Video Discussion: Why Do Good Welders Get This Wrong? (Inspector Fabricator) - Questions & Comments Thread

but these parts werent cheap , 300 dollars labor for 15 minutes of labor and zero fixturing


Man, dont argue with idiots. You will never win


I think the biggest issue we are seeing in this series is that these shops do not understand their own processes. They are also extremely eager to accept work without understanding they should ask you for spec relief or really any questions at all. If you went to those same shops and gave them a tolerance of +/- 1/128" they would still accept the job even though they would have no hope of ever hitting that spec. For the morbidly curious amongst us, is there a table of all the CMM measurements lol. This series is very well put together!


The biggest problem I have is I’m not professional and can’t justify the table. If I could justify it I would have one!! I think they are the best!

1 Like

I think the issue is not the steel table, but that these shops just don’t really care about quality/ hitting tolerances. The steel table is not the cause of the errors its just a symptom of the mentality. A shop that is focused on quality is going to have tools and setup capabilities that allow them to meet what the drawings require, this could involve a fixture table, but it could not. As you showed you can hit the tolerances on a steel table, you just have to try.


@James_M Can I ask a question then, if you were a customer and needed parts made, and couldn’t see the type of work the fab shop produced, what would tell you that they cared about the final part? I think a fixture table and fixtures is a pretty good signal that they care about the work they want to provide, if that’s all I can see walking into the shop. Thoughts?


@Fireball_Jason I agree that it can be a good indicator. But a stronger measure would be to talk with the shop about what their inspection process is. Maybe include a measurement checklist with the drawing package. These questions will communicate to them the importance of meeting the tolerances and their response is all you need to know.

You mention it in the video that fabricators need to know what they are able to build with their tools. I would be willing to bet that a lot of these shops projects are not sensitive to tolerances and some of these guys probably have not focused on it in a long time.

Edit: added the @
The flip side is that a lot of customers (myself included) don’t know what tolerances are required. Its not the fabricators job to deviate from them, but there are a lot of dimensions that are over tolerance for the use case.


you could test by lending a fixture table to the same welder so that they can try with the table so that we can see if that table can really make a difference


I used to work with fabricators in an R&D shop for CaseIH/Fiat Industrial and they all used fixture tables similar to the one you have, only way heavier. The tolerances these guys could hit was amazing. If you’re fabricating for dimensional accuracy, you either need fixtures or jigs to counteract as much of the weld draw and distortion as you can, and you need to have a vast understanding of welding to realize when too much heat is going to affect your final product. Now that I work on our family farm, having the welding shop make a piece that’s 1/8" out usually isn’t a problem. It all depends on the industry really.


I am not a welder. I am a student engineer. I have been enjoying these videos a lot. Even though I came here for entertainment at first, your content is incredibly informative - I would have never thought that distortion due to welds could be so significant! Also, when bending the frame back, isn’t there a risk that you will bend it too much so that it becomes concave?

For further testing, I think it could be interesting to see the job done by the shops who actually use the fixture table (like, the point so that the person who would do the work wouldn`t know that that’s for testing). If these defects are caused by negligence, this test will show that, because these flaws will reappear.

Anyway, I think your content is amazing, and keep up a great job!


That question about what’s a heuristic for quality is really important. It got me to sign up for the forum :rofl:

I’m not nearly as skilled as you / the people here (I do social science & design for a living), but I was lucky enough to grow up working on HVAC with my dad.

Tolerances on duct fabrication are much looser than on tables, but there were two things that I saw that said “quality” to me:

  1. the condition of a shop:
  • Are the tools well maintained?

  • Is the space super messy?

  1. the culture of a shop:
  • If you’re there at the end of the day, are the fabricators just throwing their stuff down or are they cleaning , etc?

  • Do they take safety seriously?

  • Do they keep up on new tools / laws / trends?

There’s other questions in these buckets, and while no single question is a pure disqualifier, enough bad answers and I’ll take my business elsewhere if possible.

And that’s the thing I wonder about these shops: how much competition do they have in their craft? Do they have customers coming back and asking for better quality, or is it easier for customers to just make do because of timelines or travel?

Thanks for coming to my TED talk, and thanks for the mental workout this morning!


I don’t know so this is my genuine wonder. What would happen if the fabricator in this test were to get everything clamped square and then preheat the metal to 1,2,3,400°, etc before welding it? Would this alleviate the warping and twisting? I recognize that you showed a solution of experience with weld distortion telling you that you should prebend the material in the opposite direction by “x” amount but is there another solution doesn’t require this knowledge and will remain constant with the same materials?

1 Like

Love your vids, I learn a lot, picked up a few tools. Been fabbing out of 2 car garage forever. Would love a table, no room, not going to happen. But I know enough to never promise more than I can deliver.

1 Like

Im no welder or fabricator, I can probably count on 2 hands how many times ive done both. However I have done a lot of carpentry and framing ect. and what these fabricators/welders seem like are just that in terms of tolerances, a 1/16 or 1/8 maybe even as much as a 1/4 off your measurement (depending on what you are doing) is no big deal. It cant be that important to be a little off right. But the point I can see you are trying to get across and it doesnt seem like these guys are understanding is that those tolerances matter, it doesnt matter what you are going to use the part for it still needs to be within that tolerance. I think next video you should have all of these guys try out one of your tables and maybe just maybe sponsor a table and tooling for them just so they can see the difference first hand.

