Finding the Right Fit with Inventor Tolerance Analysis [Lab]

PAUL MUNFORD: It's time, so I'm going to carry on. Good afternoon, everybody. How are you? It's Thursday afternoon, the last day of AU. So I'm pleased to see there's so many people here. I appreciate you turning up when you could be back up in your room with your feet up enjoying a little sleep after a big lunch. Let's try and keep some focus. So thank you very much for coming to this class. How many people came to the tolerance analysis class we did this morning? So a few of you've had an explanation of what tolerance analysis is.

So we're going to kick the tires on this and just see how it works and have some fun playing with it. Just to remind everybody, this is what we turned up for. So these are our class objectives. Let's just make sure nobody is in the wrong room. I always say this-- there are so many good classes that you never want to be in a class that is not what you expected. So if you need to go somewhere else, it's absolutely fine. But that's what we're going to do today.

I've got a few people to thank today. We've got a number of lab assistants. Stephen Wurst, from SIGMETRICS. So SIGMETRICS are the company that supply the technology that actually drives 3D annotation, NBD, GD&T, and Tolerance Analysis inside Inventor. So if we have any specific questions on the application of Tolerance Analysis, Stephen is the right man.

We've then got Steven Gao and Steve Dennis. Those guys are part of our Inventor team. So they actually write Inventor. So if you have any questions about how tolerance analysis interacts with Inventor, they're good people to ask. They're there to help you out. We have Steven de Strijker-- Peter de Strijker-- from our technical team, from Autodesk. And then we have Carlos, who's from the technical team at CAD in Australia. Those guys are in consulting, so they have a lot of experience with actually going into companies and helping them change their workflows, to take on board new features like this.

So this is the way we're going to run things today. I'm going to give another short explanation of what tolerance analysis is, just to level-set and make sure everybody knows what we're here to do. Then, we have five exercises to complete today. At the beginning of each exercise, I'll do a demo. So please watch the demo. Don't try to follow along. Because otherwise when you look at the screen and look down, you'll miss some stuff. And then you'll look up, and then you'll get lost. So you can just watch the demo.

When we're finished, I've got some timings here. So it's my job to keep an eye on those. So you'll get like five minutes to actually have a go. The PDF handout has everything step by step by step, so you should be able to follow along. Steve and Stephen did find a couple of indiscrepancies, so don't worry if things don't work out exactly correct, but I'm sure we'll get there.

And at the end of the five minutes, I'll say, OK, it's time to start. We're going to take a look at the next part. If you're not finished, it's fine. You can just close down the data set. I've saved the data set at each stage. And you can just go to the next version, open that one up and catch up. So if you have any questions as we go along, I might be able to answer them in those five minutes when we're working.

But if you need any help, the gentlemen in the back-- just put your hand up and one of them will come over and help you get back up to speed. If you're finding your moving along faster than the class, I've put a few extra-credit exercises in the PDF. So you can scroll down and see what else I've given you to do, and kind of do those things, too. Is that OK with everybody? Make sense?

OK, let's just do a quick level-set. So what are we here for? We're here to answer the question, "Will the parts that make up the assembly always fit together?" That's what tolerance analysis is always all about. That's what we're interested in.

How are we doing it right now? Potentially, we're doing it with an Excel spreadsheet. Maybe, we're even doing it by hand. Just wait till you get the print and then kind of figure it out. Maybe you're doing in your head. Maybe you're not doing it. Maybe just throwing it over the wall to manufacturing and just hope. But the opportunity here is to reduce cost. So every time we increase the tolerances on our components, we're increasing the cost. And in this example here, if you look at the surface-finish time, doubling the accuracy of the surface finish doubles the cost of the component. So we're trying to find that balance between quality, and getting the quality we need, without it costing more than it needs to. And if we can do that in Inventor before we pass it on to manufacturing, everybody is going to very happy.

So here's tolerance analysis. And I will skip past that one. So this is a really, really simple example, I just put together to kind of explain to anybody who's not familiar with the concept. On the left-hand side here, we've got our nominal. This is our CAD model. It's perfect. Everything fits together the way it should do. That's what we're expecting to get. It's not what we're going to get. In the middle, we've got our minimum-material condition. So in the case of these two blue box, the minimum-material condition, it means they're smaller. But in the case of the orange block, minimum means the gap is actually bigger.

