Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
#283 - Big Time Crashes Cause Big Time Problems
244 Plays3 years ago
TOPICS:
- John & John crashed the Willemin, Okuma & Kern, yikes...
- Grimsmo getting the Willemin to measure probe length.
- How to calibrate your spindle probe length.
- Saunders is finishing us a big fixture plate order.
Recommended
Transcript
Machine Crash Lessons
00:00:00
Speaker
Good morning and welcome to the business of machining episode two hundred eighty three my name is john grimsmo my name is john saunters and this is the weekly fun podcast where we talk about machining and manufacturing and crashing william and machines.
00:00:16
Speaker
It's good though. It's fine. I did bump it last night. Then this morning I checked it with the spindle test bar. I put the 180 millimeter long test bar in there and I indicate the front, the side, and the run out, everything. Everything's super square. It's still in good shape. The tool holder is trashed, so I need another solution there.
00:00:40
Speaker
Otherwise, all it was is I was kind of distracted. It was the end of the day. I was trying one last thing.
E-stop Button Usage
00:00:46
Speaker
I was texting the Wilhelmin guy back and forth, and there's a G-806 move.
00:00:51
Speaker
that kind of orients the tool holder, like a turning tool holder, gets it lined up right, and it moves it to a preparatory position, except I didn't type in the XYZ coordinates. I was just like, G806, tool 39, send. And it tries to go to XYZ zero, which is like the center of the bar at the collet face. And I'm like, that's why I'm an idiot. Like, why are you hand typing code, John?
00:01:19
Speaker
Oh, you have to for testing stuff. Yeah. Oh, I guess. Okay. Um, so was it a hundred percent? No, it was slow. I don't know how slow, but it was fairly slow. It probably took like two seconds to move from home position down to the thing. So it was slowish, but I was, I had my phone in one hand. I wasn't really watching it that closely. And like,
00:01:41
Speaker
It took me a second to start stabbing at the control to turn this thing off. And then I eventually punched the E-stop and it was already touching at that point. That is our half-serious, half-joking discussion with all of the training students as we go out, we talk about safety. And look, it's great. A lot of these students have never run machines, let alone machines that have capabilities, capabilities of cutting a lot, moving fast, of damaging themselves or whatnot.
00:02:12
Speaker
I've done this now across hundreds of students with dozens of e-stop-worthy incidences. What happens every time is you hear a noise, or usually it's a noise, but sometimes it's a visual and something doesn't look right or sound right. You immediately go into that, okay, I need to stop the machine.
00:02:34
Speaker
But before I hit that E-stop button, I'm going to look just a second longer to see what's going on. And we try to talk about like slam the E-stop button. We'll talk about it later. We'll figure out
Machine Safety Protocols
00:02:43
Speaker
what happened later. Slam the E-stop. And I mean, look, I'm kind of guilty of this as well. You're always like, what just happened there? Yup. Yup. And I mean, I'm.
00:02:52
Speaker
For years, I've been very good about you hold like the feed rate knob and the stop button or the stop or something like while you're testing something and you watch it slowly and you dig in slow and you're ready to stop it before anything happens. This was like the one out of a hundred times when I didn't do that and I paid the price. Lesson learned for everybody listening.
00:03:15
Speaker
Yeah, just a little bonehead move and I'm not that comfortable with the machine yet. So I started stabbing where the reset button would be. And then the control panel like flipped out of the way because it's on a hinge. Yeah. So it moved out of the way and it still didn't stop. And then I'm like, and I saw the big red and I slapped it. Um, so that was, yeah, it's an interesting lesson. So the, you had a piece of material in there and the tool and the tool holder moved, just moved into the part.
00:03:44
Speaker
Yeah. So crashing the part, not into the machine spindle or chuck or anything. Okay. And it was brass. So the brass ate up quite a lot of that carbide to, I don't know, it's soft.
00:03:57
Speaker
What happens on a Wilhelmin or a Fannock when you, quote unquote, crash? I mean, do you get like a feed mismatch or a servo overload? I don't know if the machine had already stopped by the time I hit the E-stop. I'm sure it overloads because all my other machines do. I don't know. But yeah, usually it'll like E-stop itself kind of thing.
