Transcript
John Saunders: Good morning. Welcome to the Business of Machining, episode 446. I'm John Saunders.
johngrimsmo: My name is John Grimsmo.
John Saunders: Catching each other up on, ah you know, what we're trying to learn to stay sharp as manufacturing entrepreneurs.
johngrimsmo: sure
John Saunders: What's on your list?
johngrimsmo: Today, my list includes Speedio Bevels Automated, past tense, past tense as of minutes ago.
John Saunders: Yes.
johngrimsmo: um The scrap blades that I made the last week, talk about that. um I have a note that says, be careful what you tell me, and I'm talking to you directly.
John Saunders: Talking to me?
johngrimsmo: well Yeah, it's just's a note. We'll get to that.
John Saunders: Okay.
johngrimsmo: And quick tour, rev numbers, and calculating radius.
John Saunders: Quick tour.
John Saunders: Okay. Can I take a guess at the, be careful what you tell me?
johngrimsmo: Yes, sure.
John Saunders: Was it, was it the banana metals?
johngrimsmo: No?
John Saunders: Oh, cause that was, uh, that was technically not public information. But after we hung up, I, I DM I was like, Hey, I told Grimso that on the podcast.
johngrimsmo: Oh yeah.
John Saunders: He's yeah, i don't care. So I was like, okay, great.
johngrimsmo: Okay, fine. Perfect.
John Saunders: Yeah. Okay. Interesting. um Okay. My list, ah brother to life, coolant dripping with, and fixing it with check valves, AI agents, some the huge automation wins over here at Saunders, read sensors and a quick, quick, fun share on Johnny five.
johngrimsmo: Yes.
johngrimsmo: Yes, fun, fun. all right, you start with something.
John Saunders: um Okay, look I have been a technician this week. I have been running UMC 400.
John Saunders: And it's a new machine to us. We've had it for a few months. I hadn't run it at all, I think. And I got to say, um very happy with it. I'm only using it for three access work, but the interpolation I'm doing is one hundred percent inspection stuff that's coming off and it's hard milling work.
johngrimsmo: Good.
John Saunders: And I'm just like, it's work that we were doing on the horizontal, which needs to come off both because we're selling the horizontal, but also it's,
johngrimsmo: Yeah.
John Saunders: It is busy right now. So I thought, well, the long-term goal would be to run this on a brother. Let me see how it does on the UMC 400. Cause I kind of went into it skeptical, like, Hey, look, this may not be the right machine for it.
John Saunders: Um, and I'm not sitting here giving anybody a sales pitch other than to say, I'm really happy with the, the, the tool life, the finishes.
johngrimsmo: good
John Saunders: Um, and there's also something to be said. It's nice to be back on in the, well vetted ecosystem happens to be Haas, but like I can go back to using some probing the way I wanted to versus Okuma.
John Saunders: It was always an uphill battle to get stuff to, you could do it, but like, you know, you're doing cause you are stubborn and want to do it. Not because it's like, Oh, this is nice, but sorry.
johngrimsmo: yeah
John Saunders: I brought all that up because the thing I don't like with the UMC 400 is it has, um, almost like this lathe style turret coolant jets that are in the spindle around the spindle housing.
johngrimsmo: Yeah, you kind of aim them.
John Saunders: Um, Yeah, they're little balls that have a little one inch pipe on them.
johngrimsmo: Yeah.
John Saunders: And the amount of residual drip is not only ridiculous, but as soon as you jog and access it, sloshes more coolant around, which causes it to drip again.
johngrimsmo: Mm-hmm.
John Saunders: And at best, it's just inconvenient for looking at your part. But at worst, it's like, well, my hands are getting wet. The part's getting wet after blew it off. Metrology tools are getting messy. So I ordered...
John Saunders: Again, just yeah pick up ChatGPT. And i think this will work. um A check valve, which somewhat counterintuitively, the check valve is allowing flow, of course, in the direction of the coolant exiting the nozzle. But the check valve is a spring-loaded check valve. So if the coolant doesn't have one or two PSI, it shuts off. That should, knock on wood, stop but stop the drip.
johngrimsmo: Yeah, my first question for you is going to be where is the check valve currently?
John Saunders: Yeah.
johngrimsmo: Is there one? I don't know.
John Saunders: There certainly doesn't appear to be one upstream given how behaving.
johngrimsmo: Yeah. If it's too far upstream, then you have you know three feet of hose that has residual coolant that's going to dribble.
John Saunders: o
johngrimsmo: um
John Saunders: Yeah. These are, i believe, quarter-inch NPT fitting, so I'm just going to pop out the existing thing, add the check valve in line, and put them back in like done, be done with it.
johngrimsmo: e
johngrimsmo: Oh, cool, cool.
John Saunders: It's not particularly hard.
johngrimsmo: Yeah.
John Saunders: Yeah.
johngrimsmo: Yeah, and then re-aim them, obviously, because they're going to be lower.
John Saunders: yeah Oh yeah, but they're already, um they're already like eight, 10 inches off. So the angle of that will change is de minimis and they're not even really angled right now anyway.
johngrimsmo: Got it.
John Saunders: So um that's a nice quality life improvement.
johngrimsmo: Yeah, yeah. Cool. Cool.
John Saunders: I'll keep you, I'll let you know if that works. Cause I was first thinking like, Oh, I could do like a little torsion spring with a flapper that would like close.
johngrimsmo: That's what checkbox is.
John Saunders: Well, I'm thinking like, you know, the exhaust on ah on a semi-truck that you see going down that stops rain from going in the tube.
johngrimsmo: Okay.
John Saunders: And then I'm like, Saunders, stop. what This has to be a ah commercial off-the-shelf of Kotz. And I don't think all check valves work this way, though. um I don't know how they are all designed, but you want one that's spring-loaded so that when there isn't sufficient pressure, it...
johngrimsmo: Yep.
John Saunders: Okay. Yeah.
johngrimsmo: Yeah, they're all flow dependent, like they only flow in one direction, they will flow nothing in the other direction.
John Saunders: Yeah.
johngrimsmo: And they have a one PSI or whatever they're rated for break setting, whatever it is. um And an application specific could introduce a flow restriction could whatever, whatever, but it could accumulate with chips if there are fines in your coolant, if it's not filtered super well.
John Saunders: That's good point.
johngrimsmo: So that's just like another restriction to keep an eye.
John Saunders: That's
johngrimsmo: But
John Saunders: a good point. Are you using these for that purpose? Okay.
johngrimsmo: Um, no, on the speedio, as you've seen on your speedio, I and don't know where the check valve is, but I think we put them in the lock line in the two coolant arms, the lock lines that like aim at the spindle.
John Saunders: Yes.
johngrimsmo: I think we added little check valves in the end there, like lock line product.
John Saunders: Uh huh.
John Saunders: Yeah.
johngrimsmo: Um, if we didn't add them, I know we bought them. I don't know where it's at, but it does dribble a little bit, but it's not bad enough to like care about right now.
