TIG welds cracking

admin

Hi everyone, this is my first post so be gentle. This might be a little long winded but I want to provide as much detail about the job as I can. I have these air injection adapters that I welded keep cracking all the way around the weld.  Two  4" flanges 1/2" thick 304 SS.  there is a 1.75" section of pipe between the two flanges.  Original pipe are suspected to be 304SS but not positive.  New pipes are 316 SS.  The pipe is .145" wall thickness.  In the side of the pipe, there is a hole for a 316 SS Swagelok 1/4"FNPT x 1/4"FNPT 90* fitting to connect to an air supply to inject air into the exhaust stream.  I have 4 or 5 of these adapters and they are are cracking around the weld between the swagelok fitting and the pipe.  I used 308Lsi filler rod on them, which is what was recommended for 316 to 304.  One last piece of useful info on these parts, they are run on dyno engines that see inlet temps of around 950-1000C for extended periods of time.  I don't know if the welds are cracking due to improper filler rod selection or if it's just the nature of the beast in the test environment.  Thanks for reading. Attached Images
Reply:Start over with some 309 for filler.......zap!I am not completely insane..Some parts are missing Professional Driver on a closed course....Do not attempt.Just because I'm a  dumbass don't mean that you can be too.So DON'T try any of this **** l do at home.
Reply:According to several charts I have found 308L is the proper choice for 304 to 316.  Other than that, I'll let other more experienced weldors chime in.    DOH! beat to the punch! by Zap. 1st on WeldingWeb to have a scrolling sig! HTP Invertig 400HTP Invertig 221HTP ProPulse 300HTP ProPulse 200 x2HTP ProPulse 220MTSHTP Inverarc 200TLP HTP Microcut 875SC
Reply:First question....everything is stainless?  Flanges look torch cut?  Plasma maybe?Not an engineer here but one thing to consider are the stresses that develop during tack-up and welding.    The argument could be made that because of the curvature of the pipe, the pulling from the first pass could produce cracking in the initial tacks.  Possible solution to build up and flush the area on the pipe which will receive the fitting
Reply:I've had moderate luck with using 309L for joining dissimilar steels.  In your case, I am wondering if the the difference in thermal mass between the 2 pieces being welded could be causing you some trouble.  Stainless expands 50% more than typical carbon steel.  When the weld and your FNPT fitting are expanding, then contracting, the thick pipe you are welding the bung into is just sitting there like a vice - it's not expanding and contracting with the rest of the weldment.  As the welded area contracts down, it will develop a lot of stress as it tries to pull away from the surrounding area, and that could be causing your longitudinal crack.  If you could preheat the entirety of the part, you might have better luck.  Just an idea.Miller Dynasty 200 DXMiller CP-300 with 30A feederHypertherm Powermax 900Oxy-fuel w/Harris torchesScotchman Glide-in bandsawMonarch 10EE latheEmi-Mec Autoturn latheDeckel FP2NC milling machinePro-Tools 105 Bender
Reply:It is cracking right through the middle of the weld, which leads me to believe that one of the items the weld is on is not recommended for welding. 303 perhaps?Weld like a "WELDOR", not a wel-"DERR" MillerDynasty700DX,Dynasty350DX4ea,Dynasty200DX,Li  ncolnSW200-2ea.,MillerMatic350P,MillerMatic200w/spoolgun,MKCobraMig260,Lincoln SP-170T,PlasmaCam/Hypertherm1250,HFProTig2ea,MigMax1ea.