00:04:22
Speaker
Yeah, like do you know, I'm just being nosy. Do you know, was there a specific alarm like servo mismatch or position there? Oh, really? I just kind of went home after that. That's fair. Yeah. So I didn't look for alarms. That's a good question though. Yeah. It made me question whether it stopped or I stopped it or whatever, but
00:04:45
Speaker
Yeah, I kind of can't believe that you could, like you would think that the machine should stop it first. Like just should. Yeah. With enough load to damage a tool holder. Yeah. It probably was already stopped by the time I got there and
Crash Detection Technology
00:04:59
Speaker
hit the stop. But, uh, yeah, I didn't know panic mode set in. Yeah. Well, the way, um, well, I wish I could remember the exact nomenclature, but the way the Haas
00:05:10
Speaker
When you have a Z-bump, it usually has something to do with, it's not an overload as much as it is a mismatch. It's a closed loop service system. I told a machine to move another 10,000 and I commanded that to the motor. The motor got the command, but the feedback loop didn't go. It's mismatch. Why did I just bring that? Like what the label is? No, I was just trying to think about
00:05:41
Speaker
Oh, it'll come to me. The Akuma apps guy that did a lot of our training from Gossacher who was superb. I absolutely loved how much he helped us and his attitude and it was really a great experience. He did something that I'd never done personally or I guess seen done, which is
00:05:58
Speaker
And he only runs programs this way. I don't know that I can get to this point, but I certainly did it as I was learning the Akuma and running test stuff, which is that he turns the physical feed knob and rapid knobs to zero. Then he hits cycle start, and that lets you... It's nice because it doesn't mean you don't have to go to single block, because single block is... Generally, I'm not interested in single block. I've been using it more and more, I gotta say.
00:06:28
Speaker
So to that point, if you're looking to just parse through a couple lines of code, that's fine. But a lot of times in our code, there's 17 lines before the first actual move, and single block just becomes button pushing.
Crashing vs. Aggressive Cutting
00:06:40
Speaker
But what's nice about this is it'll go through all the code until you get to an actual machine moving. And then what you could do is you can move the knobs up, it'll start to move. If you like where it's moving,
00:06:52
Speaker
you can kind of relax until it gets there more you can turn it back down. And it'll wait again. And he had a really good flow working where you don't like to your point about fat finger fit for getting a line like you
00:07:06
Speaker
Because you said like lesson learned or whatever. I kind of humbly disagree. We're all going to do this. There is no like, OK, I should be more focused. No. Yeah, it's both. I mean, lesson learned on the intricacies of G806 command.
00:07:23
Speaker
Yeah. Now I know. But yeah, it still happens. And that's what you mentioned is what I've been doing on the Maury forever and now on the Kern a lot too. Turn the feed rate to zero, hit cycle start, it'll crunch through all the preparatory code to get to a movement and then it'll wait for you and then crank up the feed rate nice and slow and then you wait for it to do what it wants. The Kern especially because I have so many
00:07:46
Speaker
pre-calls and quick sub routines, checking tool life and stuff like that before anything moves. So yeah, turning it to zero. Usually I'll turn, I'll move the axis just away from home. I'll turn the feed rate to zero and I'll hit cycle start and that'll go through hundreds of lines of code and a couple of sub programs before anything. And then I hit single block and then I turned the feed rate up and now I'm moving through what I want. That's nice. Yeah, which works really well.
00:08:13
Speaker
I remember what I was going to say. I was having a spirited discussion with somebody who was posturing kind of like, why can't machine tool builders just simply detect crashes? Kind of like this is ridiculous. Like beforehand?
00:08:28
Speaker
Well, yeah, like the difference between, so I'll share my limited knowledge of this, which is I remember talking with the engineers at Haas at IMTS four years ago when they came out with safe mode, which is actually awesome. And we did a video on it where if you run a tap into a hole that doesn't exist, it recognizes that the hole in there and it stopped and it didn't even break the 1032 tap. Now I think the material is aluminum. It did dent the aluminum, which kind of just shows like, okay, it's actually
00:08:58
Speaker
trying here. And it's really cool. I wish my request at the time was, man, let me command in and out of safe mode with like an M code. That way I could actually handle it from my post if I want to run things in different environments. And that led to this conversation about
00:09:16
Speaker
And I can only relay what I was told, but it kind of makes sense. The differences between a quote unquote crash and actual cutting, especially aggressive cutting are not that different. Like cutting is crashing. You are crashing the cutter into the material as you're shearing it. Totally. What we've learned about like quote unquote big data and machine learning and AI over the last few years, I got to think somebody should be tackling this idea of like, we're going to feed lots of data to these machines. And you could pretty quickly start to understand
00:09:45
Speaker
what makes things a crash and stop a crash when it's very quickly into the crash. You can have some sort of setting where it's like, how crash tolerant do I want the machine to be? And that could be something you change based on how aggressive you're cutting. But we don't wail. Even our horizontal
00:10:03
Speaker
the face mills at like 30% and everything else is single digits and that's fine. So I would value. Imagine somebody punching one inch holes in steel with a carbide end mill. Like that's a crash upon a crash upon a crash upon a crash, right? Like that's a lot of power instantly, quickly. And that's probably a more of a spike than a tap going into a hole that doesn't exist.
