Spintronics Progress Update #6: Tying up Loose Ends
over 2 years ago – Fri, Dec 03, 2021 at 02:00:28 AM

Friends,

We're getting to a point now where we're just tying up loose ends. Unfortunately, there is one significant delay that will push out delivery from January (our original estimated date) to March or April. I'll get to that below. The upside is that it gives us extra time to smooth out the puzzles, clear up explanations, and create some good online content to supplement the game. I've already added more puzzles and tutorials to help with difficult concepts.

I thought I'd spend this update giving you a clear picture of the final loose ends that are left to tie up. Starting with the resistor!

Resistor

You might remember in the last update, I asked about ways to test the resistor - to make sure it won't leak silicone oil during typical use conditions or during shipping. You had many excellent ideas. So far I've done two tests:

Test #1: Running it over with a car

Some kids explore the world with a delicate touch. Others poke, squash, bite, and step on things. It's hard to know exactly what a kid might do, so I ran it over with my car to see what happens.

It doesn't work anymore, but it didn't leak! The chamber that holds the silicone oil stayed intact. It's a sturdy little cylinder.

Test #2: Temperature test

In some situations, the temperature of the environment could get quite high, like inside a hot car. Silicone oil expands as it gets hotter, and it expands more than the surrounding ABS plastic, so at some temperature, it will leak out. But at what temperature?

I tested two resistors: one with low viscosity silicone oil (a 100 spin ohm resistor) and one with a high viscosity silicone oil (a 1000 spin ohm resistor). The 1000 spin ohm resistor showed the first signs of leaking when the temperature reached 150 °F (65 °C).

I think these results are acceptable. If you left a resistor in a scorching hot car, it might leak, but even then, it wouldn't be a big mess, it would just let out a little drop or two that would probably stay contained in the top of the resistor.

So with that, the resistors are essentially done! LongPack is currently flying someone out to the factory where they make the silicone oils. They're going to see how they're mixed to make the correct viscosities and they promised to take pictures and video. I'll be sure to share them with you. 

I'm curious about this myself - these silicone oils are extremely viscous. I have no idea how you'd mix a 55 gallon drum of them. I have enough trouble just mixing a little 20 mL vial. Does anyone have experience with this sort of thing? How do you do it?

Battery 

After using the new injection-molded batteries for some time, I opened one up and found that one of the parts was disintegrating. I opened up the other three and they all had the same problem:

This is the part that the string wraps around inside of the battery. Evidently, it experiences significant stress from the string squeezing tightly around it. One wrap of string doesn't squeeze very hard, but the string wraps around about 20 times, so the total force squeezing is 20 times stronger.

Still, even that doesn't seem like enough to break the plastic. This is the same type of plastic that is used in Legos! And when have you ever seen plastic break in that way? There are little strands connecting the broken pieces - I sure haven't seen that before. After talking with LongPack, we figured it out: It's the metallic pigment they add to the plastic. The pigment is at high enough concentration that it weakens the plastic. 

Fortunately, I haven't seen anything like this in the rest of the parts. They all seem plenty sturdy. So to fix the problem, we could simply remove the metallic pigment from this individual part, but I worry that even without pigment, it would eventually break. So I designed a new part that seems to fix the problem. It fits over the vulnerable part and should handle the stress with ease.

You can see a CNC milled version of the part in the battery below. The string is wrapped around it:

It seems to work well so far. LongPack is in the process of adding a cavity to an existing mold to make this new part. It will take roughly one month to complete.

Chain

That new spool for the battery is one source of delay, but the biggest source of delay is the chain. If you remember from previous updates, the chain has one big problem remaining: the holes that fit over the sprocket teeth are 0.5 mm too narrow.

I think if it was a simpler mold, they could fix it with some tricky mold-modifying techniques, but this mold has 80 individual cavities for chain links. That's a lot - most molds only have a few cavities in them. LongPack tried their best to fix the mold and decided it would be better to simply start over. They're estimating it will take another 1.5 months to finish it.

Ammeter

After more playtesting, it became clear that the ammeter needs to be louder, especially when it's turning slowly. The ammeter works by sliding a "needle" over a bumpy surface, causing the needle to vibrate. The needle is touching a thin plastic diaphragm, causing the diaphragm to vibrate with the needle. Since the diaphragm has a lot more surface area than the needle, it can more transfer the vibrations to more air. Then the horn gradually spreads the air vibrations out over an even larger surface area, allowing the air vibrations inside the ammeter to transfer to the air outside the ammeter more efficiently.

Our design for the ammeter ended up being sort of a hybrid between the old See 'n Says from the 1960's...

...and the talking Ninja Turtles from the early 1990's. 

The Ninja Turtles used a plastic needle like what we're using, and the See 'n Says had a similar spring-loaded needle assembly that touched the edge of the diaphragm. It wasn't hard to get the ammeter to make noise, but it sure was hard to get good sound quality and volume. 

