Why Do Good Welders Get This Wrong?

in order to properly tell you this story I need us to rewind back over a year ago to October of 2022 I released a video where I tried to show how difficult it can be to fabricate simple shapes and the obstacles Fabricators need to be aware of at the end of that video I suggested a fixture table to help solve all these problems look at the accuracy that you have with a table like this this is incredible and some people had a few things to say about all that things like Jason you are an idiot I can do all those things without that table that table is pointless if you have enough skill you don't need a fixture table with that exact line I decided to test that theory so I created a little experiment and that brings us to video number two so 5 months later in March 2023 I paid three professional Fabricators to build me the exact same frame that I had made in the first video and made sure none of them had a fixture table because if they have enough skill then surely they would wouldn't need a fixture table right yeah I got a steel table as I tested them I was surprised to find that they all failed and concluded that none of them were able to match the drawings that I gave them this experiment has got me curious I don't know why all six of these frames are failing something very strange is going on here and I think we need to discover [Music] why [Music] this mystery has me thinking we need a new approach after all you guys on the fireball Forum had some suggestions as to why these frames turned out the way they did so let's concoct a new experiment some of you said we need to communicate better with more detailed plans others said the job wasn't taken seriously because the part was too simple so the next part needs to be a bit more complex and finally it was stated that our testing method was biased so I'll make sure to test these new frames with something extreme so for this new challenge I'm going to cook up something spicier three new Fabricators will have to make this plan twice we're going to start with our original square frame that we had last time and then add four stub legs that gets welded perpendicular to the original box frame kind of creating a table but not really a little bit later I'm going to explain why this test is difficult and what makes this the perfect challenge hopefully the welders can identify the problems with this test and solve them let's take a look at the drawing I made sure that we have weld callouts weld sizes materials lengths Heights and even a cut list for the Fabricators to follow if they can't make it from this design sheet then there's some serious issues we're going to have to dress later if you'd like to get a closer look at the drawings and follow along with the build I've made the plans available to download at the fireball tool Forum this brings us to August 2023 so only one more thing left to do get the plans in the hands of the Fabricators I'd like to go to three different Fab shops than I did last time this is going to increase our sample size like the previous video I'm going to keep the Fab shops and welders identities hidden because I don't want to make anybody uncomfortable we're going to drop these drawings off in person this is going to give us the opportunity to ask the Fabricators a few questions like what strategy would they use to build this and did the drawings look okay to you yeah it shouldn't be that big of a deal flat table put it together and put your legs on Square them up and you're done sounds easy enough to me that's easy and they actually gave us a cut list oh it's really surprising oh they never give us good stuff like I like whoever made this PR challenge accepted I want to remind everybody at this point that all the Fab shops and Fabricators are using a plate steel workbench table so that got me thinking how would I build this simple test I think there's three different methods so let's talk about the first one I think I would start by building the square frame first and then when that's all welded I would start standing all four legs up and get those welded to the frame so let's talk about method number two which is a completely different approach this method starts with the legs and gets it welded to the sides first at this point you've constructed four T joints and now you combine all four T joints to make the box frame method number three puts the emphasis on making two identical parts and you can do this by making two box frames welding all four legs together creating one big large weldment and then coming back and cutting it apart in the center whatever method they choose they have to make make two identical parts so here's what makes this test also difficult so it's not uncommon for a welder to build one part and then try to make a copy from it this weldman is difficult to do that because of its geometry you cannot stack the parts on top of each other and recreate one from another and then whatever method they choose they're going to be fighting weld distortion on this test you're probably noticing that the leg is not in the corner well that's on purpose because what I want to see is how the welder is going to handle weld Distortion so we've place the leg right in the center of the tube and this is going to add a whole another element of trouble so as we wait for the Fab shops to build the frames I think we should do a little experiment right here on the table and see how much Distortion these guys are actually dealing with now I've set up a little sample right here of this exact length tube 20 in and we're going to weld this little stub piece all the way around as a control to see how much Distortion we're actually going to get I'm going to put weld seam opposite of our little stub shaft so everything's consistent here we don't