00:10:27
Speaker
Maybe, I don't know. I don't know, yeah. But it was interesting to hear about that. And look, I know some of, was it Micron has some pretty advanced crash detection stuff. And Hermla has the Z-axis crush washer, but that's a physical crush. Yeah, Micron does, too, with their step-tech spindles. They're supported in some unobtanium magic that when it crashes, something cheap will break that is easy to replace instead of a $30,000 spindle or whatever they cost. Yeah.
00:10:58
Speaker
Did I tell you I crashed a face mill on the tool setter in the Akuma? Did I tell you I crashed the Kern the other day? Oh my God. Get you first. This is getting dark, dark. I would encourage everyone to pull over for a moment. This could be difficult. It's actually a really great time to mention this because I was like you last night. It sounds like I was in the zone. I was at the shop. I wasn't alone, but I was not distracted. I actually now...
00:11:27
Speaker
take my wash out of my pocket and I take my Apple watch off. Nice. I was so focused because I was hand loading a five inch face mill. Akula can actually tool change that, but I just didn't want to.
00:11:44
Speaker
It's the only program I was running, and the only reason this story is actually harder to tell than it is, is that it happened when I was decking the DIY tombstone, which still otherwise has been an absolutely great project. So I hate to give any food to the critics here.
00:12:02
Speaker
I simulated this toolpath. I checked it. I did all my stuff correctly. What got me was the final pass happened at the lowest point in Y. So when the spindle was closest to the ground and the lead... So normally too, once the machine's moving, you're kind of like, okay, we're good. Sounds good. It's cutting the right Z depth. The tool looks good.
00:12:24
Speaker
And I still stayed there, but it's kind of
Simulation and Setup Awareness
00:12:27
Speaker
the example of why horizontals can be scary because this tombstone is all the way up near the spindle, so I couldn't even see the toolsetter, which is this little hut.
00:12:37
Speaker
on the backside of the tombstone, if you will. It's actually mounted to the tombstone carriage. And when the face mill finished, I had it extending beyond the part and the edge of the face mill hit just the sheet metal housing.
00:12:54
Speaker
of the lifter lid that protects the toolsetter. So toolsetter was fine. It marred up that housing, and it sounded horrific because it was sheet metal flapping around. So huge, 10 out of 10 ego brews, but a one out of 10 in terms of a crash causing real damage. But yeah, that's a lot of questions here.
00:13:19
Speaker
First of all, you've done this before, haven't you? Sorry, I already mentioned it on the podcast. This is like not if it happened recently, maybe six weeks ago.
Toolsetter Mishaps
00:13:30
Speaker
Sorry. I'm just wondering if this is a recurring issue because either that's a problem in your process or a problem with the machine or something. I did mention it because I mentioned on the podcast that other people have warned me about this and I just got bit by it. Yeah, for sure. And then what's my other question?
00:13:48
Speaker
The, I don't know, you said the toolsetter is mounted on the pallet. Like it doesn't change with the machine it, but it stayed on the pallet base. So your tombstone, sorry, your horizontal has the Z axis is in the tombstone.
00:14:05
Speaker
Think about a manual lathe or you have the cross slide. The cross slide moves in your Z for you, if that makes sense. That's what a horizontal is. The tool setters mounted on the backside of that fire. You need the five-inch face mill to be able to clear the entire size of this tombstone kind of thing? I need it because it is a 26-inch piece of aluminum and I have 22 inches of Y.
00:14:36
Speaker
Oh, yeah. You can't even leave the face mill off the part.
00:14:41
Speaker
No, no, I have plenty of X, so I could lead off, it's 12 inches by 26, so 12 is my X, so I had plenty of X travel to lead in and lead off, not with the same... Right, right, yeah. But I needed the extra radius for the Y. Now, I thought about minimizing the crash risk by using a two and a half inch for all of the part except the very top and bottom, but that would have meant, and I could have still done that, that would have meant I potentially had different blend lines between two different tools.
00:15:08
Speaker
And I just thought, OK, now you're getting cute. Like, John, just stay focused. The five inch hazel will work. And yeah. And it can. Sure can. And unchecking extend before retract on the facing up and Fusion fixes it, because then it just gets to the center line of the tool and it pulls off. So I have that saved. So I will have to deck a couple more of those, and I'm not worried about it.
Machine Programming Insights
00:15:30
Speaker
But that leaves a potential blend line somewhere as you're coming off. No, it's fine. I tried it. Not important. Yeah, OK.
00:15:38
Speaker
What happened to you? Interesting. So when we make our, when we grind our rask blades, there's some milling ops first that hard mill the meat of the bevel away, like the cutting bevel. So I machine 90% of it away. I leave a few thou for the grinding wheel to come in, kiss it all.
