Tiging a nipple to thin wall aluminium pipe

[scottmk1]

Hi folks, I need to weld a nipple for a coolant temp sensor into a piece of thin wall aluminium pipe. The pipe wall thickness is 1.5mm and I dont know what flavour it is, and the nipple about 5mm wall thickness and is 6082 T6 flavour. I have a couple of pics below to illustrate what Im after.

What settings and rods/tungsten would I be looking to use to do this job? I'm not very experienced with alloy welding hence the questions. Also would it be an idea to leave the sensor in while welding or should I remove it?

 

[wheresmejumper]

have you got a foot pedal?you can have more room to manouver with amps setting
its gona be tricky either way,lighting up on the 5mm will take a good few amps,then wash the puddle down to catch the pipe.its easier said than done on aluminium.
2.4mm tungsten

 

[voipio]

The job will need AC and pure argon, and Ali prep is all important. Surfaces need to be as clean and fresh as possible. Electrode material for a job this size shouldn't matter too much, though some folk have a preference for Zirconiated tungstens, which will ball up less than pure tungsten. Example Ali TIG welding rods for 6082 would be Sifalumin No15 or No27. No15 has better flow characteristics whilst No27 rods have better corrosion resistance. Both are spec'd as suitable for 6082, but the tube material may ultimately decide it for you. It would be worth trying a test weld first joining some scrap from both materials to see which rod works best. I would suggest taking the sensor out, as the "nipple" is going to get pretty hot during welding. The sensor looks like copper or brass which is not good to have in contact with Ali when there's water around; likely to result in corrosion. That may not matter too much if this is being used for racing and regularly drained/stripped down, but worth being mindful of.

 

[scottmk1]

Thanks for the replies. I have an Rtech 160amp ac/dc welder and have gas too. The sensor came out of a cast alloy housing as oem, There isnt too much option for me either as I dont want to be making this out of steel.

 

[dan.taylor.1]

Deffo take the sensor out, they're fine at 0-100 ish temps but anything more and the innards cook hence why they often go when an engine overheats.

Welding wise I've no clue as I can't tig!

Dan

 

[scottmk1]

Cheers Dan I will make sure its out when I weld.

 

[hotrodder]

The tube will be 6063 or 6082, pretty much the only flavours commonly used for any extrusions (tube/angle/tee etc)

SIF aluminium 15 & 27 are just trade names for 4043 and 5356 respectively, 4xxx series wires wet out better/more easily due to the silicon which lowers/widens the melting temp range. Corrosion resistance varies with service environment, 5356 is not always more corrosion resistant (elevated temps for example) but with fillet welds it'll always be stronger and less brittle than 4043. All pretty much mute for a sensor bung i.e. as long as it doesn't leak all will be good

The mass of an ally part plays as much, if not more of a role than thickness with regards to the amperage needed. With a small bung like that there's nowhere for the heat to go and the joint is also set up with the edge of the bung exposed which'll help i.e. the edge of the hole in the tube is covered by the bung. Personally i'd have ~100A, floor the pedal for the initial tack, run the weld from the second tack backing off as needed. Maintaining a nice tight arc length while the weld progresses is the bit that takes practice when talking small diameter fillets on round stuff as the angles of attack are constantly changing

 

[scottmk1]

Cheers hotrodder, I have a foot pedal so will use that. What about ac balance and frequency settings?

 

[hotrodder]

Different machines are different, need to determine balance settings yourself really. Run beads on an offcut at different balance settings noting the effect on the weldpool and tungsten- too far one way (too much EP, electrode positive) and the tungsten will melt excessively, too far the other way and the weldpool will get sluggish and eventually dirty/with 'bits' in it as there's not enough 'cleaning'.

Frequency is as much personal preference as anything i.e. can do that weld at <50Hz or >200Hz. Higher freqencies constrict the arc and produce a narrower weldpool which generally helps on small dia/thin stuff and fillet welds but the arc is 'harsher' at higher frequencies too and things tend to happen faster. Again, experiment on some offcuts

 

[scottmk1]

Ok, thanks for the advice. I will report back once I've tackled the job.

 

[zardoz]

if you had a few spares to practice on that wouldnt hurt.?

Good luck.
.

 

[scottmk1]

Hi yes I have been practising on bits of pipe tonight, my tungstens keep going black though. Ive tried red tipped and grey tipped both 2.4mm

 

[chevy2]

Make sure you have enough gas post flow to prevent tungsten contamination,
with that amperage about 10 seconds should be enough,
the grey tungstens should work, sharpened to a point and then slightly blunt the end off.

 

[scottmk1]

Right I have been practising on some spare pipe I have and here is a couple of pics of the results. I was using a grey tipped 2.4mm tungsten sharpened with the point flattened off, 1.6mm filler wire and the amps set from 70 to 100 amps, can anyone suggest where I'm going wrong? I thought the gas maybe ran out, but the bottle is still showing 50 bar and the flow meter still shows 8 lpm when I operate the foot pedal.

 

[chevy2]

Theres virtually no cleaning etch around the edge of the weld,
you do have it switched to AC ???

 

[wheresmejumper]

pic of the front panel settings would help

 

[scottmk1]

Yes it's on AC, I will nip out and get a pic of the front panel

 

[scottmk1]

Here are the settings on the front panel

 

[Paul.]

Nothing much wrong with your settings there, increase the post flow a bit, it needs to be a bit longer than the downslope, I would reduce the upslope too but thats just my preference, I like the current to come in quickly.
Looking at your earlier photos and the description of burned tungstens I would be looking for a gas leak somewhere, check the seal on the backcap and if possible check for proper flow at the torch with a peashooter type flowmeter.

 

[scottmk1]

Thanks for the advice Paul. What is a pea shooter type flow meter?