Elbow joints for 4mm (OD) copper tubing

Hello Timothy, welcome to the forum.

I think you might struggle to find finished elbows.

What about using 5mm square bar with rows of stepped holes at right-angles? You could use a very close spacing.

Neil

You could possibly try push fit pneumatic fittings. That will depend on whether or not appearance is important, and whether you are piping a liquid or a gas. Also, they will be expensive.

Hello,

Thanks for the replies and suggestions. I don't think I will be able to use square bar with stepped holes as i am not familiar with this nor have the equipment. And all the push fit pneumatic fittings I have seen will be too big for my purposes.

I am planning to pass liquid through a copper tube that is sitting on top of a peltier device (surface area of 40x40 mm). The idea is to pass the liquid across the cold side of the peltier device in order to cool it down. To do this I want to pass the copper tube across the peltier surface as many times as possible, similar to the way a copper pipe passes across the flame inside an combi boiler. For this, I therefore plan to spiral the copper tube inwards with right angle turns (link):

**LINK**

To keep the copper turns tight, and the copper tube close together to allow maximum turns, I need the elbows to be as small as possible. For the corners, I figured the easiest way would be 4 mm right angle copper fittings such as these (link) but smaller:

https://www.google.co.uk/search?q=spiral+right+angles&rlz=1C1VFKB_enIT639IT639&espv=2&biw=1366&bih=662&source=lnms&tbm=isch&sa=X&ved=0ahUKEwj0vcLI9aPRAhWPdVAKHemPCEwQ_AUIBigB#tbm=isch&q=copper+elbow

Would it also be possible to bend the pipe somehow? Not sure if you if I would be able to get good right angle turns with copper tubing that is so small.

Thanks again,

Tim

RS do compression fittings. You could lose the compression nuts and solder the pipe to the fitting. You could even make the endfeed elbows yourself, out of small blocks of brass or copper, very easy to do.

Steve

Would it not be possible to buy a coil of 4mm tube and form that into a a zig zag or squared off spiral, its easy enough to bend in your fingers or over a simple form tool

Timothy, you can buy miniature unmachined elbow castings from PM research in the USA, they are not dear and you can put your own thread/hole in them. they do 1/8" and 3/16" inch versions, your 4mm or 5/32 as near as dammit would certainly be possible in the larger size mentioned.

Chris Gunn

https://www.pmmodelengines.com/

As you can't drill a square bar, i'm assuming you don't have a pillar drill so anything other than an off the shelf solution is no good.

I'd agree with Jason's suggestion, in fact simply winding a spiral of a suitable tube would do the job. It will be less likely to kink if you fill it with fine sand before bending and bend it around a former with a rounded corner.

Neil

If you go the sand filling route, kiln dried sand or the type of sand used for block paving is good.....

With the minimum of moisture it flows like salt....Places like Wickes sell it..

5/32 is very close to 4mm you can get 5/32 elbows here also Macc Model eng supplies and Live Steam Models do them

H

Can't you use copper brake pipe and actually wind it in a spiral?

If I understand your original requirement correctly you are trying to create a square spiral to cover a Peltier module 40mm square?

How will you get the coolant out of the centre of the spiral?

Also how will you thermally couple the discontinuous surface of the tube matrix to the face of the Peltier?

Ian P

If I were doing it, I would simply wind a spiral - no corners - to cover the required area and soft solder the pipe to a copper sheet, for laying adjacent to the peltier surface (with heat transfer compound on the surfaces). No corners, or connectors means less pumping losses, so a smaller inner

diameter tube could as easily be used?

How hot is the hot side going to be? How much electrical energy are you expecting to generate? Apparently one needs a lot of peltier devices on a car exhaust, in order to replace the alternator!