00:15:57
Speaker
So with the end mill, I do roughing passes and then cutter comp to finish the roughing pass. And then the probe comes in and measures the surface. And then it offsets the tool radius wear compensation. And then that same end mill comes in with updated values and finish passes the end mill section so that ideally I want to leave whatever it is exactly 2,000 for the grinding wheel to always have a consistent amount of meat to eat off.
00:16:28
Speaker
So I was trying something out where I copied that probing code into a separate file. And I was testing test probing a different area of the blade over and over and over and over again, like five, six different times, kind of running the code manually, whatever. What I didn't realize was I was applying a wear offset to that tool every single time, every time I test probed. So that tool ended up with a quarter inch wear offset.
00:16:57
Speaker
No. On a three-eighths tool. And I didn't know that. So I, Friday night, go back, run the machine, go home for the weekend, log in from home Saturday morning, Saturday afternoon, whatever it was, to see a C-axis overloaded bunch of alarms, basically. And I've seen that before. And that's where the
00:17:21
Speaker
Because the machine is tilted 90 degrees sideways, and your rotational axis, your C-axis, is only so strong. And if you overcome it, it'll just kick away. And it'll be like, ah. And then you stop the machine, the lights turn off, everything turns off. And the C-axis goes limp so that it can kind of move away from the cut or whatever. Oh, really? Yeah. Interesting. So that's kind of nice. I wouldn't say it's by design to move away from the broken cutter kind of thing. But it just goes limp and moves.
00:17:51
Speaker
So I'm sitting there at home going, OK, I know I crashed the machine, but I don't know how or why or what. I don't have a video camera on it right now. So I'm just looking at the screen, logging in remotely, trying to see the error codes and see what the heck happened. I'm like, OK, this tool is in the spindle, yada, yada. Eventually I looked at the tool table, saw the wear offset was stupid. And then I'm like, how on earth did that happen? And it took me some logic to figure out I did that. And then I came into the shop.
00:18:21
Speaker
on Saturday for a couple of hours and figured it out. No damage. Everything's good. Oh, that's great, John. The end mill was crunched but not broken off. Really? I see a couple of flutes, yeah. Little ding, like that blade that was grinding was toast. Little gouge in the fixture behind it. The clamp, it blew right through the clamp, but otherwise everything's fine.
00:18:44
Speaker
Oh, that's great. Then I not only modified that test code so that if I do run it again, it won't happen again, but put a couple more safety checks in place so that if that tool has a wear radius bigger than whatever, stop everything. I was just going to say, you could design a processor on this, but it's also like, look, you're going to figure out a different way to crash in these scenarios. You would think, no, the machine shouldn't accept a wear offset bigger than
00:19:12
Speaker
You know, 50 that or something like that. Exactly. And I know that that tool, as it gets dull and wears, and it's auto comping every blade. So it's constantly writing and reading from that wear offset. And that's a lot of power to have in a process that can run away from you. So I put a hard limit on that whole probing process that says, I know when the end mill is this far, I replace it anyway.
00:19:38
Speaker
So if it's a little bit more than that, for any reason whatsoever, like stop the machine and put a big fat, you can do alarm pages that bring up a little pop-up message on the screen that tells you what's going on and you can custom write it to be whatever you want. So I wrote some notes in there, like something is weird if this is, if you see this message.
00:20:00
Speaker
Actually, I think I programmed it to skip the rest of that program, the grinding operation, and skip to the next thing that doesn't use that tool at all. It can keep going if that happens.
00:20:12
Speaker
I love that, but I'm also at the point where I'm like, I'd rather err on the side of like, let's just go make sure we're not doing more damage by skipping ahead with a compromised fixture or whatever.
Ensuring Measurement Accuracy
00:20:24
Speaker
Ideally, if it hits that message, nothing bad has happened. It just sees a high wear offset and it just skips everything.
00:20:33
Speaker
When we were talking a few months ago about a similar thing with the Okuma and how I was like, I refuse to allow work order systems to be constantly updated without returning to some baseline. Because to me, I just don't like that idea that you have the potential to add 2,000 every time you broke something. Because all of a sudden, it can become 200,000 off, and you can shove a- Which is exactly what happened to me. Yeah.
00:20:57
Speaker
I thought we crashed it last night. It didn't make sense, but I logged in at nine o'clock to shut the compressors off and from home. And I looked at the machine on the webcam and I'm like, wait, why is it? It looked like the probe was in the spindle.
