Manufacturing Processes for Composite Tooling and Composite Trimming

JON CALIGURI: Thank you to everyone who turned up at 8:00 AM at AU. That's no small feat sometimes, so we do appreciate it.

[INTERPOSING VOICES]

AUDIENCE: Especially after a night [INAUDIBLE] this year. [INAUDIBLE]

JON CALIGURI: I know, I know. Yeah, it's a little bit of a-- it's a really terrible time. But a little bit of a roll of the dice to have the 8:00 AM slot because-- but, to me, I think the turnout's pretty good.

We kind of knew that we had a niche presentation being composites. So it's good. So we'll go ahead and get started here.

So, again, thanks everybody for coming. My name's Jon. Mike and I are from Design and Software. We are a former Delcam reseller and now the Autodesk DMG master reseller for North America.

And we've been working with these products since 1997 or 1998. And we've been doing a lot of work-- specifically, with composites-- with our co-speakers here, Ed Hilligrass and Patrick Bollar from DMS.

So DMS make five-axis machine tools for composite trimming and tooling, and a wide range of applications. Sir Patrick is the founder and the CTO, and Ed is the Executive VP. So thank you, guys, for joining us for this.

They have a wealth of knowledge in composites and we've done tons of work together in this field. So we wanted to put this presentation together to just sort of outline some of the best practices that the best composite manufacturers follow. And some things that we've developed for them, specifically.

So the agenda that we're going to go through. We're going to talk about how to speed up and more efficiently program and manufacture composite tooling.

We want to talk about a process that we call adaptive fixturing, which is going to show how we can eliminate manual parts setup on the machine tool, increase the accuracy and repeatability of the part setups, and reduce-- or, hopefully, eliminate-- scrap and rework.

So I'm not sure. How many people in this room currently work in the composites industry in some form or fashion? OK. Good. So because we have a little bit of a mixed audience, we'll go through a couple of just basics so that we're on the same page. And then we'll sort of get started.

So composite material-- obviously, you think carbon fiber. That's really the one that we're going to focus on here. But there's really many, many different types of composite materials.

It's not really anything new. But certainly carbon fiber, specifically, is making its way into more and more industries. It's not just what we typically, think-- aerospace. But all of these are composite materials.

So why are composite materials used? So, obviously, their physical characteristics-- a lot stronger than their metal counterparts in some ways. There's a lot of design flexibility. You can make a really wild shape.

And that really leads into the problems that people face as they try to manufacture composites. So because of the nature of these materials, simple operations become a lot more complex.

Just drilling a hole, for example. There's a lot that goes into manufacturing, milling, and drilling these materials that makes you-- you to have to think about a lot more aspect of everything else that surrounds that process.

So the right cutting tools, the right machine, machines that are designed to manufacture composites. And we'll get into a lot of these problems-- delamination and part location.

So because composite parts can take a very complex shape-- and many times they are organic shapes, but free-form surfaces. And that does present a problem when trying to fixture, hold, and locate that part for machine processes on a machine tool.

So like that vacuum fixture that we have showing up there is an example of one that's-- that part is-- it's a challenge. If you've never manufactured composites and your job is OK, fixture that on a machine tool. It's not so straight forward.

So on the tooling side of things, we want to talk about a couple key points. When it comes to manufacturing molds-- especially in the industries that we find composites in-- one tool that we utilize is the ability to mirror left hand and right hand parts.

So if you just have a lot of time in programming something for-- see what you know. Again, we think aerospace, automotive, a lot of these things are left hand, right hand.

So we can spend time programming one mold and completely mirror that project to save up to 50% of our programming time. We can also use macro and template-based programming to really automate the programming of the molds.

I would say the main focus of what I want to present about this is as companies go into manufacturing composites, it is-- a lot of times, we get involved when it's maybe their-- the first machine that's going to do this.

And the key bits of functionality that I think to focus on for evaluating is if your CAM tool is the right choice for doing work like this-- is dealing with large data sets, being able to automate that programming, and being able to simulate on the machine tool.

So if you cannot know what your program is going to do until you load it on the machine tool and you have to run it in the air. Or you have to use some third-party validation because you don't know-- again, we're talking about five-axis machines-- very, very commonly.

