Spintronics Progress Update #11: Production Continues, Translations Are Complete
almost 2 years ago – Tue, Jul 12, 2022 at 12:02:47 AM

Friends,

This last month, LongPack continued production of Spintronics. On June 1, the lockdown in Shanghai was lifted, which allowed them to get more employees in to help with production. They're still in the process of making the plastic parts, but they've also begun to assemble the individual spintronic parts and print some paper materials. It seems they are still on schedule. That October delivery date is looking good.

In this update, I thought I'd give a quick progress update and then I have a special update from Jodie about the process of translating Spintronics into three different languages at once, which she just finished and which was a monumental task. Even if you're not getting a translated version, I think you'll find it interesting.

Progress

A lot of progress has been made, but I thought I'd mention some important points.

• First, at the beginning of June, we received a big shipment of 10 complete samples of Spintronics Act One, 10 samples of Act Two, and 10 samples of the Power Pack. Carefully looking over the samples, I made a short list of 12 changes to make. Most of the changes have already been made now.
• In the last update, I mentioned how LongPack hadn't yet made accurate resistors. They have done it now, but we both learned independently that +/-10% accuracy might be unrealistic. Now we're limiting error to +/-20% and if they do better than that, we'll narrow it a bit. 20% might seem like a lot of error, but it's not very noticeable.
• Spintronics passed safety testing! Yay!! I wasn't losing sleep over it, but it was always in the back of my mind. LongPack does a good job of only sourcing materials that will pass safety tests.

Remaining Issues

 Just three quickly and easily solvable issues linger:

1. The spring inside the capacitor doesn't have long enough legs. It slips out of place.

2. The diaphragm in the ammeter still isn't cut close enough to its edges. The extra plastic around the outside makes it not fit into the hole made for it.

3. The resistors still have the wrong bearings (shielded instead of sealed).

Jodie's Update: Translating Spintronics

At this point, we have 8 employees in Upper Story (including Alyssa and me). Jodie is our Translation and Localization Specialist. Among other things, she translated Turing Tumble into six other languages. You'd think the job would require knowing a bunch of languages, but it's actually all about finding and coordinating a bunch of different people to translate and review, and having a keen eye for detail (though some knowledge of different languages certainly helps). She's *remarkably* good at what she does. Here's her update:

For this update, we time-wind back to an office conversation one day early last summer. We were all excited about the successful Spintronics Kickstarter, and I was looking forward to starting these translations someday, maybe early in 2022. But then Paul casually asked me if I thought it was possible to have the translations done when Spintronics delivered at the end of the year.

“Wait a minute, I thought the puzzle books weren’t finished yet,” I said. 

They weren’t.

“WAIT a minute, Paul! How many translations do you want?”

“How many do you think we could do?” he asked, “Could we maybe do German, French, and Dutch?”

“Wait a MINUTE, Paul! Spintronics has not one, but two 100+ page puzzle books, times three languages is like 600 pages!”

“Yes,” he agreed. “And there will be a Power Pack, which I haven’t written yet. But that one will be really short.”

“WAIT A MINUTE, Paul! When do you want these done?”

“Well, early fall would be great...if we could do it.”

And that’s kind of how we decided to try to deliver three translations with the Kickstarter of Spintronics.

Now hold on a second! Let me just break in here for a moment to defend myself. That's not quite how it went.

*sigh* 

Actually that's exactly how it went.

The first challenge was to find the right people for the job. The Spintronics puzzle books have a graphic novel, which needs creative translation, but of course there is also a lot of very technical information in the puzzles and tutorials themselves. For example, here's a challenging tutorial to translate from Act Two:

And these need to be written in such a way that kids can understand them. We could have given the puzzles/tutorials to a technical translator and the comic to someone who specializes in creative content, but I really wanted the whole puzzle book to be translated by one person in order to keep a consistent voice. Fortunately, we had a great start on our teams after translating Turing Tumble. For some languages, we work directly with a professional translator and reviewer. For other languages, we use an agency. We have been very lucky to find people with a great blend of skills. One example is our French translator, who specializes in both technical and game translation, and is a science educator. Another example is our Dutch reviewer, who is both a science and technology journalist and STEM educator for kids. She even did some of the Spintronics play-testing.

We started translation on the graphic novel portion of the puzzle books, since that part was done. Right away, we came to some unique content that was a challenge to translate. In the Spintronics story, Natalia and her parents have a special ability to manipulate time: time-winding. This invented phrase works perfectly in English, but it doesn’t translate so easily into other languages. The teams put a lot of thought into this, and there were a lot of email discussions. One option for “time-winding” (in a language that will remain nameless) could possibly have been misconstrued as “time-farting.” Of course, we decided against that one!

