3D Printed Flip Dot Display #3 - Magnets. How do they work again?

I have updated the design of the flip dot to remove crosstalk between adjacent dots. This required me to drastically reduce the size of the magnets and change the location of the electromagnets. I also have prototypes some control circuitry to make flipping easier and more microcontroller friendly.
Check my linktree for links to my Instagram where I post updates and my Github where I post all my design files and code!
linktr.ee/larr...

[@DCRofoJR]

You may want to find a different way to mount your magnets in the dots because heating them above 80°C (176°F) can start to permanently demagnetize them quite quickly. I'd suggest gluing them in with super glue

[@larrybuilds5359]

You are so right! Didnt even think about this until I noticed the flip dots werent responding how they should. Updated the design and went with your suggestion. Its a press fit with a bit of glue to hedge my bets. Great catch!

[@deathbydarwin1985]

I made that same mistake trying to set a magnet in a 3d printed part just the other day.

[@JamilKhan-hk1wl]

@@larrybuilds5359 super glue would be sufficient

[@blow0184]

Came here to say the exact same thing!

[@jkobain]

Oh, so I wasn't the only one to see a soldering iron on a magnet and start screaming in pain.

[@BenVonHandorf]

Found this as a random algorithm suggestion this morning. I'm a sucker for flip dots and I really like your video style. Under-stated, conversational and showing the bumps and process along the way!

[@danthegoodman_]

I would love to make a flip dot magent domino line, where by flipping one dot I can flip the whole chain, would look super cool

[@AnthonyHortin]

Really interesting. Flip dot displays are so cool. Looking forward to seeing how your project goes

[@yoctometric]

Wow your channel is criminally underrated! I love this video

[@larrybuilds5359]

Haha thanks! Glad you are enjoying, more fun/random builds in the works

[@olivierconet7995]

Excellent, very simple. No energy needed to maintain the dots. Just use a array with one diode for each coil, to drive the complete matrix, incuding set/reset dots.

[@21x9Ratio]

Just a suggestion, you could always use heat shrink tubing instead of electrical tape as an insulator between the threaded rod and enamelled wire. Might be quicker and easier for a larger scale.

[@larrybuilds5359]

Ooo I didn’t think of that! That would be quicker, Ill have to give that a shot, thanks for the suggestion!

[@geekbinge]

I've been interested in flip dot displays for some time but have been put off by the cost of commercial units. Thanks for the work you're doing! This gives me hope and ideas to try my own build of a flip dot display!

[@musemeghini]

Your explanation is crystal clear, I love how you edit the videos and how informative the are! New sub for sure!

[@danieltonheim995]

keep making videos you are a legend good design my man

[@randycarter2001]

You can cut down on the number of wires and power requirements by placing the 2 coils in series. Of course you want to observe polarity.

[@jkobain]

And on the single U-shaped core.

[@DerSolinski]

Normally the flip dots are just thin sheet metal, putting magnets into them is way to expensive. They have a cut out for the electromagnet pole that is not magnetized on one side. The actual flipping is achieved through the effect that the magnetic field wants to be as dense as possible using the metal as field guide. A need side effect is that the metal pole gets magnetized and the display doesn't need to be powered to keep the display. (But also a problem if the flip pate gets magnetized)

[@larrybuilds5359]

Interesting, so the magnetic fields will cause the non-magnetized metal dot to rotate when the fields alternate? I figured there had to be some amount of magnetization of the dot itself in order to properly repel/attract during actuation. It would be great to simplify the dot design, but not sure how I could manage to mass produce sheet metal parts. Would have to get an outside fab house to laser/water jet them. Maybe a diy metal stamping machine is needed? Im always down to hack something together haha

[@DerSolinski]

@@larrybuilds5359 Well it sounds easy but there is some engineering behind it. So it needs some trial and error to find the right angles etc. I'm more baffled that there are virtually no resources out there on how to DIY flip dots. All resources I found are on flip dots/flaps with magnets. Or on how to drive them. Also it seems that the flip dot display with the metal dots is a special German concoction lol. It was from a very old bus. But for starters you could try some (really) thin washers, maybe even magnetizing them with a strong electromagnetic pulse could give better results.

[@avivharari1362]

This is some great work you did there. Cant wait to havr some school kids replicate that...

[@justinalvarez4556]

I really like what you're making here! I wish the videos were just a little tighter on the editing side but I love the building process you're showing, keep it up!

[@larrybuilds5359]

Thanks! I am learning as I go, will keep on improving!

[@isaacgraphics1416]

hmm, I wonder if you could take it a step further, moving the magnet right into the pivot corner, and having a single electromagnet with a wider surface to effectively cover both orientations of the flippy magnet. It would be quite similar to how the 0.3g micro rc actuators work.