Reply:UnkleBuk,Are the welds cracking in use, or are they cracking immediately after or even during the actual welding process?Centerline cracking like that shown in your photo, if it happens during welding, is generally a sign of either contamination(by sulfur or phosphorus, for example) or stress due to excessive joint restraint.If either piece of base material is not what you think it is, then you've got troubles.  For example that Swagelok fitting is machined with threads.  Is it a free machining grade of stainless steel?  If so, these are generally considered un-weldable.  They are made easier to machine through additions of sulfur, lead, or a couple other elements with low melting temps.  These additions make cracking almost certain as the weld cools.If you're sure the base materials are weldable grades, then the other possibility is excessive joint restraint.  Basically, the fitting may be getting much hotter, much faster than the surrounding pipe.  As it cools faster than the surrounding pipe, the fitting shrinks and this applies stress to the weld.  The fix is to preheat the parts prior to the start of welding and then slowly cool them to room temperature.  Don't heat excessively, 300F is plenty.  If you get the stainless too hot, there is increased chance of damaging the corrosion resistance of the weld.If, on the other hand, these parts are cracking in service, then the problem could be the service conditions.  High temperature exhaust can cause stress corrosion cracking from the inside of the fitting out.  It can attack the weld preferentially under the right conditions.  If the parts are failing in service, most likely the only way to figure out the root cause for certain will be to send failed parts to a metallurgist for examination under a microscope.Benson's Mobile Welding - Dayton, OH metro area - AWS Certified Welding Inspector
Reply:one last idea.  Is the Swagelok bottomed out against a shoulder in the bung?  Almost like a blind hole?  If so, just as when welding socket fittings, you have to have a gap at the bottom between the fitting and the threaded insert.  If not, the pressure created as the parts cool will tend to want to push the Swagelok out of the hole.  This pressure can easily crack welds.  Your weld looks a little concave, making the centerline the thinnest part of the bead, and therefore the most likely place for this type of failure to occur.  The fix, don't bottom out the swaglok, if you are.  Weld a bead with a convex face by adding more filler metal.  And slow cool the welds when they're done by wrapping in insulation and protecting from a draft.Benson's Mobile Welding - Dayton, OH metro area - AWS Certified Welding Inspector
Reply:Probably just a mechanical stress fracture. Is there a heavy or solid tube connected to it? If the things made entirely out of the same alloy and the line going in is secured from vibration or the bung is gussetted it shouldn't crack. Also ensuure the weld has some filled piled on it, not just a thin autogenous type weld.Welding/Fab Pics: www.UtahWeld.com
Reply:If those flanges are designed to see temps of 1000C, I doubt they are 304; more likely 309, 330, or Inconel. I don't have charts handy, but there are Inconel fillers meant for Inconel to stainless and other dissimilar metal joints. Might want to give them a try.I also second the motion for pre-heating the flanges to avoid restraint problems.JohnA few weldersA lot of hammersA whole lot of C-clamps
Reply:Thanks for the replies.  Lots of good info to consider. The new parts I have to weld are 316 (the pipe and fittings).  Would you still recommend to use 309 and not 316?  The flanges are cut out on a water jet.  99% sure they are 304 but I will double check.The welds are cracking in service.  Usually last about a month before they start to crack but depends on the test schedule (how many hours @ specified exhaust temps).The fitting is not bottomed out against a shoulder.  I looked at the dyno setup to see the air line and how it's connected.  It's a short piece of stainless line ~6" then to a SS flex line.  Not a lot of mass hanging out there for vibration.I'm leaning towards the fact that cold air (room temp) is being pumped in through this fitting while hot exhaust is flowing through the piece in question.  The cooling of the fitting vs. the hot gas heating the pipe is causing the welds to crack.  I made this post so I could find out if I was doing something wrong when welding the parts.Or, if there was something I could do different in an attempt to prevent this from happening again.  I'm going to add more filler to the fitting welds to get a more convex bead.  I'm considering a pre-heat process.  I'll let you know how it works out after about a month in service.  I appreciate all of your responses.
Reply:Can you add a bent tube extension to the fitting that would allow the cold air to be introduced into the exhaust stream further way from the fitting? I'm thinking a 90 degree sweep into the center. The point where your compressed air is expanding will be a lot cooler than the rest of the fitting. Another option would be to angle your fitting and extend the tip a ways into the stream. Is it possible that using a bung with more mass (larger) could counter act or resist the cooling affect? Just throwing out ideas. Keep us posted this is an interesting one.
