[Koot]

You are correct, the contacts in a tiny ice cube relay are just too close to have both low and high potential voltages on the contacts. Contact arcing, bounce and fatigue over-time are definitely a concern with the current scenario. He should install a second relay (with a 120 volt coil) to control the piezo buzzer, thus completely isolating the low [voltage] potential from the high [voltage] potential. It would also be desirable to have the low and high potential wires separated by spaghetti sheaths.

You are also correct that using ground instead of the neutral is a definite no-no. He should be able to tap into the neutral wire in the PC instead of using chassis ground.

[cwwozniak]

Originally Posted by Who's Me

... I stuck my finger in the coolant while touching the case and found out that I was running 120V AC through my cooling system ....

That pump motor or the wiring feeding it has a major safety problem. At no point should any of the AC power path have any connection to any metal parts of the pump or motor frame even if hot and neutral are reversed. A three wire grounding cord should be feeding power to the pump motor. All exposed metal on the pump and motor needs a solid connection to cord's ground wire.

[Who's Me]

paisanol69,

I looked at my last post and realized the way I worded it gives the impression that I may or may not take the advice.

""I will look into another way to design the relay failure circuit...maybe.""

I am NOT going to build the circuit that way, the "maybe" was there to indicate that I may or may not use A coil failure circuit. I think A better way will be to design a coolant flow detection circuit which will tell me if ANYTHING has stopped coolant flow. maybe even just a visual indicator like a little red wheel on the pump return hose that I can look at.

I had considered A relay contact failure but with the low currents I am using I was thinking
a complete contact meltdown would be very unlikely. But considering Murphy's Law it is not worth the chance.

[Noyb]

I think the Smart Strips use a solid state relay.
I have my Smart Strip wired so it controls my 4 led desk lamps.

So, in your application .. My desk lamps would serve as a failure indicator.

I'm thinking the contacts of a relay are less reliable that the coil.

A 2P2T relay could be wired so you wouldn't have to get extra life insurance

[Who's Me]

cwwozniak,

The pump is plastic I was touching my tower when I stuck my finger in the coolant.
I think the impeller Shaft is connected to the neutral wire
which is why when reversed, current passed through coolant.

[cwwozniak]

Originally Posted by Who's Me

I think the impeller Shaft is connected to the neutral wire

That should not be the case for a properly designed motor that is wired correctly and not damaged in any way. The shaft and any metal parts of the motor housing should be electrically isolated from any part of the AC mains wiring and any of the wiring inside the motor.

[Who's Me]

You are saying if the pump is submerged I should not be able to get continuity from the water to either wire?
I am using a cheap Chinese 2-wire fountain pump, maybe it has some internal problem.
I think I may have another two wire pump I will have to test.

[cwwozniak]

Originally Posted by Who's Me

You are saying if the pump is submerged I should not be able to get continuity from the water to either wire?

That is exactly what I am saying.

[Who's Me]

When I made my relay box I cut the pump cord in half and spliced the male end of an extension cord to it and the relay.
I just looked at the plug I was using before and found moist corrosion building up on the plug terminals. It looks like the epoxy in the pump must have failed (or was never sealed right) and coolant has been slowly seeping through both wires to the outlet, which explains continuity between the wire and coolant. The weird thing about that is, I cut and spliced both wires in the middle two days ago and found no corrosion or moisture in there. but the ends of the wire at the plug tell a different story. I am sure it did not travel down the outside of the wire as I have always used a drip loop with it. I guess it's time to get another pump.

And for failure detection I totally forgot about ASUS pc probe. I set it to run at startup and set the alarm for 40C. If the pump is not running for any reason my CPU will get above 40C and start blasting an alarm through my speakers.

[cwwozniak]

What are you using as your cooling liquid? Could something in it not be compatible with the materials used to seal the pump?

[Who's Me]

I have been using this diluted with water,
http://www.redlineoil.com/product.aspx?pid=74&pcid=10
and yes I think it is very likely that it may not be compatible with the pump sealing material. Last month I had my 9600GT fry because my acrylic resivour had cracked across the bottom and one of the corners was seperating. It had been sitting in my tower for a long time and developed the crack and seperation all by itself. When I replaced it I noticed the pump nipple that the hose goes on had also cracked. when I replace the pump I will probably just use some Automotive antifreeze. Maybe BMW it's a pretty blue color.

[cwwozniak]

I am no expert on the subject but you might want to look into the type of anti-freeze that is used in the potable water lines of RVs when the RV is stored for the winter in a cold climate. They are also supposed to reduce corrosion of the plumbing.

[Koot]

You might want to look into mineral oil. You can buy it at any grocery store.

Mineral oil has a low viscosity and does not conduct electricity. Oil-filled transformers of all types (e.g. pole pig transformers on telephone poles and huge substation transformers) use mineral oil to absorb heat and conduct it away the transformer's coil and core. The [electrical] windings are submersed directly into the mineral oil. Just a thought...

[Who's Me]

On the RV antifreeze, they designed it to be non toxic and not freeze. I don't know if they cared about heat transfer. I read that it damages Acetate plastics and I have no idea what I am using. It would probably work fine but I just don't know.

Mineral Oil, I had thought about that initially, and I would probably still have a 9600GT if I had been using it. I just pictured myself trying to get Half a gallon of mineral oil out of my carpet and decided to go with water based coolant.
I have a piece of 4in. aluminum square stock about a foot long that I want to make a proper reservoir out of. (with a screwed down lid and screw in fittings And maybe even heatsink fins) If I get around to building that, and have a proper sealed system I will probably use mineral oil. It's just a lot of material to remove and I have a small mill that can't take much off at once. It could take me days to cut it.

For now I think I am just going to switch to universal clear automotive antifreeze,(Considering the carpet) buy a new pump, and use the relay design in post 1 which has been working great, if I do get around to making a proper reservoir and use mineral oil maybe I'll just mount that whole thing in there to keep the relay cool and stop the contact arc.

[Noyb]

Originally Posted by Who's Me

...keep the relay cool and stop the contact arc.

If it were me ...
I'd get an optically isolated solid state relay ... And not mess with a mechanical relay.



Could even control it from the USB 5V power ...

This one predates computers and is still running strong ... Extra life insurance not required