Contact US

Log In

Come Join Us!

Are you an
Engineering professional?
Join Eng-Tips Forums!
  • Talk With Other Members
  • Be Notified Of Responses
    To Your Posts
  • Keyword Search
  • One-Click Access To Your
    Favorite Forums
  • Automated Signatures
    On Your Posts
  • Best Of All, It's Free!

*Eng-Tips's functionality depends on members receiving e-mail. By joining you are opting in to receive e-mail.

Posting Guidelines

Promoting, selling, recruiting, coursework and thesis posting is forbidden.

Students Click Here

How much weld metal is needed in a weldolet connection?

How much weld metal is needed in a weldolet connection?

How much weld metal is needed in a weldolet connection?

Consider the following situation.
A 20" sch 120 main pipe fitted with a 16" sch 120 branch pipe using a weldolet.  Should the weldolet to main pipe prep. be fully welded out?
I have seen many connections made in this way on several plants and I regularly find the weldolet to main pipe weld prep. not fully welded.  This is so frequent that it is dificult to believe that it is error or individual poor practice.  I can imagine that a B31.3 reinforcement calculation might not require the full volume of weld but I see no calculations to substantiate this (ignoring the fact that the less than full weldolet often provides a notch for stress concentration).  The difficulty seems to stem from the difference in cross secional area of the weld prep in the axial and circumferential direction (pipe surface flat at the 12 o'clock and 6 o'clock positions and curved at the 3 and 9 o'clock positions relative to the weldolet).  I have heard this justified by welders by suggesting that large volumes of weld will distort the main pipe.  However my instinct tells me that reinforced fittings (weldolets) would not be used if reinforcement was not desired and so it follows that the weld prep. should be fully filled to provide this reinforcement.

All opinions and experience welcomed,



RE: How much weld metal is needed in a weldolet connection?

As a inspector I have always made it a point to at least fill to the first bevel. Some of the weldolets I have seen have a multi bevel for contour reasons. But like I said the first bevel all around should be sufficient in all applications. The issue of distortion of the pipe during welding can be controlled by the welding technique. I haven't had a client problem using that theory. Hope this helps and I will dig further into this for some more input.

RE: How much weld metal is needed in a weldolet connection?


Your question has been asked many times before....

This question always comes up with large diameter "weldolets". The extent of welding required for the smaller ones is obvious because of the shape.

If you indeed have a "weldolet" (manufactured by the Bonney Forge Company) they will supply you with a guidleine paper on the extent of welding required.

If you have some type of "copycat" fitting, ask the manufacturer...... most typically, the bevel at the piping interface must be filled out with weld.......moseley above is correct.


RE: How much weld metal is needed in a weldolet connection?

This is what is specified in the norwegian offshore sector:

Ref. http://www.nts.no/norsok/m/m60102/m60102.htm#ch67

"6.7 Welding of O-lets

The weld bevel of O-lets shall be completely filled up to weld line on the O-lets. Smooth transition between the pipe and the O-lets is required. Notches below the weld line shall be avoided. Prior to welding, sufficient root gap shall be ensured. "
end quote..

Ideally each combination of branch/header & OD/thk should have its own unique weldolet to provide sufficient (and sufficient only) reinforcement area.

The problem is that most weldolets (Bonney Forge and "Copycats") are fabricated in groups covering several combinations of OD/thk for branch and header in order to reduce number of variants to fabricate and to have on stock. Consequently some of them (but only some of them) will have a thickness at the bevel which is way beyond the necessary branch reinforcement area.

The options are then:

1. Fill out the bevel completely for the weldolet purchased

2. Buy a "copycat", specifically to the combination of OD/Thk of branch and header (remember to order it to be design approved) and fill it out completely (which will be less in many cases)

3. Do your own branch calculation for the weldolet actually purchased and fill bevel to calculated level. (Normally almost impossible because most suppliers won't supply the full geometry) Accept also in this case the stress intensification applied.

4. In some cases, especially with thin walled SS piping it may be preferable to go back to the reinforcing pad solution, which is easier to calculate and requires less heat input.

My opinion is - if you use a weldolet - fill it out or use another form of branch reinforcement.



RE: How much weld metal is needed in a weldolet connection?