To make our ammeter louder, I asked LongPack to modify two things:

1. The mold for the black "record" needs to be modified to make the grooves deeper. Instead of 15 μm deep, they'll be 75 μm deep. I made a prototype of one with 100 μm deep grooves and it sounded...well maybe a little bit too loud. I think 75 μm will be perfect.

2. The little diaphragm inside will need to be modified.  As it is, the diaphragm barely touches the needle at all, and not very reliably. This will be an easy fix.

Other Loose Ends

• We're making the rubber on the bottom of the parts a little softer so that it doesn't slide over the metal base as easily. We're waiting to receive samples of that.
• LongPack is looking for a good vendor of the R3ZZ bearings. They're thinner than the usual R3ZZ bearings (4 mm thick instead of 5 mm thick) and that makes them harder to find.
• LongPack is looking for a vendor for the metal surface of the base tiles that can offer a little bit thicker metal.
• The springs for the transistor require somewhat tight tolerances on the wire diameter and the spring length. LongPack is looking to see if they can find a vendor that offers better tolerances.
• We're continuing to do playtesting, fixing bugs and unclear explanations. This isn't slowing anything down, though. When all the parts are ready to go, the books will be, too.
• We're waiting to receive samples of the base tiles and connectors. The molds were just finished and we've seen samples, we just haven't received them in the mail quite yet.

Timeline

Given how far along we were when we started the Kickstarter, we thought for sure that January would be a safe bet. We're so sorry! The delay from the new molds pushes us into Chinese New Year, which delays things even further. It's looking like March or April is a more likely estimate for when Spintronics will reach warehouses and begin shipping out.

Thanks again for everything! We can't wait to begin production.

Paul and the team at Upper Story

Spintronics Progress Update #5: La Résistance
over 2 years ago – Fri, Nov 05, 2021 at 10:44:46 PM

Friends,

Things are still coming along well. I've spent most of my time working more on the puzzle books and LongPack continues to make progress on the spintronic parts. It's always scary when I receive a package with new samples from LongPack. Will the parts work? Will they look right? Will we have to make an expensive, lengthy change? Fortunately, we haven't had any serious problems, and now that we've had several of these shipments go well, I get a little less nervous. 

We reached one very big milestone: all three puzzle books are written! Act One, Act Two, and the Power Pack. Here are the numbers (of course, after we do some more playtesting, things will change a bit):

Act One:

• 65 puzzles
• 15 tutorials
• 31 comic pages
• 131 total pages

Act Two:

• 77 puzzles
• 12 tutorials
• 23 comic pages
• 127 total pages

Power Pack:

• 11 puzzles
• 2 tutorials
• 18 total pages

I'm elated with how they turned out. They look beautiful, the progression is reasonable, and there's a LOT of good content. I've spent a lot of time reading other books and playing with other toys that teach electronics. I can tell you objectively that this is unique. Electronics has never been taught this way or any way like it.

Injection Molded Parts

Good news! Molds are finished for all of the parts except the base tiles (they are in the middle of being made now). LongPack sent us enough injection-molded parts to do some serious playtesting. The parts arrived late last week.

I went through the parts very carefully and returned a list of 32 issues that need to be addressed. It sounds like a lot, but most of them are easily fixable, like the dimensions of a spring being wrong or the colors still not being quite right. In fact, many of these parts actually work significantly better than my prototypes.

The most serious issue is still the chain: it looks fantastic in a sort of dark titanium color, but the gap for the sprocket teeth is too small by about 0.5 mm. That's a pretty large error, but I'm sure LongPack will be able to fix the problem.

There's a lot I could write about in this update, so I thought I'd focus in on one part that I think you'll find especially interesting. Which part would it be that you might guess is so very interesting?

Wrong! 

It's the resistor.

The resistor?

I realize it looks relatively boring compared to the other parts. And what it does is kind of boring, too. I mean, it just slows things down! Nonetheless, getting these resistors to work right took me down a windy path that involved a surprising amount of chemistry and...volcano science.

A resistor's job is simply to resist the flow of current. If you try to turn a spintronic resistor with your fingers, it doesn't turn easily. You can feel it resist the movement.

There are a lot of ways to make something resist turning. Think of pickle jar lids or rolling up a car window. Add some friction and you've got resistance. In our case, though, the challenge is that the resistance needs to be constant,  just like it is in electronics. It needs to be the same when you apply a little force or when you apply a lot of force. Like this:

The design below is the first idea I had for a resistor. It has a crank on top to make the explanation more clear, but it would be a sprocket if it were a spintronic part:

It's simple enough: there's an orange axle on top connected to a black rubber disc. Another black rubber disc (connected to the green plate underneath that can't rotate) is pushed against the top disc by a spring. If you try to turn the orange handle, the two rubber discs have to rub against each other, creating friction. 

Ok, now here's the rub: Remember how we want the resistance to always be the same, regardless of how hard you turn the handle? Well take a look at the resistance of this device plotted against the force on the orange handle:

Yikes! That's not constant at all! What's going on here?