So the combination of these things mean that instead of we get a half a millimeter gap, we're getting a 2-millimeter gap. That's our minimum-material condition. The maximum-material condition, where the two blue blocks manufacture their biggest tolerance and the orange block manufacturer the smallest, we're getting a clash. We're actually getting a half-a-millimeter clash. They're not going to go together. So this is just a brief explanation of what we mean when we say we're doing a stack-up analysis.

So when we're doing stack-ups, we're really just taking the tolerances we've applied to a set of dimensions and comparing them to the tolerances we applied to a second set of dimensions, and seeing the effect that has on some gaps on some dimension we're analyzing. The two types of tolerance analysis we're going to do today, we're going to look at worst case, and we're going to look at RSS. Worst-case scenario is the standard one. This is the one you probably all do. This is one you can probably do in your head, if it's just a few dimensions. And it's just comparing the two. But as we've put here, it's testing the limits of acceptability-- saying what's acceptable. But it's not saying what's likely, what's really going to happen.

So I put in here a quick explanation of the RSS. This is a statistical approach. Statistically, what will happen is that in a worst-case scenario, imagine that everything's manufactured to the worst case, or the minimum-material condition, or the maximum-material condition. And we're taking all those components, and we're trying to assemble them. But in real life that's not going to happen. In real life you've got a bin of components, some at the worst-case one end, some at the other end. And you're taking all the components from all the different bins, and you get a completely random selection.

So what's the probability that you're going to get the worst-case scenario? It's probably not going to happen, right? If you've got more than a few dimensions, it's probably not going to happen. But to illustrate this, I did this really simple bar graph. On the left-hand side, I've got the minimum-material condition-- so everything manufactured to the smallest possible tolerance allowed. And on the right-hand side, I've got the maximum-material condition, with everything manufactured to the biggest tolerance allowed.

On the top there, with these two red bars, I'm indicating everything that's acceptable-- everything that's within the tolerance limit. And you can see there it's most of them. So what's the likelihood that it's not going to work? It's really unlikely, because there's the two combinations that don't work, and all the rest of them do. And that's only three components. So we can see that RSS can be a really useful way of analyzing our stack-ups.

So when we do the RSS, we'll see this kind of Gaussian bell curve. In the middle we've got the average-- the nominal that we're expecting. Then every time the machine is machining a new component, and maybe there's a little bit of wear or some heat or some change, we're getting ones that are not acceptable. So the ones in the middle are within tolerance, and then the ones on the outside are the rejects. And at that point the machine-- it's a stop machine-- do whatever they need to do to recalibrate, and off they go again.

Now, one thing about the RSS, or the statistical approach, is it assumes that all parts have the same quality. Now, we do have a couple of methods within tolerance analysis to look at quality metrics. So if you're following something like Six Sigma, then we have the opportunity to change the quality value. This is a parameter called CPK. And really, in my head, the way I kind of understand this is that if we're doing a standard RSS plot, we're talking about predicting what's going to happen in the future.

So we're making some assumptions about how things are going to work. If we're doing one of these plots, we have the opportunity to feedback actual data. So it could be data from a supplier-- what kind of quality they're actually achieving. Or it could be from our own shop floor-- we can put those metrics back in here. So we're not going to cover this today. But I just wanted to let you know it's available in the software. This is the kind of thing you probably need to do with some consultancy and figure it all out.

Good. Does that make sense to everybody so far? Some nods-- good.

So a couple of assumptions we're going to make-- accuracy-- no part is ever accurate. It's just a question of how well can you measure it. But if we can measure it well enough, we'll always find an inaccuracy. Variation of size can lead to poor performance-- in our case, things not fitting. We can use tolerances to control the variation and which variation is acceptable. And tolerance analysis allows us to balance between the performance or the quality and cost. Everybody up to speed? Cool.