So if you don't know what that rotation or what the axial movement is going to be ahead of time-- and as you're programming-- you can waste hours and hours and hours of dry running a program or being very inefficient.

The operator is going to be nervous if he's uncertain if something's going to collide or something like this. So a tool like PowerMill from Autodesk desk has really, really robust machine simulation. So you don't need to rely on something like Vericut or some third-party validation.

And the critical bit is you can find problems or inefficiencies in the programming stage. So it's not a post process. So you have time to rapidly make decisions and correct things inside of the programming environment before you send code to the machine. So just a couple of key points.

So this gets into the adaptive fixturing. So eliminate manual part setup on the machine tool. And what we really mean by that is a complicated process by which you indicate and locate, accurately, a complex shape on the machine tool.

So we've encountered customers with very long setup times-- really elaborate fixtures with tooling balls all over. Because the problem is if the composite part is not exactly lined up correctly and you go to tram or mill on it-- and if it's nonconforming-- that part-- at the time when you're putting a tool to it-- has a lot of time in it. A lot of value. So it's a high-value part.

So it's very critical that the setup is done accurately. Now, most machine tools-- especially that you would use for an operation like this-- have in-spindle probes.

But one of the problems that we find is the way that the probes are typically used is very, very rudimentary. And it's actually a really capable tool if you have the software that can appropriately deal with that data.

So what you do in this process of software fixturing is effectively do a rough alignment. So have a simple fixture that the part sits in rigidly, and that we can locate-- to some degree of accuracy-- really quickly.

So once we have it roughly aligned, the probe-- so we'll define some free-form surface points and we're going to show you that bit in just a minute.

And the probe will come in and touch some points on the part. And it will actually offset and do sort of a fine alignment to your work offset on the machine.

So you can now have someone really quickly loading parts on and just hit Go. And the probe will find the part and sort of set its G54 for you. The critical part about that is that it will translate the work offset in x, y, and z. And it will rotate compound angles around a, b, and c.

So the reason why composites are difficult is it may be sort of out like this. And that's really-- no master machinist can look at the results of the probing and do the math in his head of how to fix that.

So this is where the setup times can be quite long. And they also take a really skilled machinist to set those up. And this is really a nice answer for that.

So we inspect the part, we find the location, and the machine adjusts itself, basically. So we do this in composites, obviously. Also, forgings and castings have similar problems. So any near net shape part, really-- this is a really nice process.

The other thing I would say is we're getting into a lot of hybrid additive processes with this. So hybrid additive machines that can additively produce the mold-- the thing gets finish machined. So these processes are just inaccurate by nature.

Or if you're taking parts off of a powder bed metal printer maybe-- that that's your mold. Aligning that is a bit difficult because it's not exactly matching to the CAD model.

So the main benefit of using this technology is drastically reducing setup times. So you can begin to use much, much simpler fixtures and it doesn't take as skilled of a machinist to set this part up.

So if I had to say of all the industries that we deal with-- all the customers that we have-- the main problem that they all have is finding skilled labor. And it's a problem that's probably going to get worse before it gets better. So this is really a key part of things.

So using this process, another benefit is that the accuracy and the repeatability of the parts setup goes way, way up. Right? So any time we can eliminate a human intervention, we're better off.

So you still need a human to change from part to part, but the critical bit of making sure that that part is accurately aligned is done by its software. So it's more accurate and it's also traceable.

So if you have a non-conforming part, you can go back. We can actually produce what kind of equates to an inspection report of how far the software adjusted that part to align it for machining.

So the other condition that we see is that from the same part but-- so multiple parts do not sit in the same fixture exactly the same.

So if you have composite parts that do not repeat in their fixture very well, this is a great solution because we're aligning each part individually. So the work offset-- the G-code is done. We're not re-posting any G-code. The work offset is being adjusted for each individual part.

So drastically reducing or eliminating scrap. So scrap, obviously, is always a problem. In composites it's a problem that we still see quite a bit of. So even this year, a large aerospace account had a specific part that our customer started producing for them because their scrap rate was over 50%, which is a huge amount of cost waste.

So with PowerInspect, aligning the parts correctly every time-- first of all, if the part is significantly skewed or wrong and a good part cannot be machine from it, you would identify it in the initial setup process when we probe a part.

