What's new
What's new

How to accurately measure a fillet weld size?

turnworks

Cast Iron
Joined
Dec 12, 2018
Recently failed a weld test cert for welding being oversized by .002" on one leg of the weld.

The material is a nickel based alloy .030" thick T joint. Spec is AWS D17.1, tested by metallographic.

I had built some weld gauges because I couldn't find any the proper size. The spec calls for 1.5T min. and 2T max leaving me a .015" to play with.

The way it was measured is a cross section is taken and put under 25X mag then what looks like a computer program was use that makes a right angle triangle from the base metal to the lowest point of the weld throat.

I tried to cut a section out of a practice piece and use it on an optical comparator but its hard to perfectly guess where the place that the weld test is sent to is going to put that triangle.

Any help would be very much welcomed.
 
+-7 thou on a weld fillet, well thats significantly outside the range of my unassisted vision. Does not sound like the people that wrote the spec expected it to be used on such thin shit. Have never seen gauges bellow 1/8" either.

Hard to imagine how your going to ever do that by hand and not with a automated process, is my first thoughts.
 
0.002" is not much and hard to see how it would affect a weld in use, that tolerance is more in the machining field ( and lower).

Maybe he knocked you back for some other reason and used it as a excuse or justification for same.

All i can say is try again.
 
It’s honestly more of an audit thing. We weld a pipe to a flange where the pipe is .020 wall thickness nickel alloy to a .250ish stainless flange. Customer requires a super smooth no ripples or filler metal added weld. They don’t even want to be able to figure out where you stopped at.

After a fair number of tries the process is nailed down now and it actually is real fun part to weld, just no blinking or breathing during it. It’s the in house books that is making me do the test not the customer.

I picked .030 thick for the test to cover the .020. The nickel alloy is some weird special junk which really complicates things. There isn’t any weld wire for it. We use .025 thick sheet sheered to as thin as we can. Only reason I bring that up is trying to automate the process would be a heck of a challenge.

My eyes can’t see that tolerance either with a standard style weld gauge.

Any chance I can change the size requirements on that spec? The 1.5t min and 2t max is if no size on print is stated. From rereading it I’m getting the impression that the size is print supplied.

Anyone know that spec really well?
 
Recently failed a weld test cert for welding being oversized by .002" on one leg of the weld.

The material is a nickel based alloy .030" thick T joint. Spec is AWS D17.1, tested by metallographic.

I had built some weld gauges because I couldn't find any the proper size. The spec calls for 1.5T min. and 2T max leaving me a .015" to play with.

The way it was measured is a cross section is taken and put under 25X mag then what looks like a computer program was use that makes a right angle triangle from the base metal to the lowest point of the weld throat.

I tried to cut a section out of a practice piece and use it on an optical comparator but its hard to perfectly guess where the place that the weld test is sent to is going to put that triangle.

Any help would be very much welcomed.

.
i could never manually using my hands weld to that tight a tolerance fillet weld size. my speed control and ability to hold or control a feed rate and height is not that precise. sounds like you need a robot
 








 
Back
Top