Now, in the handout, I've also put some information about how you get hold of tolerance analysis. So just a prerequisite-- how many people are using Autodesk Inventor 2020? Excellent. How many of you on 2019? So you need 2019.1, otherwise you won't get be able to install tolerance analysis. Anybody older than 2019? So you guys are going to have to catch up.

[LAUGHTER]

If you'd like to download a trial, you can go up here-- Autodesk.com Tolerance Analysis. Click on Download A Trial. You can download a trial. As long as you meet the prerequisite, you can download to try and install it. You get 30 days to play with it and see what you think. If you are a PDMC customer-- so how many people are on a subscription? So you will get it for free-- part of your subscription.

But when you go into the subscription portal to find your software, you will not find it in the first list. So just a tip-- what you want to do is you want to come down this list. And you say [INAUDIBLE] lucky. I might get all the software to play with. And you want to find the product design and manufacturing collection. Click right here on the Find What I'm Allocated. And then you can go down and grab Tolerance Analysis and download it. It is a separate install, because it's subscription. So what that means is that you will get a different serial number with it. So you need to make sure you get that serial number. When you first start Inventor up, it will ask you to either log-in to your account, depending on whether you're using your Autodesk sign-in, or you can put your subscription in, if you're on a multi-seat license.

And you'll know if you've got it installed, because you'll find it in the Environments tab. And again, this is all in the handouts. If you want to take that away and have a look, you can do that. So we opened up the PDF handout on your machine for you today, so you can follow along. If you've got a thumb drive, you want to take a copy, feel free. Take the handout in the data set. All of these will be on AU Online. About two weeks after AU finishes, they upload everything, including, I believe, a recording of the class, so you can go back and watch the video.

I have made screencast recordings of each demo. So if you've got the handout, you can click on the recording, or take the screencast and you can watch the recording. So if you want to take this back to the office and have a go when it's quiet and you can concentrate, and you're not so sleepy, then do that. And if you want to pass it around to your colleagues and see what they think, please do that as well. Good. Any questions so far? I just see lots of people going [INAUDIBLE].

[LAUGHTER]

Right. Let me jump over to Inventor, and we'll do the first demo. You don't have to do anything yet. Just watch the demo, otherwise you might miss something. So you should all have tolerance analysis set as your project file. If you haven't, wave your hand and get one of the lab assistance to come help you.

So when we go up here to the open dialogue, we should find that we go straight to the workspace for this project-- Inventor Tolerance Analysis. And you can see in here we've got a folder for each exercise. And there's a quick link on the left-hand side, that you can use to navigate the exercises as well. So in each case, we'll just go to the folder for the exercise. Do close the assembly you're working on before you open the next one. That's just to do because I did a pack-and-go on each one. So sometimes the tolerance analyses interfere with each other.

So I'm going to pick the next size one. I'm going to pick the shaft assembly. And I'm going to open this one up. We're just going to check that we've got tolerance analysis. Go into Environments. I'll find Tolerance Analysis right here. And so I'll select it. So we can tell that we're in the Tolerance Analysis environment, because we have this blue tab. We have the familiar green tic, that says we can finish the analysis. We have a series of standard tools for analysis. And then we have the analysis panel over here. And we'll see what they do at the moment.

So the two things that we need to find before we start this is we need to find the gap-- the gap that we're going to analyze. And then we need to find the loop-- the loop of components that are going to take part in this analysis. So I'm going to show you that the gap that we're going to analyze is the gap right here at the front, between this washer and this big, purple wheel-thing here. So what I'll do is I'll come up to New Stack Up. I'll select New Stack Up. And it's important that we pick these in the right order. So what I'll need you to do is pick the front face of the purple wheel and then the back face of this golden-colored washer. And that selected our gap.

We can see here now Tolerance Analysis is asking us for an annotation plane. And where should I put this annotation? Well, I'm going to expand out my origin folder. And let's go with the yz plane. Yeah, the yz plane. So I'll just select that, place my gap annotation. And what I'll see is that Tolerance Analysis is analyzing the constraints I've used to put this assembly together, to find the loop.