You can also use the probing software and process that we're talking about to actually do a quality inspection on the machine tool. So a lot of composite parts that we deal with are really large. So you could fit one in this room if you're lucky.

So there are not CMM devices that large. So then you rely on things like maybe Leica trackers or FARO trackers. Or handheld CMM devices with a person who has to move around the part to inspect them.

And that's fine. But if you are able to inspect that on the machine tool, it saves an entire process. And on smaller parts where a CMM process is still necessary, I still recommend using the on-machine probing. Because it does not take much time and you can ensure that the part will pass quality.

So the composite part-- if it is out of tolerance and you go to the CMM and they check it-- and it's incorrect-- now that part-- after you've machined it-- cannot be held in the same fixture the same way that it was before. Right? So rework is really difficult or impossible on a lot of these parts.

So if you identify that maybe something was not made to nominal or whatever-- the part was not made to conform-- as it's a fixture on the machine and you can see the report, now you can make a decision.

It's still in the machine space. We can go in and maybe adjust an offset or cutter comp or something, and rerun a trim pass or something like that. And we can solve those problems on the machine and we can still get a good part. So I think that's critical.

The other thing is that machine tools have come a long way. The accuracy of the machine tools that are typically used to manufacture these is really quite good.

So the DMS machines that we work with are accurate to, what, a thou or better-- a thou a 1/2? And they repeat to a half of a thou.

So with that accuracy and the probe technology-- and everything else-- there's really not a reason for a lot of applications not to do the inspection on the machine tool. Because that also allows customers to take the non-critical bits that-- they need to be inspected, but it doesn't necessarily have to be on a CMM-- and pull that work out of the CMM department.

So the bottleneck is, typically, at quality for most companies. So this allows you to let the machine tool that's producing the parts inspect what it can, and leave the CMM for maybe some parts that absolutely need it.

It also helps when you have a certain budget and you need a new CNC machine tool-- but to get the work for that machine tool it requires inspection, which requires another $200,000 or better investment in a CMM and software and things.

With a very, very, very small investment in software, the machine tool can do both jobs. And that helps you to spend money on what's going to make you money, which I think is important.

So here's a case study of a project that was done by Delcam a few years ago with BAE. And they even-- obviously, it's a very large company. They had the same problems that we see a lot of smaller composite manufacturers having today.

So the original process was sort of the traditional way that these will be made. So I'll show you the part in a second. So they would mount on the fixture-- hand adjust the location to make sure that things were aligned. And that takes an absolute ton of time and it's not repeatable.

They would machine the features, turn the part over, relocate by hand. And there were three fixtures that were involved to do all the machining that was involved in this part. So it was very vulnerable to scrap and, obviously, it's a high-value part.

So the adaptive machining process that we're talking about was implemented. Ooh, sorry. So there was a new, more simple vacuum fixture that was created. We can locate the part automatically using the probes.

The part, actually, sort of overhung the fixture so that only one setup was needed. They would do the best fit alignment to adjust G54 like we spoke about. And the results of that were dramatic.

So the total manufacturing time for that component was reduced by 60%. There is really no manual rework that was ever needed after that was implemented and the setup times were much, much faster.

So the other thing about using this process-- saving time is, obviously, great-- but you get a very predictable and repeatable setup process. So for scheduling this is really a godsend. Because if each part is reliably set up in the same amount of time and-- in manufacturing we know how much the machine time is going to cost, but the setup time can be variable.

I'm kind of throwing a lot at you guys quickly here and I'd like this to be interactive. So does anyone have any questions about anything that we've talked about? I can-- Yup.

AUDIENCE: So you were talking about Power Inspect, right?

JON CALIGURI: Mhm.

AUDIENCE: What controllers does that support?

JON CALIGURI: Realistically, any modern controller. So Siemens [INAUDIBLE], [INAUDIBLE], rotors. We've sort of done all of them.

So is there a specific one that you're?

AUDIENCE: Probably a Hynek one.

JON CALIGURI: Yeah. [INAUDIBLE], Siemens, and [INAUDIBLE] are the most common that we see in [INAUDIBLE]. And it's pretty seamless with all of those.

AUDIENCE: Do you do any initial calibration with the drills? [INAUDIBLE] and stuff. Cause 3D [INAUDIBLE] programming tools [INAUDIBLE] machinist.

