Accuracy vs Repeatability - Automation Insights

Video Transcript


My name is Jason Dove, and welcome to our live session here where we will be discussing the differences between accuracy and repeatability. I want to welcome everyone to the live stream here. This is a live session, so we do have a comment section here, so if you want to send us a message on the chat - feel free to do that, and we will try to answer questions as they come.


First off, I wanted to say that this session, basically, we're going to be talking about the importance of accuracy and repeatability when it comes to automation applications. So again, my name is Jason Dove. I am the East Regional Sales Manager for Destaco. We have three other panelists on for the discussion today. The first one is Mr. Jim Gary. He is our Senior Engineer for Destaco Robohand. Welcome, Jim. Next is Ravi Shivanni. He is our Global New Product Development Engineer for Destaco Camco. Welcome, Ravi. And lastly, but not least, we have Mr. Ray Tynzik. He is our Applications Engineering Manager for Destaco Camco. Thank you, Ray. Thank you guys, again, for all for joining us in this discussion here. So again, what we're going to be discussing is the importance of accuracy and repeatability when it comes to automation applications.


So one of the first questions, the first thing I want to kind of go over, explain what accuracy and repeatability is and how it's actually measured. Well, with respect to indexing products, and probably in general, there's probably a lot of confusion out there with the difference between the two. The accuracy repeatability definitions. We certainly have a particular definition of how it applies to rotary and linear indexers. Many times a customer will use the term repeatability, where after discussing with them, turns out they're really talking about the accuracy, and with indexing products, that's typically what they're looking for when a particular station on the indexer comes into a fixed station.


How will that differ from one index to the next? That's the same. It's defined the same way with whether it's a rotary or a linear indexing table. Follow-up question to my original question. Within history, everybody seems like they have kind of a different way of how they how they actually catalog the accuracy and repeatability, how they test for, you know Ravi and Jim. Do you mind kind of going over how we do that for both Camco and for Robohand? Yeah I’ll let Ravi go first. Thank you, Jim. So from the indexer's point of view, from the application data, it's very important that the customer define what is this end, you know accuracy, that is expected out of our indexers - or even on the conveyors.


So this actually leads us to have the selection of either mechanical indexers or servo positioners, depending upon the kind of environment which they're working in - whether that's a closed loop or an open loop, asynchronous motor-driven mechanical indexers. So based on that, we actually measure the accuracy of the indexers we’re using. Our end offline testing machine, where we determine the output accuracy of the indexer, and also what is important is the axial runout and the radial run out from the indexer output dial plate itself. So if these are critical to the customers, that needs to be defined well along with the number of stops, so that we determine what is the accuracy that can be output from the indexers - to make sure the indexes meets the criteria that is expected from the conveyors.


Our conveyors are not the regular conveyors. It's mostly a procession link conveyors. We have many kinds. One is the table top modular conveyors. We have two frame conveyors and we have a V2D conveyors also. So we usually refer to the accuracy, which is measured from station to station, and that depends on what is the center distance expected from the customers. So going by the terminologies that is required for an indexer, in the application point of view, is what kind of dial application is it and is it the dial application or is it an actuation indexers? What is the loading typically, the torque demand, workload, the moment load, and the pressing station speed? And how many stops are there? Is it a parallel indexer, is it is it a roller gear, or is it an oscillation oscillating moment?


A motion that actually determines the accuracy of an indexer. If I may jump in for a second too, I can explain specifically how we test our rotary indexers. In our linear indexers, and that might give some explanation to our customers, on why they really need our rotary indexers. We'll put the indexer in what we call our home or starting position. We will zero-out an indicator on that reference station and then we will start indexing, one index at a time, and each time we will take a measurement to see what that deviation is from the theoretical zero or the zeroed out station. And then we'll do a complete revolution and record those measurements and we may take two or three complete revolutions and that will give an indication of repeatability. But typically, the customers are more concerned with from one index, one station, coming in into position to the next one.


What's the maximum deviation? And that's kind of what we do for the indexers or the conveyors, too. We'll set-up an inspection station at the far end of the conveyor, being the worst possible position for accuracy, we will zero out the indicator and will index one link at a time and take deviation measurements to do a complete full revolution of the chain. That's how we measure it and that's what we will convey to the customer. Okay, to give you kind of an idea for the gripper side of things, it's quite a bit different. Grippers are relatively small, and we have the ability to inspect them with pretty simple tools, but I want to discuss what the difference is between what the Camco side of the company does with conveyors and grippers.


