOS chips increased linearly with clock rate and by the square of the supply voltage. So double the clock rate, double the heat, Double the supply voltage 4 times the heat, Double both and it's 8 times. These relationships still hold but these days with .13 -.09 micron dimensions the leakage currents are very significant and so add another dissipation factor on top of the previous 2 so you can't so easily extrapolate from known dissipations.

Another major factor is that with increased Vcore the gate drive voltage is higher and the on resistance of the devices goes down (which is the whole idea of increasing Vcore to make it run faster) and the totem poles spiking current increases. This is the current where a gate is switching and both N and P devices are briefly on at the same time as they transition and essentially shorting the supply. The totem pole current increases with Vcore. It also increases with the reduction of switch resistance from increasing Vcore which means it goes up even faster than just 1:1 with voltage. And as the devices heat up the turn on thresholds of the transistors decreases so the totem poles now short out for a longer periods of time on each transistion.

What all this means is that increasing Vcore increases dissipation at a very high rate.
Increasing Vcore reduces switch resistance which makes things switch faster which is why we overvolt, but it produces a lot more heat. More heat produces higher temperatures which increase switch resistance which slows things down, which is why we try to cool as much as possible. At some point increasing Vcore produces less speed gain than the increased temperatures reduce the speed.

So use the minimum Vcore possible that will do the job.

ferdb, thanks for that. While I was not going to question the opinion from my betters (and still don't), my intuition from watching the temps over several months at different VCore ranges suggested that the heat curve from increased VCore might be more like a 'hockey stick' graph. I've got a suspicion that the heat might be more like 145-150 watts at 2.15-2.2 actual. Nothing but a gut feeling, though.

As for measuring, I have a true RMS DMM. I'm told the true RMS is important for accurately measuring AC current. My fuzzy thinking is centered around measuring the AC current at stock settings and accepting the mfgs heat numbers as a baseline, then measuring again at the higher VCores. I'm wondering if you, or anybody else, thinks this might be a crudely acceptable method for finding useful information on heat generation.

[DDogg]

Ph, not a push, more curiosity if you still plan to do these tests. If it would be risky in your mind, then it is probably not a good idea.

Plus, a bump on the question above regarding measuring the AC current.

[pHaestus]

Sorry I didn't have time to mess with it before I went out of town Friday. I'm going to sleep tonight and not mess with computer stuff at all. Maybe I'll get it done this week though. I have no real fear of hardware damage... I'll have to modify the mobo to get big multipliers because the 2500-M I have is only 11x I think.

[Blackeagle]

Thaks for your efforts pH, and add me to the list interested in your results.

[DDogg]

pH, just another polite and curious bump.

[DDogg]

and another