Well, at first, when we lightly push on the handle, it doesn't turn at all. That's because of something called "stiction". Stiction is that initial, higher friction that must be overcome to get something to begin sliding on a surface. It's especially prominent with materials like rubber and sandpaper. Imagine putting your finger onto a rubber surface and then trying to slide it across. At first, it won't slide at all. Then, once you push hard enough, all at once it will start sliding. That's stiction.

Then there's a second problem: As we apply even more force to the handle, the resistance begins to drop. That's because the more force we apply, the faster it turns, and the faster it turns, the less time the rubber discs have to conform to each other. Like a rock skipping on the surface of water, the faster the rubber discs slide past each other, the less resistance they feel.

In short, this is a bad design for a resistor. It creates friction, but it doesn't behave anything like an ideal resistor. 

Since stiction was my greatest concern, instead of creating resistance by sliding surfaces over each other, I decided to try generating resistance by shearing a fluid.

In this design, there's an outer ring (green) that's fixed in place and an inner ring (orange) that can rotate. The fluid touching the surface of each ring is stuck to it, so when the inner ring is rotated, it places "shear stress" on the viscous fluid, making friction. The great thing about generating the friction in a fluid is that it doesn't have any stiction! So this design creates friction without any stiction.

This is, in fact, how the current spintronic resistors are designed:

I used grease for the viscous fluid, tested it out, and...well...here's what I got:

Not good! There wasn't any stiction, but the resistance still depended a LOT on the force applied to the resistor. The more force, the lower the resistance.

In chemistry, there's a term for this. It's called "non-Newtonian". The grease is a non-Newtonian fluid because its resistance changes with the shear force applied. (A Newtonian fluid doesn't change.) Grease gets runnier the more force is applied to it. Fluids can also be non-Newtonian in the opposite way. Think of Oobleck - that's the stuff that's made from cornstarch and water. It actually has more resistance the more force you apply to it. If I put Oobleck inside the resistor, the curve would be in the opposite direction - it would go up instead of down.

What we need is a viscous fluid that is as close to Newtonian as possible. That's surprisingly hard to find! There are viscous fluids and there are Newtonian fluids, but hardly any are both. And on top of that, it can't be some exotic chemical. It needs to be relatively low cost, safe, and non-volatile.

After a lot of searching, I found the answer in an unlikely place: a review article about modelling volcanos. People who make accurate models of volcanoes use silicone oil to simulate magma because it's a viscous, Newtonian fluid. It did the trick! It's also non-toxic. In fact, it's one of the main components of Silly Putty.

Here you can see my tests with a whole bunch of different viscous fluids, and then with silicone oil:

Silicone oil was the best by far. The curve was virtually FLAT! This was a very exciting moment.

You can tell silicone oil is unique just messing around with it. Here's some highly viscous silicone oil. You get a sense of how it flows when I put a stick into it. 

Measuring out viscous fluids like this is tricky. You can't suck it up into a dropper or pipette because it's too viscous, but you also can't use a scoop. It turns out that sticks are the easiest way to transfer fluids like this. You wrap some around a stick and transfer it to a different container.

As another demonstration of the Newtonian behavior of silicone oil vs. the non-Newtonian behavior of grease, I put a dab of grease and silicone oil on a vertical surface. The force of gravity on the grease is small, so its resistance to changing shape is very high and it doesn't move. On the other hand, the silicone oil slooooowly drips down the surface. Just like lava. :)

What about leaks?

Indeed, this has kept me up at night. Yet everything seems to be sealed up well enough that it doesn't leak at all, even with very low viscosity oil, and held upside down or sideways for months of time. What sort of test(s) would you recommend to make sure that it doesn't stand a chance of leaking in the real world?

Next up

Things are looking good.  Despite the fact that the factory is only able to use electricity for 4 days a week and that there's a silicone oil shortage at the moment, we're still making good time. Specific things we plan to do in the near future include:

• LongPack is in the middle of making the final injection molds for the base tiles, the tile connectors, and the tile connector tool.
• Playtesting! We finally have two full sets of parts and the puzzle books are ready to go. I've never had this many parts before!
• Finishing the back of the box designs.

By the time the factory has everything 100% complete, we ought to be done with playtesting and have final changes made to the book. Then we can begin production. It's getting there!

Thanks again. It's fun to make this knowing that you all will get to enjoy it,

Paul and the team at Upper Story

Meet the Illustrator: Aleksandar Zolotić
over 2 years ago – Wed, Oct 06, 2021 at 05:42:26 PM

Friends,

We've made a lot of progress since our last update, buuuut I'm going to save it for the next one. In this update, I thought you might like to meet the illustrator for Spintronics: Aleksandar Zolotić.

I didn't really know how to find an illustrator before Spintronics. With Turing Tumble, I didn't have to look far to find one. Jiaoyang Li was an undergraduate student doing research in my lab. She showed me some of her art one day and I asked if she'd be interested in doing the art for Turing Tumble. She agreed and that was that. But with Spintronics, it was harder - we were going to have to hire someone that we had never met. 

