shop-talk
[Top] [All Lists]

Re: welders MIG vs. TIG

To: <shop-talk@Autox.Team.Net>
Subject: Re: welders MIG vs. TIG
From: "Karl Vacek" <KVacek@Ameritech.net>
Date: Fri, 31 Dec 2004 14:03:07 -0600
Loooong boring reply follows.....

> > I have
> > heard OA--under some circumstances, at least--provides a stronger weld
due
> > to more even heating/fusing of the surrounding metal.  Obviously,
> > TIG/MIG is plenty
> > strong since it's used in critical apps (airplanes, bridges, etc.)
> > but is there any
> > inherent differences in the "quality" or strength of welds produced by
the
> > different methods?
>
> I don't believe there is any inherent difference ... you can make strong
welds
> with any method.  However, I've heard it said that it's easy to make
pretty
> looking, but poor penetration (weak) welds with MIG; while it's less apt
with
> other methods.  You can make bad welds with any method, but with TIG & OA
it's
> more obvious (once you know what to look for).



A properly penetrated weld is pretty similar whatever process is used.
However, one very important consideration is normalizing the weld and
surrounding area after welding.  The intense heat of welding causes lots of
differential expansion - swelling metal surrounding the welded area, and
compressing the weld itself.  As the structure cools, shrinkage then pulls
at the weld and surrounding metal.  OA tends to do this as part of the
process, partly because there's less heat to start with and you can only
heat the area fairly slowly, while the heat has time to spread ouyt through
the surrounding parts of the structure.  TIG and MIG tend to apply lots of
heat in a narrow band and also remember that MIG goes so much faster, so the
rest of the structure stays stone cold.  If you look at a recently-welded
piece of steel (it helps if it was really clean first) you'll see narrow
bands of distinct color variations parallel to the weld.  In each area the
condition of the steel is different, and the properties of the steel are
different.  The unheated steel has the typical mill properties, while the
heated areas, depending upon the extent to which they were heated and how
rapidly they cooled, are likely either harder of softer than the unheated
steel.  Stresses, then, will deform the softer areas and not move the harder
areas.  Vibration, however, may well crack the harder areas.   The stresses
locked into the structure by the differential expansion while welding help
this along.

With torch welding the weldor has an ideal opportunity to normalize the
welded area by playing the torch over the weld and surrounding area as it
cools, the goal being to heat the entire area to a uniform dull red (with
chrome moly, similar alloys, and low carbon steels at least, other materials
may require different temperatures or techniques).  Then slowly decrease the
heat (i.e. pull the torch away very slowly, playing it over the entire area
as you move it away) and avoid rapid cooling, breezes, etc. just let the
welded area cool slowly.

Alternatively, after doing all the welding, the welded areas may be
re-heated as above, or perhaps the entire structure may be heated in an
oven, and then slowly cooled as above.  Either way, the object is to let the
crystalline structure of the steel align itself properly and avoid a sharp
zone of transition from heated to non-heated areas.

Many people MIG or TIG and leave out this important step.  For aircraft use,
at least, torch welding is still the first choice.  It's cheap, flexible,
and safe, and bad welds USUALLY look bad, so inspection and quality control
is simpler.  You naturally tend to normalize the welded area by just being
careful, even if you don't do anything else afterward.  However, lots of
manufacturers are using MIG and TIG on plain old chrome moly, for higher
production speeds.  For that matter, even in the 1930's Piper used stick
welding on some of the Cub structure - and I really don't know whether they
normalized the structure, but I bet they didn't.  My Cub is still OK, many
decades after it left Ponca City.

Happy New Year

Karl






<Prev in Thread] Current Thread [Next in Thread>