I Made A Mechanical Seven Segment Display Clock

Today we continue our saga as we attempt to build a convoluted mechanical 7-segment display clock. On the last video, we were able to successfully complete a single digit. Now I haven't touched this thing in a few months, but I seem to remember it having all sorts of problems which I'm sure are patiently awaiting us. So first things first, we're going to get the original digit working reliably, then apply some electromechanical power, make three more digits, hook them all together, and in theory we'll have ourselves a working prototype. Now where was I with this stupid thing? Oh yeah, the linkages keep popping off and the numbers count backward. Well I ain't trying to make no funny clock. First things first, let's see if we can get an easy W here by flipping these cams around. I think this is the third or fourth or fifth time I've messed up the cam orientation on this project. Well, third or fourth or fifth time's a charm because I've finally wised up enough to start I do a lot more explaining on how cam timing works in the last video, so go check that out if you want to go a little deeper on this project. All right, unless my hippocampus is on backward, I'd say this is counting forward. Wuh. You know, all this clock direction sh** has got me feeling introspective. What if we actually could wind the clocks back to a time before our time and take a little look-see into the past? Give our ancestors a handshake and say, what's up and stuff, and in doing so, learn just a bit more about ourselves and our genetic heritage. Well, all that's impossible, but the sponsor of today's video offers the next best thing. Now, I did not see that one coming. I recently partnered with MyHeritage to take my very own DNA test so I could get to know myself a little bit better. And with my discount code, MUJIN, you can get your very own MyHeritage bundle, which not only includes the DNA kit, but also a 30-day free trial, free shipping, the Ancient Origins feature, and so much more for the unbelievable price of just $19, a limited time offer. Taking the test is quick and easy. Give those inner cheeks a swab with this Q-tip thing, snap it off into the biohazard tube, and send it on its merry way. In a few short weeks, you'll get an email revealing your results. I'm often told, you don't look like you're from around here, and so on. That may technically be true, but we're about to find out once and for all. My great-great-uncle is a genealogy nerd, and I have the family tree put together to prove it. Now, according to him, this side of the family is mostly English with a kiss of Irish. Lucky me. As for the rest, well, I guess we'll find out. All right, here goes nothing. Jesus. Okay. 27.6% Scottish and Welsh. 19% Germany. Germaniac? That one I kind of knew. English, just a little bit of everything Eastern European, and even more ethnicities. Mostly Scottish and Welsh. Now, I did not see that one coming. So another part of this is DNA matches, which shows you relatives you have wherever they may be. Looks like I have some family up in Canada here, and a fourth cousin in Ireland. I'll be darned. Well, none of you are getting added to the Christmas card rotation, I'm afraid. And if you want to go way far back, there's this ancient origins feature, which can trace your DNA back up to 10,000 years and connects you with historic populations like Vikings or medieval Europeans. And MyHeritage has committed in its privacy policy to never sell or license your genetic data. So you're covered there. Click the link down in the description below or the pinned comment below that and use my discount code MUZHIN to get your very own MyHeritage bundle. This includes the DNA kit, free shipping, a 30-day free trial, the ancient origins feature, and so much more, all for just $19. Significantly cheaper than any other DNA kit on the market. But you gotta hurry, because this offer only lasts a limited time. And now, on with the show. These linkages keep pooping out because the current wire gauge is too big and can't be bent the way it needs to be bent. I've got some 1 16th inch brazing rod here, with which we'll pull the ol' switcheroo. This should be much easier to grapple with. I had attempted to make some bending fixtures for that fat-ass wire, all of which failed. I got by bending those by hand, but I resolved early on to make one that works for this new gauge. This is the Mark 7, which features square keyways for bending points for tightest by a mile bends. That is trademarked, and I will sue you. Getting here was a lot of trial and error, and at times it felt like watching the 3am specials on Animal Planet. But I'm getting consistent, repeatable bends, and as a result, the linkages have finally learned a little respect. The digit runs smoother than ever, with no derailments yet. The segments are still a bit jacked up, because they're not really glued or fixed in any meaningful way, but I'm not tripping over all that right now. I'm spinning this much faster than it would ever spin under power, but I'm still a little worried about the amount of force required to spin a whole chain of these digits. Still, I would not consider the emotions currently regulating my behavior to qualify as... tripping. They call these Synchronosity Motors. Just kidding, they're made for clocks. This one will spin at 5 or 6 RPM depending