So we'll see here it says, one path has been found. And if we pick, on the dropdown with a left-click, we can see what the path is. So it's basically analyzing these make constraints. It's saying, oh, that must be the loop you want to analyze. In this case it is. We're all good. So we're just going to left-click. And to finish we'll click on the big green button that says, OK, and we're done. That's the first thing we're going to do. Everybody happy?

AUDIENCE: Except for nobody's going to be able to do it because it's requiring them to log-on.

PAUL MUNFORD: Everybody?

AUDIENCE: I [INAUDIBLE] not letting everybody [INAUDIBLE]. I know where he lives.

[LAUGHTER]

PAUL MUNFORD: Right. Does anybody know of any jokes?

AUDIENCE: [INAUDIBLE] He did stick his head in earlier. [INAUDIBLE]

PAUL MUNFORD: OK, so while we're waiting for the IT guys to come and log you in, I guess probably the best thing you can do is go into the PDF. Maybe click on one of those links and see if you can watch the video play. And we'll come back to it in a moment.

AUDIENCE: Where is it? Do you know?

PAUL MUNFORD: Thank you very much, everybody, for your patience. Put your hand up if you can't find the log-in.

AUDIENCE: [INAUDIBLE]

PAUL MUNFORD: So everybody is logged-in now? Hey. Thank you very much for your patience, everybody. Woo. So we're a little bit out of time. We might have to rush through. But go ahead and do that exercise. Does anybody remember what we're doing? You were going to start up Tolerance Analysis. Let me change screens again. Start up Tolerance Analysis Create New Stack Up. Click the face of the purple object first, and then the back face of the golden-colored washer. Find your origin plane to place the annotation. Once you've placed the annotation, you should get something that looks like that.

If anybody's managed to do that already, one of the extra-credit exercises was just to do a bit of tidying. So you can do that by hovering over the annotation. I found it actually works, but if you hover over the leader line. And then you can click and drag these annotations around, position them in a way that looks a bit more OCD, if you're like me. Remember, this isn't a drawing, so it's not as critical that we place these dimensions in a nice place. But I do like things to look tidy.

AUDIENCE: If I click New Stack Up.

AUDIENCE: There's no MTB. It's like the command failed.

PAUL MUNFORD: You're also [INAUDIBLE] panel. So why don't you click Exit?

AUDIENCE: [INAUDIBLE] Of Inventor and try again?

AUDIENCE: Yeah. Let's try that.

AUDIENCE: Because you ran that back.

PAUL MUNFORD: So how many people have successfully managed to do that step? OK, that's over half. So I'm going to go on and choose the next step. So I apologize if I'm leaving anybody behind, but I'm feeling a bit under pressure to kind of cover a bit more. OK, now I'm going to close this file. So just down at the bottom, I'm just going to click on the little crossing-bullet button on the tab and close this file down. I won't bother saving any changes.

And I'm going to come back up and choose Open. And I'm going to go to Exercise Number Two and open up the file called Shaft 02. I guess you could continue this on for number one. The only difference here is that I've shown some of the 3D annotation for this component. And what I want you to take note of is that when we adjust values in Tolerance Analysis, it's going to adjust the values in the 3D annotation at same time.

So let me come out to the Environments, Tolerance Analysis. So we can see the stack-up has already been created here. Now, I'm going to draw your attention over here on the panel. So the panel is where we see all the dimensions that are forming part of the stack-up.

And if you take a look, some of these have a little link symbol next to them. That link symbol means that the tolerance we're seeing in the panel is linked to the 3D annotation in the file. You don't have to apply 3D annotation to use Tolerance Analysis. It will just apply a tolerance anyway, and you can adjust it from here. But if you have applied 3D annotation, you could even maybe use that 3D annotation-- you can put it through into your 2D prints. And it just keeps the whole thing coordinated, so you don't have to get out of step.

Now what we're going to do on this exercise is we're going to scroll down to the bottom. We're going to take a look. We currently have an objective of greater than 0. We can stick with that for the time being. So we can see that if we stack up all the tolerances on this component, we're going to have that problem. We're going to have that problem, that in the minimum [INAUDIBLE] condition things are going to clash. We can see it's red on the graph here. So that's an issue. So we want to adjust some tolerances, but which tolerances do we adjust? Well, to find out, we can come down here to the Contributors tab.