This is a good question. On many occasions in the past I have come across Weldolets which have not been fully welded out to the weld line. One in particular resulted in a near catastrophic failure. In general I find that contractors do not fully weld out weldolets to the weld line generally with the mis-conception that provided they have the same thickness as the branch wall they will be O.K. However what they forget to understand is that in many situations Weldolets have been specified to meet Thermal Stress Analysis requirements and not just Pressure reinforcement. As such "Code" stress intensification factors have been used to calculate the stress levels. These "Code" SIF's are lower than set-on branch SIF's since they are based upon the assumption that the Weldolet is fully welded out to the weld line. In not fully welding out the Weldolet then the actual SIF's are far greater than the "Code" values. The problem is that in many cases the Pipe Stress Engineer has assumed that the Weldolet is fully welded out whereas in fact the inspector approves cases where this has not occurred. This is a recipe for disaster. Weldolets which are not fully welded out in my opinion are worse than Set-on branches due to the inherent "Notch" that is in-built in the fabrication. I have even come across Weldolets specified for a 20"nb branch on a 40"nb line. Again these were not welded out fully, to the extent that shart "notches" were formed, and the Pipe Stress engineer had assumed the "Code" SIF's in his analysis.Scary!!!. My opinion is that Weldolets should be fully welded out always and if the parent pipe is so thin that distortion is a problem then use some other type of fitting.

RE: How much weld metal is needed in a weldolet connection?

Hello all,

This is just a bit off topic but it is a good review for branch connection design.  It is from one of the copycats.


Regards, John.

RE: How much weld metal is needed in a weldolet connection?

Hello all,


While the referenced paper given above may proove to be useful, I really meant to give this reference from the WFI web site:


Regards, John.

RE: How much weld metal is needed in a weldolet connection?

Hey John

When are you going to break down and become a member so we can mark your posts as helpful?

Patricia Lougheed

RE: How much weld metal is needed in a weldolet connection?

Hi Patricia,

I have been a member from the beginning but I can never remember my sign-in password.  Thanks for the positive response though.

Regards, John.

RE: How much weld metal is needed in a weldolet connection?

In response to mgp's post on 5/9, specificly his #4 comment.  The manufacturers have fittings designed specificly for thinwall SS applications and here's a little insight as to saome of my experiences concerning their use or lack of:

I'm retired from one of the very largest chemical companys in the world but it took from 1987 until 1996 to convince my site management and the global powers to be that they were specifing and using the wrong branch connection fittings for thinwall SS in their capital and maintenance specifications.  In addition it became clear that little, if any engineering was taking place governing their use for intended services, just lots of cutting and pasteing in the spec books.

The company had made a big push to mandate the use of intergrally reinforced branch connections but the piping engineers and design engineers failed to understand the principle and difference between wall thicknesses when specifing them!  They kept specifing sch 40 fittings for sch 10 services and runs!  Heck, I even made a mock-up of a fully welded 1" sch 40 fitting on a 2" sch 10 run to show what happens when using the wrong fitting while maintaining the welding requirements.  Even with the fitting manufacturers catalog inhand, all I got for that little venture was a wasted fitting, 4' of well bent and squashed 2" sch 10 SS pipe and blank looks.

The wheels of progress do seem to turn ever so slow within a giant company but after nine years my recommendations were finally heard/understood and the global engineering specs were changed so that sch 10 fittings were spec'ed for sch 10 runs and sch 40's for sch 40 runs and so on.

RE: How much weld metal is needed in a weldolet connection?

Hi Seldom,

THANK YOU for your input!!!

Seems like just about the time we learn what we are doing wrong, there is a "changing of the guard".  It is a darn shame to watch all that hard earned experience march out the door on retirement day.  It would pay ANY company to have their "junior engineers" shadow the senior guys for at least the last year they are in the harness.  But if we do not learn from history, we are destined to repeat it.  So, somone else will figure it out just before he/she retires.

Now I have vented and I guess I feel better.

Regards John.

RE: How much weld metal is needed in a weldolet connection?

To Seldom

I agree totally with your input about using branch reinforcement fittings specified to suit the actual header and branch (as I also wrote in my response)

My point about the fourth option was that most manufacturers don't supply these "suitable" o-lets off the shelf.

Where do I find a Schedule 10 weldolet off the shelf?
They're not in Bonney Forge's standard programme and most other suppliers tend to have the same schedules as Bonney Forge.