00:21:11
Speaker
but the probe tip was missing, because the probe tip on that thing is four inches long. So it's big enough to kind of see, but it's still pixelated. But the wrong tombstone was in the machine for even having a probe. And so I thought, oh man, we must have probed the wrong, somehow. But I have a program that won't allow code to run on the wrong tombstone. It's like, this just doesn't make sense. So I thought, I told Yvonne, I was like, hey, I'm going to run into the shop real quick. Because if I crashed it, I at least want to
00:21:37
Speaker
I don't want to sleep overnight coming into a crash, but I really thought, even though it didn't make sense, I really thought we did something and I came in and it's actually great. That machine is fussy. When it's in pallet mode to run automatically, if you have anything other than the cycle
00:21:54
Speaker
I don't know what it's called, but the main program screen on the monitor display, it won't keep running the code the next program. So Garrett had left the where comp page open. I don't fault him for a second because yeah, I just don't, but it basically stops or pauses the machine, which I guess
00:22:13
Speaker
You can see it as a design feature, but annoying. But anyway, as soon as I moved off of that, what was funny was, as soon as I moved off that, the machine resumed. And what I thought was the probe was actually like a ER, it's super weird. It's like the only ER25 holder that we have that had a short chamfer tool in it. And I never see it because it normally gets used for like two things and it's gone. But it looked like the probe from a distance. Right, right.
00:22:39
Speaker
That's awesome. Yeah. The current will do that too. If I'm editing the tool table and on the current, you actually have to click edit on edit off to allow access to change the tool table, which I kind of like that feature. If you're edit on and you're highlighted on a certain tool, like say tool 100, um, when the machine calls that tool and the tool tables are open for editing, it will stop.
00:23:02
Speaker
but it can call any other tool, any other tool. Okay. Yeah, that's so better. I've done that before a bunch of times. Angelo has done it too, you know, cause we're right before we go for the day, you know, we update the tools, we replace the broken ones or worn out ones. And then if we leave that tool table open on any one tool, eventually at some point in the night, it's going to need that tool and it's going to stop at some point and then you just come in and you're like crap. Yeah, that was me. No big deal. But we just lost eight hours of production cause I forgot to click one button, you know? Yeah. Yeah.
00:23:32
Speaker
What were you doing with the gauge blocks on the Instagram post? A couple of things. First of all, I've been reading the 1950s book by the Do All Company called The Science of Precision Measurement, which is fantastic. Meg got it for me for Christmas and she actually couldn't find a print copy, so she found
00:23:58
Speaker
a random company in India that can find the PDF somewhere and print you a bound copy and send it to you. So I got it. It's this really cool looking bound book. It wasn't that expensive, she said. So I've been reading that and learning more and more about gauge blocks and I have a set of gauge blocks and I was like,
00:24:19
Speaker
I need excuses to use these and play with them and ring them together and clean them and measure the stack and all that. But what I was actually doing with them was on the Willemon because I've been working on that a lot lately.
00:24:34
Speaker
There's a procedure to calibrate the probe length that is different than the procedure that we normally use in a milling machine where you like take a known end mill with no length and you cut a flat and you take the probe and you probe that flat and you know your information kind of thing. On the Wilhelmin you put a tooling ball in the turning spindle up at a 45 degree angle.
00:24:54
Speaker
get this, you remove the tool holder and you put the spindle nose down onto the top of that ball and you put a test indicator under the ball so that any ball movement you see on the indicator as movement. You're flexing it, like pushing down it. So you knock the spindle down at micron intervals until the indicator just moves and you're like, that is zero. And you look at the machine coordinate positions and you're like 79 point whatever millimeters is my
00:25:23
Speaker
current, like that's the number I need right now. And you type that into your calibration program, you let it call the probe and then it probes that surface and yada, yada, yada, it establishes the tool length from there. But I was getting a tool length like nine millimeters different than what I physically measured the tool holder, the probe as. And I'm like, that's not right. Something I'm doing is wrong, something in the code is wrong, something, something.
00:25:51
Speaker
So we sort of used it as an excuse, Angelo and I, to use a bunch of different measurement styles to measure the physical length of this probe. We put it on the optical comparator. You know, we measured from light to light, from spindle phase gauge point, whatever, to tip.
00:26:09
Speaker
Angela put on the CMM, like put a block, a one, two, three block. He slid it against the tip of the tool because the tool is sideways. He CMM like face to face and I'm like, let me throw the gauge blocks in there too with a height gauge. I made a stack of gauge blocks with the number that he CMMed the probe was.
00:26:29
Speaker
and then height gauge measures the top of the gauge stack and the top of the tip, and I was within a thou or so. I'm like, okay, so we have all these methods that tell us it is exactly this, yet the Wilhelmin is still telling us it's significantly different. I was talking to Wilhelmin about it.
00:26:47
Speaker
I don't know what I did wrong, but I just kept trying it and then I got within like tenths of the right answer. I just kept doing it again. Like that where you touch the spindle face against the ball and you see for deflection, I just did that a couple more times and it's got similar results and I don't know. I think we were trying to shortcut something or I don't know, but it works now.
00:27:15
Speaker
Good. Was it the ball? Would it be compensated for the radius of the ball? Sort of. You have to do that, I assume. Yeah, you do. You tell it. That was like a 19.5 millimeter ball or whatever three quarters is, 19.05 millimeters. And the offset was almost like half my sphere diameter, but not quite.