So “accuracy” for us is based upon the measurement from the dowel pin holes in the body through the dowel pin holes in the jaws. And every gripper that we make has gall pins for accuracy and what that does is it basically says how close is that gripper to those holes being perfectly positioned with respect to one another, as they're actually drawn in the catalog. And what that does is that gives you a couple things. It tells you, if you replace a gripper, how far off that gripper could possibly be with respect to the gripper that you just removed. But it's more useful to know how close you'll be when you're actually going to machine together. You're not going to have to knock the nest around, or you're not going to knock the gripper around, to get it into the location when we do measure. That can be done.


Also it's one of the things that we try to do, is do it under air pressure. That way, any play or anything, any looseness in the gripper, is taken-up and it's measured in the same way that a customer would use it repeatability. For a gripper, again, we have a little bit different way of doing that. For us, our repeatability is over the life of a gripper in a normal application, so after 5 million cycles, you would be able to have the same repeatability that you had when you originally got the gripper. And the way we measure, that is essentially, you have a gauge part and fingers you run - you'll cycle the gripper under load and you'll compare a position of that gauge part from when you first set it up to know. You'll probably do it every million cycles to check it but then you compare it to five million.


The reason for this is our customers need repeatability. They want the same thing every single day. That's why it's helpful. The words “accuracy” and “repeatability”, we have to know the differences between these two. Sure - on the left hand side we have something that's both accurate and repeatable. So “accurate” means it's close to the bull's eye. It's what we intend to be. We want to hit that particular location. We then have not accurate, but repeatable. When we look at that, we're not on center but every time we hit it - we're hitting it in the same place. So we can make it repeatable every time, and it has the ability to come close together, but we're not getting it in the center where we really need it.


Then you have accurate, not repeatable. Your results would be scattered or equally around the center of the bull's eye. And then your worst case is not accurate not repeatable, and that's where it's all over the place, so we're always striving for the situation on the far left one. What I want to kind of go over is, are there any specific markets or types of applications where accuracy or improvement is more important? Well, it's an interesting question because there is no industry that doesn't need high accuracy and high repeatability. It's really based upon what they're doing with the part. You can have situations where you have an extremely tight fit, that the parts have got to be aligned extremely closely, so you don't have binding or wear in the equipment and it literally goes across every single industry.


So usually, things with tight fits, is it tends to be one of the main places, but the other thing is just there are some situations where accuracy of picking and placing is super critical. Even though they may not have a constrained fit, but the location is absolutely important, so again it really depends specifically on what product the customer is doing and what he's trying to do with that. Yeah - like Jim was saying, it depends on the part process. If for example, we're doing a welding operation, accuracy or repeatability may not be nearly as important as if we're handling syringes moving them from one station to the next.


So we have to look at, although all our products are fairly accurate, there are some that are much more precise than others. And we also can discuss whether or not we want to do it with a servo motor or a fixed station, or a fixed stop cam, so like Jim said - we do have to look at every application individually. But for the most part, whether it's a rotary or linear index, they're very accurate inherently, just because of the design itself. Yeah - and I can say some of the applications, that I know personally I've run into, were within grippers and indexers where it may not be more needed and sometimes where need be. So like this, for example, if you've got a gripper that's going to be loading a part into a lathe chuck, you really don't have to be that exact, accurate, or repeated much in that application because typically you're just presenting a part, and typically the lathe chuck itself is going to be centering the part every single time.


Now if you're trying to like do like a needle valve, or something like that, of course you have to have accuracy and repeatability in the gripping situation. That's kind of the same thing with indexer, for a lot of times, whenever an indexer or dial plate's turning around, and you're doing assembly work on that - you need it to be a lot more accurate. You need to have the repeatability there, but if you've got a station to where you're doing like a reject or you're basically offloading the part, sometimes you don't really need the accuracy and repeatability in those particular applications. And so for us, all of our grippers, there are various levels of precision that we have in our products. But a customer may not need it, for as you say, they're taking a rejected part.


You don't really need much accuracy in position to do that, but for all of our products, we have a very tight tolerance. Again - everything is on a dowel pin, but at the end of the day, what really comes down to what a customer is looking for is a gripper that's going to grip the same way when they set up the machine in the beginning until they run their last part; whether that be a couple years down the road or if it's a short life cycle within a few months. But they just want that gripper to do the same thing over and over again, and I think for Camco, it's the same thing as well. It's consistency, because that keeps the consistency of the parts that they're making at the highest level possible. One other question - are there any variables that you guys want to go over that that can determine what product is needed?