John Saunders: Yeah. Okay.
johngrimsmo: And on the current, actually, because the check valve, I think it's actually a solenoid fairly close to the spindle. In one of the, um forget what I'm doing, but there's, oh, i'm when I'm pushing putting in the ceramic bearing ball and I'm pressing it in,
John Saunders: Oh, yeah.
johngrimsmo: um I don't want coolant ever to drip on that hole before it presses in. So I actually do like a Z move, like down, up, down, up, down, up, down, up, down, up to literally knock off any stray coolant drops.
John Saunders: missing
johngrimsmo: And we do that.
John Saunders: Please tell me that's like a NC program that's called urinal shake.
johngrimsmo: No, but i like that. um And it works great.
John Saunders: Yeah.
johngrimsmo: works great.
John Saunders: Don't use that program on your willy.
johngrimsmo: Yeah.
John Saunders: Willimon.
johngrimsmo: Nice.
John Saunders: is that either uh carryover note about detent dropper is that the same thing
johngrimsmo: ah Different issue. So we, ah many years ago, you and I talked about this and we kind of put it out into the world and a few people reached out with ideas and one guy sent me a prototype.
johngrimsmo: um Welles Design Co, guy in Portland-ish, I think, um made me one. And it's, the design of it is a totally different thing than I'd ever considered before. It's like fairly simple, fairly mechanical. um And we installed immediately and it started working.
johngrimsmo: And then I gave him some feedback and three weeks later I had Rev2, which was like way better. And it's amazing.
John Saunders: Can you remind everybody what it does?
johngrimsmo: So, it So let's see, in a knife, what keeps it closed when the knife is closed is a little bearing ball that is spring loaded and it kind of falls into the blade and keeps the blade closed. Most knives work this way. It's a detent. Installing that bearing ball, the height of which is, I found is absolutely critical. Like we want it within a 10th or two of repeatability.
johngrimsmo: And doing it by hand has problems and it's time consuming. And like in the early days, you literally squish it. You have your handle, you have, you push the ball in with tweezers and you have a shim with a hole in it.
johngrimsmo: That's your height. And you squish it in a vice.
John Saunders: Mm-hmm.
johngrimsmo: You squish it in a vice and it works okay-ish, but, um, Basically, I want to do this on the machine. For years now, we've had the machine drill the hole and then wait for us. And we go in with our fingertips and we put in the ball and then the machine pushes it in The probe comes in, probes the surface, probes the ball, it says it's actually 28.
John Saunders: Mm-hmm.
johngrimsmo: We want to be 26. So then come back and press it again, another three thou. And then after that, probe it again and confirm the difference between surface and top of ball. um and then print the result. And that's been working great, but I'm like, i don't want to sit here every, I want to be able to automate this and just have it run.
johngrimsmo: So this detent ball dropper has a hopper. Um, the way Brent made it was with 200 bearing balls inside. Um, And it automatically drops one into the chute.
johngrimsmo: And then there's a ramrod pretty much that spits it down and presses it as you're pressing it onto the surface.
John Saunders: Okay.
johngrimsmo: So it like presents one ball in the way of the ram and then it presses it in. And then it literally presses it into the handle automatically.
John Saunders: Cool.
johngrimsmo: And now you have a half press automatically done and it's reloaded for the next time.
John Saunders: Yeah.
johngrimsmo: um He made us two. Over time, everything breaks. um Sometimes a ball explodes, sometimes a ball flies off to the side. And this is very tight tolerance, little finicky little things. So we've relatively destroyed the first one. The second one's on its last legs.
johngrimsmo: It failed again the other day. And we thought we were kind of hooped. We're like, oh no, what's going on? And then we were able to use some of the parts from the first one into this second one and like get it running again. So the big panic for me that I mentioned it last week was ah I thought we were dead in the water without property, didn't it but we actually have a functional system.
John Saunders: Oh, interesting.
johngrimsmo: But I will be reaching out to Brent again. And I owe him some goodies as a very gracious thank you, because this has become literally a part of our process now.
johngrimsmo: And yeah, I'll be sending him a little care package. And whether we make the next ones ourselves, because I've got some tweaks to make, or whether he's willing to make another one i don't know yet.
John Saunders: Yeah.
johngrimsmo: um I don't want to assume that I can just keep leaning on him. But he did send me the CAD files at one point, but I can't find them anywhere. So I need to like politely ask for them again.
johngrimsmo: okay But yeah, otherwise, it's it's been the coolest thing ever.
John Saunders: what
johngrimsmo: 1 16th.
John Saunders: What's the diameter of the ball bearing? 60,000. Okay, so it's pretty small. Like it's not like a, you told me ball bearing, I'm gonna picture a marble.
johngrimsmo: Very
John Saunders: This is quite small.
johngrimsmo: No, it's a, yep.
John Saunders: And this whole this whole thing that was designed runs, is it um part of like, is it integral to like a tool holder that gets tool changed in and out?
johngrimsmo: It's got a half inch shank and it it is in a RegoFix tool holder. um
John Saunders: Okay, yeah, so it moves, the those ball bearings,
johngrimsmo: It's a tool change.
John Saunders: come in and out of the machine with the tool changer.
johngrimsmo: Yeah, yeah, it's it's super thick and it's in the current, yeah.
John Saunders: That's insane. It's Kern though, or is it the Mori? Okay, yeah. that's it So it never has to go sideways. I don't know if that matters, but yeah.
johngrimsmo: Correct. It could.
John Saunders: Interesting.
johngrimsmo: Yep. And then because machines have general mist and general oil that kind of mist around and gets everywhere, um it does get gunked up every now and then.
John Saunders: It's really cool.
John Saunders: no
johngrimsmo: So we have to like flush it out with IPA and, you know, maintain it, keep it clean.
John Saunders: Yeah.
johngrimsmo: But...
John Saunders: Yeah.
johngrimsmo: It's, I've always wanted to do like a full deep dive video on this concept and show it and show off Brent's ingenuity. And it's, I just haven't yet.
John Saunders: Yeah.
johngrimsmo: Cause it's really cool. I'm, I'm tickled with it.
John Saunders: That's awesome.
johngrimsmo: yeah And I had all, I had all these designs that had like a little sliding lever and all these little micro parts and his is just, it's brilliantly simple.
John Saunders: Oh, and avoiding.
John Saunders: Isn't that the the best when you're like, oh yeah.
johngrimsmo: yeah I couldn't have come up with it. Like,
John Saunders: Yeah, that's awesome. Okay. Sorry. i I stole your thunder though with a carry for what's, so what's your update on speedio?
johngrimsmo: um Okay, so the something I've been working on for a few months now is like, you know, tracking the thermal growth of the machine, tracking what happens, tracking the wheel where is the laser accurate, all these things.
John Saunders: The rules are automated.
johngrimsmo: um And then the latest issue is I've got five fjell blade pallets. So when we're grinding fjell blades and there are inherent differences in the width of each pallet, these are tapered surfaces. They're kind of hard to probe.
johngrimsmo: They're like undercut tapered.