want the seam to add any extra weird stresses into our test for this first one I'm going to tack it all the way around weld the sides and then do the fillets last if you're already a welder this is just something we deal with every day and we can predict how much it's going to warp but I think this is a good demonstration for those of you guys who haven't seen this before these aren't overly sized welds the plans call out for an eighth of an inch minimum I could definitely TIG this but if you want to make money you're probably going to be wire feeding it so we're going to let this cool for a second and then we're going to test to see how much Distortion there really is as we wait for the weld to cool there's a lot of heat concentrated in this area and as it slowly cools it just slowly pulls shrinking on each side and as that does it pulls this up this is my first rodeo boys I know she's going to get Rowdy over here going to do buck and Bronco WEA okay I think she's cool cooled down enough to go test on the surface plate to see where we ended up now I'm just going to Teeter it has bowed up quite a bit you can see some light through there so how much Distortion are we talking here I have my Fireball shims let's see if a 16th will slip underneath the edges of that almost a 16th of an inch 32nd maybe so not quite a 16th but definitely more than a 30-second so if the Fab shops don't do anything to their material they're most likely going to get at least at least a 302 to 16th of an inch of distortion now let's see some other ways they can fix this let's clamp this next one down and test it weld goes down part goes on top and we clamp it to this 2 and 1/2 in thick plate steel table that's going to fix it isn't it not going anywhere now let's weld it same thing as last time tack it first then weld it it's cooled off now so let's take it over to the surface plate and see what we got let's put her down here and see how she does I'm just going to Teeter it here's a 32nd of an inch goes underneath here's our 11/16th here's the control here's our first weld second weld I think they're almost identical maybe this is a little bit better the one we clamped down but in all reality they're the same same piece of material not welded as you can see it's perfectly flat as flat as a piece of tubing gets that's what makes this test so valuable as I know that this was going to happen every welder's nightmare is just dealing with Distortion and it's very predictable I am inevitable so you're probably saying Jason you didn't leave it clamped in long enough a minute is too short you should have left it in until it completely cooled we can do that test too if we want to but in all reality if you're have to wait more than a minute to do one weld at a time it's going to take you forever so we know clamping doesn't work we know just welding it and the open air doesn't work the next option is to anticipate the Distortion only way I've known to really fix this is to pre- bend it now I'm going to use the fireball shims here I got all assortments we just showed that it at least has a 132 of uh Distortion we know it's probably a little bit more than that but I think this is going to be pretty close so what I want to do is pre- Bend this tube that same amount so I'm going to put these two shims together and they're basically going to be positioned right underneath the fillet welds and when I put this tube steel down you can see that it it Teeters about just as much as it distorted last time see the Gap over there I'm going to bend this tube the other way got to have big heavy substantial table now we got a good bending it let's weld this guy down right over the top of those shims and then let's see what happens okay we'll let that cool for the next minute or so let's pull them out shimmed pre-bent piece of tube steel I can barely see it moving has a little wiggle in it that would definitely itely be in Spec I don't know you probably can't even see it come up on camera it is barely wiggling but that is a big Improvement so we know the prebend works really well before welding what happens if they did bow it and warp it do they have to start over or can we fix it so in order to fix this we're going to need something flat I would use this big piece of plate steel to straighten this piece up but it's really not perfect as you can see here my plate steel table has an eighth of an inch dish in it so unfortunately it doesn't make a good surface for inspecting my fabrication throughout this experiment I've been using the granite surface plate to check flatness but I'm unable to fabricate with it so instead I'm going to use something that I can do both with the fixture table so let's go over to the fixture table and I'll show you how we can do it there so now I have a good reference surface I know this is flat I know this is straight so I'm going to do the same thing I did before but I'm not going to be able to use these 13c shims that I'm going to have to Beef It Up I'm going to need to R really stretch it so I'm going to use these E8 in shims I'm going to go underneath here where I had that before remember this is our control this is the one that's bowed and has all that Distortion in it we're going to try to straighten it out clamp it all down I know this tube is pretty stiff so I'm going to get it clamped and then I'm going to add just a little heat back into it and we'll straighten it up I'm just going to warm up the fillets a little bit cuz I know that's where all the stress is at this is the long way don't make this mistake okay let that cool down Okay so we've taken all the tension and stresses out of the welds even though this is still a little warm we've relieved it so it should stay and it did it stayed pretty flat so those are some things we'll look for when we get