Tig welding aluminium radiator

[pmb5]

I'm trying to weld up cooling tubes on an aluminium bike radiator i've cut up to fit round the exhaust.

I'm new to welding aluminium so got some 0.9mm and 0.7mm off cuts, got some ok welds and was even able to weld up the edges

I had some practice on an old radiator i'd cut up and it wasnt too bad... even though i didn't bother cleaning it that well, but with the actual one i need to weld up, i'm really struggling.

Anyway my question is really about the balance control. I've got an r-tech tig and the balance goes from 30-70% but i've read conflicting info on what it does. At 70% balance the tungsten balls up and i get lots of penetration, and 30% lots less penetration and the tungsten keeps its point. So is the lower percentage giving me the most cleaning? I read on the miller site that if you use too much cleaning the tungsten balls, but this is the opposite of what my r-tech manual says and at the welder does.

When welding the rad i was using a pointed 1.6mm tungsten, 18amps, about 35% balance and 250hz frequency to try and keep the weld focused.

thanks in advance for any advice/tips/.

 

[hotrodder]

Approx 2/3 of the arcs heat occurs at the +ve electrode, this is why DC TIG is almost always DCEN (torch negative) as having the torch positive results in melted tungsten. DCEN puts more heat into the work, less into the tungsten. The 'cleaning' bit of AC is related to electrode positive which puts more heat into the tungsten, less into the work.

The Miller guide is correct, more 'cleaning' = more tungsten balling and less penetration but... while most machines set AC balance as %EN, some do it backwards (%EP), sounds like yours is one of them which would mean 70% EN would be a setting of 30%.

You've cut the tubes on a radiator and are trying to weld them up. Ambitious. What's happening to the coolant flow? I can't see how you'd have access to do anything other than cap the ends which is effectively reducing cooling capacity

Old, used radiators are often about as hard as it gets. Nightmare to get them clean enough for TIG, tank repairs bad enough but tubes... i'd be brazing with oxy, flux and 4047 personally or if that's out of the question maybe (shudder ) that monkey metal aluweld/solder

 

[pmb5]

Thanks for the help. It had really confused me, I'd have the balance up, the electrode would ball and I get lots more penetration... confused the hell out of me.

Yes, I wondering if this was really good idea now. Welding up this rad was the excused I used to justify selling my DC tig and getting an AC/DC one. Basically this rad is bigger, even with the cores cut, I needed to cut it to get clearance between the front wheel and the pipes.

Originally I was just trying to weld up the cores with fins still attached, this worked on the test rad, but no chance on the actual rad I need to weld. I've cut the cores all the way back to the end caps but still not having much luck, gunk is still coming out as I try to weld and i cant get a puddle to form.

I've got some of those soldering rod things, so i might give them a go next.. then if that doesnt work I might cut the end of the tanks and try to weld that, at least then I wont be trying to fill a huge gap as i weld.

 

[Shox Dr]

Welding old Alloys comes with experience, I found almost impossible at one time, these days I don't seam to have half if any problems. But then with experience comes the understanding that cleanliness IS GOD

 

[chap]

When i've welded heads and other scabby material in the past, i've done a number of runs with no filler rod and no fusion; wiped and wire brushed between each pass to clean the crud that comes out, then done the weld. Just be carefull that the ali doesn't melt and peal back!
I have found that the contamination sits on the top of the weld, but doesn't affect the integrity of it too much, so they haven't leaked.
If you only think the weld might have sealed and it's a bit dodgy, you could plate over the join. Use fewer amps so you don't break through with the penetration on the lap joint, then the crap shouldn't come through and contaminate.
Lastly, remember that you can't weld a sealed vessle.. the hot air inside will just blow out of the last little bit your trying to fill!

Hope this helps
Good luck!

 

[pmb5]

Thanks everyone for the help. I did the passes on low amps then scrubbed the weld after as per Chap's advice. Then once I got my bottle up and set the amps high enough it wasn't too bad. I remember making this mistake when I learnt to tig mild steel, just need to get the amps up and move faster!