JON CALIGURI: So you can use like the [INAUDIBLE] calibration routines. We do a 3D calibration on like a datum ball when we install this.

Mike's actually our tech who installs this stuff. So he could probably answer it better than I can. But the calibration on the machine tool can be done using the onboard cycles, or you can calibrate it inside power and spec.

So it's up to you.

PRESENTER: So [INAUDIBLE] expand, though, because I think that's a good question. A lot of machine [INAUDIBLE] come with probes in many cases. An example [INAUDIBLE] because they're the most common. Probably 80% of [INAUDIBLE] come looking for [INAUDIBLE].

Most of those are mechanical probes, which we can use support for [INAUDIBLE]. But [INAUDIBLE], especially if you [INAUDIBLE] 3D probe [INAUDIBLE].

So if you get a [INAUDIBLE] gauge probe, typically a Renishaw, or the higher end probe [INAUDIBLE] back where the gloom or [INAUDIBLE] or whoever it might be, if you get their higher end probes that support 3D probing, you're better off.

The probe that you recommend absolutely is the Renishaw [INAUDIBLE] gauge [INAUDIBLE]. So typically, in these applications, it would be like an [INAUDIBLE] 600. That's a radial probe. But it could be an LLV 600 just the same.

So with those probes, you'll likely have a problem. There's really not a big cost difference between the two. The street and retail price [INAUDIBLE] website, it's probably a couple of grand. It's probably the difference between $5,000 to $6,000, or $8,000 to $9,000, between the products difference.

The difference is the results are going to be far more accurate to [INAUDIBLE] as far as calibration and things like that. It eliminates a lot of those problems.

As far as the controller compatibility question, pretty much any control that has a [INAUDIBLE] print function, generally speaking, that's the first fundamental thing we need, which is what, once we take a measurement, we all [INAUDIBLE] right out and resolve what that measurement is. It's a stack of several measurements, [INAUDIBLE] what you want to measure.

But the controller has to come with you to push data back out of it. And that's the single biggest thing that needs. Beyond that, you start getting into individual controllers, and CRDs, and software levels of control. But that's the single biggest thing.

Most modern controls have that function on it.

AUDIENCE: And then on [INAUDIBLE] I don't think the [INAUDIBLE] stores 3D data calibration [INAUDIBLE].

PRESENTER: Yeah. Well, all of them do. The [INAUDIBLE] works pretty much the same, like [INAUDIBLE]. They all store it kind of the same way. So those are all really pretty standard.

So pretty much any of the [INAUDIBLE] controllers work, whether it's the next gen or the traditional one. The oldest machine probably on a [INAUDIBLE] was 8 or 9, 10 years old. So it goes back quite a ways.

If you have a machine that was much older than that, you'd really have to have that serious conversation about whether you want to implement probing on that machine anyways, because the old [INAUDIBLE] are less accurate than the new [INAUDIBLE]. And if it's been in service that long, it's probably lost a lot of accuracy over time.

AUDIENCE: Can I just go [INAUDIBLE] and get another coffee.

[LAUGHTER]

PRESENTER: [INAUDIBLE] legit.

JON CALIGURI: Yeah.

PRESENTER: But those are great questions. Cause those are actual concerns that [INAUDIBLE].

JON CALIGURI: Yeah. And also, we do not use any onboard probing cycles.

AUDIENCE: Just all [INAUDIBLE] plain, like [INAUDIBLE].

JON CALIGURI: It's basically just posted out like G-code.

So we do not use [INAUDIBLE] probing cycles, or any of these things. Renishaw macros, or any of that stuff. It's completely independent of all that.

So any other questions about what we've talked about? And then I want to get into some what to look for in a machine tool. And TMS have actually done some really clever things about creating machine tools that are designed to manufacture composites, dealing with carbon dust that can get into everything, and be really problematic to some ultrasonic technology and some other things.

PRESENTER: I have a couple questions before you do that.

JON CALIGURI: Yep.

PRESENTER: How many people here are doing parts for aerospace or have maybe controlled aerospace? [INAUDIBLE] or whatever the case may be.

OK. So a couple of them.

So the reason why I asked that is aerospace is always the one that we kind of put on the shelf a little bit, because everything's so controlled. But if we take a look at the control that's there, even if we're making parts for [INAUDIBLE] performance and cosmetics.

