View Full Version : Ultra compact water cooler for Tig Welding?

26-12-2008, 08:54 PM
Hello all,
My first post. I hope I've finally arrived at the proper forum for this stuff. I've been everywhere and been met with mostly silence. This winter a basement workshop project has turned into a bit of an obsession.

I have recently purchased an inverter tig machine that is very compact. I learned tig on bigger machines with water cooled torches, so although this new machine is only a 200 amp ac/dc unit, I ordered it with a water cooled torch and built a water cooler for it shortly after it arrived at my door. I strongly prefer the lighter weight torch and the ability to weld with thinner leather gloves. I'd water cool even if it was only an 80 amp machine.

The parameters for this range of welding cooling is 1-2 l/min at 20 to 50 psi. 20 psi is the working pressure with 50 psi design pressure for overhead. There is no way for me to determine what the range of real heat produced/removed at the torch head during normal use.

Commercial welding coolers range from 6000 to 15000 BTU heat removal. Even that lower range is probably well beyond the requirements of a single medium sized tig torch at 200 amps. My homebrew cooler I've estimated at about 6500BTU capacity. It would be VERY useful if any of the torch manufacturers could tell me what this range is but they claim propriety!

I built my present cooler around a "dual 120 mm" CPU "BlackIce Extreme" water cooling radiator from an overclocking store online along with a pair of 120mm 100cfm muffin fans. I got a MicroPump micro annular gear pump at a lab surplus house for 30USD. I used a 1/2 gallon poly reservoir. I managed to pack everything into an 9x10x20 inch box. It works very very well for my application.

But I've challenged myself to see if I could build an even more compact version for my hobby. The first catch was locating a compact positive-displacement pump of the MicroPump type. I soon discovered that my 30USD MicroPump was purchased for 600USD new and that there is literally NOTHING available to fit the size/performance category for less than 100USD. I've looked everywhere!

It must be a gear/vane (strictly volumetric) since it must supply that 1.5 l/min reliably across a range of pressures induced by long runs of very small diameter hose (approx 1/8) and through the tiny passages in the brass head of the torch.

I could go on for many pages about this, so basically I'm wondering... with what is available today, how would any of you approach the challenge of cooling a water cooled tig torch seeing no more than 300 amps arc in the smallest package?

Microcompressors using R134a... microchannel radiators... etc? Basic, though updated, tube and fin radiators such as are used in CPU cooling? It needs to be "off-shelf" componentry that is already in use fairly broadly in some other industry or else I'd never be able to afford it.

Why is it that (consumer class) CPU cooling has not graduated to the use of miniature high pressure (positive displacement vs centrifugal) pumps and micro channel radiators? I have not found such a pump in that industry yet. Volume would have brought the cost down to what I can afford.

Thanks for any help here.

26-12-2008, 11:10 PM
Adding a refrigerating compressor to your system greatly increases the weight and size of your system. If you find it acetable with a 30C something water temp, then I would go for a tube and fin water cooler coil. Use a coil with a very small fin spacing and small tubes.

I recently saw a 1 cirquit condenser for a smal domestic cooler for a "corner fridge". Tube size was maybe 7mm. 30-35mm thick and 18x30cm long/wide. Fin spacing very small, maybe 1.5mm. Performance could probably easily be in the 1kW range with a reasonable temp difference between cooling air and water.

I dont remember the make of the system, but I believe it was either Italian or German.

If you could get hold of such a heat excanger you could probably build a more compact system

27-12-2008, 12:57 AM
The compressor idea was actually born of a link from this forum that showed a compressor the size of your fist. ...link is in this forum here .. showthread.php?p=102821
I know nothing about this sort of phase change cooling so something like this might be ridiculously inadequate.

Yes, I do believe the key to this is going to be some form of 'microchannel' exchanger in a standard water loop. I realize that 'microchannel' is a loosely applied term. True microchannel is probably micrometer level, but in between that and commonly available air/water exchangers is considerable room. I get the impression that most commercial welding coolers simply use exchangers "off prescription"; refrigerator condensers or automotive heater core "radiators" that were either intended for far higher pressures or far higher flow rates. None of them appear optimized for 1 l/min at 20 psi of water/glycol. I'd really appreciated more links from members here to compact exchangers that would be better suited.

What I'd really love would be some link to an affordable compact volumetric pump! The closest I've been able to find is a Procon CMP-7000. ..which can be had for a little over 100usd pump and motor. That's a really good price for such a unit. But at 4.5 x 9 inches, and with a pretty bulky capacitor hanging off the side in addition, it's not nearly as compact as what I'd expect to be available to deliver 2 l/min at 50 psi.

Please... can anyone point elsewhere?

27-12-2008, 01:37 AM
There are, as well, compact water chillers used for reducing the temp of large aquariums. These utilize very compact versions of compressors and heat exchangers in what I imagine to be pretty efficient units (ridding 100 gallons of water of excess heat being no small feat). Adaptable to cooling a welding torch? Any opinion?