We did get the comic and existing puzzles translated and reviewed by fall. Then, there was a break while we waited for the books to be 100% finished.

That was also my bad. Those books took a lot longer to finish than I expected.

Early this year, we got back to translating the puzzle books. We now had completed books, but there was over 30% more content interspersed throughout. How do you combine the work we had already done with the new parts that still needed translation? This is where Computer Aided Translation (CAT) tools are a lifesaver. This is NOT the same thing as machine translation, where content is translated by a computer. CAT tools are special software that help track and manage the process of translation. There are over 20,000 words in the puzzle books, but many of those words and phrases are repeated over and over. CAT tools help make sure you translate consistently, from puzzle 1 to puzzle 149, and they create a translation memory (TM). So, we could load the new, completed puzzle book files into the software, and the TM automatically filled in everything we had already translated, leaving just the remaining 4892 new words that had been added since last fall.

Finally, we were ready for the last step: the graphic design work. We could download the files from the translation software, which automatically put translated content in the book. But, it looked something like this:

German, Dutch, and French translations can all be up to 20% or even 30% longer than the same phrases in English! So, adjusting for text expansion is one of the most important things that happens at this step.

The graphic designer has to make adjustments to the text boxes, speech bubbles, diagrams, tutorials…there’s A LOT of detail work!

Notice how much longer the German version is once the graphic design work is done!

By the time we get to this point it feels like we’re almost done, but the list of little things to tweak keeps growing. For example, we forgot that all the decimals need to change to commas, which is the standard format for most European countries.

This whole process of translation, revision, graphic design and final proofreading happens for Act One, Act Two, the Power Pack and each of the boxes—for each language. The puzzle books alone, multiplied by three languages comes out to 972 translated pages, which is more pages than The Brothers Karamazov! 

This has been our biggest translation undertaking so far, but I'm happy to report that the German, Dutch, and French versions of Spintronics are all complete. And I'm confident that our translation teams have done a fantastic job!

Next Steps in Production

In the next weeks, LongPack will continue to make more plastic parts and assemble more of the spintronic parts. By August 1st, they plan to begin assembling the complete kits and they should be done by August 15th. I expect those three lingering issues I mentioned above will be taken care of in the next few days.

Thanks for your patience, everyone! We're getting close now.

Paul and the Team at Upper Story

Spintronics Progress Update #10: Production Continues!
almost 2 years ago – Sun, May 29, 2022 at 01:42:12 AM

Friends,

I'm happy to say that LongPack is deep in production! Even with the continuing lockdown, they've managed to make progress. They recently sent a detailed Excel sheet with production milestones and their completion dates. What would you expect to be the slowest part of the whole manufacturing process?

A. Acquiring the outsourced parts (e.g., bearings, springs, screws, magnets, etc.)

B. Producing the injection molded plastic parts

C. Assembling the spintronic parts (the junctions, diodes, transistors, etc.)

D. Assembling the kits (making the boxes, putting the spintronic parts in them, and sealing them up)

Well, fortunately it's not A. It looks like the outsourced parts will all arrive before they're required. 

B takes about 35% of the total production time

C takes about 50% of the total production time

D takes about 15% of the total production time

It's assembly of the spintronic parts that takes the most time: C! I was surprised by that.

Updated (Firmer) Timeline

In addition, we were given a firmer completion date for manufacturing. It will be August 15, which is 80 days from now. We'll get the games on a boat as quickly as possible and should have them in the mail to you by October. I did not expect manufacturing to take so long. The lockdown has really slowed things down in China. That said, I get the feeling that the August 15th date is quite firm. I don't think they're going to miss it.

This last month, our work has been all about details and fine-tuning. Things like getting the spring dimensions just right, finishing up and reviewing translations, building another ohmmeter, and designing the cartons. In this update, I thought I'd tell you about three interesting things: one answer to a mystery, one point of progress with the resistors, and one frightening moment. 

The Mystery of the Missing Pegs

In the last update, I showed how the pre-production sample of the battery was missing pegs that used to be there in a prior sample. In the picture below, you can see that only 3 of the 6 sides had the little pegs sticking out. It used to have all 6 sticking out! How is that even possible? I couldn't imagine a situation where the mold would have been somehow modified to remove the pegs. Also, that would be a hard thing to do. They'd have to somehow fill in part of the steel mold that they'd already cut out.