Many times despite the fact that you specified the schedule of both branch and header for the weldolet, the fabricators will go to the supplier and he will be offered a Schedule 40 weldolet because this is the lightest one they have in their standard programme. Back in the shop he realises that it is far too heavy and assumes he can get away with filling it out partly. You end up in a discussion (probably one of them initiated this thread).
Eventually you have to decide if you can wait maybe 4-6 weeks to get a taylormade weldolet (with approved documentation because the geometry is not covered by any code) - or if you should maybe accept a reinforcing pad solution instead just to get on with the project.

My point was, rather than going through the above again and again, it could be an option in these cases to accept a "poorer solution" as fex. the reinforcing pad. Then at least you know what you get. After all this solution is still in the codes isn't it?

It would be nice to see an (inter)national code for weldolet geometry.


RE: How much weld metal is needed in a weldolet connection?

Hello Mogens and all,

Thank you, the issue of standard fitting geometries is a good point for our discussion.  

It surprises me to find how few piping engineers are aware of HOW LITTLE of the piping fitting's geometry is "standardized" in ANSI B16.9, Welding Fittings.  The end-to-end dimensions and tolerance for "square-ness of angles" is standard.  The wall thickness is not.  The wall thickness at the weld line location is standardized - but not through the rest of the fitting.  The fittings simply must be shown to be able to pass a burst test in which the straight piece of pipe welded to the fitting will fail under internal pressure before the fitting will fail.  MANY providers of welding elbows provide one schedule greater thickness so that there will be enough thickness in the extrados.  In the case of welding TEE's, the external geometry is not standardized.  If you examine the products of several manufacturers you will see some "barrel shaped" Tee's, some really rater spherical shaped Tee's, and variations on these.  When you think about the Stress Intensification Factors (SIF) in the B31 Codes you must remember that the testing done (in the early '50's) used only one manufacturer's welding fittings (it is advisable to read ALL the notes in B31.1 and B31.3, Appendices “D”).  In Tee's, the crotch radii varies GREATLY from one manufacturer to another.  In the case of Weld-on and Weld-in "self reinforcing" fittings (AKA "o-lets”) we have the additional complication of what the final weld geometry will look like.  

This piping engineering field ain't really an exact science (I guess that is why we have safety factors (AKA "indices of ignorance")).

Regards, John.

RE: How much weld metal is needed in a weldolet connection?

Hi mpg,
WFI has the sch10/sch10 fittings and I think they're called Pipets.  Once you have one in-hand, you'll notice that the amount of attachment weld is very similar to that of welding on an equivalent #3000 coupling.

"There's no such thing as a welding problem, there are only welding puzzles of assorted sizes!"

RE: How much weld metal is needed in a weldolet connection?

Hi Seldom & John

Thanks for the link and view on the fittings.

This is the first time I've seen anyone actually offer these fittings for SCH 10S etc.

John, I must admit I never looked in the codes for wallthickness away from the end of the fitting although I knew they were thicker. I assumed there were standards specifying these minimum thicknesses similar to the german DIN standards (e.g. DIN 2605) where you can choose between a weaker fitting (with wallthickness as pipe) or a thicker "full utilisation" fitting.

Would this then mean that the SIF's used for bends in stress programmes are based on the pipe wallthickness (due to no code data being available), and that the SIF for the purchased bend will always be lower, i.e. the stress analysis will ignore notes 6,7 of B31.3 appendix D?

Sorry GGh for not sticking to the subject.


RE: How much weld metal is needed in a weldolet connection?

Hi mpg,
Hey, don't feel bad about not knowing about them.  Heck, I had a whole company of engineers that took about nine years before they would look at a WFI catalog!

"There's no such thing as a welding problem, there are only welding puzzles of assorted sizes!"

RE: How much weld metal is needed in a weldolet connection?

Hi Mogens.

Perhaps we should start a thread for fittings discussions.  By the way, Phil Ellenberger (from WFI) is now writing a book for CASTI ("Practical Guide to...") on the subject of piping fittings as specified by various Standards.  Should be good - Phil is a Code Committee member.