ours back as if they had to fix it cuz you can see a little heat Distortion around it there's also a few more reasons why I chose this table shape for this experiment and the first is that all four legs are completely unsupported at the end what this means is that it's going to be difficult ult to keep everything straight Plum true and perpendicular to the base worst case scenario you have welded the legs crooked and now you want to adjust them unfortunately if you try to straighten the legs it's going to cause some problems for example leg condition one legs are SPL out evenly well you could put a clamp on them and pull them straight same goes if the legs are splay in you can get a clamp reverse it and push the legs back out what you're probably really doing is twisting pulling or bending the base unfortunately that's going to make make the base even less flat than it probably is to start with here's condition number two one leg tilted in one leg tilted out unfortunately the clamping method to fix this is not going to work because as you clamp yes one leg is straightening but the other one is getting worse here's another compounding issue yes you've straightened the legs up but you twisted the base but did you think about the leg length changing at the same time that's going to make things teeter totter so we're stacking up compounding issues here I'm sure you're thinking Jason you're making a bigger deal out of this than it really is well I guess it's time to find out let's pick up the parts and see how these Fabricators did when I arrived at the first shop I was immediately asked really for so I explained to him that he was a control for a big experiment and that he was representing welders with plate steel tables and I was going to be looking at how accurate this Frame was welded and put together I also told him I was going to be asking questions that you guys left me on the fireball Forum from the last video so the first question I asked was how difficult he thought it was to weld together why would it be difficult so I explained to him the concept we explored earlier well Distortion I mean tacking it up isn't a problem but yeah when you weld it out then things move on you yep now that we were on the same page about the existence of weld Distortion I asked him how he made sure the part was right and correct to the drawing well tape measure square so if you you got a little weld sticking out but if you check your squareness okay if it was upside down so this one pulled out a little bit supposed to be 20 and a quar wait a second did he just tell me this is wrong even though he said yeah it shouldn't be that big of a deal and this is right at 20 okay so the the bottoms and the top right the same theoretically should be right and then I asked him this do you have a way to check flatness well when I put you kind of level those four pieces okay when when you tack them together okay and if they're level yeah when you tack them together theoretically they should stay huh I mean you know you don't see any gaps there right I wanted to know if he felt he had the right tools to check if the work he's creating actually is flat is that table good enough to test for flatness um I'll take that as a no I told him the Twist and flatness is important to me because what happens if I wanted to put a piece of glass on top of this Frame well if if you were going to put a piece of glass on there you would you would have for cushions of some sort rub so glasses rub absolutely but but wouldn't the cushions be the same height so no matter what if one's lower you would have one that would teeter totter on this conversation went on for a little bit and it was great picking his brain and I don't want to waste any more of his time cuz I'm sure he's busy so on to the next Fab shop when I got to Fab Shop number two I also took the opportunity to ask a lot of questions what was his technique to get the frames as close to tolerance as possible I go off the table uh for I mean for something this small uh I i' put a i Center it up in the table uh my table's kind of kind of like that a little bit okay so something like that if I put around in the middle uh it get bigger projects I'll uh I'll put angle finders on it uh it's a digital reader uh within like half a degree so like if I got two pieces out I'll zero this one out and then match this one to that uh but for something with small I just put it right in the center of the table basically what he's saying is that because his table's so crowned that it's really difficult to to use it as a reference surface or use a square so he's decided to use the level method and use that as the reference I have a hard time believing that this is the fastest way to build this because you know in Fab shops time is money and this is where you make your profit with my parts in hand made with a level it was time to go to shop number three so how were they were they hard to build No Easy Peasy easy peasy perfect how flat is your table do you know at this point I mean uh it's one those things that typically I throw a level on anything if I'm like okay this is something that actually has to be precise right uh we've talked about it but it's just one of those things of getting an actual fixture table so how are you holding stuff in the middle of the table are you welding it down uh if we have we put it in the middle or we just do everything off of an edge as best we can we get creative with uh clamping and all right we got to do what we got to do to make it happen can I ask you how long it took you to to make them about two hours two hours for so an hour a piece uh basically yeah so here's how the cost breaks down generally materials are about 130 bucks labor is about 2 hours and you can vary between $100 to $170 an hour depending on the shop