We spent a lot of time looking around for an illustrator that matched the look and feel we were aiming for. We wanted art that was stylized, attractive to kids and adults, and highly detailed. We were going to have illustrations of complicated clocks, spintronic machines, and clockmaking tools, and for all of those things, detail is paramount. We also wanted someone who gave their characters plenty of personality in their facial expressions and postures, and someone who would do the necessary research to make the images accurate for the time period and location. In the past, I tried UpWork to commission art and I didn't have much luck - it was hard to predict the quality of the work and the nature of the platform didn't encourage artists to ask questions. After a lot of searching, I ended up using the Bright Agency. Honestly, I wasn't excited about the idea of going through a middle-person to work with an artist, but Robbin (the agent there) was surprisingly helpful in narrowing down an appropriate artist for us based on things we could not have known like the personality of the artists. She had us make a list of all of their artists that we were interested in and then she narrowed it down. We chose Aleksandar and held our breath while she asked him if he'd be interested. He said yes!

Phase One

Once he was ready to begin, I had a video meeting with him where I explained as much as I could. He came back with rough sketches of the main characters...

an important background...

and a spintronic machine...

We then gave several pages of feedback. Here's one of the pages, to give an idea what our feedback looked like:

Phase Two

Aleksandar took another shot at those same pictures, this time with more detail. The characters...

the background...

and the spintronic machine...

Phase Three

We were getting close. Aleksandar was receptive to our feedback and responded perfectly. He sent a final family picture and we gave our final feedback before he began sketching out the book.

Phase Four

Now that the main characters and the general look and feel was fleshed out, Aleksandar started sketching the comics.  He did them in three batches. After each batch, we gave feedback. This way, he didn't have to go to all the effort of coloring in the pages before we asked him to make changes. Here's one page from the sketches:

Phase Five

Once Aleksandar finished the sketches and made revisions, he started coloring the pictures. This was a massive job. Frankly, I expected that over the course of the 6 months he worked on this project, the quality would sort of deteriorate. I mean, I think it would if it were me? But instead, it started strong and just kept getting better!

After a few rounds of revisions (which were minimal - Aleksandar did an exceptional job), he sent the final artwork:

Aleksandar Himself

Aleksandar was a dream to work with. He understood what we were looking for and just nailed it on the first try. And even after he made our revisions, he kept going back to make the art better and better on his own, without us even asking. He would find a shadow or a proportion he didn't like, and he'd go back and improve it.

One thing I didn't expect is how much creativity an illustrator has to have, even when you lay out what you want pretty clearly. There are a million little details to decide on. Just in that last image above, Aleksandar drew a whole bunch of things I didn't even discuss with him: the type of ship, the slight fog, the boarding apparatus, the clothes, the type of luggage, the boat in the background, the type of ground, the age and hairstyles of the other people. Aleksandar figured all that out himself, and he made choices for each and every image that make the images more compelling than I ever envisioned. 

I asked Aleksandar if he'd be willing to share a little about himself for this update and he agreed. 

Me: Where did you grow up? How did you come to illustrate for people around the world?

Aleksandar: I grew up in a small town near Belgrade, our nation’s capital. From my early age drawing was really important for me. My main occupation, next to playing soccer. At some point of my childhood I stopped dreaming of becoming famous footballer and started seeing myself as an artist. Primarly a painter, of some sort… Children’s books illustration is something I became aware of during highschool. Upon graduating I illustrated books for Serbian publishers, but since the market of children’s books publishing in Serbia isn’t a huge one it was necessary to apply for a full time job in gaming industry, for example. Animation industry is almost non-existant in Serbia, so full time jobs for illustrators are really, I mean, really rare outside of gaming industry. In order illustrating to become my full time job and proffession it was essential to join an international illustration agency at some point. So, after couple of years of game development work I started working as freelance illustrator, represented by The Bright agency, that provided me with enough projects so I can declare illustrating my full time job I always aimed having.

Me: How did you learn your art?

Aleksandar: I went through art schools, where I learned to draw and paint in a classical way, with traditional techniques first. I gained bachelor’s degree in Illustration from Faculty of applied arts in Belgrade in Serbia. Programs were based on learning traditional print making, caligraphy, drawing and painting, animating, and how to illustrate and design books studying all that. There wasn’t too much attention given to digital tools at that point, but it was great having knowledge of classical art making as a background. Digital painting and drawing was something we learned on our own. Much of what I know about digital illustration today came from watching works of others, gathering information online, and receiving advices from my illustrator colegues.

Me: What was it like working on such a large project with Spintronics? How long did it take you? What was your favorite/least favorite part of working on it? Which is your favorite illustration? Which was the hardest to illustrate?

Aleksandar: It was certainly one of most demanding one, since it features a graphic novel with more then fifty, fully colored pages. Important part of the work was researching the looks of that era and gathering references for the look of the town and the way people dressed. Also, getting to know about clockmaking tools and parts, as well as becoming familiar with Spintronic parts and the way they work and operate. It was trully a journey, that lasted around 6 months. Favorite part of working on it was staging the dialogue between characters as it was really nicely written and… painting the buildings. I wouldn’t call it the least favorite but the hardest part was working on Spintronics made machines, but in the same time fullfilling. My favorite illustrations are from the page where Natalia, the protagonist of the story, is racing her parents with Spintronics made cars. Also a birds view of the town covered with network of Spintronics parts providing new homes and businesses with power.