So if we select Contributors, we can see the biggest contributor is the bearing. And that's not surprising, because there's two bearings in this file. So what we want to do is we want to go and adjust those tolerances. So I'll come back to the results. And I'll use my scroll bar here to scroll up and find the bearings. In the bearing, I can see that's 0.01 of an inch. So let me see if I can adjust that-- maybe 0.001 of an inch.

And as we adjust the tolerances here, I'm expecting you to see two things. The first thing is we'll see the change in the graph at the bottom, thus allowing us to see the effect of that tolerance on the stack-up. The second thing is that we should see the change-- somewhere down here in the bearings, we should see the change to the tolerances happening at the same time. So that's the next step. So you can all have a go at that now. Any questions? Does that make sense to everybody? Have you all done it already? I'll give you a few minutes.

I think a good one to see the dimensions updating is the spacer in the middle. So I'm currently looking at the spacer, which is the dimension down here. It's just highlighted in blue currently. So this dimension here, if I change that dimension, you should definitely see that update in the model. By a show of hands, how many people have finished their exercise? Yeah, most people. Good, thank you. So I'm going to move on.

So once again, I'm going to close this data set down and go to the next one. So I'm just going to close Shaft 02. I'm not going to worry about saving it right now. I'm going to go back to my Open and go to the Exercise 3 folder and open up Shaft 03. So this has already been generated. We don't have to worry about the tolerance analysis here. But the purpose of this exercise is to generate a report.

So if we've worked hard to make sure the tolerance is all correct on this component, we really want to share that information and show people we've done our due diligence. But we also want to look for opportunities to work with manufacturing and production, to ensure that we haven't missed anything. So we want them to see our thinking in what we did here, so that they can make their contribution, and we can figure out the most-efficient method for our company to manufacture this component. So we want to be able to create a report, so that other people can see what we've been doing.

Now a couple of things before we create our report. I'm just going to ask you to go to the top of the stack-up details. There's a little Chevron arrow there. And if we click on it, we see this summary report of all the stack-ups that are in our file. This is a good place for us to rename. So I'm just going to left-click where it says, Name. And we can call this Shaft Stack Up.

Once I've renamed, I'll expand this. And to make it really easy to read, we can go through and start renaming some of these attributes. I'll just rename a few, just to demonstrate. So I can say this is the bearings. So I'm going to call this Bearing Face. And Dimension Number Five isn't too descriptive, so I will call it Bearing Width. And then, Face 11, so it could be Bearing Rear. So I'd like to spend perhaps 30 seconds or a minute just running through some of these and making some changes to the names, to something a little bit more descriptive, a little bit more useful. So I'll give you a little bit of time to do that.

So we've taken a bit of time to rename things, to make them a bit more user-friendly. So the next thing we're going to do is we're going to take a snapshot. So up here in the Report Panel, there's a button called Take Snapshot. This is a little bit like a Save View. So I'll just click here once. We didn't see anything happen. But I can show you that it has taken a snapshot by reorienting my model-- and clicking on the button that says, Show Snapshot, and it will just re-orientate it back to the snapshot we created.

So we know we've captured that snapshot. And the reason that's important, when we click on the button here on the Report Panel that says, Generate Report, we'll just click. We'll give it a name. I'm going to call it Paul's Report. And it will generate this report as a HTML file. And it should open up your browser. I'm hoping each of these machines has got a browser. And we see that saved snapshot view we created is the first page in the report. And if we look down the report, it will give us the summary, that we saw in the Summary tab. It will give us the dimensions, that we allocated and their tolerances, and it will give us the results of the report.

So this is now available for us to save somewhere on a server, point everybody else to it, and say, hey, check out this stuff we did-- is that going to work for everybody? Does that all make sense? Did you all Generate Report already while I was talking? Most people? [INAUDIBLE]. Good. Any questions? Great.