I'll take that question. For us, and I think for Camco, it's a little bit different because the kind of customer that goes to Camco is usually being driven by an end user. But for us, we have machine builders and then we have the end user. The end users may also build machines in-house, but essentially, an integrator is building a machine for a customer and normally they come with a one year warranty, so they're looking to try to get something that works but it doesn't have to last forever. Because they want to keep the price down and they want to be able to deliver that machine and make a make a profit on it. End users, however, have a completely different viewpoint. An end user wants that perfect gripper. They want that gripper to repeat over the life, of whether they're making a million parts, five million parts. They want that repeatability, over and over again, and they don't want to have to swap out grippers due to wear or due to loss of repeatability, so it's important that that perfect gripper is going to do the same thing that it does on its first stroke as to its last stroke - and not give you any problems during that time.


That's really what that end user is looking for. With indexers, it's pretty much the same. We also sell to machine builders, and sometimes end users, but probably more often the machine builder. And with indexers, it's really critical that we get all the application information you could possibly get. Give us more than we need, we'll filter through it. But the problem, if it's not sized properly, we're usually at the heart of the machine. We're down in the center, and sometimes it's tooled up so much, you can barely even see the indexer. So if something goes wrong, it can be really catastrophic to try to dig it out to fix it or rebuild it. So we're looking for, first off, just simply how much weight it needs to support.


We have axial and radial load values that we have to stay within and then we're looking at, extremely important, is how much inertia that is, and what today's CAD systems, if it's modeled - if you’re at the point where you've got everything modeled, then give us that CAD data. As long as you got material properties assigned to everything, it's a very valuable tool. So the weight, the inertia, and if there's any moment load on there in any external workforces - the more we have up-front, the better we can do some ballpark sizing if we're just early in the process. But we always recommend come back to us once you've finalized your design. We'll run the numbers again. Better up-front before anything is submitted than when you're sitting at the machine and something's failed because you made some changes and now you've got twice as much weight as you thought you might have.


Just to remind our audience here, so again, this is live if you guys have any questions - please feel free to use the chat session here. We actually do have a question that has come in that I will see if we can do. The question we have is, if we can kind of again differentiate between accuracy and repeatability? I think we did that just a couple minutes ago, but if you want to kind of, Jim, if you want to go over again - if we can pull up that slide that we had a second ago that kind of shows the difference between the accuracy and repeatability again, where it's showing the target there. If you don't mind, Jim. If you don't mind going over that one more time here. Okay - actually i'm going to start on the right side this time and we'll move left. So on the right side, we have not accurate not repeatable, so that's kind of the worst case. That's where you're not near the bull's-eye and your shots are not close together.


Moving to the left, we've got accurate but not repeatable, so your shots are separate from one another but they tend to congregate around the center. Then you have not accurate but repeatable, so now your shots are all coming at the same place but it's not on the bullseye, so again it will repeat. They'll come together every time but they're not coming together in the right place. And then, finally, on the far left - is the accurate repeatable situation and that's what everybody is looking to get. The idea there is that you're always going to hit the bullseye and your positions will always be very close together with one another. Again - that's what customers are generally looking for. They want to make sure that they're on target as far as where they're going to be placing that part. In the case of a gripper, usually a pick and place situation, replacing or assembling that part with respect to how repeatable it is from every time you place it to the next to the next.


Thank you again for that. So I want to also say that since this is the live session we're going to have a poll here, so you guys should see here at the bottom of your screen. And with this particular poll here, we just want to kind of get a poll from our audience of, “do you guys have experience programming servo motors?” So again - you guys will see that here at the bottom of your screen, if you don't mind clicking on that, if you guys have experience programming server motors. We're just going to kind of take a poll on that. Moving on a little bit, want to see if any of you guys have any examples to share where accuracy and repeatability is very important. So I’m going to turn that to you guys, but before you do, I just want to say you know my personal experience - I will say I can't give the customer name, but I will say where they were actually using our grippers to install a needle for a syringe basically inside the inside the syringe.