John Saunders: Yeah.
johngrimsmo: um
John Saunders: Oh, yeah.
johngrimsmo: But I figured out a way to do it. And now I'm comping for each palette in situation. And that adds to a modifier. And then there's a manual modifier that we can come in if we find the whole thing is off, like generally off by a thou.
John Saunders: okay
johngrimsmo: So when we're grinding the bevels, we want accuracy and repeatability is the most important thing. And... Angelo just hit go. So not only have I made it working, I've now handed it off, like the new process to Angelo. And he just hit cycle start for a, i don't even know, like six hour run of totally unattended, like all hard milling hard and hard grinding.
John Saunders: And so all five fixtures use the same, assume zero point coordinate system.
johngrimsmo: Yeah. Yeah. It's the Arroa base.
John Saunders: And so then, yeah. And so are you then, you're doing a modifier that could say change up to three axes, but then it resets it back each time or what?
johngrimsmo: It does, yeah. yeah
John Saunders: Okay. I like that.
johngrimsmo: Basically, I have my default G59, which is the center of that work coordinate system. And then I have whatever it is, G54.3 or something like that.
John Saunders: Mm-hmm.
johngrimsmo: That is for side one.
John Saunders: Mm-hmm.
johngrimsmo: And then 0.4 is for side two of the side size of the fixture. And I ran a test last night. I was here till 2 a.m. last night. And I hit go. And I came in in the morning. And they they all worked. like They all work great.
John Saunders: Good. That's great.
johngrimsmo: Yeah, so it was a lot of testing, tuning, math, a lot of theory, and trying to find all the little variables so that it just works.
John Saunders: That's awesome.
johngrimsmo: you know And we've been doing this on the Speedio for like two plus years now, more manually, one at a time, trying to chase our tails.
John Saunders: Mm-hmm.
johngrimsmo: And I'm like, everything we're doing and trying to anticipate is logical. It's all there. it's relatively doing the same thing every time. Just what are all those variables? And it's, it, the the rabbit hole is a lot deeper than I thought it would be.
johngrimsmo: Um, And then not only that, that's for the side bubbles. We also have a blade on top of the fixture for hard milling the critical features, the the pivot hole to the exact diameter, the slot where the stop pin goes. So we're we're slotting out hardened steel with a high feed end mill and doing a lot of critical tickling machining where we're just like, I want to hold one tenth.
johngrimsmo: on the speedio and i am both Z depth and also in X, Y, and I'm deep printing all the results and I'm tracking it in graphs automatically. And it's just, I'm tickled pink. It's the coolest thing ever.
John Saunders: Yeah, awesome.
johngrimsmo: Yeah.
John Saunders: Good.
johngrimsmo: so now it seems to work. And then as Angelo keeps reminding me, it's like, what's the failure mode and how do we recover from it?
John Saunders: o
johngrimsmo: you know, if we have to rework something, is all this probing like going to get in the way? Can you start a program in the middle or do you just have to air cut it from the beginning? Like what's the safest way to do it? And I'm like, I don't have all the answers for you right now, but we will figure that out over time.
John Saunders: Right.
John Saunders: Yeah. And it's funny because I, the more we go through this journey, the more I'm like not interested in building, putting extensive amounts of time and effort and tribal knowledge into rework process procedures I'm not saying the Angela's wrong for sure. and And we've had to do rework before every, every shop does and will, but like, man, I really, um, I'm the like, I'm the way we're coming, building up automation now is like, I'm much more fine with it. Just stopping when it can detect something real time versus like, okay, let's know that we have x amount of parts to rework.
johngrimsmo: Yeah.
John Saunders: Um, you
johngrimsmo: Yeah. So what's your speedy automation point?
John Saunders: We had a phenomenal hurdle reached and coincidentally two things happened on the same day. Pure luck. One, the Speedio with the Lights Out manufacturing, UR10 system is working and running parts.
johngrimsmo: The day you guys turn the lights out is going to be a funny day.
John Saunders: We start. but
johngrimsmo: Yeah.
John Saunders: We're only running during the day right now, but it made, i think the first test and hurdle was five in a row, which doesn't sound like a lot, I get it
johngrimsmo: Yeah. Sure.
johngrimsmo: Oh, there it is.
John Saunders: But we started with probably the hardest part, certainly the hardest part from a tolerancing standpoint, because it just made sense to say, if we can do this, then the rest will be fine.
johngrimsmo: Yeah, yeah.
John Saunders: There are times where I'd flip it around and say, hey, let's build baby steps, get some wins under our belt, and then tackle the hard thing. I still think we did it. did it right doing it this way. And the biggest thing was was the OP2 pushing the part down with the robot on a a three ball, three carbide, I think it was just hardened, excuse me, not carbide, balls, and then actuating a vise and not having the part lift and able to achieve a ah spec that we're looking for that way in five in a row.
johngrimsmo: Amazing.
John Saunders: And they they're they're beautiful. And it just it just ran. so of course it's
johngrimsmo: Yeah, I remember you talking about that that ball theory a few weeks ago.
John Saunders: Yeah. And they're they're working.
johngrimsmo: That's awesome. Yeah.
John Saunders: um And it will fail or hiccup. We we know that. But it's a huge step. The same day that that happened, the Willowman, we came in the morning, the Willowman was still running.
johngrimsmo: No.
John Saunders: Yeah. And, you know, I'm a little bit defensive because we should have, we could have, or should have had that happen earlier in some respects, but like, no, no, like whatever it happened and it's awesome.
johngrimsmo: No, fine.
John Saunders: And it is, fact, it was really funny. I ended up coming in early to get a head start on some parts and, I was sitting at my portable desk out by the UMC, which is kind of adjacent to the Willem.
John Saunders: And it's like a seven 30. I looked behind I was like, I think the Willem is just moved. i'm like, whatever, don't like focus on what you need to do And then grant came in.
johngrimsmo: focus
John Saunders: He's like, the Willem is still running. I'm like, no way. Yes. That is awesome. And it's just one of those, I think it was plug jacks, but I don't know.
johngrimsmo: what part what What part are you making on it?
johngrimsmo: Yeah.
John Saunders: And,
johngrimsmo: Like a relatively easy, like stable part.
John Saunders: a
johngrimsmo: That's my question, I guess.
John Saunders: No, it's a pretty tricky tolerance part, to be honest with you.
johngrimsmo: Yep.
John Saunders: Yeah. um The only part that we make that I would say is is like, just don't care all day long, no matter what. You you know you could put a nice chest next to the machine and be It would be the Modvice washers. Those have no criticals.
John Saunders: But most everything else has ends up having our half-inch... tolerance, which is usually, think it's like 5005 plus or minus two tenths on the, it's a turned feature. Sometimes we interpolate it sometimes we turn it or mix of the two.
