Galvanic Corrsion test

[Lox]

In the name of science I figured I would see how much voltage is created when using a aluminum rad. and copper water blocks.

Test Method:
Unplug computer power from source (wall socket).
Wait 15 minutes for energy to dissipate/equalize (yes I know there are capacitors).
Using a multi meter put positive probe on unpainted surface or a spot on rad where there is continuity. Eg screw going through bracket into rad. bracket.
Put the negative probe onto a bare spot on the copper water block (may need to scratch surface if oxidized)
Measure DC voltage.

My System, Thermaltake Armor LCS case with integrated water cooling.
Aluminum Rad
Copper Water blocks

Results:

My CPU water block creates 51 mv, VGA water block 2 mv.
This isn’t a whole lot of voltage but it seems to prove that aluminum and copper don’t mix and was just interesting

I’m not sure why the VGA block voltage is so much lower than the CPU’s. However it’s interesting to note that I can see more products of corrosion (white salts) through the clear top on the CPU water block as opposed to the VGA water block.

I'd be interested to hear if those of you using copper/copper get any voltage readings.

 

[Arcygenical]

Well, in my corrosion worklog (which i'm reviving later this summer) my aluminum stock bar and copper penny were pushing almost 300mv! My DMM is a little bit off recently (too lazy to calibrate it, as it takes a full day) so that number should be taken with a grain of salt... buut.

Now, I'm not too sure about this, but since the rad and blocks are connected in a ground-sequence, AND you were measuring voltage with the DMM, I'm sure the actual voltage would be much, much different. Just as running 2 resistors in parallel pretty much halves the resistance, having two electrical connections could quite possiby change the voltages.

I'll go measure my copper, copper loop right now

**EDIT** I get about 6mv, which is in the error of margin for my DMM... Considering it measures open air at around 10mv...

and FYI, copper is the cathode (+) and aluminum is the anode (-) so you should hook your DMM up accordingly.

Now that you can see voltaic action, are you going to try and take steps to prevent it?

 

[Lox]

Interesting that you got 300mv as that is what I read on the CPU block if I hook the probe up to the liquid that is leaking from the water block (The Acrylic top is leaking).
I'm a little rusty on my electrical theory so on Monday I'm going to talk to my electrician buddies at work to see if my method is proper. However I believe the lower voltage reading on the VGA block is due to more resistance to ground as the path is through the PCI card.

Bottom line for me is that in less than six months the cooling system is corroding BADLY due to Galvanic Corrosion! Thermaltake won't even respond to the issue of corrosion, when I ask how the system is supposed to last for the three year warranty period when I already have lost a lot of metal ( as evidenced by the salts of corrosion in the system).
When I do change out the system I'm going to strain out the corrosion products and weight them on a lab scale.
I live in Canada and found an online store called NCIX.com which has an awesome selection of water cooling products.
I would like to have changed the system out yesterday as $$ is not an issue, getting another copper rad to fit where the current one, is.
Arcygenical, you've been around for a while so can you tell me if the Black Ice GT Extreme is much better then the GT Gen2 Stealth rad (240 series)? Thanks

 

[ghost6303]

My CPU water block creates 51 mv, VGA water block 2 mv.
This isn’t a whole lot of voltage but it seems to prove that aluminum and copper don’t mix and was just interesting

could it be that the heat spreader of the CPU is grounded? if your using a core2 they use low temp solder between the core and the IHS, making a possible path to ground. the VGA block usualy only touches the plastic RAM chips and the silicon top of the GPU, so there is no path to ground preventing the reaction.

 

[Arcygenical]

could it be that the heat spreader of the CPU is grounded? if your using a core2 they use low temp solder between the core and the IHS, making a possible path to ground. the VGA block usualy only touches the plastic RAM chips and the silicon top of the GPU, so there is no path to ground preventing the reaction.

Yeah, I really want to check to see if the top of the IHS is grounded. I'll be doing this when I get my board back

 

[mwarps]

Should you not remove the loop entirely to test for voltages? The CPU heatspreader conducts, as does the thermal compound, and the heatspreader is most likely grounded. (makes sense for it to be grounded, anyway.)

 

[Lox]

Should you not remove the loop entirely to test for voltages? The CPU heatspreader conducts, as does the thermal compound, and the heatspreader is most likely grounded. (makes sense for it to be grounded, anyway.)

If the loop is removed there would be a electrical path, so I wouldn't expect to see any voltage.

 

[Arcygenical]

If the loop is removed there would be a electrical path, so I wouldn't expect to see any voltage.

The Voltmeter would serve as an electrical load connection... At least, it does for my isolated tests (IE, 2 pieces of metal in the solution).

Arcygenical, you've been around for a while so can you tell me if the Black Ice GT Extreme is much better then the GT Gen2 Stealth rad (240 series)? Thanks

The GTX radiators are much better than the GTS radiators with a 1200rpm fan... But I still wouldn't buy it. Just get a swiftech MCR220 QP series which performs better than the GTS with a low-noise fan setup.

As for NCIX... if you go to their sister site (directcanada.com) and shop there, you can save a bunch of cash. NCIX will also price-match directcanada's prices if you have to do it that way.