1 Like

It’s an entertaining video, and a good ad for your products, but for it to be a true experiment about fixture tables you’d need to give the same job to a bunch of welders with fixture tables. Then imagine the jobs you got from the “fixture table guys” are better. Does it mean it’s the fixture tables, or is it that people who invest in their own business to get best tools possible are in general more precise about their work? Who knows? To control for that you’d need to take half of the welders from the fixture and non fixture table groups and swap them (so guys normally working on a fixture table do this job on a steel table and vice versa).

It is an undisputable fact its going to be easier to be precise with a fixture table. But just as you’ve shown this job can be done properly without.

Still some people just don’t really care. I’ve project managed building my house I know how many shoddy workers there are in all trades…

From your results that only 2 out of 6 frames passed on size it’s obvious none of these welders really cared enough more than to eyeball the cuts with a tape measure probably.

Then the twist… I think they just took the plans, and used them as a set of instructions. “oh they twist? Of course they twist, it’s your fault for spec ing such a big weld” no doubt that’s what they’d say if pressed.

Finally the “matching pair” do you think any of these guys even knew what you meant by that?

Sadly, if you need something made well and you can’t make it yourself you put penalties in the contract for various faults. That’s the only way.

Coming back to the tables. I’ve built straight and precise weldments with no table at all. I tacked the parts together in a way they could be moved with a hammer. Then I bent them into the shape I wanted checking with straight edges and big squares. Finally I tacked them properly and I welded them slowly. About half inch length weld at most. If you design your part we’ll so your twists cancel out and you go around it doing small welds like this in a correct order you can make really precise parts in a dirt floor using an old door on two sets of cinder blocks as a table. I know I’ve my 200 ton hydraulic press that way. How do I know everything lines up great beside measuring it? My table moves up and down in the frame without binding or looseness. Also all 4 pins slide in properly at any table position. This frame was made with 1x5 in mild steel bar for the long bits and 8in C channel for the top and bottom rectangular sections. All welded with dual shield wire on about 340 amps. No twist. No bow.


It is like taking your classic Aston Martin DBR1 to Earl Scheib (I’ll paint any car for $29.95) and expecting Concours results. :upside_down_face:

@Fireball_Jason In my opinion as an engineer in the field, the biggest thing I take issue with in this test is the drawing.

  1. at 22:50 Jason says “that 1/16” at every corner", If he wanted flat he should have put the proper GD&T callout for flatness on the drawing. He gave them 1/16" (0.062") as the tolerance right on the print, no shop is going to spend more time or money than they have to, if you are putting 1/16" tolerance on the print, you should expect the part to be 1/16" off from what you asked for.
  2. The drawing has no tolerances listed at all for angular dimensions, we only get 90deg call outs in several places. With no tolerances listed, even in the title block the builder would assume that the tolerance is +/- 1deg, acceptable range being 89-91deg. Over the longest dimension 20 1/4" doing the trig reveals that side could be 0.353", almost 3/8", away from perfectly square and flat and still be within the assumed tolerance of 1deg. Over the 15" dimension callout for length of the leg from the assumed flat bottom plane (that is not explicitly called out) that would be 0.262", over 1/4".
  3. The drawing is not explicitly fully defined. For the desired apples to apples comparison between shops to be valid you cannot rely on everyone interpreting the drawing the same way.
  4. In other videos especially about Jason’s fixture table he talks about aerospace quality. If that’s what was expected here, a table that would sit on a granite plate and not rock at all, have all legs line up etc, the print should have been dimensioned/toleranced as such and the part would have cost thousands of dollars more for the additional operations like grinding/machining that would be required. If this would have been done I would be willing to bet that all the shops would have turned down the job based on the true dimensional tolerances required.
  5. as a final point on the drawing, in location F6-F7(upper left) of the drawing there are two 9 1/8" callouts to the outside of the perpendicular leg instead of dimensioning to the centerline of the leg. This does not take into account the variability of the 2x2 square tube OD at all. If you cut the item #2’s both at 20 1/4", then pull off the tube in the center when the tube OD is thicker or thinner than 2.000" you will have conflicting dimensions. The same is true of the 9" dimensions on the other side not being edge to centerline.

Conclusion: In large part I believe Jason got what he asked for on the print more than proving or disproving the premise of the video that a fixture table would have resulted in a better or worse final product.


I agree completely. I replied as such. there is also no tolerance on angularity at all. Average welder would assume 90deg to mean +/- 1deg. over the 15" of the legs dimension shown on the print that 1deg is over 1/4"

1 Like

Totally agree. The last part of this experiment should be

  1. Provide feedback to each fabricator on the accuracy of their original job.
  2. Give/loan each of them one of your fixture tables with appropriate accessories.
  3. Make sure they have links to your videos on best practices and provide access to the forum for q&a.
  4. Run the original experiment again and show the results.

For the industry as a whole to improve, one has to start educating and demonstrating the ROI for a small fabricator… with small fabricators. I would love to see each of the fabricators “pass” the test! When you have an expert (you) showing how it’s done with all the “fancy tools” it’s not always clear if it is the expert’s experience or the tools that make the difference.


@Airick That would fail the dimension call out. Its clearly printed distances at the end of the legs.