John Saunders: um But the fact that it was just like, it's just one of the things like you could talk about it, you can hear about it. When you leave work and you come the next morning, it's still running.
johngrimsmo: Yep, I knew that would happen to you one day and I'm so glad that's actually you're there now. It's um amazing.
John Saunders: Yeah. So now that the sort of thing, the debate is, um we hit this awesome milestone with a brother and it's currently running the adjustable side of the mod vice.
John Saunders: I'd like to let it, you know, stretch its legs. I'd let it, like feel like to let it go to work, but we're actually kind of hitting pause and moving over to get the next part set up and running, which feels kind of like a, Oh man, I wanted to let it, you know, start itself, but it's a good thing.
johngrimsmo: to Yeah, yeah.
johngrimsmo: Yeah. Yeah. And that's where I'm at with the speedio now is like, I need data to make sure that, you know, I've made one part. Great. I made three parts overnight. Great. Let's make 10 in a row and see if they're all the same.
johngrimsmo: See if, see if there are any more variables that I'm not anticipating that I'm not seeing any failure modes, any, you know, chips under the pallet or whatever, the washdown is not good enough, or somebody forgot to tighten a clamp, which doesn't really happen anymore.
John Saunders: Yeah. Right.
johngrimsmo: Um, And the problem I'm going to run into next is tool life, like predictable tool life. Because if I'm running for, you know, if you're running one part at a time and it's, you know, 10 to 40 minutes, you'll see the life warning pretty much before you run out of tool life.
johngrimsmo: And then if we're running 6, 10, 12 hours in a row, you're going to like run out of tool life mid-cycle. And unless you have a way to predict that or to manually look at the tools every time before you hit cycle start and run your own math, which is probably the short-term solution,
John Saunders: yeah
johngrimsmo: um you' just it's going to stop mid-cycle because you could be like, oh, I'm out of life. You should have replaced tool one yesterday.
John Saunders: You have that predictive tool life on other machines, just not on the brother.
johngrimsmo: On the current I do, yeah. yeah
John Saunders: Okay. thought you had on the Mori, too.
johngrimsmo: It just has... um
John Saunders: Years ago.
johngrimsmo: Kind of, I do. It's sloppy, but it yeah, I do kind of. um It just hasn't been a priority on the Speedio yet, but it was like now now that automation is good, I will figure out how to tackle that next.
John Saunders: Okay.
John Saunders: Yeah, sure, sure.
johngrimsmo: Either I go full nerd with like an external Raspberry Pi kind of calculation system like I have on the Kern, which has been absolutely phenomenal, or Angelo and I were just talking about it. He's like, dude, just give me the numbers.
johngrimsmo: and say, these are all the tools and these are all the cycle times of that tool on this program. And we can manually calculate this for now. We like, okay, we have six hard blades, tool one, six minutes each, six times six is 36 minutes.
John Saunders: Yeah, sure.
johngrimsmo: I need 36 minutes to run the cycle. Oh, okay. I have enough. And you can just do that. should take the operator like minutes for now.
John Saunders: Yeah. Yeah.
johngrimsmo: So that's what we'll do today.
John Saunders: Dude, awesome.
johngrimsmo: Yeah, it's good.
John Saunders: That's huge.
johngrimsmo: It's been like, mm. probably close to 100 hours in the past few months of me like working on it.
John Saunders: Yeah.
johngrimsmo: But now it's good.
John Saunders: It's there.
johngrimsmo: Yeah.
John Saunders: Yeah. Yeah. And then it's like, now you and unceremoniously move on to the next, to the next thing.
johngrimsmo: Exactly.
John Saunders: Yeah. That's awesome. Um, we had uh, you can talk about the extremes. The, the two nights ago, the Willman ran overnight last night, Grant texted me. He's like, Hey, are you swinging by the shop? He's like the Willman doing something weird. So I, I was literally driving by. I popped in, don't know why, don't even care at the moment, but it was just frozen. Like it was,
John Saunders: Spindle on, in the cut, coolant or oil on, and it was just not moving. And Grant's like, just hit reset. I'll figure out what happened in the morning.
johngrimsmo: How did he know?
John Saunders: Chatter, actually. Chatter told him. Er, he checked it via chatter.
johngrimsmo: okay cool. Got it.
John Saunders: And we have a web we have a webcam on it as well.
johngrimsmo: Nice.
John Saunders: um But I was swinging by. No problem for me to hit reset. But um I was like, there's no reason we should have to do that. So I just ordered the SwitchBot IoT Switch. Have you seen one of these?
John Saunders: but like a little It's like a little, it's the size of a half the size of a pack of cigarettes, but it has it like in Velcro or suction on and it's an IOT and it has a little plastic thing that protrudes out.
johngrimsmo: Googling now.
johngrimsmo: What?
John Saunders: It can flip like a light switch or it can flip, push a button. um And so we're just gonna put that over the reset button.
johngrimsmo: No way.
John Saunders: So Grant on his iPhone from anywhere in the world can just hit reset to reset the machine remotely if we need to.
johngrimsmo: Dude.
John Saunders: Right?
johngrimsmo: That's cool. Yeah, we've taught you've talked about that for years. I just want a little little button to just hit this one button.
John Saunders: Yeah, there's a whole argument of like, i kind of want like remote control over everything, but, um you know, I want to go see the alarm state or I want to even hit cycle start.
johngrimsmo: Of course, but
John Saunders: But those are all obviously much bigger cant can of worms.
johngrimsmo: yeah.
John Saunders: This is this is a no brainer.
johngrimsmo: Yep. That's cool, man.
John Saunders: Yeah.
John Saunders: What's next on your list?
johngrimsmo: I had a guy come by the shop on his way up to Ian Sandusky's shop, Lakewood Machine. but
John Saunders: Oh yeah, Brad, Brad, Jacob.
johngrimsmo: Yeah, Brad, Brad Jacob.
John Saunders: Yeah.
johngrimsmo: He stopped by and he said, say hi to John.
John Saunders: Yeah. Look Brad. Yeah.
johngrimsmo: And he yeah, he said he was at your open house in 2017 and gave a talk on machine repair.
John Saunders: Brad's a good first good dude
John Saunders: dude. Yeah. Can't believe that that was nine years ago.
johngrimsmo: Nine years ago, exactly. Yeah. Yeah, so it was great to show him around the shop. He's got his own shop. um He's building up his from side hustle to now full time machine shop.
John Saunders: Mm-hmm.
johngrimsmo: He's got a Tormac lathe and a Haasmill, I think. And it's just crushing him and his wife. Like they work together in the shop and they both machine parts. It's beautiful. And he's re-listening to the podcast from the beginning.
johngrimsmo: And he said he's on episode 80, something like right now. And and like you and I 2018. ah So that's where he's at in the in the chronology of this podcast.
John Saunders: That's interesting. Oh my God. Did we start it in 2018?
johngrimsmo: Yep.
John Saunders: I guess I don't think about the math that way.
johngrimsmo: No, before that, before that. No, we started in like 16, maybe 15.