The reason is that the mold had TWO cavities. All of the earlier samples came from just one of the two cavities, but this time they sent a sample from the other cavity. The person cutting out the mold simply forgot to cut out those three pegs in the second cavity. They've since fixed it.

Overcoming the Ohmmeter Lockdown

Last time, I mentioned that the resistor samples all had incorrect resistance values (e.g., the 1,000 ohm resistor was only 350 ohms). I learned the reason is because the office building that hosts LongPack is under lockdown. No one could get into the building to get the resistance measuring device I sent to them. So the resistances were incorrect because the factory was shooting blind. They asked that I send them another ohmmeter to a location outside of Shanghai and I did.

I was a little concerned that they would have trouble making accurate resistors, but they appear to have nailed it. Here's a video they sent of a 1,000 ohm resistor that measures about 1,050 ohms - only 5% error! I'm hoping for +/- 10% error in production.

It's a great relief to see that they can create accurate resistors on their end.

Ammeter Freak Out

The ammeter is a part that measures spintronic current by the pitch and volume of the sound it makes. The faster it turns, the higher the pitch and volume. You might remember that it makes noise from a needle sliding across a "record" with a bumpy surface. The bumps vibrate the needle, which vibrates the air around it.

Well, everything seemed to be working with it until I received the latest samples. I spun the record and...no sound! The needle couldn't quite reach the surface of the record. Something about the way the parts were made was slightly different this time and now the needle couldn't reach. I freaked out. Would we need to modify the molds to make the needle longer? We already started production - that would slow things way down! And what if that wasn't even enough? What if we needed to make a bigger change to one of the other parts?!? Aaaaaaaaah!!!

This could have been very different Kickstarter update, but thankfully, after a couple hours, I came up with a better solution. I could just move the diaphragm! Instead of it being under the needle, I'd put it on top. Like this:

It worked brilliantly. Now the needle can go plenty far down and the quality of the sound actually improved beyond what it's ever been! I should have always had it like this.

Ok, so what is the point of that "diaphragm" anyway? Why not just have the needle by itself? 

What exactly does the diaphragm do?

I'm sure many of you already know this, but first let me back up and explain a little about sound waves. Any sound - a voice, a note played by an instrument, or a clap of thunder - is carried by vibrations in the air. Like ripples in a pond, sound waves move and spread out through the air. But unlike ripples in a pond, where the water moves up and down as waves move through it, air pressurizes and expands as sound waves move through it.

To make sound, you need to transfer vibrations (pressure ripples) into the air. The needle, by itself, is pretty bad at transferring vibrations to the air. It's small, and it only touches the little bit of air that's right next to it. In fact, most of the sound the needle makes isn't directly from the needle vibrating the air around it, but from the needle rattling the plastic body of the ammeter, which vibrates the air around it. Here's what the ammeter sounds like without a diaphragm:

Yuck! You can't pick out much of a tone from that. It's just a lot of rattling and noise. (Note that it's actually a lot quieter than that in reality - Kickstarter's website automatically normalizes the volume of a video to make it louder when it's quiet.)

Ok, now before we go on, I need you to understand how simple sound waves join together to make more complicated sounds. If I play a low frequency note on a keyboard and a high frequency note on a keyboard at the same time, the two sounds "superimpose" on each other, making a more complicated sound wave, like this:

With simple combinations of two or even three frequencies, your brain might be able to pick out the individual tones making up the sound. But if you superimpose lots and lots of different frequencies, your brain can't pick them out anymore. Instead, we interpret them as sounds with different character. For instance, if you combine enough pure tones, you can create something that sounds like a guitar string. Of course, the frequency of the specific note being played by the guitar string will be a very important contributor to the overall sound you hear, but it will also contain countless other frequencies that give it the character of a guitar string. Combine other frequencies and you can get something that sounds like a trumpet or a violin. In fact, any sound, even a human voice, is just a combination of different frequencies of sound, all superimposed on each other.

Now with some math, you can do the reverse: you can take apart a complicated sound wave and see the pure tones that make it up. It's a technique called "Fourier Transform". I applied Fourier Transform to the sound from the ammeter without a diaphragm and here's what I got:

The x axis shows the frequency of each pure tone that makes up the sound and the y axis shows the loudness of each frequency. Based on the speed the ammeter was turning and the number of bumps on the record, it should be producing sound with a frequency of 437 Hz. Even though I can't really hear it (maybe you can?), you can see a peak in loudness at roughly 437 Hz, but it's overwhelmed by noise at the other frequencies.