Peltier, or thermoelectric, exchangers are, from what I understand, not yet ready for prime time in comparison to air/liquid exchange in terms of capacity.

Just what is the "optimal" method of shedding heat from water at 1.5 l/min at 20psi?

27-12-2008, 09:00 AM
Could you first connect domestic cold water to your welding equipment, regulate the flow to required (1l/min) and record in and out water temperature during your welding session to establish what is generated amount of heat for your project, as a first step.
Then you will have starting point for further dimensioning and part selection.

27-12-2008, 03:30 PM
That's actually what induced me to wonder about the actual needs of such an intermediate TIG setup in the first place. I went all out on my current cooler with F temp coolant output, and F temp coolant return, as well as system PSI. I had originally set everything up for bench testing and then when the project was half completed decided, what the hell, and put the led readouts on the front panel.

Accuracy is very good on the pressure, but only within about 5 percent on the temp. But I can run 185 amps arc on heavy aluminum plate at the edge of the duty cycle for this welder (30 %... or three minutes of every ten) for an hour and never reach 105 degrees returning to the cooler after cooling the torchhead and cables. (secondary to cooling the head itself, the coolant cools the conductor thus enabling the use of a very light gauge wire to carry enormous current). Coolant output never rises above room temp.

It might be possible that with 100percent duty cycle machines used in heavy industry, that the same arc would shed heat to the coolant by some logarithmic increase. Otherwise, I just cannot understand the overkill designed into these coolers. The average dissipative capacity, from a cursory scan of the web, is about 12,000 BTU. That's more than my woodstove is cranking out! This capacity for a torch head that has a small (albeit extremely hot) arc about half an inch from a tiny brass barrel with coolant channels in it.

Like I said, my understanding of thermodynamics is pretty seat of the pants, but there's just an intuitive disconnect somewhere here. Seems like you could cool a 0-300amp 50% duty cycle tig welding torch (about the max in non industrial settings) with much, much, less capacity than is built into even the smallest of these coolers.

Are there any members of the forum that are engaged in this industry?

28-12-2008, 08:52 AM
Why is it that (consumer class) CPU cooling has not graduated to the use of miniature high pressure (positive displacement vs centrifugal) pumps and micro channel radiators? I have not found such a pump in that industry yet. Volume would have brought the cost down to what I can afford.
This is simply because their cooling systems costs more than the whole computer.

You have to see all this more as a technical challenge between enthusiastic techs to achieve the lowest possible temperature and then direct related to it the fastest processor speed. These CPU coolers like they made these aren't used in real life, they just put some together or buy a more powerful one right of the shelf.

I prepared exactly the same first advice as Nike 123 described, determine first your heat load. I bet your supplier doesn't know it exactly because they figured it out the same way as Nike 123 described.

I don't think you will find anything better than a plate heat exchanger and you don't need high flows nor high pressures nor microchannels to achieve good results.

We used in the past jabsco pumps. Look once here http://www.jabsco.com/products/marine/index.htm

If you want to make it small for a high load, then it will not be that small at all. If you can have a small ice bank, then you can use smaller components but you will loose space and gain weight due to the ice bank.
Don't forget these CPU coolers delivers only 40 or 50 Watt at very low temperatures which is extremely low for your application.
L'unithé Hermétique has a flat small horizontal scroll, Emerson Copeland also but these are the bigger ones.

28-12-2008, 03:04 PM
I did look into the jabsco pumps earlier. All of their volumetric pumps are diaphragm. I have never seen a diaphragm pump utilized in a weld cooling loop. No doubt they would work. There must be another reason why they are not considered by the designers of these coolers. Perhaps excessive noise/vibration. Perhaps longevity at sustained pressure.

The gear pump commonly used on welding coolers is made by Oberdorfer. The vane pump often utilized is made by Procon or Fluidotech. In every of those instances the manufacturer utilizes a narrow range of pump bodies and simply swaps internal parts to achieve flow rate per given model. Unfortunately anyone who needs a pump at the lowest flowrate is forced to sacrifice bulk and weight in their application. I have not found a single example of a gear or vane pump dedicated to similar purposes and very compact. Vane style fuel pumps come closest, but they are not designed for water based fluids.

I don't think you will find anything better than a plate heat exchanger...

Would a plate heat exchanger be appropriate for this project? You aren't using the term to generically include "tube and fin"? Is there a plate heat exchanger that you could link me to that is superior to the common tube/fin for this physical space, water glycol, at 1.5 l/min?

...and you don't need high flows nor high pressures nor microchannels to achieve good results.

In this application, the flow rate and pressure are fixed at very near 1.5 l/min and 25 psi. Some torch heads are more or less restrictive, and cable runs vary from 10 to 30 feet. That might require 50psi in some instances. In any case, the cooler could be designed using atmospheric pressure at the exchanger on the pump inlet, or make use of system pressure (if it was beneficial) on the outlet.