so the total is around $470 unless this happens uh if they come in and they say it's wrong we fix it it's what it is you know if we they gave us a so then if it's fixed on their time they're out the labor cost and maybe even more material costs too but we won't know if that's going to be the case until they're tested but before we do that let's take a quick look at them after all I find it strange how each fabricator said this was going to be an easy job to do and what's inside there's always stuff inside of these tubes what are they packing in there I'm just going to clamp this to the table of this tube I'm just going to look at the roll but these are just visual inspection stuff that you can do before the customer gets gets it you can see that this tube is extremely Twisted it might actually still be in Spec it's just strange to me in order to know whether or not these actually follow the drawings we have to test them the last time I tested frames I used my fixture table and set up some positive stops use a dial Caliper and use the surface plate to do some measuring but as was suggested on the Forum I'm taking it a step further so this time we're going the extra mile I'm going to be using a CMM or coordinate measuring machine let me show you how it works a CMM is a giant camera that looks towards an object and then a probe is used to touch various points on the object this machine can measure down to the thousandth of an inch the camera sees where the probe is touching and creates coordinates these coordinates are fed into a computer to create a perfect 3D Recreation of the object and this is what's going to tell us the frames pass or fail the drawings so who's going to be running this machine well I'd like to introduce to you Caesar hello he knows the ins and outs of this machine this CMM is one of the tools we use when we run quality control on our Fireball products I use to test all the fireball squares all the tables a lot of different things throughout the shop so whenever we need to test something this is our go-to baby over here we have the dylus it's going to be the Ruby Crystal basically really hard material won't scratch it's going to let us get accurate measurements in 3D space uh it uses a set of sensors camera system tracks me as it moves as long as I have clear line of sight it should be giving me a really good measurement this right here really reliable if it tells me something is not flat something is not Square something is off I trust it today we're going to be using on this table frame to check if these schematics are in fact correct we're going to be testing overall Dimension so length height width so it's going to be from bottom of the weldment all the way to the top we're going to be testing if these legs are twisted in any way so if it's off or if it even is in the correct place we're going to be able to tell that with this machine so probably throughout the whole part uh while I'm testing it it's probably going to be like closer to 250 300 points in total so you can imagine it's going to give me a really good picture of how this thing truly is making a perfect frame is virtually impossible but how far from perfect are these frames allowed to be well the drawings indicate 1/8 of an inch tolerance around the weldment except for the twist of the frame which is 116th kind of like the game Operation if any part of this exceeds the tolerance it fails this machine is going to be able to precisely identify each piece down to the thousandth of an inch completely removing any human error so the CMM will make sure this weldment is not a parallelogram a trapezoid or an irregular quadrilateral but that's just for measuring the frame since we added legs to the drawing there could be a lot of things keeping them out of alignment things like are they tilted to the left to the right or are they leaning inward or outward how twisted are they are they centered front back and side to side at the T joint and are they even the correct length all these questions are only taking the dimensions into account so what about the twist of the frame luckily the CMM will be able to measure the tube Distortion by finding the highest highs and the lowest lows then we will use the numbers it gives us to see what frames pass or fail as long as green light is on you're good to measure and then you can just lightly place it on Top push the button it'll give you a beep and then you can go crazy so just light pressure as long as the light is on hear the beep you got a point you can go upside down all the way around one of the good things that this cmn machine does is you can actually take a picture of what you are essentially measuring it should pop up right there you can actually see in real time where the stylus is on that picture so if I start making a plane on this Edge right here or this surface it will show up on that picture with the right perspective to actually seem like it's on that surface and I'm trying to be be very careful I do not want to move this if it moves even just a tiny bit all my measurements are going to be off so I'm very careful just letting the Ruby hit it took about 16 points like I said nothing too crazy so right here you can see that plane we just got in relation to the picture you know it looks like what I just measured and it's at a 6,000 flatness so if that was one of our tables it would fail and Jason would not be proud this is just a quick demo of the CMM and how it works and what we're going to be measuring I got a lot more to go so I'll let you know when I'm finished just let the Ruby do the work while Caesar's collecting all these data points let's see if we can visually identify any problems with the frames remember how we talked about weld Distortion earlier I have noticed that not a single frame has addressed