Think I would be soldering two squares of brass or copper sheet together with a strip of metal around the edge to form a flat 40mm x 40mm box and have pipe connections at opposite sides, add an internal baffle or two if needed. Just pass water through the box and you will also have 100% contact with your surface to be cooled

Re: filling the tube with sand, - there's a special plastic for that, it's cheap, infinitely reusable, melts at around 60 degrees C, and is soluble in water (presumably so you can wash any left overs out)

It's from BendAlloy.co.uk and it's called polyBend, it's a cheap alternative to proper bend alloy. You just have to be smart about how you heat and handle it so you don't dissolve it in the water by accident.

The OP stated that he 'does not have the equipment' to make his own elbows it sounds that he has very limited facilities to fabricate this cooling device.

My suggestion would be to investigate the water cooling equipment that is made for PC 'modding'. Whilst much of the stuff is cosmetic (illuminated fans, coloured wiring etc) the watercooling kits for the processors look well engineered and are designed to thermally couple to a flat surface about the size of his Peltier.

Ian P

Tim, re reading you second post properly......

Your going to get poor results with any of your listed methods. The best way to achieve high thermal efficiency is to create or use a standard copper water block from a company like EK.

Inside this block you'll find a solid copper substrate with many parallel <0.4mm slits. Water enters the top (centre) through a changeable laser cut baffle plate. If the top entry is no use you can also get GPU cooling blocks which typically have their ports on the side. There are also cheap Chinese blocks which are only a few quid and are much smaller with copper hosetails which could be cut and soldered to perhaps. This will still perform significantly better than your proposal.

Something like an EK supreme copper waterblock will handle 100-150 watts on a surface area smaller than 40x40 and maintain a temperature differential of sub 10c, whereas the Chinese one would likely be twice as bad with half the heat - which would actually work quite well with a 40mm peltier.

I'm speaking from many years of experience of liquid cooling silly competition computer systems using water, TEC plates, phase change units and even liquid nitrogen. You'll seriously struggle to DIY this but if it's possible, you'll want some square hollow copper from Chronos, filled with a bend alloy/polymer, zig zaged or bent 180 in the centre then bent around its own back (so it forms a bigger and bigger U shape each time. Then you'll need to fill all the gaps and bond it altogether with solder of some sort, then sand and lap the base flat.

Hi everyone,

Thanks again for all the suggestions. Apologies for the delay.

To give a little more detail about my project, I am trying to create a water bath that is at a constant -10 degrees Celsius. To do this, what I panned to do is pump water to and from this water bath via a peltier system. The main tubing will most likely be plastic tubing (about 2 mm ID, and 5 mm OD, and is also insulated) but will then feed into copper tubing (4mm OD) to pass over the peltier device. I am using two peltier devices staked on top of one another and with this I can get a stable -15 degrees on top of the cold side of the peltier. I have not measured the heat on the hot side of the peltier but I am using a system of heat sinks and fans to dissipate the heat.

Also, just another point, I add about 2 molar salt to the water to hopefully stop it freezing when being pumped around the solution. If this does not work, I will look into using some other form of coolant (not looked into this at all yet).

I initially planned to use a water cooling block on top of the peltier to cool the solution, but, I want the solution to be on top of the peltier for as long as possible (to transfer as much coolness as possible) and the water cooling blocks I have seen only have about 3-4 turns, therefore likely to pump the solution across the peltier for only a very short time (the pump rate will be about 25 ml/min). This is why I thought it better to use small copper tube that makes as many passes as possible across the top of the peltier. Also, all the water cooling blocks I found had large diameter openings (at least 8 mm) - I was not sure if this would work with my plastic tubing with an ID of 2-3 mm.

Once the copper spiral had been made, I was planning to weigh it down on top on the peltier. I would put insulation in between the weight and the copper tubing to stop the cold temperature being transferred to the weight.

Further details: heat-transfer paste will be used between the copper tubing and peltier. The pump rate will be about 25 ml/min. The copper tube system will be covered in insulating material as much as possible. If possible, I may even use copper tube with a smaller diameter (to increase the number of turns on the peltier)

Hope this makes the project more clear. Thanks a lot once again for the suggestions.

Tim