Heatpipe investigation

http://www.phoronix.com/scan.php?page=article&item=892&num=1

Enjoy

OCZ, Thermalright, etc etc investigated

I wonder how long it would take for notlag to come in and tell us that this proves that the ultra 120 xtreme has bad heatpipes

lol, anyway, heatpipes should contain a bit of liquid no? a miniscule amount is more then enough...not liters of fluid

Thats interesting to read that most of the heatpipes they opened had no liquid in that at all. i wonder if they put any in, or does the liquid escape over time? wtf?

I am annoyed now at my V1 Ultra. Oh well, I still say the temps kick butt for onlt $15

Very interesting.

Where's Notlag?

Drinkyoghurt said:

I wonder how long it would take for notlag to come in and tell us that this proves that the ultra 120 xtreme has bad heatpipes

Click to expand...

AHAHHHAHAHAHHAHAHHAHAHAH ohh that made my morning.... im sure he will use it in his pethetic arsenal of excuses why his ultra120's are crap...

i think the liquid used in heatpipes is so little that it evaporates upon opening.

A most excellent read for sure. Makes me wonder as well if there's any liquid in the heatpipe coolers.

I was contemplating getting an HR-03 for the video card. Hmmm, makes you think doesn't it?

Now I wonder does Zalman put any liquid in their coolers?

Drinkyoghurt said:

I wonder how long it would take for notlag to come in and tell us that this proves that the ultra 120 xtreme has bad heatpipes

lol, anyway, heatpipes should contain a bit of liquid no? a miniscule amount is more then enough...not liters of fluid

Click to expand...

If this was a test of a U120X I might On a bit of a tangent though, were you thinking about validating your CoreTemp readings, Drinkyoghurt (ie - turning the clock speed and voltage down as low as possible to boot, then verifying the temps were still over ambient)?

And yes, heat pipes should contain a small amount of fluid. If there's too much or too little then they won't operate properly. They also won't operate properly if the mix of the fluid is wrong. Ideally the fluid needs to be able to boil at the temperatures created by a CPU even when idle. But they also need to have a high enough boiling point to remain liquid at room temperature and to be able to condense in the upper parts of the tube even when the CPU is overclocked at and at full load. Different fluids (other than water) are typically used to achieve those conditions, and sometimes pressure within the tube can be controlled to adjust the boiling point.

As for no fluid coming out of the pipes he cut... The heat of machining might be enough to cause the fluid to become gaseous. For that matter, the heat of the human body may be sufficient too. We're around 37C (even if our skin temp is not). He may have caused the liquid to boil just by handling it for an extended period of time. On top of it, there could have been pressure inside the tube used to raise the boiling point and when he released the pressure by cutting the tube that's the gas he observed escaping (and once the pressure was let off the boiling point went down and the heat of his hand boiled off the rest of the liquid).

Who knows.

EDIT: Admittadly, he didn't observe pressure changes on all of them, but the change may be subtle. I don't honestly know how much pressure or fluid there should be.

nd4spdbh2 said:

AHAHHHAHAHAHHAHAHHAHAHAH ohh that made my morning.... im sure he will use it in his pethetic arsenal of excuses why his ultra120's are crap...

i think the liquid used in heatpipes is so little that it evaporates upon opening.

Click to expand...

Funny.

Anyways, for those who are interested, here is a reference on heatpipe operation in the context of CPU cooling. It's not too bad and fairly concise:

http://www.heatsink-guide.com/content.php?content=heatpipes.shtml

And I'm sure other folks can decide for themselves whether I'm competent enough to have reasonable explanations for my U120X's don't perform well. It's not my fault if you can't follow the reasoning in my Thermalright thread or have gotten to a point where you can no longer poke reasonable holes in it and have to resort to taking jabs at me instead. People can form their own opinions.

ty1er said:

Thats interesting to read that most of the heatpipes they opened had no liquid in that at all. i wonder if they put any in, or does the liquid escape over time? wtf?

Click to expand...

Bare copper pipe of the dimensions of a heat pipe would have too low of thermal conductivity to achieve good cooling performance. Heat pipes can transfer heat orders of magnitude faster than bare tubing. In theory the liquid could escape over time when it is gaseous, especially if the tubes are poorly sealed, but one would notice a definite increase in temps over time as it did so.

However, in some ways, perhaps getting the heat pipes' pressures and fluids just right may be difficult to do cheaply and consistently on assembly line, and that may be where fluctuations in heat sink quality / performance come into play within a model line? If there was no fluid the heat sink would perform very badly, but if the fluid was there, but not quite right...

Notlag said:

If this was a test of a U120X I might On a bit of a tangent though, were you thinking about validating your CoreTemp readings, Drinkyoghurt (ie - turning the clock speed and voltage down as low as possible to boot, then verifying the temps were still over ambient)?