John Saunders: Oh my God.
johngrimsmo: Yeah.
John Saunders: Wow. Well, 446 divided by my math weak. Divided by 52 is 8.5 years. divide by fifty two is eight point five years So yeah. Holy cow
johngrimsmo: yeah And he has recently heard about the Linda wheel. So he's sitting in my shop. He's like, do you still have it? I'm like, yes, I do. You want to see it?
johngrimsmo: So I showed him the Linda wheel and I showed him how we do it now. And I explained it and I put it under the microscope and showed him all the details. And he's like, I mean, I had no idea. Like hearing it on the podcast is one thing, but seeing it's another. um
John Saunders: These things become larger than life, John.
johngrimsmo: So that I know apparently, apparently.
John Saunders: It's really funny, right?
johngrimsmo: So the the seven people out there that remember about this Linda wheel saga.
John Saunders: Yeah.
John Saunders: Of which Linda knows nothing.
johngrimsmo: Exactly.
John Saunders: ah Well, did you see what happened to Lockwood?
johngrimsmo: No.
John Saunders: Rob? Maybe, i don't know if he DMed me or was in the group chat. He was like, he like messaged me. He's like, wait, hold on. And apparently his like podcast player started the bomb from the beginning. and he's like, this does not sound good.
John Saunders: And none of this makes sense.
johngrimsmo: Yeah, a
John Saunders: I don't know what's going on
johngrimsmo: it doesn't matter. okay
John Saunders: on. Yeah, good. Brad's a good dude. I believe his shop is called Maple Lane, if that sounds right.
johngrimsmo: Yeah.
John Saunders: Not to be confused with Maple Leaf. Yeah.
johngrimsmo: Yeah, was really nice. And he bought a fuel.
John Saunders: Oh, awesome. That's cool.
johngrimsmo: It was really great. He's like, so do you have any knives I can buy? I'm like, oh, yeah, absolutely. Come into our showroom.
John Saunders: Awesome. But that made me think you never talk about the Mori. Is it just still the Dura Vertical just still doing its thing?
johngrimsmo: Crushing.
John Saunders: Yeah, it's awesome.
johngrimsmo: Yep. it ah What was the problem it had the other day? It was minor, but oh, the guys had to do like an ATC spindle oil flush, something like that. We've been putting it off for about six years and it's going great. So it does Norseman handle top sides and Norseman hard milling blades.
johngrimsmo: So the the lines on the Norseman blade, it's all hard milled on that machine.
John Saunders: Yes. Yeah.
John Saunders: Got it.
johngrimsmo: Yeah. So it probably runs four hours a day, give or take, but pretty steady.
John Saunders: Got it.
John Saunders: Yeah. Sweet. Sweet.
johngrimsmo: Yeah.
John Saunders: Okay. I have a very funny transition that goes directly from my garage door project at home to the Okamoto surface grinder.
johngrimsmo: I told my wife about that.
John Saunders: are you ready for this?
johngrimsmo: Yeah.
John Saunders: It's a actually a great example of like the pitfalls and perils of of the current state of all these automation stuff. So I bought a SwitchBot garage door opener, which has both the relay that just basically fakes pushing your garage door button.
John Saunders: So that opens or closes it. But of course it doesn't know whether it's opened or closed.
johngrimsmo: OK.
John Saunders: It just triggers it to fire. But then it also came with a small relay, excuse me, a small read switch, which will then tell you whether the door is open or closed, which is actually quite helpful because you don't really want to build an automation routine to open the door if you're kids your kids or wife had already opened it for you or whatever.
John Saunders: um That SwitchBot app, works as an app, but I don't want that. I want it to be run through home automation. And that happens through a protocol called Matter. Now this is where I start to quickly become a beginner or novice at best, but Matter is this new protocol that has a bunch of momentum behind it, which is a platform or a way for these devices from different manufacturers to talk to home automation.
johngrimsmo: Yeah.
John Saunders: And what's nice is it does not require you to own a physical hub like Alexa or SmartThings, et cetera. Matter doesn't expose, or sorry, Switch bot those turds don't expose the state of that switch through their matter API or their matter Lingo so I can control this the garage door via home automation easy already Sorry be a home assistant already, but I can't home assistant can't see whether the darn garage doors open or closed and so that doesn't work so I bought a Third-party relays read a switch actually gonna get my tongue untied here
johngrimsmo: Hmm. chi
John Saunders: um You know, kind of annoying. It's another 20 bucks. It's another device, but not a big deal. um So I'll install that third party reed switch on the garage door. Home assistant sees that directly. And now home assistant can easily sort of say, okay, if the door's open, don't open it. But if it's not open and the car's coming home, open the thing. Okay, done.
John Saunders: Here's the eureka moment. I'm running the Okamoto grinder this morning. It's on the far corner of the shop. You know what's happening? I'm ordering read switch today. And when the table parks itself in the far left and it stays there for more than you know five seconds, that means the grinding cycle is done and I get it i get an email or alert or a light, whatever you want.
johngrimsmo: The cycle's done.
John Saunders: Then then the the world is your oyster.
johngrimsmo: Okay. There you go.
John Saunders: that's That's going to be awesome.
johngrimsmo: Okay. And a read switch is just a little magnetic actuated switch, right?
John Saunders: Yeah.
johngrimsmo: Yeah.
John Saunders: Yeah. Yeah.
johngrimsmo: How long is a grinding cycle? But it's like you walk away from it.
John Saunders: Oh yeah, it's it's probably, i don't even know, seven to 15 minutes.
John Saunders: What's nice about those read switches is that they usually will actuate if they're within about an inch of each other.
johngrimsmo: Okay.
John Saunders: So you don't have to have a, I don't want, you know, contact switches or or proc switches that need insane proximity to the other two.
johngrimsmo: he
John Saunders: So that's just a nice quality of life thing.
johngrimsmo: Sweet.
johngrimsmo: Yeah, the thing with all these automation and and Raspberry Pis and ESP32s and all this is... There's so much you can do, but you have to you have to like focus your laser. you know You have to make sure you're doing worthwhile things. Otherwise it's just all fun, kind of pointless, useless stuff.
johngrimsmo: um But some of it is amazing. So I'm finding myself pulling back like, wow, I could do that.
John Saunders: Yeah.
johngrimsmo: Yeah, but there's no reason to do that.
John Saunders: That would be a great segue to my AI agents a speech, but it's your turn. So what's next on your list?
johngrimsmo: All right. I will say, be careful what you tell me, because last week you're like, we were talking about the interferometer and you're like, dude, I figure within like, within like a solid day, you could probably have this wrapped up. Right.
johngrimsmo: And those words were stuck in my head and I put something like that.
John Saunders: Did I say that?
johngrimsmo: You're like, it doesn't sound like you're that far away.
John Saunders: Okay.
johngrimsmo: And so I worked on it quite a lot over the past week, which has been super enjoyable. And I guess, thank you for that, for the little push. Because i can I can slap something together that will function great, but is not the final rev.
johngrimsmo: um But it will it will answer a lot of questions. Or I can try to design and build like the ultimate, the final like final design, which will take longer and cost more and things like that.