[@wombora]

Maybe you want to make the dot a bit thicker and have a layer of plastic on both side of the magent and push it in from the side - so event if the magnet shatters it will stay in place

[@larrybuilds5359]

Ive actually needed to add some plastic spacers to the magnets in order for the electromagnets to overpower the attractive force between the flipdot magnet and the steel rod. Haven’t had one of the smaller magnets break yet!

[@JessHull]

This is really interesting. Nice job.

[@mad_circuits]

Really, really nice. I also love flip dot displays. Here my notes: I understand the principle, but it would be highly preferrable to have less than 2 coils per pixel from a cost and energy efficiency perspective! 1 coil per pixel might be impossible, but what about 1.5/px or a fraction below that, made possible by using another pixel‘s coil as a guide/helper for the first pixel‘s turn. Maybe use springs to get away with less coils. 2nd remark: Try make the coils smaller but use higher voltages to reduce costs and production time/effort.

[@BenjaminMarshallScienceMan]

While watching this I just assumed I was watching an established channel with lots of subs. I was suprised to see your channel is as small as mine! I love scratch building things like this, you've definitely gotten me to subscribe. Keep the quality content coming!

[@BerkErkul]

Hey, great video! Looking forward to the rest of the series.

[@larrybuilds5359]

Thanks! Stay tuned!

[@nerddub]

Since your Z Height is very valuable, already having the e-mags sticking out from the frame, you should look into maybe 35-40mm (19mm rod projection + standard width electrical tape (19mm) = 34 + top nut/washer) M3 Button head screws, the head will function as your bottom nut/washer/combo and eliminate all the rod stuck out the back (button head is ~2mm thick), allowing you a little more standardization and saving you several to many millimeters by the look of it. Love the project, subscribed for more!

[@larrybuilds5359]

Great idea! I hope to use some of that extra length to attach the control boards once the design gets to that point, but standardizing to something I can just buy from mcmaster is probably the smartest option, especially once I start fabing a bunch of these. 2000 of them for my planned wall display, going to get tedious for sure.

[@TDOBrandano]

You can trade the number of transistors for the amount of copper used, if you wind two coils per electromagnet and bias them as opposites. Energizing one winding will flip the dot in one direction, energizing the other will flip it back. It might also make it simpler to drive as a set of rows and columns

[@wesleythomas6858]

Put the magnet closer to the point of rotation. It will work faster, utilise greater magnetic flux and also interfere less with the surrounding magnets as change in position is smaller. Agree also with those suggesting the U shape magnet

[@prozacgod]

You could probably eliminate having two electromagnets, if you use a horseshoe design. But then you'd have to wind the center of the horseshoe... Could be a design choice you could Chase when you make a winding machine. Would also be a much more compact design. You may not really need to actually have a complete horseshoe you could just insert the electromagnet between two steel rods. It would lose some efficiency in the air gap but it really wouldn't be that bad you're not relying on huge forces here

[@leoputchinski8768]

what a great work of engineering!! I look forward to future videos! Best regards

[@larrybuilds5359]

Thanks! Its coming together better than thought.

[@bbeerrkkuutt]

Try u-shape solenoid + switching polarity instead two separate coil

[@unknownhuman1000]

My thoughts exactly.

[@randomelectronicsanddispla1765]

Raw/address selecting will need diodes or three state drivers to avoid alternate current paths

[@rodneyolinger7955]

You might consider trying a electromagnetic core that is a better magnetic conductor (permeability). I've found that bolts are poor magnetic conductors. They retain some magnetic field due to the current in the coil and the permanent magnetic. When you apply the current in the winding it must overcome the residual field before it can reverse the field (hysteresis). The metal used in these bolts have a high hysteresis which will increase the power lost to heat. They are cheap to buy or it is easy to salvage them. with a stronger field you should have less problem with flipping the dot. Very good work! You must have learned a lot and others can learn from you (including me). Thanks

[@yorgle]

Awesome! I somehow only just found your channel, and am now a subscriber. :D

[@claudehebert3131]

To simplify electronics; why not an u-shaped electromagnet, both sides of the "u" would come in contact with the magnet; one side in the "white" position, one side in the "black" position. If you properly align the magnet's poles, you might need only a single transistor per dot (or a pair), instead of 4.

[@boazmulder1934]

Wow how do you only have 939 subs???? This is the same quality as a channel with 10's of thousands of subscribers. Keep up the good work and you will grow fast! I think I found you early haha

[@larrystone654]

Fantastic! Please keep these coming!

[@Laparedsurfshoponline]

The people want more Flipdots!