00:27:40
Speaker
Yeah. It was just confusing. I just needed as much data as I could to figure out what was going on. Yeah. I know you're not going to agree with me, but this is one reason why it scares the pants off me that you're going to switch the units because who knows what measurement is like mixed units now. It's like millimeters of Z plus an inch ball radius.
00:28:03
Speaker
Yeah. Yeah, fair. But that's what I'm doing right now anyway, because every time I run the machine, I'm switching between metric and inch and inch. Okay. Touche. It's all right. Like I'm already, and I can imagine, you know, Angelo or Pierre or myself or whatever running three layers, two of which are inch, one of which is metric. Yeah. And you need
00:28:24
Speaker
two tenths wear offset. Oh no, that was two microns. Oh, it's backwards or it's two millimeters. Something bad is going to happen.
Lessons from Machining Mishaps
00:28:32
Speaker
And with the actual conversion that I'm doing through Willimon, it's like the machine is now inch. It's not a fake little offset thing where it's like, yeah, the machine values are still in metric and your display is an inch. That's easy to do. But
00:28:47
Speaker
I'm converting it from four digit metric to five digit inch so that they're the same accuracy level. Yeah, I hear you. It's a call, right? It's like I'm making the call. Can you explain the way you just mentioned touch the quote unquote simple probe test, cut a data, measure it, and then adjust your probe length? Sure.
00:29:14
Speaker
Yeah. And I'm certainly not the expert at it. I know some of our buddies are very comfortable doing that, but I've done it enough times. So like on the current, it all establishes down to a, where does your measurement accuracy start? So on the current, I have a tool holder that is a calibration pin.
00:29:35
Speaker
It's an HSK40 holder that's machined into it with a cylinder at the end. It's all one piece. It's all one piece. It is like from the factory, they tell you it is exactly this diameter and exactly this long. The cylinder at the end has a neck on top of it, so the height of that cylinder is exactly this much, comes with a certification report. That becomes the ground zero of measurement accuracy.
00:30:01
Speaker
From there, on the current, you calibrate the laser toolsetter with that tool. Okay. So now your laser is accurate. Okay. Actually, first you do a five-axis kinematic so that your kinematic movement is accurate. If you don't mind, explain it on a three-axis because I think what I'm asking is simple. I'll just skip that part. But basically, so now I know the toolsetter
00:30:25
Speaker
is accurate based on the calibration pin. So now every tool that I touch off should be accurate, theoretically.
00:30:32
Speaker
Relative well, yeah, no relative to The master you're using which is fine because that doesn't it doesn't matter if it's wrong as long as everything is relative Exactly. Okay, so then you touch off a tool like I use a an eighth inch end mill Dennis was saying like three eighths or half or something bigger, but whatever I just have this three eights and or eighth inch Aluminum cutting end mill that I never use It's like my tool number two in the current
00:31:01
Speaker
And then I put anything in a vice, chunk of aluminum, whatever, call device and skim cut, even just manually jogging, whatever. Um, I make a clean cut and how do I define the length? I forget two or the part height, the datum, the part height and the datum and all that. Maybe I forget, I don't know.
00:31:27
Speaker
I'm sure it's easy. I just can't think of it right now, but anyway, so you're, you're cutting a clean surface with a known tool length and then you come in and probe that and somehow use that for your tool length.
00:31:39
Speaker
Okay, so let's talk this through, because I think that last part is what I'm missing is important. So let's say you have a rock piece of aluminum extrusion. So the Z isn't like perfect, but it doesn't matter. So in tool two is quote unquote perfect, because you touched it off with your toolsetter, which you know is perfect, because it's calculated master. So then you come in and let's say you probe. And let's say you know, your probe is wrong by up to three thou. Like it's, it's close, but you're trying to get it perfect. So, um,
00:32:07
Speaker
Let's just say it's off 2,000, 3,000, whatever. You come in, you probe the top of your extrusion with this bad probe length, but you're going to take a 5,000 cut. I guess it does matter because it's the depth of the cut. Not the depth of the cut because you don't care the top of the raw surface. No, that's what it is. Actually, you shouldn't start with the raw surface. You should deck it. I'm talking out loud. You should deck it so that you know you have a good plane probe it.
00:32:39
Speaker
No, no, you should take, you should machine a step.
00:32:42
Speaker
No, I think you should machine the flat with the end mill. We're going to get a lot of DMs about this. You should machine the flat. Machine the flat and don't move the cutter or don't lift up and tell the machine that is now my zero, my datum zero, my G54 zero, whatever offset you don't use. That way when you come in the probe and you probe that surface, just measure the surface, you'll see your difference.
