Crack on the weld due to Heat stress

[ngt]

OK I found an exhaust manifold that was tig welded a long time ago. And it cracked as many times as it was repaired (a long time ago too)
So i was observing this mani today, and i was wondering, what made some welds crack right next to the welds, while the next weld had cracked right in the middle. And why overall it cracked ? I only know that this manifold is 304 stainless.
IS this all because of :....
1) the famous carbide precipitation, (sensitization), chrome depletion, etc ?
2) Exceeding the heat stress capability of the parent metal ? (ok why it didnt crack somewhere along the pipe ?)
3) Poor root of the weld ?
4) Poor filler deposition (undersized welds) ?
5) Hardening of the weld ? (why?)

Here are some pics.

 

[shenion]

That last bit with the reinforcing bars probably made it worse. If that is mild steel, it will have a different expansion rateh than the iron.

my vote is carbon dilution. Shoul dhave been done with NiFe rods. Clean up the welds an i bet they look crystallized.

 

[ngt]

Believe it or not everything on the manifold is stainless.... Except the flanges that are mild. Yup thats a pretty cooked stainlesz!
And at the top picture you can see some rust ? I dunno where that came from. But it had some turf of heat wrap that had turn into glass (woops :P)
Can we use Nife rods on Stainless ?

 

[shenion]

Probably thermal expansion. Stainless is about 2x that of mild steel.

 

[ngt]

Propably. But At the last photo where you see the reinforcing bars. You cant say expansion did it because we have no expansion forces at the neck. The bars were welded after the first crack happened there.
Plus the manifolds' flange had the mounting holes enlarged to handle the expansion because some bolts were sheared before.
So what would you suggest for flange material and pipe material considering expansion and heat tolerance ?

 

[shenion]

What filler wire was used?

 

[ngt]

I wish i knew... ! I suspected too that a 304 may have been used but... Who knows ! The guy is a welder for years into welding various structures (carts containers) for the food industry.We bought and cut the 304 pipes, he only welded them.

 

[ngt]

Ok i searched some, and, In this thread
http://www.mig-welding.co.uk/forum/showthread.php?t=479
The first most obvious answer was lack of penetration. (maby they meant grinding the weld shows up the gap?)
Its a good answer but.... there are some cracks at my manifold that are not ON the weld.
Maby my problem is a combination of expansion for the side-weld cracks. Lack of penetration - wrong filler rod - lack of back purging for the welds cracked in half.

 

[ngt]

Ok So you guys to get an idea how this manifold is, and where the crack points are, here it is.

 

[3Rs]

Talk to TechnicAl he would know whats going on there.

 

[Autolupus]

It looks like poor quality welds and would also suspect incorrect filler rod.

 

[hotrodder]

I reckon that fatigue (thermal cycling and vibration) are responsible for most, if not all of those cracks. A manifold for a turbocharged engine lives in a pretty harsh environment. Cracks are always going to follow the path of least resistance- chain is only as strong as it's weakest link and all that. Those welds are pretty ropy, plenty of surface flaws for cracks to propagate from and probably lots more on the ID too

I've seen quite a few manifolds with similar damage and of those that i've repaired the centre line cracks pretty much always corresponded with lack of fusion/incomplete pentration at the root (the beginnings of crack right from the start!). Cracks following the toes of the weld... usually could be traced back to a crater or similar defect that was likely the initiation point

My limited understanding of carbide precipitation is that both a specific temperature range (420ish to 850ish C ?) and sufficient carbon is required. Temps mean that the effected area would be in the HAZ rather than the weld bead itself but AFAIK the 'L' grades are pretty much immune (not enough C) and the most common i.e. what you get unless something else is specifically asked for. The operating temps of turbo manifolds probably muddies the picture though as the entire thing is probably in the right range so if the filler wasn't a low carbon flavour...

Another dead turbo manifold here

 

[TechnicAl]

To be honest without chapter and verse on whats been done its very difficult to state categoriacally what has happened.
However, based on the photos and what we know I agree with HotRodder. Its probably caused by some form of defect or even grain growth in the HAZ at the toes of the weld.
These are inconsistencies in the structure of the metal and they provide hard / soft spots where stresses concentrate.
If it is 304L carbide precipitation wont be a problem, in fact, carbide formation is mostly detrimental to corrosion resistance. What happens is that when heated the carbon and the chromium react and form chromium carbide. This leaves areas of depleted chromium and its the chromium that provides the corrosion resistance. You need relatively high carbon (which 304L isnt) and time at temp (possible).
HotRodder........the 420 to 850 oC is the temp range for sigma formation which is a similar problem but not quite the same.........sigma is formed from the delta ferrite which is usually quite low in 304L / 316L. The carbide precipitation is remelted parent metal.

 

[hotrodder]

Thought i'd get it at least partly wrong . Metallurgy... fascinating yet mind numbing at the same time.

Don't know if you had a look at the link i posted Al but could this pic be an example of the result of carbide precipitation? No idea of the exact alloy but there were several areas showing similar corrosion, not all had accompanying cracks

 

[Shox Dr]

There was a thread on http://weldingweb.com about this subject last yr. some very good info in it.

(Do you remember seeing it Hotrodder)

 

[TechnicAl]

Thought i'd get it at least partly wrong . Metallurgy... fascinating yet mind numbing at the same time.

Don't know if you had a look at the link i posted Al but could this pic be an example of the result of carbide precipitation? No idea of the exact alloy but there were several areas showing similar corrosion, not all had accompanying cracks

Whoa hold on.....I wasnt correcting you. What you posted was correct I just added a bit more.
Again, its difficult to be exact without all the info but its a possibility (usual non commital Metallurgists answer). I assume it another exhaust, if so and if it is carbide. the inside would be more corroded than the outside. Can you see inside?

 

[ngt]

-I forgot to add that at the neck of the mani where the Egt probe fits.
We saw a 1150C temperature about three times. And regularly 1000C.
-Inconnel probe was melted-broken once.
-The manifold flange got warped from expansion once (sheared a couple of bolts), was fixed. Now again i see the manifold is warped again.
- Doesnt the stainless weakens if it stays too much in a ~900C state ? I see that it has beome black. Maby its only the outside. Maby it became brittle from the high temperatures ?

This is how it used to be. It had been cleaned with a solution. thats why it looks like sandblasted.

 

[Tungsten Sharpener]

The choice of material for the original part could have been anywhere from 304 to 410 and using dissimilar rods would not work either.

Did a job for turbo technics many moons ago and the exhaust temp was in the region of 920c, I had to modify the manifold and could not get a weld that wouldnt crack unless i pre heated to cherry red and post heated with wrapped cooling.

Looking at the part, I would have kept the manifold plates and rebuilt the lot.

TS