[@larrybuilds5359]

Haha, trying my best!

[@maesto]

I've been looking into getting customized smd magnet coils. Once you have final sizes go talk to a supplier on Ali that makes standing smd inductors. They can do different cores as specified.

[@AZ-74]

remember seeing one of these displays in the airport for a Winston cigarette advert the only problem I saw was that it was spazzing out in some sections

[@not_just_burnt]

youtube recommends can surprise me, thats for sure...

[@m1geo]

Nice project!

[@randomthingswithrandarawan8077]

The only thing that wasn't covered was in the last arrangements. The tiny magnets helped the intial problem but there's one problem that I theorize. In the event that you flipped the dot to the camera left of the first one you flipped leaving the original one as it sat would the second flipped dot cause the initial problem once again?...

[@TBL_stevennelson]

Thanks for sharing. Great video

[@RedFathom]

you could print a bending jig and just use one coil and bend the rod into both positions.

[@prozacgod]

I wonder if you could use stepper motor drivers to energize your coils, they are pretty ubiquitous and relatively cheap. Of course h bridge chip aren't that bad themselves, may be easier to use in a row/column activation design

[@hamjudo]

Given that you drive the coils with a bipolar circuit, how do you plan to scale up to row/column addressing? A separate H-bridge per pixel? Or positive and negative column lines each with a diode per pixel, likewise for the row drivers, leading to a total of 4 diodes per pixel. Or adding a component to each pixel that has a voltage drop that is higher than one third of the driving voltage. That way the coil you intend to energize gets close to two thirds of the driving voltage. Since all of the other paths through the array go through at least 3 pixels, the combined voltage drop will exceed the power supply voltage, and no current will flow. A green LED will drop 2 volts in one direction. 2 such LEDs wired in parallel with opposite polarity will drop 2 volts in either direction. With a 5 volt supply, and a pair of LEDs at each pixel, the intended pixel will waste 2 volts in the LEDs, while no power will flow through any of the other current paths. Or something else?

[@larrybuilds5359]

That is one way to do it, but I am trying my best to avoid that. I have found a way to power them in a row column format that greatly reduces part and I/O count. I am working on getting that video out in the next couple days

[@retrolobo]

Cool! Will you be posting updates to this project? Cheers

[@jozefnovak7750]

Super! Thank you very much!

[@patrickradcliffe3837]

Copper shielding. I don't think you need quite so many windings on your coils.

[@drgoo10]

Great concept, nice editing. 👍

[@jkobain]

I wonder why you use two coils on two rods instead of one coil on just one U-shaped core.

[@larrybuilds5359]

They are wound together to act as a single coil, but I opted for two cores for ease of manufacture at the time. I didn’t find any U-shaped cores of the size I wanted and had the rod stock on hand when I was designing.

[@mattmanandeddie]

Cool Idea. Maybe try charlieplexing for a bigger grid.

[@marsrocket]

Heat shrink tubing instead of electrical tape

[@PaoloConteDeveloper]

Looks great!

[@martintatak8851]

Looking forward for any progress! I am already hooked and cant wait to see it working. What about using heatshrink tube as an insulator for the rods?

[@paulromsky9527]

Couldn't you have used small steel washers/plugs instead of magnets? They would interfere with each other less.

[@NicolePapincak]

Flippin’ Dots!

[@PiotrEsdenTempski]

Hey, that looks like a really awesome project. Subscribed, and looking forward to more episodes. I have one question. The commercial flipdot displays use ferrite cores instead of iron cores for the electromagnets. Those cores are selected so that they have hysteresis. This means that when you energize the electromagnet you end up magnetizing the ferrite core one way or another. So the dot itself is not just held by the magnet in one position or another. This prevents the dots from flipping due to vibration or mechanical interference. As far as I can tell it should be possible to buy ferrite rods that you could use instead of the threaded rod. I bet you looked into it and decided against it, but still curious about your design decision here. :)

[@larrybuilds5359]

I did look at ferrite cores originally, but went with the threaded rod just because I had it and I didnt know yet what size magnet I would need. I might eventually use ferrite if I need more magnetic field strength, however the steel rod should provide some amount of hysteresis. It will be less than ferrite, since it has a higher magnetic permeability. Great suggestion!

[@z08840]

rare earth magnets start losing their magnetism at ~80C

[@johnadieb7586]

hi larry , i really like what you are making and would love to see your next tests , but i think you would face the same old issue if you were testing the -left bottom- cell as the magnet in this case will be near to the one on bottom right cell so it would be more likely to affect each other , what do u think 🤔 ?

[@larrybuilds5359]

Thanks! I am having a lot of fun building this up. And there for sure could be more issues with this design. You did identify a good test case, gotta put this little 2x2 through its paces.