Reply:Originally Posted by UnkleBukThanks for the replies.  Lots of good info to consider. The new parts I have to weld are 316 (the pipe and fittings).  Would you still recommend to use 309 and not 316?  The flanges are cut out on a water jet.  99% sure they are 304 but I will double check.The welds are cracking in service.  Usually last about a month before they start to crack but depends on the test schedule (how many hours @ specified exhaust temps).The fitting is not bottomed out against a shoulder.  I looked at the dyno setup to see the air line and how it's connected.  It's a short piece of stainless line ~6" then to a SS flex line.  Not a lot of mass hanging out there for vibration.I'm leaning towards the fact that cold air (room temp) is being pumped in through this fitting while hot exhaust is flowing through the piece in question.  The cooling of the fitting vs. the hot gas heating the pipe is causing the welds to crack.  I made this post so I could find out if I was doing something wrong when welding the parts.Or, if there was something I could do different in an attempt to prevent this from happening again.  I'm going to add more filler to the fitting welds to get a more convex bead.  I'm considering a pre-heat process.  I'll let you know how it works out after about a month in service.  I appreciate all of your responses.
Reply:Originally Posted by forhireCan you add a bent tube extension to the fitting that would allow the cold air to be introduced into the exhaust stream further way from the fitting? I'm thinking a 90 degree sweep into the center. The point where your compressed air is expanding will be a lot cooler than the rest of the fitting. Another option would be to angle your fitting and extend the tip a ways into the stream. Is it possible that using a bung with more mass (larger) could counter act or resist the cooling affect? Just throwing out ideas. Keep us posted this is an interesting one.
Reply:What about making the part from round bar? Bore the ID diameter. Drill and tap ends for bolts (or studs). Cross drill and tap tap for the fitting. If your remaking the part every few hundred hours I'd think it would make sense. I know this is WW but it's not really a very large part to make without welding.
Reply:UncleBukKudos to all posts . . . . . they're genuine and focused.UB -  My technical knowledge has immensely benefited from your problem  - unfortunately the recommendations mostly apply to my past problems, and maybe not solve your current one.If I read your condition correctly  -  your weldment has two .500" flanges welded to  -  .145" wall  /  1.5" SCH 40 pipe  -  with a Swagelok colostomic welded bung  /  .250"+-  wall  -running at 1000C  /  1832 F.Even with a perfect weld  -  this sectional mix will suffer thermal inequity, and prematurely fail somewhere.I would take the heuristic approach.http://en.wikipedia.org/wiki/HeuristicBeyond a successful weld  -  in service, you have to share the thermal stresses via sectional mass, or the piece is short-lived.  I would up-size the thru-pipe to SCH 160  -  .281" wall; and apply all of the weld and procedural recommendations posted.Opusp.s.  Weld first, but SCH 160 might be thick enough to be directly taped, thus omitting the welded bung.
Reply:If that thing sees 1000C (~1900F), you are WAY outside the box of good temps to use Austenitic StainlessCarbide Precipitation, especially during cycling could be part of the problem.  1900F is actually around the solution HT temp for ***Also, the material strength at these temps is super low.Could be that the temp differential and resulting stress is cracking the fitting weld, epsecially because the metal is pretty weak at operating tempsExperience is something you get right after you need it
Reply:Exactly- If it's really running at 1000C it shouldn't be made of 304 at all.A few weldersA lot of hammersA whole lot of C-clamps
Reply:Lots of good info; send it to a lab to get the pieces checked out.. Even a metallurgy type place can put the welds under microscope and figure out what's going on
Reply:It looks like there is a thread-o-let or something welded to the pipe rather than the Swagelok fitting itself! What's that made fromExperience is something you get right after you need it