Me: What pieces of technology do you use? When did you switch from paper to digital?

Aleksandar: I’m now using S size of Wacom Intuos4 digital drawing tablet as a tool to create illustrations in Adobe Photoshop. The switch from paper to digital happened the moment I felt confident enough to draw and sketch entirely with digital pen. Drawing tablet I first came across hasn’t been good enough to capture dynamic of free hand drawing so I mainly used it to paint the illustratos. It was either that or it just required a whole new approach to drawing in order to be happy with the results you see on the screen. So, I would have my roughs drawn on paper, then scanned, and colored inside Photoshop leaving a little or no linework at all in my final illustration. During that time I also created some personal artwork where I combined scanned drawings and textures painted with watercolors on paper with tools and features that could be found in Photoshop. Later on, as techology behind drawing tablets improved, I switched to using no paper at all. Drawing and painting apps today are so advanced that it’s sometimes hard to tell apart what’s created traditionally from what’s created digitally. Most of them are now really good in mimicking the looks of artwork created with traditional tehniques, such are guache and watercolor.

Me: What makes you unique among illustrators?

Aleksandar: It’s hard to tell from personal perspective. I usually don’t think of my work as significantly unique. I don’t use any amazingly different way of creating illustration. Speaking of style, I stand among those who like to see their illustrations detailed, with a sense of existing space, movie like compositions and character expressions.

Me: What is it like to be an illustrator?

At this point, it’s a job I’m really happy with. Working on children’s books and educational products aimed at children and young ones feels rewarding, because mission of the illustrator is to make all those things, necessary for their development in many ways – more attractive, clever more interesting and entertaining. Occassionaly you get invited to participate in workshops and bookstore promotions so getting publicity for your work affects motivation and self confidence in a positive way. Profound organizational skills are required too, because projects usually overlap each other. It doesn’t differ too much from many other proffessions, especially now, during pandemic, as jobs are moving to working spaces at home. And so as in many other proffessions, you have to work a lot and with all your hearth in order to get better with each new work.

Aleksandar also made a video for us that shows his process. It's over two hours of him illustrating an image, and it makes you truly appreciate just how much work goes into each image. With the thought and care he puts into each image, it's no surprise that he keeps improving. I sped it up to 20x the actual speed:

Aleksandar is truly amazing and we were lucky he agreed to work with us. We certainly hope to work with him again in the future.

A Clever Trick

Ok, so I didn't know this before - maybe you did. Take a look at this picture. Do you see anything...abnormal?

Hint: It has something to do with their bodies...

Did you find it? Everyone has four fingers! But...why?!? There are actually a lot of reasons for it. There's a video from Channel Frederator that explains it well. It's worth watching if you have a moment.

That's it for now. Thanks again everyone! I'll write more soon about our progress in the other parts of the project.

Paul and the team at Upper Story

Spintronics Progress Update #4: The Transistor, the Ammeter, and an Ohmmeter
over 2 years ago – Thu, Sep 09, 2021 at 05:01:04 PM

Friends,

Things are still coming along surpriiiisingly smoothly. (Knock on wood.) An awful lot has happened since the last update so I'll jump right in. Some of it gets a little complicated, but I'll try to explain things as I go. If it gets too far in the weeds, just skim to the next section and it'll get less weedy.

Overall Progress

Since the last update, we received a sample of the transistor and the ammeter (I'll get into those shortly). Next week, we will receive four more injection molded parts: the resistor, the inductor, the capacitor, and the switch. That keeps us right on schedule. We had LongPack make the parts in decreasing order of difficulty, so I don't expect any real problems with the rest of the parts, especially considering the quality of the last parts they made. After that, all that's left then will be the tiles, the connectors, and the fixes to the chain. I've been working hard on the books the last couple of months and they're getting close to completion. Parts of them are even being translated already. It's coming together!

Spintronic Ohmmeter

With the spintronic resistors almost ready, the factory needs a way to measure the resistance of the resistors accurately. I'd like the resistors accurate to +/- 20%, so we need a tool that can measure resistance with at least that much accuracy. Measuring resistance isn't particularly complicated. In fact, you do it with your fingers without even thinking about it. With a spintronic resistor in your hands, you would simply apply a force to it with your fingers (i.e., a voltage) and feel how fast it turns (i.e., how much current). If it has a high resistance, it wouldn't turn very fast. If it has a low resistance, it would turn quickly. You calculate resistance the same way: resistance equals voltage divided by current.

To measure resistance quantitatively, we need to be able to provide a constant, known force to the resistor and then accurately measure the current going through it. How do you provide a constant, known force to the resistor? I thought of using a spintronic battery, but the spring would get slightly weaker over time and it would become less accurate. I thought of building a constant torque motor, but that would get complicated quickly, it might need regular calibration, and it might be temperature sensitive. Then it hit me: I'll use gravity to provide a constant force. It's pretty much the same wherever you go and it doesn't change.