John Saunders: Okay. Mm-hmm.
johngrimsmo: So I decided to parallel process and to kind of start ordering the parts for the final design, but figure out how to slap it together on my current division system microscope that I have. And so the design is there.
johngrimsmo: um All the parts in fusion are in place and I have all the components. Now, last thing I really need is just to like design 3d printed housing to kind of hold everything where it is in fusion.
johngrimsmo: Um, which is kind of not that hard and then so i've got like a print there and i've got i'm waiting on another small order from thor labs to come with a precision pinhole um which would be ah a critical thing so it's like a little one inch disc with a 50 micron hole perfectly circular hole in the middle which will be important
John Saunders: Oh.
John Saunders: Yeah. So like alignment?
johngrimsmo: um Currently I'm using a piece of aluminum foil with a pinprick in it, which functions, but it gives me a messy fringes. um
John Saunders: Got it.
johngrimsmo: Apparently the precision pinhole will just give me crisper fringes, easier fringes, easier to find, things like that.
John Saunders: Interesting.
johngrimsmo: So that should be here tomorrow and build that into the system. And once I have this 3D printed thing, I should have a perfectly functional white light interferometer.
johngrimsmo: I'm still going to work a bit more on the software side, but Yeah, it's cool.
John Saunders: Yeah.
johngrimsmo: And then it's it's super exciting.
John Saunders: Dude. Awesome.
johngrimsmo: And I just can't wait to start sharing the results of this, like like the scans that I'm going to do. Like not only when, um let me think, I think when Cyrus showed some scans, not only can you see the feed forward lines of a machine surface, like where each flute is hitting the surface,
johngrimsmo: you can see the grind from the flutes, the animal manufacturer.
John Saunders: Dude. insane john insane yeah have you been posting about this yet at all like on insta okay
johngrimsmo: Yeah, which normally we can't see with a microscope. um So yeah, to say I'm excited is an understatement.
johngrimsmo: Not really, no, not the interferometer. I've posted about the vision system a little bit, which is your backlit, you know, black and white contour outline circle thing.
John Saunders: yeah yes the watch gear
johngrimsmo: Yeah, the watch gear. um I don't know if you saw the story I put up two, three nights ago of of showing that system in use real quick. You might have missed it, but missing out, man.
johngrimsmo: No, it's I have a little tiny.
John Saunders: you yeah Yours yours is the stuff I want to see, but I'm sorry.
johngrimsmo: I know. i know. That's OK. This I chose to do a story and just let it disappear, whatever. So I have a little tiny um circuit board and I don't need it perfectly modeled in CAD, but I'm like, my vision system is on right beside me at my desk at home. And I'm like, let me just put it on here. It perfectly fills within the field of view. I can't see much bigger than that.
johngrimsmo: And I was able to scan line to line and it auto snaps the square of the sides of the board.
John Saunders: Sick.
johngrimsmo: And I can hit three points on each circle, of the through hole of the circuit board, and I can automatically determine the circle size.
John Saunders: Sick.
johngrimsmo: So I know each circle is 21,000ths of an inch and all the circles match within about a tenth or two. And this is calibrated, so it works. And I know the big mounting screw holes are 89,000ths of an inch and they're both within a tenth or two each.
johngrimsmo: And I was able to kind of... measure that and transfer it into Fusion, and then even take a screenshot of my thing, because I know it's a telecentric image, it's perfect, it's not distorted, it's anything like that, and then bring it into Fusion as a canvas, lay it over my CAD model, and if I mess with the lighting and the exposure and stuff, it's no longer just a black and white silhouette.
John Saunders: Yeah. Yeah.
johngrimsmo: I can actually see the features of the part with a little bit of top lighting, so I can see the resistors on the front, I can see the text on the thing, and now in Fusion, I have a canvas with my
John Saunders: Oh, pull out that stick.
johngrimsmo: like it looks like the circuit board
John Saunders: John bravo. Well done.
johngrimsmo: it is so cool
John Saunders: That reminded me, I did go through and watch parts of the diffraction limited video on his micromanipulator and scene accuracy. And I'd forgotten one of the the biggest flexes is he very, just like cavaliers, just like, okay, let's prove this by putting a integrated circuit on this micromanipulator. And he has the appropriate vision lens, whatever to to look at the, um what would you call it? The, uh,
John Saunders: the ASML, like the, like the serial lithography etching on this transit on the integrated circuit. And he's moving the thing around and showing that it repeats on these transistors, which are, you know, these are the things that there's 30 billion of on it.
johngrimsmo: Insaneless.
John Saunders: I mean, just like unbelievably cool. So anybody needed a nudge to actually watch that, that's like, yeah.
johngrimsmo: Yeah. if If you haven't seen it and you're interested in anything we're talking about, yeah. Diffraction limited micro manipulator. Just watch. It's like short 10 minutes, 20 minutes, something like that. it's
John Saunders: Yeah.
johngrimsmo: It's just enjoyable.
John Saunders: Well, that's my, that's where I feel like I'm left out. Like it's like hardcore FOMO. Is that like, yeah I've heard you talk about this so much. I don't know what it looks like like.
johngrimsmo: mean
John Saunders: I want to see you've gone so far compared to like, like bring us along.
johngrimsmo: easy
John Saunders: I know you're sharing it with me, but like, I want to see more of the John, but if you don't want to share it, I totally get it.
johngrimsmo: Yeah.
johngrimsmo: Yep.
John Saunders: um
johngrimsmo: Yeah. For the vision system, I have been filming YouTube video about how I've built it. And then for the interferometer, I'm absolutely going to film a progress report like a YouTube video.
John Saunders: Yeah.
John Saunders: Yeah. Sweet.
johngrimsmo: Yeah.
John Saunders: um Okay, AI agents, i'm going to keep this pretty quick because it right now, go ahead.
johngrimsmo: I don't really know what an agent means.
John Saunders: That's all I'm explaining.
johngrimsmo: Yeah, please.
John Saunders: Yeah. So, um and I did some research and I thought about, Hey, how can I present this to our audience in a way that really shows like how awesome AI agents are? Cause I hear all this buzz and hype and there's a lot of YouTube videos that you can watch. But I've come to the conclusion right now, these are very interesting. They're very powerful for many folks. I can't think of legitimate reasons.
John Saunders: They're useful for me or even you.
johngrimsmo: Interesting. Okay.
John Saunders: so So an AI agent, Also, the term that I had to laugh at being agentic, fancy word agentic is like acting like an agent. So I had to like, I had to click with me.
John Saunders: um
johngrimsmo: Oh, agentic agent. Okay.