00:33:09
Speaker
You'll see that because it's to be the tool height with the machine offset or position position or something. Yeah. But that would, if you probe the tool and it's different or you jog until it touches or whatever, and the value the probe gives you is 7,000, you are exactly off by 7,000 and then you can manually comp it, I guess, or maybe there's a routine that does this. That's how I've done it before. Got it. I think I do something similar on the Maury, but not as often or as like carefully.
00:33:38
Speaker
And I don't have the calibration pin in the mori, so there's probably some error. I think I've definitely done because I've got the OTS with the little.
00:33:48
Speaker
A little hockey puck. Yeah, the hockey puck thing on the side of it. I think I've definitely gone spindle probe to OTS, touched, whichever one clicks first is the mast. I'm not saying that's the way to do it. What confused me is the cutting. You should cut, I think if I'm understanding this correctly now, but in theory, this is not as accurate, but it would help me.
00:34:10
Speaker
It'll help me explain it and other folks are confused maybe because of your whatever. Let's just take a really good gauge block. It's a good flat surface. Let's put it in a vise and assume you could hold it in the vise really flat. What you could then do
00:34:25
Speaker
is bring in your master, like your master tool, or even tool two, but not cut with it, but take a precision piece of shim stock, come down and touch the tool off with the shim stock in and out. This is where it gets subjective because what are you calling?
00:34:40
Speaker
the correct theories. But let's say you could do that perfectly or you could use the electronic conductive style. But then if you're basically telling the machine, I know the top of this gauge block is a really perfect datum, set it there. And then you come in and you kind of probe it. It's just a way of measuring how much different the probe length is from the node master. And you're using tool two as a surrogate for the master because
00:35:04
Speaker
told you was set by the tool setter, which was set by the master. Okay. I'm with that. That would give you the same relative result as doing a test cut. I'd say almost a little bit less accurate because you're using shims and just- Oh, way less accurate. Sure. But similar concept. You're establishing a reference with a known tool.
00:35:24
Speaker
and using the probe to see the difference on that. Yeah, yeah, yeah. Right. Yeah. Thank you. And that all comes down to dimensional accuracy and the order of events and how everything is interrelated. And this book from the 50s talks a lot about that kind of stuff. And it's really, really fascinating.
Current Projects and Challenges
00:35:46
Speaker
Cool. Well, other than the crash, how's the little man?
00:35:51
Speaker
Ah, really good. I was posting some code last night. I made a code because I wanted to, I have two turning tools, like a cart-off tool and a turning tool that I crashed last night. I made a quick code so that I could use variables 999, 998, whatever.
00:36:11
Speaker
to face and then come up, do a quick little radius and turn the diameter. And then I can command like 14.5 millimeters diameter and it'll face, it'll turn it. And then I go 14.4 and it'll face and it'll turn it. And then I can measure it with the mic and I can be like,
00:36:27
Speaker
Wow, that is insanely accurate. Oh, that's awesome. I need 10 microns of tool wear comp. So I learned about tool wear comp, and which way to go, and how to apply it, and how to get my turning tools to do exactly what I want. And then that brings up the question, how do I touch off Z between different turning tools? I haven't quite answered that yet. It's not hard, but yeah. So that's where I was at. OK.
00:36:55
Speaker
So once you would get a buy new holder and then I have some stuff so I can make it work. Cool. Um, so yeah, I'm just getting my feet wet, posting programs, sending them to the machine through, um, filezilla. Oh, good. It is ethernet capable. I got it. Ethernet. It's, it's weird, but it's good. It works. Okay. That's great. So that's good. Yeah. I'm just getting comfortable with it. And I'm at the point now where it's like,
00:37:22
Speaker
It's time to make parts on this thing. I have two parts in mind that I need to make now. And, uh, I'm like, what am I waiting for? Nothing. Okay. Full focus. Let's go. Yeah. Agreed. That's great. Super cool. So I'm, I'm excited. It's going really well. Good. Good. Yeah. What about you?
00:37:49
Speaker
We're just finishing up the substantial order for Langmuir that the folks that make that Crossfire Plasma System came out with a MR1 CNC machine and we partnered up with them to make fixture plates for it. Sweet. So they did a pre-sale and we made quite a few fixture plates, mod devices, et cetera.
00:38:14
Speaker
It really threw a curveball at us because the horizontal, you know, was designed to help Saunders Machine Works do what we need to do. And really, since that machine came online three months ago, most of its focus has been helping meet the deadline for lightning, which we are going to be we will be on time for, which I'm really proud of.
00:38:36
Speaker
But what it meant was we really kind of I was a back burner, but we have we I got soft shawls running on the horizontal last week I've got pallets and mod vice bases, you know Potentially hours away like both of them are actually well pallets have not been made on there But we did a mod vice test part, but my view right or wrong. My view is kind of like look don't risk
00:39:00
Speaker
testing a new product, new fixture, new process right now when we don't need to do it. Let's get this order wrapped up. It's one of those like we might actually pull the last parts off today or tomorrow.