And yes, heat pipes should contain a small amount of fluid. If there's too much or too little then they won't operate properly. They also won't operate properly if the mix of the fluid is wrong. Ideally the fluid needs to be able to boil at the temperatures created by a CPU even when idle. But they also need to have a high enough boiling point to remain liquid at room temperature and to be able to condense in the upper parts of the tube even when the CPU is overclocked at and at full load. Different fluids (other than water) are typically used to achieve those conditions, and sometimes pressure within the tube can be controlled to adjust the boiling point.

As for no fluid coming out of the pipes he cut... The heat of machining might be enough to cause the fluid to become gaseous. For that matter, the heat of the human body may be sufficient too. We're around 37C (even if our skin temp is not). He may have caused the liquid to boil just by handling it for an extended period of time. On top of it, there could have been pressure inside the tube used to raise the boiling point and when he released the pressure by cutting the tube that's the gas he observed escaping (and once the pressure was let off the boiling point went down and the heat of his hand boiled off the rest of the liquid).

Who knows.

Click to expand...

can't back it up with coretemp for now, as i said my computer was only available to me friday when i borrowed a videocard from a friend of mine, i have to wait till the 8800gt cards are available here before i can do some tests

and yes i've tried pci cards, i only had one, but it was busted

If the amount on liquid is very small it could evaporate, but there should be a pressure differential (the article made note of these when they occurred).
However, one way to help determine the existence of liquid you could cool the tubing as much as possible before cutting; liquefying the vapor.
This might help decide if they actually had water in them.

Drinkyoghurt said:

can't back it up with coretemp for now, as i said my computer was only available to me friday when i borrowed a videocard from a friend of mine, i have to wait till the 8800gt cards are available here before i can do some tests

and yes i've tried pci cards, i only had one, but it was busted

Click to expand...

Doh That sucks. I feel your pain.

Interesting. If there was water in it, it should fill the heatpipes, and so therefore drip out when cut open, but it didnt.

If there is not an amount of liquid to fill the entire heatpipe, then what is the point of it? Obviously we could be getting better performance if they actually filled these pipes.

Ya know what's next don't you guys?

Better head down to home depot to get a saw, a heatgun and a syringe! Little home modding!

nd4spdbh2 said:

AHAHHHAHAHAHHAHAHHAHAHAH ohh that made my morning.... im sure he will use it in his pethetic arsenal of excuses why his ultra120's are crap...

i think the liquid used in heatpipes is so little that it evaporates upon opening.

Click to expand...

And if you continue to keep insulting him or any other member of these forums, you will incur further infractions which can and will also result in you losing your posting privileges for a period of time and/or permanently. If you find that you don't agree with someone's viewpoint and can't reply rationally without attacking them, you need to set them to Ignore in your User CP so that you aren't tempted to insult or flame them. That way you don't incur further infractions resulting in your loss of posting privileges here.

Imo, that method it self is way too stupid.

They could have tested the suspected HSF at a running CPU or other heat source and measure the CPU temp or a probe at the HSF's base, and then drill a hole or even cut the pipes like those and see if they are any changes in the cooling performances between a sealed pipes and leaked/cutted pipes.

Too bad, many heat pipes HSFs are wasted just like that, again just plain stupid.

NotLag brings up a good point about what kind of fluid is being used in these heatpipe and why the linked article didn't find any liquid in the heatpipes when they opened them up. They could very well be using a liquid with a much lower vaporization pressure than water and with the minuscule amounts being used in the heatpipes, you don't even hear the little bit of pressure relief when the heatpipe is first penetrated even though the fluid contained within is in vapor form.

A client of mine was complaining his Dell XPS system was very loud, calls to Dell told him that it was "normal". I checked it out and the base of the HS was burning hot while the fins on the heatpipes were cold. After I searching around on the net I found other stories and from the sounds of it Dell sold approximately 250k systems with faulty heatpipe coolers. One a positive note, a call to Dell by me had a new cooler at the door in 2 days and now the system actually sounds like a Dell. I'm really surprised that 3.4 Pressy survived nearly 2 years like that!

If someone wanted to, maybe they could use a small torch or soldering iron to heat up one end of the heatpipe while they touched P - or used a thermal probe on) the other end to see if it heated up quickly. In theory, if the heatpipe works, the heat should transfer to that end VERY quickly (and then likely overheat as CPU heatsink heatpipes aren't meant for that temperature range). There's too many variables for it to be accurate though (how much heat ends up in the fins before it gets to the end of the heatpipe, how long it takes to overheat, how fast it really should transfer, how fast it would transfer on a bare copper tube, how much heat ends up transferring through the base into other pipes, etc...). But might be a quick thing to check.

The bigger issue to me though isn't whether they fill the heatpipes with anything. It's whether or not the heatsink performs adequately. Does your heatsink perform as it should?