So we still kind of have the same problem. It just doesn't require all the paperwork. Which we lay out two or three parts. We can put them through the process to do that.

We ultimately [INAUDIBLE] trim. In many cases, what customers that implement this find is that before the part's even trimmed, it's actually scrap.

So in that meeting, because in the course [INAUDIBLE] lay out [INAUDIBLE] through the auto plate and all that, the parts were [INAUDIBLE]. So in many cases, the machine operators are everywhere from-- I'll be as polite as I possibly can-- incompetent to semi-competent.

But more importantly, they don't have an engineering background to know if that part just looks correct the way it's sitting on the fixture.

So in many cases, it's not their fault. They put it on there. They get vacuum. Vacuum cools it. It works a little bit differently. And now that it's sitting on the fixture in a position that's ready to be scrap, even though we haven't cut it yet.

JON CALIGURI: You're not mic'ed up.

PRESENTER: Oh, sorry.

So the biggest thing is is that with the OMV solution, you can go in and do a cursory inspection of the part, and validate that at least where it's sitting on the fixture before you trim it is, in fact, correct. Because in many cases, what has to happen is that you might have to release vacuum, put the part back on it, hit it with a mallet a couple of times. Then pull vacuum. Then hit a couple more times with a mallet.

Now it's actually sitting on the fixture the right way, cause the part's kind of relaxed and can form to the fixture. Then now if we do the adaptive alignment and trim it, now we get it successful parts.

So whether you're doing something like an aerospace eliminates the part being red tagged and have to do all the conformal checks to see what went wrong. And by the time it gets into the quality lab, now you have to start all the way back to the beginning and say, OK, where did it go wrong?

Well, now, if we put a stop gap and say, OK, we're going to put on a trim fixture before we trim it, and say, has everyone else done the job correctly up to this point? Because that way maybe this part can still be saved before it's then scrapped.

And whether or not it's aerospace, or just making consumer products, if you lay up three parts, you want to get three good parts out of it. Because to go back in and add in a fourth lay up just to get three parts, from a scheduling standpoint, materials, and all the waste that's involved, is pretty heavy. I think that's a big takeaway from all this as well.

JON CALIGURI: Yep. So I'm going to kind of ask Patrick some questions, because there are some things that I think that really bear mentioning.

But the reason why we worked with TMS so much on these is they're really sort of more innovative and agile than a lot of the maybe larger companies that make machine tools that we typically see for this. And they constantly are introducing new technology, redesigning machines to perform better, and to perform specific tasks in key markets like composites.

So one of the most interesting things to me now are these ultrasonic machines that they're producing. So maybe Patrick can elaborate about the benefits of a ultrasonic for composite manufacturing.

PATRICK: Good morning, everyone. Again, just a quick overview. TMS manufactures three and five axis machine tools. I've been doing it personally for the last 35 years.

Our forte is that we actually build custom machinery. So we'll build small machines as low as a two foot by two foot, but we can build a 20 foot by 120 foot with 12 foot of Z for the yacht making industry.

We have thousands and thousands of footprints throughout the world, cutting composites at all the big name customers you can think of.

This year we'll ship about 160 machines out into the world. And I'd say that at least 80 of those machines will be the cut and composite. And at least 60 of those will be cut in carbon fiber.

So this has allowed us to focus on the carbon fiber industry. I've made every type of machine for that industry. But we've focused in on what the customers are going to want, what they want in their plant, how they're going to load the machine, what type of dust [INAUDIBLE] collection they're going to need for this, and all the protection that come with carbon fiber, as far as boxes, and even the right type of wire. We're going through that. The air conditioner makes a difference on the box when you're in the carbon fiber environment.

So what you saw up there-- I'll get to ultrasonic in just a second. But you saw one of my fixtures up there. And I wanted to tell you about a study we did on that quick change fixture. We built it out of necessity for a customer.

We try to be the best solution provider to our customers as we can. So as CTO, I'm constantly looking to the customer. How do I save my customer money, make them more productive, and make them buy more machines?

So that little blue fixture off to the side was for a startup recreational aircraft company. They came to me with a task that says that they have 356 carbon fiber parts, and to do a complete chipset on that particular aircraft. And they wanted to do that in an eight hour period. They wanted to know how many machines it would take to cut that one chipset in eight hours.