John Saunders: that Yeah, I'm glad that you had the same kind of like, okay.
johngrimsmo: Yeah. Yeah. yeah
John Saunders: Okay. um And I wanted to come up with simplest example of what an agent is or could be.
johngrimsmo: pretty
John Saunders: So right now we all know what ChatGPT is as in a large language model, LLM. Like I can go into ChatGPT and I can ask it just about anything. For example, I can ask it, hey, write a one paragraph essay on today's weather in Zanesville or Toronto or wherever.
John Saunders: And it would do that. Here's what an agent would do. you could kind of build or program or set up an agent. There's different tools and WYSIWYG and ways to do it, where you can say every Monday morning, I want you to look at the next five days of weather in both Zanesville and Toronto.
John Saunders: And I want you to summarize that in the bullet points. and I want you to do that using, um, using ChatGPT, I want you to create that summary. Then that summary I want you to pass into Claude, I'm making this up. In Claude, I want it to write a long form essay on what the weather will be like. Then I wanna pass that essay from Claude into Gemini. I want Gemini to write a LinkedIn you know Business Pro snippets that's LinkedIn savvy.
John Saunders: And then I wanna pass that into Copilot to proofread it and see if it has any edits that would help make it social media and SEO friendly. And then it will rerun that thing and it could even potentially automatically post it. It could rerun that every week.
John Saunders: So just think of it as autonomous,
johngrimsmo: That's crazy.
John Saunders: Yeah, and there's like way better examples with like booking flights or creating content, but I want to give a simplest example.
johngrimsmo: yeah Sure. Sure.
John Saunders: It's basically, ChatGPT right now requires me to prompt it. An agent, and there's a question, I guess, online about how autonomous an agent needs to be to be an agent, but you could obviously see, and the closest thing I could think of for Saunders would be you could feed in um Shopify and ShipStation and Freshdesk in every week, it could write a report and summarize it and pass that summary into Copilot.
John Saunders: They could create an Excel file and it could automatically start creating KPM, whatever information.
johngrimsmo: KPI is here.
John Saunders: And that could be real, but I'd be doing that for the sake of trying to prove its usefulness. So for me right now, it's a pass other than I want to keep learning about it. But if other people are using it in more cool ways that are practical, I'd love to learn because I am curious and the the power of it is no doubt incredible.
John Saunders: Yeah, that's my spiel.
johngrimsmo: Interesting. And that's, is that different than ah like the guy up in Edmonton, confounded machine. He said he built his own AI on a local PC.
johngrimsmo: I don't know what that really means, but.
John Saunders: I don't know that example. I do know, and I've looked into this for Johnny Five, thank you for the segue by the way, which is, there are ways you can sort of download kernels, models, things to low, to run it without needing like the internet.
johngrimsmo: Yeah. Yeah.
John Saunders: um I don't think people are like rewriting their own AIs, but sure, sure.
johngrimsmo: No, but yeah. but But have it it. It sounded like he wanted it stored locally so that for all of his hundred podcast episodes that he's recorded, it can it has them right there. It summarizes. and But without internet, anything like that, I don't really know. Yeah.
John Saunders: Yeah. Actually, might you may have just said the best agent Like you and I are probably, i don't want to speak for you, but like we're kind of behind the curve. I have no interest in creating additional revenue for Saunders by automating AI.
John Saunders: And I don't know why, it's just a personal decision. It just feels scummy, but I think I'm naive to how much that's happening. So you and I could have an agent read the, listen to the bomb and create all this social media stuff, push it to Reddit and social and LinkedIn and promote it.
johngrimsmo: Yeah.
johngrimsmo: Mm
John Saunders: I'm just, it's not me.
johngrimsmo: hmm. Well, Zencaster, the platform we're using, does auto generate, I think, up to eight or ten video clips from our thing.
John Saunders: Uh-huh.
johngrimsmo: so those are the things that we post online. um And you can go in and tweak them and be like, oh, that's a little cut off. Let me extend it by five seconds or something like that. But for the most part, the clips are like fine.
johngrimsmo: They're like good enough. So you're saying an agent would take those clips and like do something, like purposely post them to different platforms and like take next actions like a like a flight agent would, like a travel agent.
John Saunders: Yeah, and it could create, and like extrapolate.
johngrimsmo: It would do something.
John Saunders: Again, you think about like Instagram has a different voice than LinkedIn and um it could write it could create content and posts that were were specific for the right audiences.
johngrimsmo: So, I mean, chat GPT will tell you answers and we'll even like write code and do some stuff for you, but it sounds like an agent does stuff for you. Like we'll, we'll do your work.
johngrimsmo: We'll do more things. We'll integrate a lot of different aspects, lot of different external programs and websites and apps.
John Saunders: Yeah.
johngrimsmo: Okay. Interesting.
John Saunders: Even being able to, you know, the idea of an agent being able to take the output from something, from an from a LLM or somewhere else, and feed it into a different LLM for a different point. is i will i will respect that as being powerful. It's like, hey, check this, proofread this, like cite this, find out what could be wrong on this. That that that is intriguing.
johngrimsmo: Isn't that ah Obama patting his own back? That meme? Like when he gets the the metal?
John Saunders: but with When he's giving himself the, the yeah, yeah, that's how that's hilarious.
johngrimsmo: Yeah, yeah.
John Saunders: Yeah, right, right. Yeah.
johngrimsmo: The AI checking the AI's own work kind of thing.
John Saunders: Right. Make LLMs great again.
John Saunders: um My humble brag, a big thank you to everybody. I mentioned it last week on the podcast and then we threw the Instagram clip up. It it is by far, I think we just crossed 3000 likes, which like I hope everybody knows I'm not one to like pat myself on the back on Instagram likes, but it is fun to see the passion from folks to see Johnny five running is frigging awesome.
johngrimsmo: Yes. Yes.
johngrimsmo: yeah yeah good for you man it's it's really cool to see you know yeah from a a bucket of bolts to uh yeah
John Saunders: Thank you. I appreciate Yeah. Yeah.
John Saunders: Yeah. And to be able to step back and watch him um as the, as the create, I mean, obviously to be clear, I'm not the creator of Johnny five. That was the Eric Allard and the team at, in the movie, all in the movie, but like to be, to build a replica, you get lost of the weeds and then to step back and watch him work gives you a new perspective that is, that is freaking awesome.
johngrimsmo: yep
johngrimsmo: Now, would you say that your replica is significantly more advanced than the original or is it not? I mean, this was in the eighties.
John Saunders: Yeah, I'm not going to throw shade at them. I mean, they had full-size, full-functional builds, but many of the... um When you rewatch the movie with an eye for looking at how the robot was used, a lot of the more complex animations were tight shots of partial builds.
John Saunders: So, like, his eyebrows at one point, his um eyebrows pick up a box-end wrench and drop it on a switch to turn it on.
johngrimsmo: Yep.
John Saunders: Like, his... they they're absolutely not strong enough to pick up a chrome plated box and wrench or he other things.
johngrimsmo: got it yeah
John Saunders: So, and the current design does use some parts that are really not available and wouldn't be worth trying to find like the major servos and arms and so forth.