on input frequency. Being of the land of the free, we're on a 60 kHz grid, so we need a 6 to 1 reduction from the motor to the Geneva wheel, a net of 1 rotation per minute. I wouldn't call it the most beautiful layout, but we have our whole lives ahead of us to make it pretty. For now, we just need to know if this thing is even going to work. A few failed attempts and broken bits later, we have the parts necessary to animate our inanimate objects. I see a router table somewhere in my future, but for now I beat these tabs out like a caveman. We should probably see if this motor even fits before we get too far ahead of ourselves. Oh yeah, that's nice. Almost immediately, I'm noticing these brass standoffs are too long. Also looking like these bolt holes are too small too. I'll adjust the design so I won't have to drill them out again later, in case I ever have to make this part again. And indeed, that will be sooner than later. This big wheel interferes with the base. How did I not notice this in CAD? There should also be a hole in the side plate to catch the gear shaft. And the gears are too loose in their shafting. And now all the segments are bunged up. And now my bathroom ceiling is leaking? Alright, let's try this again. We should have all the parts necessary to animate our inanimate object. This 30 tooth gear and this 12 tooth gear need to be wed in holy matrimony. But instead of rings as a binding agent, we'll use super glue. This gear train is designed so that we get incremental movement at the last stage. This 24 tooth wheel spins the Geneva wheel 1 tenth of a turn every minute. This 10 tooth pinion needs to be fixed to the motor, which came with some grub screws for this brass thing. The shaft has these dimples that are clearly there for keying. I'm going to try just threading these grub screws right into the pinion. That looks locked and loaded to me. The wiring this motor came with is a house fire just waiting to happen. We'll use it as is as a proof of concept, but whatever you do, don't take your eyes off of it. The screw I used sticks out too much and interferes with the 12 tooth gear. It all looks very, very tight in there, so I'm not optimistic a smaller screw is going to fix this, but it's worth a try. Holy s***. Well, clearance is clearance. Okay, that's it. Moment of truth. I'm going to stay by the power switch in case something funny happens. Oh my god. It actually works. Huh. Well, okay then. Uh, I guess we design the rest of it now? Where do I even begin? As I mentioned in the previous video, clocks like this aren't exactly a new idea, but all the ones I've ever seen are actually Arduino controlled digital clocks powered by two servo motors, so I think it would be a neat challenge to make this clock truly mechanical. Shake things up a bit. I'll spare you the details of this exercise in futility, but basically the cams beyond the minute digit are split into 12 sections with 12 slot Geneva wheels incrementing the 10 minute and hour digits, and the 10 hour digit is just tied directly to the hour digit. Trust me. To keep things spicy, I've added no practical way of adjusting time. Before we start cutting all that out, we need to get some efficiency all up in this process. Until this point, I've been screwing workpieces into the CNC spoiler board like a savage. It's a sturdy enough method, but it's flawed for a number of reasons, and we can do better. I've been developing my own take on a universal workholding system, and it's finally time to put my butter where my bread is. I'm embarking on one big experiment to see what methods truly rise to the top, and what's just hype. It looks like I ordered 400 t-nuts in the wrong size. Good. Every handyman knows unused hardware is power, and now I have an excuse to buy something on McMaster Cars' perfect website. This is what $142,000 worth of t-nuts looks like. Anyway, the idea is that this will be a system that can work for any application in the wood shop, not just CNC machines. So I'm going to continue to obtain data and arrive at some conclusions before developing some opinions about it. If I find any success, I'll post plans on my website. The first test will be holding down sheet goods for clock parts. I have about 30 of these botched phone stands as a result of figuring out how to use this machine, and I think I've finally found a use for them. Obviously, I don't want to cut down into that spoiler board I just made. While that's cooking, we'll get to work on some of the other components. Bending the segment wires is quite similar to bending the linkages, except the wire is more cumbersome and the fixtures don't work very well. The segments themselves are split into two categories, topside and sideside, then to the drill press where they receive a new orifice based on their identity. It's a painful but exciting time for them. And what would a mechanical 7-segment display clock be without its segment linkages? A whole lot of nothing if you ask me. After everything gets its own private dance with the superglue bottle, the upper arms or segment linkages or what have you also pay a visit to the orifice reamer and get their tops orifice-ted. Shafting is also cut down to size, deburred, and tortured. Oh, sounds like the last of the parts are ready. Let's see what the damage is. One of the plates dislodged itself and the CNC performed a little dismembered