weld symbols, fillet weld size dimension - minimum, maximum, or other 1

[weldgooch]

I am having a disagreement with a fellow AWS CWI. I see his arguement, but every other welding engineer I have ever worked with has taken my stance. I'm just trying to understand if my view point is how the Standard intends it to be read, or if the other CWI is correct.

Background: Simple fillet weld is called out on a flange to pipe Socket joint. The designer of the system placed a "1/4" size designation to the left hand side of the weld symbol. The drawing has a tolerance block on the bottom that says that dimensions called out to .XX are +/- .01, .XXX are +/- .005", .XXXX are +/- .0005. The actual measured weld size measures .328".

View point 1: The Level III Visual NDT Examiner rejected the fillet weld size because it is not exactly .250".
View point 2: Fillet weld size callouts are always minimum, unless otherwise specified on the drawing. The part is acceptable.
View point 3: Fillet weld size callouts are always maximum, unless otherwise specified on the drawing. The part is rejected.

Which of these 3 stances is correct, and why? If you choose options 2 or 3, please specify the paragraph in the standard/code you are using to support your view point.

 

[racookpe1978]

I question the 0.250 +0.005/-0.005 "dimension" for the fillet on a socket weld for completely different reason: NO welder can maintain a 10/1000 inch tolerance on a hand weld around a pipe fitting. The generic "dimension" note based on the number of dimensions is often used, but (at best) is (only marginally) adequate for machined surfaces, and total impossible in the real world for welds. For example, is a 3/8 inch dimension 0.38, or 0.37? Neither, the 3/8 inch 0.375 in the real world, and a surface feature on a 3/8 plate will begin 0.375 away from the far surface, regardless of what is "dimensioned" in the note. (Rather, most pipe welders will create a 45 deg fillet whose leg length is just under the thickness of the visible socket weld edge.)

Further, 1/4 is very small pipe weld for pressure piping, and not strong enough for many common systems.

However, the question must be asked: What dia pipe and fitting for the socket weld, what is system design pressure and Code, and what schedule pipe and what class fitting?

 

[gtaw]

The AWS standard for welding symbols does not imply dimensional tolerances.

The tolerance to be applied is typically stipulated by the code which includes workmanship requirements or the "company" that is manufacturing the component.

AWS D1.1 takes the position that the weld size and lengths listed by the welding symbol are minimums, but includes provisions for undersized fillet welds in Tables 6.1 and 9.16 (the latest edition of the Farm Code).

The manufacture is free to tighten up those tolerances if they choose to do so. If the manufacture decides it is necessary to maintain a +/- 0.001 inch tolerance, they are free to put themselves out of business if they choose to do so.

Many years ago on a planet far, far away, I had a client call me in to help them correct a problem with distortion. They were trying to maintain an overall dimension of 10'-0" +/- 1/64 inch. They were having problems maintaining the dimension, so what could they do different to meet their specification?

I asked what the component did? It held a compressor, an evaporator, and associated electrical controls.

Where does it go? It goes under a rail car.

How is it attached to the frame of the car? Holes are drilled and Huck fasteners installed and tensioned.

Are the holes predrilled in the rail car frame? No, the frame is placed under the car, jacked into position and holes drilled and Huck fasteners installed.

Is this a case where there is limited space and the frame has to be carefully positioned? No, the rail car frame is 65 feet long between the trucks and the frame can be placed anyplace that is convenient for the install crew.

So, if the frame is an 1/8 inch too long, it will still fit? Yes.

If the frame is 1/4 inch too long, it will still fit? Yes.

If the frame is 1/2 inch too long, it will still fit? Yes.

Why are you trying to hold the tight tolerance? That's our standard tolerance.

Why? That's what it has always been.

Change the print. We can do that?

It's your design, it's your install crew, it's your equipment.

Customer's response; "We were wondering how long it would be before you smartened up. We inspect to print. If you want to hold 0.001 inch, that's what we inspect to."

Problem solved. Change the tolerance on the print to +/- 1/4 inch. It still fits under the rail car.




Best regards - Al

 

[dik]

gtaw's got the solution... change the tolerance spec to something that is reasonable... and make sure the contractor, regardless of overwelding, is responsible for final fit-up.

Dik

 

[weldgooch]

Thanks everyone. The application is a 3" ANSI flange with a 1" Sch 80 pipe sticking through it and a smaller flange welded to the end. The smaller flange on the end is where this weld size is called out. The application is a sampling probe for use in an ASME B31.3 system for a customer of ours. The .250" fillet size comes from the 1.4*t fillet size requirements in B31.3.

What's funny is that the small flange has a boss at the weld location that limits the fillet to being only .250", and the testing lab wrote it up for being oversized...The welds are slightly concave, and the welder washed the weld up the run of the pipe more than he probably should have, which is why the NDT tech wrote them up for being oversized, DESPITE the fact that ASME B31.3, Figure 328.4 shows the weld size as being the largest triangle that will fit into the weld...

Our intention was for the weld size callout to be a minimum size, so we are dispositioning the welds as acceptable, I'm just trying to determine if we are required to specify "minimum" in the fillet weld size callout.

GTAW, Can you tell me where in AWS D1.1, they say that fillet weld size and lengths are minimums? I know it's AWS and my application is ASME, but it could provide evidence of intent since ASME borrows so heavily from AWS. I know in ASME B31.3, they do not say anything about it, but all references it size in ASME B31.3 are "minimum" requirements.

Thanks, All.

 

[gtaw]

D1.1 Structural Welding Code/Steel as well as other AWS structural welding codes only specifies the minimum weld sizes required to transmit the design load or based on metallurgical considerations. An exception for lap joints 1/4 inch or thicker is included in the code. The maximum weld size in a lap joint 1/4 inch or thicker is 1/16 inch less than the thinner member. If the joint is not a lap joint, the limitation is not applicable.

Table 6.1 does make provisions for welds that are undersized provided the length of the undersized weld does not exceed 10% of the total (continuous) weld length.

I know of no provision limiting the weld size except for the case of a lap joints as already mentioned.

I would let the opposition show you what provision limits the size of either partial joint penetration groove welds or fillet welds in cases not involving lap joints.

I am not aware of any provisions in the ASME code sections that limits the maximum size of fillet welds. If my memory serves me, the dimensions for the fillet welds, specified by the piping codes, that are deposited against the hub of slip-on flanges, socket flanges, integrally reinforced branch fittings, etc. are usually minimum throat dimensions (check the sketches closely) based on the thickness of the branch. The designer has to multiply the calculated throat dimensions by 1.414 to determine the minimum leg dimension of the fillet weld required.

Now, if the organization involved is silly enough to apply machining tolerances to manual welds, well it is time to let Darwin come to the rescue and put that organization out of their misery. Companies have been known to die sudden, painful deaths because they listened to idiots in high places. How difficult would it be to make a notation that the tolerances in the title box are for machined surfaces?

Best regards - Al