The results are very messy, I tried to use that solder stuff before hand, but ended up just melting the aluminium and so had big holes to fill up. Now I just need to knock up some bungs on the lathe and pressure test it.

thanks for all the help, here's a pick of the horrible mess

 

[hotrodder]

Really need to make new side pieces for the rad, just a piece of 1.5 - 2mm ally bent into a shallow channel ] shape. Will protect the fragile core and add some structure tieing the top and bottom tanks together. Any reason why you didn't narrow the entire rad i.e. cut through the tanks too and then make new side pieces and tank ends? For some reason i was picturing a hole cut into the side of the core rather than a narrowed core

Probably know already but don't pressure test with air. Can use 2 or 3 psi for an initial leak test (go over the welds and adjacent tube-tank joints with soapy water) but pressure test proper wants to be done hydrostatically

 

[Shox Dr]

What HR said. It would have been far easier to have cut the end off and plated it up, at least then you could have made sure the inside was clean.

 

[hotrodder]

Thanks for the help. It had really confused me, I'd have the balance up, the electrode would ball and I get lots more penetration... confused the hell out of me.

Not suprised you were confused, just had a read of an R-tech user manual...
Initially says this about AC balance under controls and settings

13. AC squarewave balance (SP%) SP% AC Squarewave balance control 30% – 70% Set at 50% this provides balanced control suitable for most AC Welding. 30% is max cleaning and 70% is max penetration
When welding dirty/contaminated material more cleaning may need to be selected. Tip. If you are welding close to the tungsten limits, I.E 200amps on a 2.4mm tungsten and the end of tungsten is wobbling and falling off, turning down the SP% will give better tungsten stability

The bit i've highlighted in bold suggests conventional AC balance display i.e. like in the Miller guide and the majority of machines displays %EN. The comment about dirty/contaminated material agrees with the bold bit but the sugestion about tungsten stability is completly contradictory

and again under tips for AC welding...

1. The AC square wave balance (SP%) can be set to a higher percentage electrode negative (decrease below 50%) which minimizes tungsten heating and erosion

which is a complete and total contradiction to the bit i highlighted in bold but consistant with your results i.e. displays %EP. Not as comical as a typical chinglish instruction manual but about as clear

 

[Delgado]

Got the 200amp model and after some initial problems (rtech, to their credit sorted it out quickly with no fuss) I've got to say that it works well, although as said, the instuction manual might as well be written in chinese. If you turn the balance dial anticlock it gives more EN and clockwise more EP. Contrary to their advice I've found that it works best with the balance dial at about 10 o'clock (65% EN), giving minimal tungsten erosion with a tapered grind and better arc starting in AC mode.

 

[pmb5]

Ah, that was going to be next question what was the max psi i could test with air

This is the clearance I required:

As you can see it is a curved rad. I didnt want to cut it all the way back to begin with, as you've pointed out i need to re-join the sides otherwise it's only the core tubes holding the two halves together!

I didnt cut the ends off as I figured once that was done, if i was blowing holes in the tank welding it unsuccessfully i'd start running out of metal to weld.... basically i wasn't confident in my ability to weld it back up.

Thank to everyone for looking into and explaining the balance, i too ended with it at about 65% EN

The R-tech welder came with a WP26 torch, it's knackering to use, my old welder had a wp17 that was much easier to use. Could i swap back to a WP17 or could 160amps AC over power it? Or is it just the heat that builds up in the torch? I've never had the skill to weld for that long for it to be a problem.

 

[hotrodder]

WP17 style torch typically rated 115A AC @ 60% duty cycle. It's a heat dissipation thing, usually get too hot to hold before damage is done but more likely to have issues with collets going soft or thread damage to the head if pushing things. CK do torches with higher ratings but will be best part of £100 for a ck17 or just swap back to the wp26 when running 120A + AC

I typically hydro rads at 2bar. If you don't have a suitable pump can brim the rad with water and pressurize that with air. Won't be enough air in the system to do any damage if something lets go. Stay out of the firing line regardless though unless you don't mind getting wet

Bear in mind that coolant will find leaks that water alone won't so worth doing the low pressure air and soap test first if not confidant

 

[Delgado]

Sure, the WP26 torch that comes with it is utter sh%t, I went out and bought a CK17FX and it's a dream to use in comparison. IIRC it's rated at 150a AC 100%, could probably go higher but as said, they're not cheap so normally use the WP26 for anything around or above 150amps.

 

[pmb5]

what are CK torches? google's not giving me much info on them?

thanks, you guys have been awesome

 

[hotrodder]

Rapid welding are the UK distributors, link to 'CK worldwide' on their site