So we had to have, with our grippers, when we picked up the actual needle itself - it had to be both accurate and repeatable; have the accuracy and repeatability every single time, because we had, I think it was, a thousandth of an inch play - if that. You know, whenever we were inserting that needle into the actual syringe itself. So that's an example there where we needed to be both accurate and have the repeatability, as well. I don't know if anybody else has another example of where we would need to have that as well. Yeah - in other gripping applications, it really is different. Again, it's based on the part.


But one of the things that we have again are a wide range of products that can meet those application needs. So we make one of, if not the most, precise grippers available. These are grippers that are actually around after assembly, so that they're able to keep these very tight tolerance, positionally, and have the repeatability be the same over and over again. Jason, as you were saying, within plus or minus half a thousandth of an inch. It's very important to have the right product to do that. That's the key thing. We make a lot of different products, but there are some that are very specific to precision applications. Yeah with indexers, it's the same. We've got a complete line of different rotary and linear indexers with standard repeatability and accuracy numbers.


We also do specials, as needed. We designed a conveyor system. Our standard conveyor is plus or minus five thousandths of an inch. This was a razor blade manufacturer where they needed plus or minus a half a thousands, much like the needle application, and we were able to come up with a shot pinning mechanically-connected shot pinning device that positions the links to plus, or minus, sixteen thousandths of an inch - or roughly half a thousands plus or minus. And this was doing it at 200 indexes a minute, so we can do specials. So if you're looking through the catalog, and you're just saying “well this just isn't going to work” give us a call anyway. We'll see if there's something else we could do and, I think with the grippers, you guys do that as well, correct?


Yeah. Switching gears - this doesn't relate to what we're talking about here, but I know one of the questions that I get a bunch is, “I’m a customer here and I’m trying to get into automation.” So this is maybe some of our audience. They may not be fully automated. Some of them may be fully automated so maybe looking to get into automation, and so one of the questions that I think we get on a regular basis is, “what do I need to consider when I am trying to start getting into automation?” Anybody will take that. Yeah - when you're moving from what's currently a manual process into automation, there's usually a couple different drivers. Generally, it tends to be quantity. If you need to meet much higher quantity levels, you go with automation - but also quality.


Sometimes there are things that you just can't do with a manual process, where you need the repeatability and accuracy that you can get from an automated process. So when making that transition, it's important to understand why that transition is being made. Certainly, the human hand in a manual operation has a lot of compliance and there's feel to it, and that sort of thing, but again - something like that isn't necessarily very repeatable. It's almost an arch, and you can't always guarantee that it's going to do what you need to do. The other thing is economics. And that usually is another driver, particularly in this day and age, we're seeing people who are just starting their first time doing automation, are coming in at the cobot level.


It's huge, because these customers can literally take a look at what a cobot can do for their application, and they don't have to do a cost-benefit analysis. They can run the numbers in their head and they know they're going to be making that money back in a very short period of time, so there's a short ROI there. But that's what's right now the biggest driver of transition from manual to automation, right now in the industry at-large, and it's a revolution. I agree. I think one of the things you touched on, as far as with ROI, that's one of the big things that's pushing a lot of customers that are looking at a manual system and they're looking at trying to go into automating, because really, when you look at automating a manual system - you can see if you increase that throughput - number one. You're eliminating human workers that that will call in sick, miss work, that make mistakes - that do different things like that.


Not that automation is perfect, but again, for the most part - you're going to have a much higher accuracy and repeatability, just better quality product whenever you're really using automation. And you know with this, it's pretty easy to calculate the return on investment that you have for an automation project, because basically - you're really looking at one point, I had this human worker. They're making this many parts per hour. This is how much it costs for me to do that. You pretty much can divide what it costs to how many parts you're making to how much it costs. And so, within automation, it's very simple. You basically can take what is this project going to cost me, how many parts per year, per month or whatever - that I’m going to get out of this.


And then you really do that math and tell very quickly how fast that this project is going to pay for itself. And for the most part, it depends on the application. You know I would say, generally speaking, most of the time where you're probably looking at it at a year or two years, depending on the actual application, but I've seen applications where we've had customers where, in six months, they've paid for the automation project. Because it's so much profitable, and they have so much more throughput. Yeah - and even small companies, maybe with not a huge budget, could benefit from automation. I mean you don't have to go fully elaborate with vision systems and all kinds of sensing. You could simply get an indexer to move parts in a position where, even you could still include some manual labor, in-addition to that.