00:39:15
Speaker
I couldn't ask for a better outcome because if we hadn't had that horizontal, we really just wouldn't been able to meet this deadline unless we worked out something where we ran two shifts. Many big manufacturing companies run multiple shifts. It's just never something that I've even thought about. Agreed. Not very interested. I have machines to run all night so that we don't have to.
00:39:37
Speaker
Yeah, exactly. And honestly, I'm afraid of a culture standpoint. Get your night shift, your day shift, the oversight training. It sounds like a mess. I'm not interested. Yeah.
00:39:51
Speaker
Yeah. You say that and I hope that's right, but man, if you look at like, I'm going to make what Spyderco or Microtech, I mean, I bet you those places are running two, three shifts. Maybe, I don't know. I can't say I've heard, so I don't know. Yeah. But ideally, I mean, the Willman has a bar loader. The Swiss, the past two nights has still been running when PR came in in the morning, so that's huge. Oh, that's great. With no oils fixed?
00:40:16
Speaker
No, not fixed, but it just works on this particular program. We're using it minimal amounts, but the new drum of oil should be here kind of any day now. It's been many weeks since we ordered it. We haven't had a night run in a while. I think Pierre's made over 6,000 screws. These are like the knife screws, the handle screws in our knife. Norseman takes seven, the Rask takes five.
00:40:42
Speaker
and it's just going really well, making thousands of them. That's great. That's fantastic. Buy the machines to run all night so that we don't have to.
00:40:54
Speaker
Yes, what you just said. We have a little bit more demo work to do next door in the training building because the floors are getting epoxy and we've got to move the machines that we put in for training class. We've got to move them next door to the room next door so that that floor can get coded. No big deal. And then when that's done, we will be
00:41:14
Speaker
in a great place. There's going to be longer term projects to do like TLC or renovation on a bathroom or
Facility Upgrades and Business Solutions
00:41:21
Speaker
other things. For sure. We've really got a lot of work underway. I want to do a video on that. We also had the ramp, the concrete pad, it's at a slight angle outside that building. Okay.
00:41:35
Speaker
It, uh, did I talk about this? We had it hydraulically raised. No, this is the coolest thing. Like, I feel like I want to spread the word. If anybody ever feels like they need this done, the ramp, it's not that steep. Maybe it goes up. I'll guess 12 to 18 inches across eight feet, 10 feet. It's just basically because the building, uh, floor is slightly above the exterior grade. So this concrete ramp that was always there.
00:42:00
Speaker
Um, it was an inch shy of meeting flush with the building. I don't want to say a bump, a one inch bump, like no big deal for a vehicle, but, um, very hard on forklifts and very difficult with a pallet jack that has any weight on it. So, um, and I don't know if they poured it that way. I think they did or settled. It looks like it just was poured that way. So, um, my dad had had this done to their
00:42:28
Speaker
back deck, which is where this company comes in. It's called hydraulic leveling or raising or lifting. It's like totally, I mean, I was blown away. They did it in like 30 minutes. Good on them. They're making good money. It wasn't that expensive. They drill five holes in the concrete maybe just under one inch. It's like halfway between tile grout and true concrete, but it's a slurry. It doesn't have any aggregate in it.
00:42:56
Speaker
And John, it was insane. They're hitting a button. And as I hit this button, it pushes in, you just watch that pad lift up an eighth of an inch 316. It was precision control like you're jogging a CNC machine. What? They lifted the pad up. The grout that they use to lift it is what stays in the hole to support it. So actually, it'd be a great solution for a foundation floor where you think as a hollow spot, you don't have to lift it, you just backfill it. And that reseals the holes and done.
00:43:24
Speaker
That's awesome. It is like perfectly flush right now. So satisfying. Yes. And now every time after having watched that, every time you drive over it with anything, you're going to be like, oh, yeah. That was a good decision. Oh, yeah. We drove the forklift in this morning to get a new barrel of quality cam that we store over there. And it was just like, oh, right over. Easy. Oh.
00:43:44
Speaker
That's cool. I didn't know about that. Yeah. So that was a win. It was $1,150, which to me was a no-brainer compared to trying to pour something, or we tried to withdraw from that. Oh, for sure. So I thought that was a really great price to solve the problem. But then when I saw how quickly they did, I'm like, those guys have a good business. Oh, they got good on them. But yeah, I mean, it's like, what are you going to do? You're not going to do it yourself. So it's fine. Yeah.
00:44:13
Speaker
Cool. Yeah. Sweet. See you next week. Sweet. Yeah, sounds good, man. Take care. All right, take care. Bye.