OK. So what we've seen so far, we've seen how to take a component that already has constraints and already has 3D annotation applied to it and run a stack-up analysis. We're now going to look at an assembly that doesn't have any constraints. It's the same assembly. I've just deleted them. And we'll see how to do it manually. Now, you might want to watch this, because it's going to take me a few clicks to demonstrate. So I'll go back to get started. I'm going to click on Open. I'm going to go to Exercise 04, select my shaft and open it up.

So if I take a look in my Relationships folder, there are no relationships, because I've deleted all the constraints. So when we go into Environments and Tolerance Analysis, we'll click New Stack Up. And we'll choose the same two faces as before-- the front face of this purple bearing and the back face of the golden-colored circuit-- which use the same annotation plane-- the xz plane. And we can place our gap annotation. But this time nothing happens, right? Because there are no constraints to analyze. Tolerance Analysis doesn't know what should be included in that loop of components, so we're going to tell it. It has made the bearing translucent, because it knows that that component is going to be part of the stack-up, because we picked a face on it. But it is now asking us, which other components do we want to include in the loop? And we need to pick them in the correct order. We need to pick them in the order of the loop.

So I'm going to pick the spacer that goes in between the two bearings. Then, I'm going to pick the second bearing. Then, I'm going to pick the shaft, and then I'm going to pick the circuit. And watch what happens. When I click the circuit, all the other components disappear. So Tolerance Analysis is saying, yea, thanks very much, now I understand what components are in the loop. But hey, how did these components relate to each other?

Now you need to tell me which faces we're going to analyze. So this time I need to go through-- I hope you can make it out on the screen. I've got the bearing. The spacer is there. It's just been made translucent. So I'm going to tell Tolerance Analysis how I want this face to be analyzed against this face on the spacer-- this face on the spacer against the back face on the bearing. The back face on the bearing is going to push up against this shoulder here on the shaft. And then finally-- I'm going to zoom in a little bit here-- this tiny little shoulder here is going to be associated to the face of the circuit.

And now we see exactly the same thing we saw before. We see these green faces, to show the loop that we're analyzing. And we can click on the big green tic button, and we have tolerance stack-up. Does that make Sense does anybody want to see me do that again? We were going to have a go. You're all going to have a go-- awesome. Off you go. I'll give you a couple minutes.

So the real benefit here is that, for example, we could bring in a STEP file, that's got no relationships. And we could analyze that using Tolerance Analysis. We could use AnyCAD to bring in a file from any other software that we've got available to us. And we can do stack-up analysis on that. Or it just may be that the way you've constrained your assembly, Tolerance Analysis can't understand how you want to loop those components-- and so you need to tell it manually.

I think we need to move on to last exercise. Now, the last exercise, it's not really one exercise. It's really a choice of two. So this is an opportunity for you to choose one of the two assemblies that we have to work with. And you can choose what kind of analysis you'd like to do. So you're going to kind of test this out for yourselves. I've put some suggestions in the handout for what you could analyze. But you don't have to. You can pick something else.

Now, I've got two examples here. One is an adventure assembly. It already has constraints and 3D annotation. The other one's a STEP file. Who would like to see the STEP file? Who would like to see adventure assembly? I'm going to show both, but I'll show the STEP file first. I think we have more hands for that one. Now at this point, feel free to work through at your own pace. You've got instructions in the handout. You've got the files. Open them up, have a go. If you're not feeling particularly confident, just watch the screen. I'll do a demo. And you can watch that and then have a go.

So I'll start with the STEP file. So here it is. So if you do need to change your file type to make sure you can see it, I've just changed it to All Files. And it's the file here called Two Speed Gearbox.stp. So I'll click Open. So you see, as we're opening this, we're actually translating it from STEP into an Inventor file.

Who's familiar with AnyCAD? You're not familiar with AnyCAD? You need to see Mike's class this morning on AnyCAD.

So we can choose here to convert the model, or we can choose reference. I think I'll choose reference, which is going to link the model in, and choose OK. So we're now bringing this STEP file into Inventor. So this file is not going to have any 3D annotations. It's not going to have any relationships, because it's STEP.