Those are probably the earlier days of automation, but it seems like even smaller companies are going to full automation, because they see the benefits and the return of doing so. Yeah - one of the things that we tend to notice out there is that when these companies do start to automate, they are in many cases going to grow, and so they're actually not laying off that worker. They're moving that worker to a more skilled position and that tends to be a standard. So it's interesting. A lot of people think, “oh - that cobot's going to put me out of work.” No - that cobot's going to probably put you into a higher paying position within the company you're at and a more interesting job, and that tends to be the rule more than the exception. Yeah - I totally agree with you. Now, let's say I'm a customer and let's say what I've decided I need to automate this system. Where do I go from there? Yeah - where do I find the right company to help me to automate?


The first thing you can do is get in-touch with us because, right now, you're probably using Destaco clamps if you're doing things manually. And the beauty is, Destaco has a full range of automation, as well, so we're a great place to start. But the other thing is, find machine builders in your area. Go to trade shows, when we start having trade shows again - after Covid. And look into machine builders. Again - it's a process. It's a transition. It's not something that you can do overnight. But it is something that will definitely pay off in the long run. Now the other thing too is work with other companies that you're friendly with and find out who they're using to build their machines. But trade shows are a good place to start for somebody who's going from manual to automation. I would definitely recommend that and I would agree with that Jim and I think for a word of advice to customers looking for a company, there tends to be companies that are going to have a reputation for making certain types of machinery, dealing in certain types of industries.


I personally highly recommend sticking with those companies. You, as a customer, you may pay a little bit more as far as just having that project quoted out from you from general people, other companies, typically it's a lot of times whenever you're paying less - you're getting what you're paying for. So yeah, you're buying experience is what you're doing you're buying experience. 100% totally agree with that. And I said those companies that are known for making, I'm going to throw out something - let's say you're trying to spin yarn or something like that. There's thousands of ways of doing it, but there are companies out there that I’m sure that have that they have mastered the art of spinning yarn.


Whatever it is they're doing, but whoever does it best, it really won't be too hard to find that. You're going to go with word-of-mouth but go after those companies that have that reputation and you're going to be better off with going there - as aside from just trying to pick someone and just trying to hope that they know what they're doing. Yeah - the same applies with indexers too. Reputation is huge. That means that you've got a reliable product. I hear customers all the time, “I’ve got this 15 year old indexer or I’ve used it for this application. I'd like to repurpose it. What can we do with it?” Like I mentioned earlier it's at the heart of a machine. If it's not going to last, it's going to leave a really bad taste in somebody's mouth here when they got to rip it all out, tear apart their whole line, and lose production.


So our reputation is important. Are you flexible? We are certainly flexible. We'll sell you a catalog product but we'll be happy to discuss every aspect of your application, and if changes come up, we'll be happy to assist you with on-the-fly changes - before and after the sale. And in the field too. We'll fully support you up-front. Reputation and helping out in the field, as well. We'll be there to support you all the way. Yeah - so you know one of the things, again on this journey of going from manual to automation, when you're working with a machine builder, as we talked about, reputation is extremely important. But at the same time, you need to think about how is your product going to be made when it's automated, and how long is that machine going to last.


Is it going to last for the full life-cycle of the product? Are the products that are being used on that machine just there because they're the cheapest or because they last the longest? Again - it's that you know you can pay me now or you can pay me later, and with all of the stakeholder products Camco, Robohand - they're all known as the highest precision and the best value over time for your money in the industry. Well guys, that's our closing for our topic here. I do thank you guys for your time. Jim, I want to thank you. Robbie and Ray, I thank you guys for taking your time to give your experience and expertise to this discussion. I want to also thank the audience, everyone for joining us here and thank you for the question that we were able to get. I want to just say a quick mention here that for everyone to check out our product pages in the platform.


You can access our literature on all of our products. Feel free to, as and you can reach-out to any of the product specialists that we have with inside. The next day ahead, today, and tomorrow - through this platform to where you can ask questions. If you have any, and please feel free to check out some of our other sessions that we have going on for the rest of the day and tomorrow. And once again, I thank you guys for your time, and I thank you to our audience.

Click Here for Customer Support

As a Global company, DESTACO has customer service representatives based in your region, no matter where you are in the world. Contact DESTACO Customer Service in your region for technical assistance, repairs, field service and more

or call us at 1 (888) Destaco (1-888-337-8226)