So with that, I built a spintronic ohmmeter:

You simply hang a known weight from the loop in front and it measures the resistance as the weight falls. You can see it working in this video:

You can see the resistor in the video measures between 1.4 and 1.5 kΩ. However, after moving this device to different locations, I quickly discovered that the resistance of spintronic resistors depends heavily on temperature. You can see the temperature dependence for three different resistors here:

I'm not sure what the temperature will be in the factory where they're making the resistors and I want to be sure it works in any location, so I added a very accurate temperature probe to the ohmmeter. 

The ohmmeter measures the temperature and then takes it into account when it reports the resistance. It now reports what the resistance would be if it were measured at 22.0 °C.

Now I just need to write a procedure for using the spintronic ohmmeter and then it's time to ship it off to LongPack!

Transistor

Our first sample of the injection-molded transistor arrived! LongPack is making adjustments to the top piece to give it a better fit, so they didn't send it. Instead, I put on a top piece from a prototype sample I made previously (that's why it's more purplish in color.)

There are two sprockets in a spintronic transistor. There's the sprocket on top, called the "gate", and the sprocket on the bottom, which we call the "path" (it would be called the "source-drain" on a MOSFET, which is the type of transistor it is most similar to.) The gate controls the resistance of the path. Like a dimmer switch, the more you turn the gate, the lower the resistance of the path. You can see how it works in the animation below:

This animation is a little old - I've updated the design a bit since then - but on the left hand image, do you see those two little arms that squeeze the orange o-ring? (There's a third arm you can't see in the back.) Those are what create resistance for the path. The more force (i.e., voltage) you put on the gate, the more it pulls the little arms away from the o-ring, decreasing the resistance of the path.

There was a lot to test and optimize here. When you look at an electrical transistor's datasheet, there are pages and pages of data describing its behavior. Spintronic transistors have just as many properties that could be measured, but the most important ones are:

•  Fully ON gate voltage: What is the gate voltage at which the arms aren't touching the o-ring anymore (i.e., the gate voltage where the path's resistance is zero)?
•  OFF resistance: What is the resistance of the path when the gate voltage is zero? (This is the highest resistance the path can have.)
•  Resistance as a function of gate voltage: What is the relationship between gate voltage and path resistance? Is it linear or...something else?
•  Resistance as a function of current: Does the resistance of the path change if there's more current running through it?
•  Consistency of resistance: Is the resistance of the path smooth and steady? How much does it vary?

First, I built a device that allowed me to put a constant voltage on the gate while measuring the resistance on the path. To measure the resistance, a Lego motor puts a constant, known current on the path while a force probe measures the voltage across the path.

First, I looked at the resistance of the path as a function of the gate voltage. I kept the current through the path for all of these measurements at a constant 2 mA.

It was wonderful to see a nice, smooth relationship between the gate voltage and the path resistance. You can see that when the gate voltage is zero, the path resistance is about 6 kΩ. That's about where I want it to be for the circuits you build in the puzzle book. And it looks like the resistance of the path would juuust get to zero with a gate voltage of about 1.1 V. I was hoping for something between 1.0 V and 1.5 V. All of this was spot on.

Next, I looked at the relationship between the amount of current through the path and the path resistance. Now a perfect resistor wouldn't have any dependence of resistance on current, it would just always be the same resistance. However, this is not a perfect resistor. There's a strong dependence of the path's resistance on the amount of current running through it. It makes sense - the faster the o-ring is spinning, the less the arms can dig in to stop it. This isn't a problem, really, it's just a property of these spintronic transistors. Electrical transistors have all sorts of their own quirks, too.

Finally, looking at the consistency of the resistance, you can see the data directly from the force probe with a range of currents (and zero gate voltage):

The force generated by the path resistance is very consistent. I was worried it would go up and down a lot from inconsistencies in the thickness of the o-ring, but it doesn't. It gives a nice, steady resistance.

I'd say the transistor is good to go! Phew, that was an important one.

Ammeter

Our first sample of the ammeter also arrived! There are some small issues LongPack is improving with the aesthetics, and they don't have the injection molds for the sprockets on the bottom made, yet, so they put on 3D-printed ones instead.

The ammeter works a lot like the old See 'n Says. A needle sits on top of a "record" with a subtle surface texture. When the record is spinning, the needle moves along with the texture. 

The needle is touching a diaphragm, making the diaphragm vibrate with the needle. The diaphragm has an awful lot more surface area than the needle, so it is able to more efficiently force the air around it to vibrate and make sound.

When I first received the ammeter sample, it was too quiet. I modified the shape of the diaphragm, the material it was made from, and its thickness until I was able to get good sound out of it.

Because the diaphragm in this ammeter is fairly small, it can only make fairly high pitched sounds, so it needs a certain level of current before you can hear anything. You need at least 10 milliamps or so to get a strong sound. You can also use a current multiplier (something you will learn to build with a junction) to hear lower currents.

The ammeter is looking good!

