Leak detection circuits
 

I had an idea earlyer for a circuit thats easy to make, which could detect just about every possible leak in a w/c system. I dont know of any product that already exists for this purpose.
Every system I have built has had very minor leaks usually from barbs unscrewing slightly when tubing is put on etc... easy to fix but always makes me continually check the connections untill its been up for a week or two.

It could be done several ways.

At every hose connection (or other place that could leak/high chance of leak) you would have two copper traces. When they get wet, the resistance is low enough to complete a circuit and activate an alarm.

So assuming any leak forma hose joint would run down the tubing, with the right positioning any leak could be detectable.

The circuit would be cheap to make, and could do other functions if a microcontroler is used (like RPM detection from pump or flowmeter).


Im going to think about the design a bit and make a few for my 4 w/c'ed pc's.
Ill probably use a microcontroler, so not something simple for anyone to make for themelves.... butim sure theres a simplecheapway to make it with LEDs or an alarm of some kind.

if anyone in the UK wants one I could make a few extra, Id like input for extra features it could do (I will include RPM monitoring for DDC pumpfor example).

...or simply run a pressure test: "inflate" your loop to 15 psi.

Make a tee adapter, like this:

sounds like overkill for bad system design to me.

use a decent thread sealant product (i use loctite thread sealer, I cant recall the number now) and hose clamps - never a single leak...

that's a clever rig you've got there ben.

You like my high tech solution? :D Goop works wonders!

BTW 15 psi would be a max; heatercores are only tested to 20 psi. (typical pressure loop rarely exceeds 5 psi). Watch out for pressure surges when you pump it.

I made one of those about a month ago using a bike inner tube valve.
I have no presssure guage to know the presure but it has still been very uefull.

Thats not the reson I thought of it. All the leaks Ive had have been very minor that were fixed by quater turn of the thread - where it had came lose when fitting hose clamps. I use PTFE tape and worm drive hoseclamps so I would not call it bad system design. Is the loctite thread sealer similar to sillicone sealant? I dont see how that would be better than PTFE..

I try to use parerlell barbs now if I can as oring seals seem to be more reliable than the thread seals from taper barbs as they dont move as easly and can be over- tightened more. The only problem then is if the oring gets squashed out which happens if the oring is not perfectly tight against the end of the thread.


Having thought more about it, I would rather not have extra wires going round my case and would prefer just to manually check for dampness each time I fill a new loop, but I expect I will think of another use for it...

No, its not similar to silicon sealant, its far superior for this application.
Its an anaerobic (active without oxygen) sealant that is rated at 1050kpa (around 150psi) and hardens fairly nicely, especially over a bit of time. Ive not had a leak yet - teflon (PTFE) tape isnt reliable in my experience, hence the change to the loctite product.

Id imagine if you're silly enough to twist your tubing to get it off, it might leak after that - but it would take a fair amount of pressure to destroy the joint.

PTFE tape is pretty damn average, I dont trust it personally after a few slow leaks like you're speaking of.

Heres a link to the product in question -
http://www.loctite.com.co/english/products/567.htm

Be warned, its not particularly cheap (but id imagine your electronic circuit wouldnt be either...)

http://www.smallparts.com/images/LOCTITE/AL-567_CL.jpg

Thats only for metal though..?

Most joints are brass-to-polycarb or acylic.

Ill be using BSPP parrelell barbs in future as they are more reliable than taper threads when used on plastics. They dont leak when tightend properly and can be overtightend a bit without risk of cracking the plastic.

All metal-metal joints in my system are soldered, becuase its a lot easyer, stronger, and reliable than any sealant could be.


The circuit would cost almost nothing, which is why Im tempted to make it although I agree it is not really needed, and having the extra wires round my case would probably annoy me more than having lots of slow leaks...

Theres a similar sealant that works fine with plastics, i believe the storm/cascade blocks were sealed with it. I can't profess to having used it, however.

I only have metal parts in my loop (barr the pump, but it obviously has plastic moulded barbs) so I use this sealant.

edit :: two options -

1) loctite 5331

2) loctite 55 cord sealer
Like thread tape, only easier to use + more reliable.

(and yes, i know someone who works for a loctite distributor, hence the loctite recommendations ;))