So we actually did a study and called about 100 of our customers, and asked what the typical turnover is from job to job, their setup times and everything like that, to go to more to John's point. And the shortest one we found was about 20 minutes. And the longest one was about two hours.

So we knew that with that type of a time, there's no way we're going to make that chipset.

So this is a quick change fixture you're looking at right there. And we got that whole turn over time down to about 10 seconds, what that does.

That one button, that fixture comes off the top of that. One button that goes back on. The vacuum's hooked up. It's in position within a half 1/1000ths relocation. And there was a barcode system on there. So when you set it on there, it actually selected the program and loaded into the controller at that point.

So that came out of necessity. That's actually a two stage one. Actually bigger fixtures can go on the bottom alone. The top one is a fixture on a fixture for the smaller parts on there.

So that was a great project for us. And we learned a lot about on that particular job, how important it is for the containment. They have about six of our machines currently now. But they wanted more of a showroom condition inside of their plant. They wanted to push people through there.

So again, changing the design. The dust collection we put on the back of there. They were hoping for a shuttle table to shuttle it in and out.

But if you're running composites and you shuttle something outside of that work area, and you blow it off, you have composite dust all over your shop in about 10 minutes.

So everything was a dual cell machine. We call it a dual cell basically because you can actually load a part inside of the machine on one half of it while it's actually cutting on the other side for the high production.

So again, we spent a lot of time thinking about this. I could talk for days on this topic. But I'm going to get through this pretty quick. But I wanted to tell you about what I believe is going to happen from a tooling perspective inside of the composite industry.

So what I'm currently working on right now, as anybody who cuts composites all day, they know tooling is a big problem, a big cost on the machines. So tooling and blow out obviously on the parts is where it all starts.

So in my particular industry, you're not using flood coolant on a lot of these projects. They won't allow it. But through air coolant is going to be very important in the future.

The tooling's coming out now. So we have essentially the same process as through flood coolant. But this is through air coolant to keep those tools cooler while they're cutting.

So that's a special spindle to do it for air. Because obviously there's no liquid lubricating the seals in there and everything.

So there are spindles coming out now that we can do through air coolant to keep the tool cooler. We are doing a lot in ultrasonic right now. And we're not talking about ultrasonic knives. Although we have aggregates that we can put ultrasonic knives on there. But we're talking about ultrasonic tooling.

So again, it looks like an aggregate on a spindle. So it'll pick up a special bit. I mean, I'd have to say that 75% of the work done on these things is a 3/16 or a quarter inch bit actually machining the carbon fiber.

So this aggregate would have that most common bit. Or you could have a few aggregates obviously on the machine for tool change. But actually ultrasonic on the actual tool is making a big difference.

We see speeds up to sometimes 2 and 1/2 times faster than what typically we're doing on that. And the tooling is lasting a heck of a lot longer on that.

So those things on the ultrasonic are huge. The through air coolant, keeping it cool.

And, again, we do a lot of work in the Honeycomb world. We have a lot of tier one suppliers with our Honeycomb cutting machines. And those blades there with the through air coolant, very important. The abrasive blades blowing air on the tip of those things.

Heat's your enemy in this case. So anything we can do to blow some air on that is what is going to help the situation.

So if you guys have any questions. Again, I know I threw out a lot of stuff there. But I could talk for an hour on probably each one of those things that I just talked about. So that's it.

JON CALIGURI: So what are the other things that, just from a standard machine tool that's going to cut composites, do you guys kind of package as far as protecting the ways from the dust, and all of those things?

PATRICK: The machine is specifically designed for that. And as an example, one of the nastiest environments I've ever put a machine to is the silicon carbide industry.

So an example there is they put a Haas machine into that environment. It lasted six days. And it was scrap.

So we typically use an overhead gantry style. Because we don't want any of this carbon fiber falling on any type of a precision component, a motor, or a lead screw, a rail. It'll get inside the bellows. And it'll get down to it. And then it'll start to wear everything out.

So we like an overhead gantry style. We like it enclosed completely on top. We can do that with a full enclosure. We can do that with a bellow enclosure. And then it's a little bit costly, but we do mount [INAUDIBLE] right to the back of the machine. That's always the best case scenario. But that's pretty pricey.