Puzzle Books

The last month I've spent most of my time working on the puzzle books. The Act One puzzle book is fully drafted with all the final artwork. It turned out fantastic! It's unlike anything else and I can't help but wonder if it will change the way electronics is taught. 

The Spintronics puzzles are different than the Turing Tumble puzzles. With Turing Tumble, there were some basic concepts to learn, but then it was all about combining them to solve ever more complicated challenges. That's how it is with computers - you have some basic building blocks and you build incredibly complicated systems with them. But with electronics, there are an awful lot more basic concepts to understand, a lot more building blocks, and things get complicated fast. Each of the little circuits we have today were big discoveries in their time. It would be too much to expect players to figure them out on their own. For example, I couldn't just give players an inductor, a capacitor, and a transistor, and expect them to figure out how to build an oscillator - it would be far too difficult.

So the puzzles in Spintronics are generally easier and more instructive. They're a little more about learning and building an intuition, and a little less about flexing your puzzle solving skills. My goal is to give players the understanding they need to go on and build their own circuits. That said, it's still fun and intensely interesting. Concepts that are flat-out boring in electronics come alive when you build them from spintronics and watch them work. I never expected it would be so satisfying to watch an active current source do its thing.

In addition to the puzzles, there are also a series of 27 tutorials scattered in between the puzzles to teach concepts directly that would be too difficult to figure out on your own. Some of them also make the connection between spintronics and electronics. Here's an example of one of the tutorials from Act One:

With Spintronics, there are a LOT of concepts to learn. There will be more puzzles than I originally thought. Between Act One and Act Two, there are currently 145 puzzles. Act One gives an introduction to a lot of basic topics like:

• Current, voltage, resistance, and capacitance
• Series and parallel circuits
• Circuit diagramming
• Voltage dividers
• Basic logic gates
• A variety of other circuits
• Some things unique to spintronics: spintronic coupling, voltage doublers, resistance doublers, etc.

I'm finishing up the Act Two puzzle book now. It covers a wide range of more advanced topics including:

• Capacitors in series/parallel
• Transistors: as switches, the active region, gate charge, amplification, darlington pair, and biasing
• Voltage and current sources
• Current to voltage converters
• Diode circuits: rectifiers, clippers, peak detectors, voltage limiters
• Inductors and inductor circuits
• AC, frequency, superimposition
• Reactance, high-pass, low-pass, and band-pass filters
• Switched-mode power conversion: buck converters and boost converters (with and without synchronous switches)
• Feedback: oscillators, active current source, and the flip-flop

I think it's pretty cool that you can learn how all these circuits work without getting bogged down by equations and calculations. It truly does become intuitive. And it's incredibly easy to debug a circuit when you can see it and feel it. We'll need to do a bunch of playtesting soon to work out issues with the puzzles, but I'm very, very happy with how it's all turning out.

Thanks again for making this happen! I can't wait to send it to you!

Paul and the team at Upper Story

Spintronics Progress Update #3: The Battery is Complete!
over 2 years ago – Fri, Aug 06, 2021 at 04:06:09 AM

Friends,

First, a quick reminder - if you haven't already, please fill out your survey. You need to complete it to receive your game. If you haven't received an email with a link to your survey, please reach out to us at [email protected] and we'll get you set up.

Fortunately, things are still coming along with surprisingly few problems! I've spent the majority of my time the last few weeks working on the puzzle books, but in this update, I thought I'd cover progress that's been made in manufacturing. I'll start with the big one: the battery.

The Battery

While the junction was the most difficult to produce because of the high precision required, the battery is by far the most complicated. With 48 parts that all have to work correctly, significant stress placed on parts by the spring, and a circuit breaker mechanism that slams it to a halt when it gets going too fast, I figured the first sample of the injection molded battery would have a number of problems to fix. I just hoped I wouldn't have to redesign anything and have molds remade.

Well, two and a half weeks ago, the first sample arrived in the mail. I opened it up and...wow, it looked great! The colors weren't quite right (the copper and gold colors are kind of drab, and just for this sample, they used copper everywhere it was supposed to be silver), but it still looked great!

Then I tried it and...it WORKED! It wound up perfectly, the circuit breaker mechanism engaged at just the right speed, nothing seemed weak or warped, it was great! Upon closer look, I found some small problems, but they were mostly in how it was assembled, which is easy to fix. There's only one little part on the inside that needs to be adjusted:

That little part in the picture needs to be 0.5 mm shorter. As it is, the little one-way ratchet mechanism doesn't engage all the way in the gear above it. It's not that that part was made poorly, it's just that small errors in the heights of the other parts in the stack added up, and the problem can be fixed by shaving off 0.5 mm from that part. 

The next day, I 3D printed a shorter version of that part, reassembled the motor correctly, and did some tests with it. First, I pulled the string and let it go a whole bunch of times to make sure the circuit breaker mechanism is sturdy.

It didn't show any signs of wear after an awful lot of that.