self-immolation before the job was done. That one is f***ed, but the others can be saved with a bit of good old fashioned manual labor and a pinch of love. Aside from one cam getting its hole blown out, these all look pretty good. But then I thrusted the shaft and to my horror, realized all the holes are too big and this all needs to be redone. Unfortunately, that goes for the gears too. The little holes seem to be fine, but the keyed shaft is going to be a reset. Double unfortunately, all that was the last of my 3-eighths inch plywood. Or as Nuke Dookum would say, I'm here to kick ass and buy more 3-eighths inch sheets of plywood. And I'm all out of ass. Well that redemption arc didn't go as planned. The Z-height must have been set a little too low or some s***. I had to terminate the run before I broke another bit. Most of them got done at least. You know what, can we just get a montage out of here? I didn't really realize just how many parts this all was until I laid it out like this. I would consider it a win if less than half of this ended up in the garbage. Shall we? Yeah, this isn't right. For one, the segment linkages are too loose. Same reason the cams were too loose earlier I'd imagine. It didn't feel that bad when I was putting them together, but I guess it is that bad. If we don't get this sorted out, this clock is going to be counting sideways. The cam timing appears to be f***ed as well. I'm going to have to consult my spreadsheets and get back to you on this one. Cams are fixed, I just had some in the wrong places. Segment linkages weren't quite as simple, but this is them and we're good now and we can forget it all ever happened. Anyway, shall we? I forgot to hit the record button for that one, but here's some footage of the last two, which are basically the same thing. There's still some tweaks to make with positioning of the segments. I was still getting some play with the segment wires, so I started bending them in the middle to adjust size more delicately. And that's that. From this point on, it's uncharted territory. I think the best path forward is to get all these shafts set up with their gears. For whatever reason, this side plate is missing a hole. Maybe I can transfer a marking and drill it out manually. Nah, that's not a very good transfer. The tolerance I have in this design is like two thou, so precision is critical. Well, it was worth a shot. Okay, new part is back with the new hole and we're good to go. Oh wait, it looks like this little hole is missing at the top. Well, I'm not making it again. This is where it gets tricky. Admittedly, I didn't design this with the assembler in mind. I need to carefully make sure nothing is binding up and everything is in the right position. F**k it, let's send it. Okay, it's all led up to this. And there it goes. Well, I guess we need to give it some time. It doesn't work. The 10 minute digit is only moving a fraction of what it's supposed to move. Something is off with the gear ratios. I need to think about this. Everyone needs a colon. Lesson of the day. Never give up. I thought I was cooked back there for a minute, but no. It does do its thing and it does it pretty well. Most importantly, it holds time accurately and watching it is just memorizing. But I'd be remiss if I didn't bring up its shortcomings. This little motor has been holding its own, but I think it's too small for this application. It's not too hot to touch, but it's quite warm and it's loud as hell. It takes a few minutes for anything beyond the minute digit to complete its cycle. Basically, the amount of rotation required for these Geneva wheels to do a full index is more than what's being inputted at once. The result is numbers that look all explodey for a little while. Viewed head on, these don't look so bad, but from an angle, they look like one of those perspective illusions. At least changing the time is easy. Simply pull all the digits out of the base, pull them all apart, index each of the digits to about 10 minutes ahead of your target time, link them all back together, put them all back into the base, and turn the clock on once your time... Okay, so clearly not perfect yet, but perfection was never really the goal. The goal was progress and to have a working prototype, and I for one think we got there. We're going to be revisiting this project at a later time, so make sure you subscribe to catch the next installment and or follow me on Instagram for more real-time updates on whatever I have going on. If you found this video entertaining, then help me appease the algorithm. Likes, comments, shares, subscribes, and so on, it all helps and it's all greatly appreciated. If you like what I do and want to support the channel, you can check out my online store mujinmakes.com, where I've got some new goodies available, or I like coffee too. And of course, thank you to the sponsor of today's video, MyHeritage. If you've ever been curious about your ancestry, click the link down in the description below or the pinned comment below that and use my discount code MUJIN and get your very own MyHeritage bundle for just $19 while this offer lasts. Thanks for spending time with me today, and for surviving all those plugs I just threw at you. We're going to be doing some work around the house for the next one. I'm going to need all the help I can get, so I hope to see you there. Until then, Godspeed.