But on every other component, as far as the rails are concerned, we're putting 5 ton seals on there, which help tremendously. So in this one company, GC is actually out in Oregon, but we have had machines out there for over five years now that we haven't touched. And it's lasted in that silicon carbide environment.

And if you've never worked in that environment, you've never seen silicon carbide get into aluminum pores to where you'd have to face off a table. And it just destroys your tooling in about two seconds. It's not a fun story.

So we do every time. Again, we are very reactive to our customers. We listen to them. We take surveys as much as we can. And any step.

And he's right. Like the biggest problem, how do I want a robot load all these parts? And because of what we're talking about, you can't get a robot to put a carbon fiber part on a fixture and go-- until it seats correctly. It just doesn't happen.

So I can pull it off there all day long. And I can get a robot to get it close. But either someone has to actually hit it. Or you have to have a system like what we're talking about today that sits there and probes it, and makes a little bit of difference.

Cause even some of these pictures, even if they're perfect fixtures, you can see it. Unless someone's drilled a hole in the mold, then they can actually set it on some pins, you cannot reliably put that in the same place every time.

So this is very important. The partnership we have with Autodesk on this, every customer is actually interested in this. We're looking on an offering that this is just absolutely included if you're getting a carbon fiber type machine.

PRESENTER: I was going to say, I think it's important, some of the bullet points that Patrick was talking about that they're offering, which is odd on user considerations that people don't use [INAUDIBLE].

So they [INAUDIBLE] machine [INAUDIBLE]. Got the accuracy or whatever, or whatever machine it is. They'll come back [INAUDIBLE] everywhere [INAUDIBLE]. And I keep forgetting about that.

But I think the take away with the stuff that Patrick said was, really these are the factors you really need to consider when picking out a machining tool. So even though the machine might have the accuracy, and it's got the range, and the machining envelope is adequate to do the trimming, if you think about the long term environment that that machine is in, and also the human impact, as far as the operator running that piece of equipment, these are all things that they've thought about. But these are things that really need to be considered, whether you're looking at buying a piece of equipment in the future, or just maybe self analyzing the piece of equipment that you have now with a more process centric piece of equipment that's dedicated for that process.

There's a lot of advantages by doing things like that. I know we have other customers that have attempted to do it with other kind of off the shelf general machines. And they're perpetually fighting maintenance issues, just because the carbon fiber gets in stuff. It starts grinding down components, because it's basically an abrasive.

Every time the table moves across something, it's essentially just grinding away, like sandpaper at it. And you're doing it over, and over, and over again throughout the day. Then on top of it, all that carbon fiber dust is electrostatic. So it starts sticking to things.

It actually conducts electricity. So it starts settling in on electrical components and shorting them out. And some machines and other sensors just start going haywire for no reason because all this composite dust got into them. And they know how to clear it off, and reset it, and reboot the machine. And then they're back up and running again. But if you start thinking about how much time they're wasting with all these things, if they actually had all these considerations when they made that initial decision, or just as their processes dictated, they could avoided a lot of that stuff.

So the big part of our thing is the adaptive alignment. But just in general, the composite thing needs a whole specific solution, looking at like yeah, I can do that on any machine. And technically you can. But there's a lot of other factors that most people don't think about.

I think that's just the important takeaway.

JON CALIGURI: So I do have sort of a technical demonstration of how this technology works.

Once PowerPoint gets a hold of the computer.

There we go.

So we'll just sort of talk through this. And it's a live video of using metrology software. It's not exactly the most thrilling thing. I understand. But this is kind of conceptual. And talking about it, if you've never seen the process, I think it does help to have a visual of sort of the steps. And I can sort of jump it through.

But this is Power Inspect. This is the software we're talking about for all of the adaptive alignment and inspection processes.

So you can load in CAD models from sort of anywhere, and [INAUDIBLE], GDNT information that's stored in the machine, or in the model file. So for all this, like the model base definition initiatives, that all comes into Power Inspect, which is more applicable on the inspection side obviously.

So maybe Mike's better to illustrate what's going on. But the process of defining the points and setting up the process for the alignment is, honestly, really, really simple. It's easier than using any piece of CAM software, or anything like that.