John Saunders: Otherwise, it's a pretty darn honest rebuild, if if that makes sense.
johngrimsmo: that's awesome but even like the electronics the voice controls the like it's all it's done differently but
John Saunders: Yeah.
John Saunders: don't know how did that. Well, the voicing was overdubbed.
johngrimsmo: Okay.
John Saunders: Actually, that's probably a good point. I don't believe, and I can't say this with certainty, but I don't believe the movie set robot would have had any voice. That was all done by a guy named Tim Blaney.
johngrimsmo: Makes sense.
John Saunders: So our, my build and the others that are building them, like they, Johnny Five can talk to you. Like he has his own speaker system and um so forth.
johngrimsmo: Yeah.
John Saunders: So yeah.
johngrimsmo: Because that's what makes it fun. like
John Saunders: Yes.
johngrimsmo: Yeah.
John Saunders: Yeah.
johngrimsmo: Dude.
John Saunders: yeah
johngrimsmo: Good man. What's what's next on Johnny five, your next critical task.
John Saunders: um I had tried a board called the Kyber board that allows individual servos to be controlled over the remote control. This sounds like complicated, but basically i want to hit a button on the remote control and that runs a pre-programmed sequence.
John Saunders: So like your mama was a snowblower or wouldn't you like to be a pepper too need input or Stephanie, nice software.
johngrimsmo: Yep.
John Saunders: um When you hit that button, the audio will play and a preset motion of servos like his head movement movements and all that will occur. um That board that i fried came back in.
John Saunders: I started to get it. It's complicated because it involves its own wifi and RC signals and maestros and S bus connections. I got it working, but it's not actually playing, but this was this morning.
John Saunders: So I just need to spend some time on it.
johngrimsmo: Yeah.
John Saunders: Yep.
johngrimsmo: Amazing.
John Saunders: What's next on your list?
johngrimsmo: um a quick thing on rev numbers so for a while we've been i've got a spreadsheet called rev history revision history where i'm tracking like norseman soft blade rev number 4.19 and that's engraved on the the blade on every blade in a tiny little spot and then in the spreadsheet i can have every little tweak every little oh i made that whole little bit bigger in cad oh i I changed that animal from there to there.
John Saunders: Mm-hmm.
johngrimsmo: And it's just like a good mental history of everything that's changing with the physical engraving also on the part. So as we have, say we have a hundred blades flowing through the shop, I might've made a change or two in between.
johngrimsmo: Now we can look at each part and know what the difference is. With the Fjell blade, we have, I have a soft milling rev number, but the hard milling program is totally different program, totally different tools, totally different tolerances.
johngrimsmo: So just yesterday I added a hard milling rev number. So there's two rev numbers now on each, field blade yeah because there's the soft milling side of it and then there's the hard milling side of it and they're different operations and different reps if they take the knife apart yeah yep yep normally my small engravings are 80 thou letters uh infusion and these ones are 20 thou yeah
John Saunders: On the blade. Interesting. And that stays on the, like the customer can see that.
John Saunders: Oh, it's hidden below the the pivot. Okay. Got it. Got Got
John Saunders: Really?
johngrimsmo: yeah
John Saunders: What are you using engrave 20,000?
johngrimsmo: I'm pretty sure it's a 1.32 ball. So nothing crazy.
John Saunders: Interesting.
johngrimsmo: Yep. And it's legible.
John Saunders: Cool.
johngrimsmo: It's like 1,000 deep. So this is also on a feature where I'm critically probing the top of the surface. So when I engrave 1,000, it is 1,000. No, um
John Saunders: Are you tipping is on the current?
johngrimsmo: it's on speedo. Three axis.
John Saunders: and Okay. Yeah.
johngrimsmo: Yep.
John Saunders: You're tipping it over to like get the better surface, whatever you'll need to.
johngrimsmo: no For most engravings, I don't do that. For surfacing, like a big curved surface, I will if I can tilt it for sure.
John Saunders: Mm-hmm.
johngrimsmo: But but yeah, so rev numbers have been great.
John Saunders: Yeah.
johngrimsmo: And then it it hit me like i've because of the 30 fail blades that I scrapped due to an end mill change.
John Saunders: That's awesome.
johngrimsmo: um I probably should have rev that, but I didn't have any rev numbers. So yesterday I had to walk around the whole shop and be like, let's quarantine all these blades because I, I need to know that they're collected.
John Saunders: It's a good word.
johngrimsmo: And but luckily when I went to the front, they, um, they already had them all relatively separated. So I just put them all in a red bin and said, all right, that's what these are now.
John Saunders: Yeah. Cool.
johngrimsmo: Yeah, it's good.
John Saunders: Do you use the stick fonts from fusion or do you go full Grimson on?
johngrimsmo: I found um a long time ago, like, Kanban fonts or something like that. Just this font package, a single line font package that been using forever. um
John Saunders: Okay.
johngrimsmo: And they work great. So there's, like, kind of one that I use for almost everything now.
John Saunders: Yeah. Secures.
johngrimsmo: But, yeah, single line fonts are a godsend compared to a double line traditional font, like Arial or something like that.
John Saunders: Oh yeah.
johngrimsmo: It just agrees better.
John Saunders: Right.
johngrimsmo: It looks better on small letters. On bigger letters, double line can look fine. But, yeah, on small letters, single line is king.
John Saunders: Yeah, sweet.
johngrimsmo: Does Fusion even have defaults? I don't even know.
John Saunders: Yeah, it includes, it ships with single line fonts. You kind of got to look for them, but they're there.
johngrimsmo: Yeah. i don't know what they're called. Be curious to see what they look like.
John Saunders: If you go to create a text, you'll just see them.
johngrimsmo: Yeah, okay. Yeah, mine are like literal fonts that I've installed on my Windows computer as fonts.
John Saunders: Yeah, yeah, yeah.
johngrimsmo: One's called like G12D13 or something like that. i don't know why, but it works great.
johngrimsmo: Yeah.
John Saunders: Sweet.
johngrimsmo: You?
John Saunders: ah That was my list, actually. No joke.
johngrimsmo: Yeah, mine too.
John Saunders: Yeah. Yeah. Oh, I ran the, o I'm running the Okamoto right now. Haven't run it a while. And it like, I do a warmup routine and then I do a test block. And usually we do have to bump it a few tents and then it's stable.
John Saunders: It was just odd, which was like, kind of was like, okay, that's why I bought this machine.
johngrimsmo: Nice.
John Saunders: That's awesome.
johngrimsmo: Yeah, yeah.
John Saunders: Yeah.
johngrimsmo: Yeah, there's no, sometimes there's no substitute for just having a really nice machine.
John Saunders: Yep. Yep.
johngrimsmo: Cool, man.
John Saunders: Sweet.
johngrimsmo: All right.
John Saunders: Sounds good.
johngrimsmo: Talk to you next week.
John Saunders: Sounds good to see you.
johngrimsmo: Okay. Bye.
John Saunders: Bye.