this issue they all have a curve to them and we're not talking any curve we're talking that exact 116th of an inch deviation in all the corners of every single tube that doesn't even take into account for Twist which we'll check out in a minute everyone is Warped flip them over I know this is accentuating the problem it's pretty lousy and they're all this way here's two frames that are supposed to be identical which what is it they're so bad I don't even know which way it goes or it has to go this way if I line two legs up the best I can here's what the other ones look like I'm going to line two legs up again the perfect world that's how they should look on this side and this side this one's pretty good this is the best out of the three but we still got some issues here that's probably with intolerance while I'm not hopeful from what I've seen so far Caesar let me know he's done taking measurements so I just finished measuring all the frames right now what I'm going to do is I'm going to go through get all the measurements I need so things I'm looking for is overall dimensions height length width I'm looking for the twist in the bottom frame seeing if that's flat we're just going to go ahead and check the flatness I have already set to show me flatness whenever I click plane so you can go ahead and see 106,000 so for this we should be within a 30 second so that's uh 32,000 it's about and as you can see clearly we're almost tripled at it is really out of spec on this so for that reason this failed I'm going to go ahead and wrap the rest up and I'll go ahead and give the results Jason let's dive into what season found because he collected so much data on each frame we've broken it down into four categories the first is dimension two out of six of the frames passed and were within the assigned tolerances for length height and width second is perpendicularity again two out of six of the frames had legs that were close enough to 90° of squareness to the frame and passed third is Twist zero out of six frames passed as all of them were Twisted far enough to exceed the tolerances and finally it was asked for a matched set to to be made and again zero out of six were able to have their frames match within the tolerance so unfortunately none of them passed just to confirm that the CMM is correct let's look at twist we know that all the tubes are bowed making them out of Tolerance but now we have a perfect plane let me show you what I'm talking about we're going back to video one remember these two pieces right here welded together make a perfect plane three points you weld this one together perfectly flat teeter totter or warp but what happens if you don't have a flat table is that this next joint this joint and this joint is where you get into trouble because if your table is crowned or dished or has any weird manipulation now you have four points one two 3 four and it's going to follow the surface of your table which means this is going to kick up these two points have now introduced Twist on this axis this is why having a flat table is very important and we're going to be able to show you that right here we're going to identify the corners but we're twisted right here and right here this is 1/16th of an inch this is what the plans call for you can have this much twist in the part so I'm going to push down on this corner creating those two flat surfaces remember and then this part's going to fly right underneath not only are these tubes bowed out of Speck the whole frame is twisted too and we can continue to test that on all of these and the results are almost exactly the [Music] same it's just wrong in so many ways okay so that's six frames that failed what about the other six frames that we did before that's 12 frames that are not made to size or flatness I do notice that nobody TIG welded these could that have ch change the outcome maybe I know they all Mig welded them for Speed so you can make money one of the things we talked about on the Forum is the tolerances on the drawing and unfortunately as Fabricators we don't get to decide that that is up to the customer all we have to do is try to build to their specifications when you look at a drawing can you hold tight tolerances I think being aware of what kind of tolerances you can hit as a fabricator is really important and that needs to be taken into consideration so what did we discover remember I added a lot of information in the drawings to improve the quality and I just did not see that in this instance I did not see this job taken more or less seriously by adding more complexity to the part I know one of your concerns was my testing methods on these parts but by adding the cmn machine I think I can confidently say that the measuring is correct once again we're left with more questions another six frames and testing that didn't pass the quality control so what's causing these frames to fail is a speed square and a tape measure like we saw in the video good enough for QC so what tools do you guys think these Fabricators need to build and quality control their parts all of these Fabricators told me that their plate steel tables are seriously warped how much of that is influencing the product that comes out of it what do you guys think about the weld Distortion in these parts is that something we as Fabricators should be paying more attention to I have lots of more questions for you we're going to put all these up at the fireball tool Forum I'd love to hear your feedback we have a lot more investigating to do to get to the root cause of failed parts thank you guys for watching and I'll talk to you at The [Music] Forum

Channel: Fireball Tool

Views: 776,734

Keywords: fireball, tool, Fireball tool, Welding

Id: 5SSUbxpCVZs

Length: 29min 51sec (1791 seconds)

Published: Sat Dec 09 2023