PRESENTER: So just to kind of give you an explanation of what's happening here, the first thing he did was just something create what we call an orientation point, which is essentially just a clearance plane, as far as how we want the initial probe to approach the part.

Then we're going to do what we call a freeform inspection. And from there, what we do is just simply double clicking on the part, we can designate which positions we wish to do the measurements.

So when selecting these things, we just want to think about the six degrees of freedom, whatever's going to kind of lock that part down.

The more data you give it, it's kind of the better the alignment would be. These are all laid up parts, so a lot of cases the part's fixtured against the good side. So as you're probing it, you're probing it on the rough side.

So you probably want to kind of have a little bit more than just the minimum, just to kind of average everything out. You can fine position these things, just to kind of show you what's happening there to kind of your looking.

It also does complete collision checking with the probe, as well as the probe head, which is what they're showing there with that orange indicator. And also the kind of crusty stars there, as I like to call it, kind of indicates that there's a collision.

So by modifying some of these values, you can kind of validate that whole process in software. Because obviously whenever anyone programs a probing path, everyone gets super nervous about wrecking that probe, just because it is expensive. They are kind of sensitive pieces of equipment, and surprisingly more robust than I ever figured, the few of them that I've whacked pretty hard.

But the end of the day, you just want to make sure you have all that stuff visualized.

Ultimately, what we do is we run that program. We essentially post-process it, just like you would like any kind of CAM or cutting routine. We would then run that on the machine. The machine then would generate what we call a results file, which is basically all that measurement data.

So basically it's an XYZ, with the approach, the vector to which we measured it. We then read that information back into Power Inspect.

And then from there, we ultimately get a quality report. That's just a byproduct of the process. But through that, we can also do what we call a best fit alignment.

And what that will do is allow us to figure out what the deviational differences are, not only from a translation standpoint, but also rotational standpoint.

And there's also some controls on how you do that as well. So do we want to do it with like a least material condition, or a maximum material condition, or just a best fit. So that's what they're going to now.

You could see, there's various different ways we can do it. We can tell it which points we wish to include in the best fit.

And then as it processes that, then that will generate what basically the delta transformation would be.

And then with that information, you could literally go back over the machine control, like on, say, a Siemens, and go to the fine alignment section in your work offset for your work piece. And just simply type those values in.

Or kind of what is the preferred way is we can then take that and post-process that back out again, and ultimately, that becomes what I would like to call just a sub routine.

So you'd load it in your control just like a cutting program, execute it. But all it really does it just loads all the values into the control. So that way you don't have to hand type any of these values in.

So this is the process of the probe actually doing the measurements on the machine. The real life kind of thing.

JON CALIGURI: Yeah. And you get the same machine simulation that I talked about with Power Mill. So it has full collision checking. You can simulate the entire machine tool, and all of those things, which just makes you feel a little bit warm and fuzzy about not breaking the stylus off on a piece of the fixture or something.

PRESENTER: Yeah. So here is where they're actually going to read the results in from the live measurement. And then once we plug that in, then we can actually process the results.

So the color coded dots there we call confetti, which is basically color coded based on tolerance. So I think it's green is intolerance. Anything that's red is above. Anything that's blue, or shades of red and blue, are below tolerance.

And then these little indicators, these pins, indicate what the deviation is at each of those positions.

Having said that, we can then output that as the alignment. So they're just basically creating a datum based on G54. And then ultimately what we're going to do is go to a delta transformation [INAUDIBLE] based on our G54 position and our best fit, we can then output that out.

So we have really three options. We can output it as NC code. So this would be for a Siemens, but whatever machine tool you have. And then that would be the sub routine that ultimately gets ran in the machine.

We can also export it out as an IGES file, which is basically just a work plane. So if you wanted to, you could then import that into your Cam system, whether it's, hopefully, an Autodesk CAM product.

Or if you're using [INAUDIBLE], because maybe you have to, cause it's aerospace, whatever the case may be, you can import that in and then post-process the CAM to that work plane.

For most people, they want to have a master program that kind of lives for that production job. So you don't really want to necessarily post it to that per se unless it was a one off job. But ultimately what we're doing is calling a sub routine in the cutting routine that basically calls in the alignment data, which is typically what we do in the aerospace higher production environments.