Then I did a test to determine the battery's voltage, capacity, and noise. Voltage is how hard it pushes. You can see the voltage displayed by the capacitor/voltmeter on the bottom left of the video below. The capacity can be determined by how much chain it's able to push on a single charge. A battery's capacity is usually given in units of amp-hours (Ah). That is, how fast the chain is moving (amps, A) multiplied by how long it's able to move that fast (hours, h). Noise is how much the battery stutters as it turns. If the gears don't mesh quite right or there's significant friction inside the motor, you see the voltmeter wiggle and bounce up and down a lot. My first prototypes of the battery were very noisy.

But everything looks excellent here! There's virtually no noise and the capacity comes out to about 0.0004 Ah (for those checking my calculations, this resistor is actually about 1260 ohms). The voltage is slightly lower - I think we'll call it a 6 V battery instead of a 7 V battery. That's no matter - the voltage can easily be stepped up or down with the spintronic parts. What really matters is the battery's capacity.

To summarize, the battery is essentially complete, with a couple minor issues left to work out. In terms of the larger picture, I've learned that Spintronics isn't too complicated. As I was prototyping the parts, I kept thinking to myself, "This is complicated. Will anyone be able to manufacture it?" The answer is a resounding "Yes!", and I could make even more complicated things in the future. :)

Chain

Yesterday I received a new sample of the chain. Note that it's not the correct color - they made it silver for prototyping, but it will be a very dark metallic gray in production.

LongPack corrected the length problem and the friction problem - it feels pretty good! But now it's clear that there are a couple of other issues:

1. The gap in the middle is too narrow. It's pretty far off. The chain doesn't fully engage because it can't fit all the way over the sprockets.

2. The snap isn't formed right. There's a dent in each snap that's hard to see with your eyes, but you can see it with a microscope:

The dents cause a number of small issues, but I'm betting they'll be able to fix them without too much trouble.

One unexpected property of the new chain is that it's 22% heavier than the Lego chain. A small part of that comes from a difference in design (it has 2% more material), but mostly it's because it's made of POM plastic rather than ABS. In the video below, you can see the Lego chain floats in sugar water, but the new POM chain sinks.

I used POM because it's a very slippery plastic, so it should have less resistance, but I didn't consider the extra density. Ideally, you'd want the chain to have zero mass so that it doesn't add inductance to the circuit, but then again, I'd rather have stray inductance than stray resistance.

Challenges

As LongPack continues to cut out new injection molds, there are two issues we're in the middle of trying to resolve. One has to do with the ammeter. The little disc with grooves on top (sort of like a record) may be challenging for the injection mold factory to cut out.

It has very, very small features. There are 600 grooves around the perimeter and they're each only 15 micrometers tall (0.015 mm).

LongPack asked to reduce the number of bumps and make them 50 micrometers taller. The problem with less bumps is that the pitch of sound gets too low and the volume drops off. The problem with taller bumps is that the needle doesn't sink into the grooves anymore. It just slaps the edge of each groove as it skips over the top of them. When it does that, it sounds more like random noise than a tone. So the dimensions need to be pretty close to what I sent them. Fortunately, they they just sent a message today to say that they think have a way to do it.

The second challenge has to do with the tiles:

The metal on top turned out to be more expensive than originally quoted. They discovered that their quote from the sheet metal factory wasn't for six tiles, but for one. So we're both trying to figure out ways to make this part cheaper as it's surprisingly expensive.

One way they figured out is to use tin-coated steel instead of galvanized steel. That was a good idea - galvanized steel has a thin coating of zinc on the surface. It gives it that characteristic "crystalized" look that you see on playground equipment and other outdoor metal that's unpainted. Tin-coated steel is less expensive and looks a little nicer. It doesn't stand up to outdoor use as well, but that shouldn't be a problem for us.

The tin-coated steel in the image is all scuffed up, but fortunately it isn't easy to scuff like that. I tried a lot of things to make more scuffs, and really only sandpaper could do it, so I'm not sure how it happened. LongPack said it wouldn't be that way in production.

They also want to make the metal significantly thinner. That causes two problems: (1) the magnetic force holding parts to the surface isn't as strong and (2) the metal bends easier - it could potentially bend upward and expose a sharp edge. I'm testing the first problem - it seems like the magnetic force is still strong enough, but I'm going to have to change the tile design to account for the second problem. We'll need two ways to hold down the metal so we can be sure it never bends upward to expose a sharp edge. 

We'll do it like this:

We'll bend the edges of the metal down and they'll snap into the plastic base. The edges will be hidden in pockets underneath. Second, we'll use double-sided tissue tape to hold the metal down tight. I just finished a prototype today. It works great and it feels higher quality.

What's Next

The injection molds are currently in the process of being made for all of the parts except the tiles and the diodes. The molds for the diodes are a little further along - they're designing them now. Now that the tile design is complete, they'll begin designing molds for that, too. We should receive injection molded samples of some other spintronic parts in the next couple weeks.

I'm ecstatic with how things are turning out. Not just the physical part of it - the puzzle books, too! (I'll leave that for a future update.) Thanks so much for supporting Spintronics! It's exciting to see it come together.

Paul and the team at Upper Story