So you can see here that this STEP file is now linked. So any changes made to the STEP file will be reflected inside Inventor. Now we're going to analyze the shaft that goes through the middle here, to make it a little bit easier to see. And I'm going to put a view through there. So I'm going to come up here to Views. I'm going to pick from this dropdown Half Section View. And I'll do a half-section view on the xy plane. And that happens to go right through the middle of the shaft. So I click on the big green tic button and say, OK. So now we can see what's going on inside this assembly.

So before we can do analysis, we have to have made a decision what gap we're going to analyze and which components are going to become part of that loop. So I'm just going to change to face selection. The gap we're going to analyze is the gap that's on the front face of this cog. And we're going to analyze the gap between the housing that everything's contained in.

So what components are going to be part of the loop? Well, we've got the cog here. That forms relationship with the shaft. The shaft forms a relationship at this end with, I think, this bearing in here, which forms the relationship with the bearing here. And then the circlip, I think actually goes around the housing. There's a gasket in here. And then back to the housing that we're analyzing the gap between. So once we've figured that out, we've got the components we can choose for our loop.

So let's go into Tolerance Analysis. Environments, Begin Tolerance Analysis, Create A New Stack Up. I'm going to zoom in here and choose the front face of that cog, and then this face right here on the housing. We can choose an annotation plane. Again, I'll go for the xy plane.

So there's no relationships here, so Tolerance Analysis can't help us. We have to do this manually. So I'll pick the shaft. What does the shaft connect to? The shaft connects to-- I think it pushes up against the bearing, which pushes up against this bearing and the circlip, the housing, the gasket, the face. And there we are. So we've now got all the components that belong to this loop. So now we have to say which faces go together. So I'm going to zoom right in here. And we've got the face here on this shaft goes to the face on this shaft. We've got the bearing here. Oh, I think I messed up. Let me just try that again.

There we go. So the face of the cog in the shaft and the face of this little shoulder and the circlip. The circlip pushes up against the bearing. The bearing pushes up against this roller bearing here. The flap pushes up against the housing. Working my way around, the housing pushes up against the gasket. The gasket pushes up against the face, and eventually we get our loop. So we can [INAUDIBLE] big green tic button. We have our tolerance analysis. And then we can start adjusting dimensions. So that's one option. Does that kind of make sense? I'm sorry, I kind of messed up a bit there. I have put the picks and clicks in the handout, if you want to follow through.

So if you're interested in that one, by all means, have a go. I'll now show the other one that you could do as an exercise, and you can see if you prefer that one. So I'll close this assembly down, go back to Get Started, Open. This time I'm going to go into the folder called Gear Reducer, and I'm going to open up this file here called Gear Reducer. It's a native Inventor assembly.

Now, this time we do have relationships in here predefined. I'm actually going to start by going to this High Speed Shaft. So I'm going to choose the Slow Speed Shaft view-- which again, I've left some annotation turned on, so we can see the effect of that. So which gap are we going to analyze? In this case, we're going to analyze the gap between the slow-speed spacer here and the retaining plate. So this will be a gap of 0. I'm going to imagine that maybe this material can crash a little bit, so we have an opportunity for some tolerance.

So if that pushes up against the spacer, the spacer pushes up against the housing. The housing pushes up against the spacer on this side, through the spacer, through the big cog in the middle, back to the spacer here again. So that's our loop. So Environments, Tolerance Analysis. And I'm going to come in here and choose New Stack Up. I'm going to be picking the front face of the spacer, the back face of the retaining plate. Choose my annotation plane.

And this time Tolerance Analysis has found us a path, because I already have constraints in here to make the path available. So I can just select there, press the big green tic button. And again, we get our tolerance analysis built, and we can start to play with these figures and try and figure out what the best stack-up is using worst-case RSS. So I can now give you about 10 minutes-- 15 minutes perhaps-- to try out one of those examples and have a play and try and figure it out for yourself. Again, there are further instructions in the handout, if you want to go back and review them. Everybody OK? Thank you.

So I was just saying that I think I'll run through the conclusions. Anybody who wants to stay can continue. So our learning objective for today was why does Tolerance Analysis cause problems in assembling components? We found out they can cause problems, because if we haven't done it, we can find when we get to the assembly stage, things don't fit, it costs us money.