[@georgmik2548]

Hey really great video. I was just wondering if you could use only one coil per dot with a U shaped core (bending the M3 rod after winding the coil). I think it should work because it would direct the magnetic field in the direction you need (this should also minimize the effect on other dots). An other advantage would be that the whole design would be smaller (and you would only need half as many coils) I'm curious of what you think about this suggestion. Keep up the great work.

[@Noxmyn]

How to manage multiple points? Is there an addressable diode strip?

[@larrybuilds5359]

I am using diodes for sure, they are helping me limit the current flowing to a single row/column combination. This will be the next video dropping in a couple days, stay tuned!

[@Noxmyn]

​@@larrybuilds5359 Are you going to use a cascade of shift registers?

[@InAmberClad]

Neat project, the sound when then magnets flip is great! On another note, what is the name of the song at 6:20 to 7:50? It sounds very familiar, possible Tycho or Helios but I cant place it (neither can Shazam unfortunately). Cheers

[@Mr0neShotAway]

really nice, subbed!

[@manishpoudel4540]

nice keep going

[@Ktrjkloo]

Hey I love your videos! I'm rather inspired to make a set myself! Was wondering what material was the electromagnetic cores you used? Read online that a material with high remanence will help retain the "memory" of the flip dots.

[@larrybuilds5359]

Thanks! Glad you are enjoying them. The cores are made from low strength steel M3 threaded rod. Could probably find ferrite or iron rods that would make for a stronger magnet, but I had these on hand. And you are correct that residual magnetism is left in the cores and hopefully will keep the dots locked in place. This is especially important once they are oriented vertically.

[@Ktrjkloo]

@@larrybuilds5359 HAHA nice! Looking forward to your next video!

[@jacobhuckins494]

The fact that they aren't flat is bugging me. Is there a reason you need that small angle? Why aren't the posts like 1mm shorter? Love the build so far tho, subbed.

[@larrybuilds5359]

I agree, some of the example commercial flip dots I saw were slightly angled, maybe to help it flip properly. I copied that design and went with the slight angle. I havent measured it or been really precise with how long the posts are once installed. I am going to try and eliminate as much of that angle as possible as I approach a final design. Thanks for watching!

[@ChourkyBob]

But can it play bad apple?

[@joelcordier6762]

Hi, nice ideas. Never thought making your electromagnet in a U-shape ?

[@larrybuilds5359]

I have seen designs that use the U shaped ferrite core. Might increase the magnetic field strength, not exactly sure why a U would be preferable, but I dont have a great grasp on what the effect would be. Both electromagnets are tied together in parallel, so they should act similarly

[@joelcordier6762]

@@larrybuilds5359 I think I did not explain my idea clearly. The idea is to replace two electromagnets buy a single U shape. In this manner you will only have one electromagnet per pixel and also divide the electronic part in halve. You only need a single H-bridge for each pixel. Also when your U electromagnet is polarized, it has one side attracting the permanent magnet and the other repelling the other face of the permanent magnet. Another idea to explore would be a way to keep a small residual current in the magnets to keep the pixel in place. 😉

[@dvi3660]

cool

[@NextLevelCode]

Managing all the h-bridges for your pixels is going to become difficult on a breadboard fast. Maybe time to create a PCB soon.

[@larrybuilds5359]

For sure, i am going to move to a protoboard first and then likely fab some boards and move to surface mount parts.

[@palpatinewasright]

No need for hbridge, I'd wire the two magnets in different directions with the same wire. They're never going to be both n or both s

[@Kalvinjj]

But wouldn't you still need an H-bridge to invert them? From what I understand what you mean, the intention is to wire both coils with reverse polarities, so whenever you power the coils, one is always north while the other is south, then, reversing the polarity again inverts both at once. For that part you would require the H-bridge wouldn't you?

[@palpatinewasright]

@@Kalvinjj I think you are right, I was totally off base. 🙇 Brain fart. I was concerned that the circuit used two terminals and an hbridge per magnet and so four terminals and two hbridges per dot, when instead you could wind two magnets with the same wire, requiring one hbridge serving two magnets.

[@yezbozz]

Where did you get the music @6:11

[@larrybuilds5359]

Its the track ‘Becloud’ from the stock music that you can use in iMovie

[@yezbozz]

@@larrybuilds5359 thanks!

[@davelordy]

At 10'00" are you playing the guitar with your feet ?

[@alberthiggins383]

great video but really didn't need 3 minutes of soldering, I was just looking for a 30 second overview of flip dot mechanisms and had to scan through the whole video to get it. Again great video but could be half the length.

[@kadircinkaya]

:100: