Aluminium - how long to leave clamped after welding?

[FASTdan]

As I have posted before I manufacture bespoke inlets for bike carburettor installations.

I use 10mm MS plates to bolt the head flange down to during welding. The head plates themselves are 10mm thick and the manifold runners are 3mm wall thickness (45mm and 50mm dia's) and all applications are 4 cylinders.

How long do I need to leave these bolted down after doing the fillet to the plate to minimise distortion? Ie, when does the significant shrinkage stop? Totally cold? Or am I ok to release once the manifold is warm but touchable?

Just trying to speed up my process.....

 

[langy]

I unclamp when cold to the touch, ally takes a long time to cool

 

[Engineer1964]

I unclamp when cold to the touch, ally takes a long time to cool

Yeps thats Right, let it cool down completely, 10mm Ally will cool in about 3 or 4 hours at least, Clamp it down to a large steel beam or Box section that will act like a heat sink and spread the heat out thus cooling quicker

 

[TechnicAl]

I think once the weld has solidified thats it. Try one and see.

 

[Engineer1964]

I think once the weld has solidified thats it. Try one and see.

It will still bend if its not flat and fixed down, I had this problem with Ally ramps i used to make

 

[tigdean]

you can make a thick copper base jig to act as a heat sink. might speed things up a bit in a production line process..

 

[hotrodder]

I'm gonna take a middle road. Agree with Al that the bulk of the shrinkage occurs as the weld freezes, beyond that it's a bit of the old 'how long is a piece of string... the shrinkage can result in stress and strength goes down as temp goes up i.e. it's more likely to move/will move more when hot but shape/size/material/weld geometry all have an effect

I know we're talking ally but take a stainless manifold... poor conductor of heat so the temp gradient across the part is bigger, unclamp it while hot and it'll move. With a one piece flange (especially if it wasn't preset) cutting through the flange can result in the thing 'springing' even when it's cold due to the locked in stresses. If the flange was preset/it was welded from the port side (closer to the neutral axis) things will move a lot less. As an aside it's also why (decent) straight 6 flanges will always be split into at least a pair of 3s rather than one piece- movement in use

With ally the temp gradient will be less and something small like a simple 4 pot inlet manifold for throttle bodies/bike carbs welded with TIG... entire thing is gonna be 'kin hot by the time welding is done. Once it's cooled down enough to be held with thin TIG gloves/touched with bare hands the temp is gonna be pretty even and it shouldn't* move. Having said that in a production environment i'd probably increase the number of fixtures as we're only talking about a chunk of something heavy with some bolt holes in it. That'd be more efficient than speeding up the cooling with heat sinks/cool water/air/giving it half an hour and checking temp etc as you just carry on working...

* as said only way to know is to try. If it does move then presetting the flange would fix that i.e. if something bows ~1mm when just clamped down then in the future you'd shim the centre by 1mm or a nats over so it pulls straight

 

[FASTdan]

I'm gonna take a middle road. Agree with Al that the bulk of the shrinkage occurs as the weld freezes, beyond that it's a bit of the old 'how long is a piece of string... the shrinkage can result in stress and strength goes down as temp goes up i.e. it's more likely to move/will move more when hot but shape/size/material/weld geometry all have an effect

I know we're talking ally but take a stainless manifold... poor conductor of heat so the temp gradient across the part is bigger, unclamp it while hot and it'll move. With a one piece flange (especially if it wasn't preset) cutting through the flange can result in the thing 'springing' even when it's cold due to the locked in stresses. If the flange was preset/it was welded from the port side (closer to the neutral axis) things will move a lot less. As an aside it's also why (decent) straight 6 flanges will always be split into at least a pair of 3s rather than one piece- movement in use

With ally the temp gradient will be less and something small like a simple 4 pot inlet manifold for throttle bodies/bike carbs welded with TIG... entire thing is gonna be 'kin hot by the time welding is done. Once it's cooled down enough to be held with thin TIG gloves/touched with bare hands the temp is gonna be pretty even and it shouldn't* move. Having said that in a production environment i'd probably increase the number of fixtures as we're only talking about a chunk of something heavy with some bolt holes in it. That'd be more efficient than speeding up the cooling with heat sinks/cool water/air/giving it half an hour and checking temp etc as you just carry on working...

* as said only way to know is to try. If it does move then presetting the flange would fix that i.e. if something bows ~1mm when just clamped down then in the future you'd shim the centre by 1mm or a nats over so it pulls straight

Thanks for the concise reply.

Try one and see - well I have previously released them at the point where you can handle bare hands but still pretty warm and not really noticed any further distortion than fully cold. I just wondered if there was a 'theory' really.

I've not explored shimming as yet, I worry that it might introduce as many problems as it solves. A consistently bowed alu plate is happy to pull in flat if tightened from the middle outwards - I'm concerned that shimming might induce multiple bends causing local high points etc.

Not really a problem to solve, just always looking for ways to improve. As has been said, ideally I should increase my number of fixtures, I'm just being lazy as I hate drilling all those holes - although to be honest it'd be more econmical to laser cut them.

 

[Shox Dr]

I'm just being lazy as I hate drilling all those holes - although to be honest it'd be more econmical to laser cut them.

Much like a lot of things: Preparation is the key, spending a bit of time before you weld saves a lot of hassle afterwards. And if these are for money then anywhere you can save time should always be a factor