How can the Inventor Tolerance Analysis help? So I hope you've seen enough that you can see Inventor Tolerance Analysis can help by allowing us to do that Tolerance Analysis stack-up on the models. And it means we can do it very early on. So we're not having to wait until we've completed doing drawing prints-- or even worse, wait until we put it into manufacture. We can do it as we design it. So as we're producing the model, we can do stack-ups. And if we go back and change the model, our stack-up is going to update. So we can start the [INAUDIBLE] analysis very early in the process, when it's going to be most effective.

We have had a quick look at analyzing a STEP file, so we can see that we can do things manually. We don't have to have constraints. We don't have to go 3D annotation. And we have produced a report, so we can share our findings with everybody else in our company. So I think we did OK-- a few hiccups. I think we covered everything.

[INAUDIBLE] when I do a Q&A, I always always ask you questions and see how much you took in. So you can tell me, so why is Tolerance Analysis important?

AUDIENCE: To find issues before building.

PAUL MUNFORD: To find issues before building. Well, thank you very much.

AUDIENCE: [INAUDIBLE]

PAUL MUNFORD: It can save us money. If we loosen up our tolerances, it's going to cost us less to make the components. But if we tighten them up too much, yeah, sure, we'll get the fit, but it's going to cost us money. Do we need relationships in three annotations for these Tolerance Analysis?

AUDIENCE: No.

PAUL MUNFORD: Nope. We don't have to have them. It's a nice-to-have. It's not a must-have. Worst case-- so right in the beginning, we talked about worst case versus the statistical methods. How can worst case lead to ovetolerancy?

AUDIENCE: [INAUDIBLE] tolerance to the absolute worse case [INAUDIBLE].

PAUL MUNFORD: Yeah, exactly. So tolerancing to the worst case, it assumes that all the components are out of tolerance by the same amount, and that's very unlikely to happen. Cool. So now you get to ask me questions. Please, if you have a minute, if you could take your phone out-- take the Autodesk app. There's a survey in there. I'd really appreciate your feedback. The marks are good, but the comments are even better, because they show me what I need to do to improve next time. So does anybody have any questions I've not managed to answer yet? Yes, sir?

AUDIENCE: [INAUDIBLE] a constraint [INAUDIBLE] but maybe whoever put the constraints together didn't exactly follow how they should have been [INAUDIBLE] those components together. Can you instead select the path of the loop that you want, even though it's already [INAUDIBLE] suggesting in the pull-down?

PAUL MUNFORD: Yeah. So the question is, if somebody is applying constraints to the assembly, and the constraints don't make sense to Tolerance Analysis, can we choose the manual method instead? And the answer is, yes, absolutely. So I'd say, actually, the biggest learning curve for me with Tolerance Analysis has been, how do we use the constraints to get Tolerance Analysis to do them automatically.

So there will definitely be occasions when you've constrained your assembly, in the way you usually do in Inventor assembly. But it doesn't make sense from a tolerance analysis point of view. So we will go to Environments. We'll got to Tolerance Analysis. We'll go to Stack Up. We pick the faces for our gap, pick somewhere to put our origin. And when it says, Path Found, if we don't want to use that path, we can come back up to the top and choose Select. And then we can go in and do the manual selection process, that we talked about earlier on. Is that OK? Cool. Any more questions? Yes, sir?

AUDIENCE: If you're going through and you get all the way through picking your [INAUDIBLE] the wrong surface, can I go back?

PAUL MUNFORD: So the question is, if I'm going through picking all the components that should be part of the loop and then picking the faces, and I mess up, can I go back? And the answer is, no. Sorry. At the moment it's cancel it. But if you've got some good ideas about how you'd like it to work, either speak to Steve or Stephen, or come down to the answer bar and give us some feedback.

AUDIENCE: Yeah. I mean, this could just [INAUDIBLE] forum. You could do an ideas [INAUDIBLE] we took what SIGMETRICS had done for this, brought it under the covers into Inventor-- and took care of stuff in licensing and install-- stuff like that-- and we still need some improvements to make. When I was going through it, I kept on pulling Mike over and saying, why is this working like this? This isn't like the rest of Inventor. Things like that, we have to fix.