The B31 Code SIF's came out of testing that was done by Markl (et. al.) at Tube Turns Corp. in the late '40's and early '50's.  These tests were comparisons of the number of cycles to failure for various piping components when compared to a piece of straight pipe with a girth butt weld.  Each of the testing machines simultaneously tested a component under in-plane bending, a component in out-of-plane bending and the straight pipe with a girth butt weld.  ALL the fittings were NPS 4 Tube Turns examples, they were all A-106 GrB material and schedule 40 and the tests were done at room temperature.  The goal was to develop factors that could be applied as a "band-aid" (Yankee term, that) to modify beam theory to "adjust" the calculated stress (in the component) such that it represented the stress the "must be" present to cause the component to fail in the correct number of cycles.  Of course the S-N data (and the "band-aids) were "extrapolated" from the tested NPS 4 size down to NPS 1/2 (not so bad) and up to NPS 72 (a stretch).  The B31 Codes try to "warn" the designer of the limitations and they say that the SIF's should be used only if the designer does not have better data (and he/she will not of course).  Also, there is a note in Appendix “D” that warns about D/t ratios greater than 70.

The software providers are in the position where they must make their software consistent with what is in the Code - and they do a good job.  So, if you use a uniform wall thickness in developing your model, the software will have to do what you ask it to do.  If you know the elbows are thicker you can develop your model accordingly.  Caesar II even provides a utility that allows you to change the thickness in ONLY the elbows in the model- good stuff, that.  But by using whatever is in the model, the software will calculate the SIF's according to the Code rules using the thicknesses you provide.  

The geometry of the Tee’s is hugely important in determining the final SIF’s (note that the Code uses flexibility factor of 1.0 for branch connection although we all know the FF is much more that that for branch connections (maybe that is conservative in static analyses???).  The SIF’s will depend greatly on the “cheek” profile and the crotch radii.  Trouble is that only Tube Turns samples provided the test data and every manufacturer does not use the same geometry in manufacturing the components (e.g., Tee’s).  Again, all the notes must be perused.  

When it comes to fabricated branch connections, we know from recent testing (by Glynn, Everett and Phil at WFI, btw) that the workmanship and final profile of the weld is very important in determining the actual SIF’s.  But that cannot be effectively guaranteed by a specification.

Turning to B16.9 reducers, we see other areas of interest.  The cone angle varies greatly from one manufacturer to another.  The “transition curve” from one OD to the next OD varies greatly from one manufacturer to another.  The wall thickness is only standardized at the weld line (typically line-bored to make spec).  The Code SIF’s for reducers are representative of the TT Co. product.  We usually use the maximum of 2.0 as a SIF for both ends of reducers (conservative for static analysis) since we do not know what the purchasing department will buy and we have not other data.

There is work now being conducted by Glynn Woods and Everett Rodabaugh to improve the area of SIF’s.

Oh well, enough for now.  Mogens, I hope that I addressed your question somewhere in all of that.

Best regards, John.

Red Flag This Post

Please let us know here why this post is inappropriate. Reasons such as off-topic, duplicates, flames, illegal, vulgar, or students posting their homework.

Red Flag Submitted

Thank you for helping keep Eng-Tips Forums free from inappropriate posts.
The Eng-Tips staff will check this out and take appropriate action.

Reply To This Thread

Posting in the Eng-Tips forums is a member-only feature.

Click Here to join Eng-Tips and talk with other members! Already a Member? Login


Low-Volume Rapid Injection Molding With 3D Printed Molds
Learn methods and guidelines for using stereolithography (SLA) 3D printed molds in the injection molding process to lower costs and lead time. Discover how this hybrid manufacturing process enables on-demand mold fabrication to quickly produce small batches of thermoplastic parts. Download Now
Design for Additive Manufacturing (DfAM)
Examine how the principles of DfAM upend many of the long-standing rules around manufacturability - allowing engineers and designers to place a part’s function at the center of their design considerations. Download Now
Taking Control of Engineering Documents
This ebook covers tips for creating and managing workflows, security best practices and protection of intellectual property, Cloud vs. on-premise software solutions, CAD file management, compliance, and more. Download Now

Close Box

Join Eng-Tips® Today!

Join your peers on the Internet's largest technical engineering professional community.
It's easy to join and it's free.

Here's Why Members Love Eng-Tips Forums:

Register now while it's still free!

Already a member? Close this window and log in.

Join Us             Close