Now have a listen to the ammeter with a diaphragm:

MUCH better! You can hear a tone from it now: roughly an 'A' note. Here's what the Fourier Transform looks like:

There's still an awful lot of noise at a wide range of frequencies, but now you can see a little nicer peak at 437 Hz, what we call the "fundamental" frequency. There are also two stronger "overtones" that appear: one at twice the frequency of the fundamental and one at three times the frequency of the fundamental. The overtones have the effect of increasing the loudness of the fundamental frequency we hear.

Looking at the Fourier Transform, you can finally see the purpose of the diaphragm: it's there to make the vibrations in the needle louder so they drown out all the rattling and other random noises. It makes the vibrations louder because the diaphragm has a lot more surface area than the needle by itself. It can transfer the vibrations from the needle into the air more efficiently.

There are a lot of ways the diaphragm can be adjusted: the material, thickness, and shape all have a strong impact on the sound, letting some frequencies of sound transfer to the air better than others. There's still time to make adjustments to the diaphragm (it's vacuum formed, not injection molded), so I'm currently working with LongPack to try different materials and improve the quality of the sound just a little more.

Thanks again for your patience and support! It's wonderful to finally have a date for the conclusion of manufacturing. I'm looking forward to giving you news of more progress in the next update.

Paul and the team at Upper Story

Spintronics Progress Update #9: We Are in Production!
about 2 years ago – Thu, Apr 21, 2022 at 01:38:23 AM

Friends,

That's right! We've begun production. You might remember in the last update, some things were delayed because of a Covid-related lockdown in Shenzhen. That lockdown was lifted, but now Shanghai has been under lockdown since April 1st. Thankfully, the injection molding, which is the first step, is being done in Shenzhen. We're not quite sure what sort of timeline to expect for production to finish, because the final assembly is planned to be completed in Shanghai. We're very happy that they've started, though! One step at a time...

Pre-Production Sample

In the midst of the Covid lockdown, Christina at LongPack managed to get a sample of everything together and sent it to us. It would be the first sample of the entire kits: the first boxes, books, vacuum trays, and they fixed a number of issues with the parts. However, Christina couldn't assemble or even check over the sample herself because she was confined to her home. She had someone at the injection molding factory in Shenzhen put it together and ship it with her detailed instructions. They assembled everything and tried to send it by DHL like they usually do, but DHL stopped sending shipments, so they got it out by FedEx instead. It took a long time to get the sample together and a long time for it to ship.

But two weeks ago, on April 7, it finally arrived.

I opened the box, butterflies in my stomach, and...

...the boxes looked great! 

The samples worked quite well, too! I went through each part carefully and made a list of the remaining problems. I sent a list of 18 things that needed to change, but most of them were very easy fixes. For instance, one piece of the battery was assembled incorrectly, the chain color needed to be darker, the springs on the transistors, capacitors, and the ammeter weren't quite right, and the resistor stickers needed to be a different size. There were only 3 problems that may be a little more work to fix, and here they are:

1. This is the weirdest problem. The battery was fine in the last sample, but now it's missing 3 protrusions from the outside edge:

I can't figure out how on Earth that could have happened unless they accidentally used an old version of the part from when they were initially testing the mold.

2. Though they were fine in the last sample, the inductors in this sample had too tight of a fit with the flyweights.

3. My stomach dropped when I opened the Power Pack box and found silicone oil had leaked out of the 50 ohm resistor onto the vacuum tray.

I took the resistor apart to see what went wrong and was happy to see that it was an easy fix. The person assembling the resistors used the wrong kind of bearing, so the silicone oil leaked right through. There are three main types of bearings:

An unshielded bearing is wide open. Dust can get in and liquids can get through. Shielded bearings have a metal disc over each side to block out dust, but there are still gaps where liquid can get through. Sealed bearings, on the other hand, have plastic covers on either side that keep out dust and keep liquids from getting through (though they don't turn as easily). In the sample I received, they used a shielded bearing on the resistor instead of a sealed bearing, so the silicone oil leaked right through it.

That wasn't a significant problem, but what WAS a problem is that the resistance values were all incorrect. For example, I measured the 1000 ohm resistor to be only 350 ohms:

I measured the 500 ohm resistors to be about 300 ohms, the 200 ohm resistor was 150 ohms, and the 100 ohm resistor was 56 ohms. Part of the problem could be that some of the oil leaked out through the bearings during transit, but I suspect they were incorrect to begin with. They'll need to figure out how to manufacture them with more accurate resistances. We're shooting for +/- 10% accuracy.

Production Begins

Christina then asked if they could begin production on the approved plastic parts (which is all but the bottom part of the battery and one part on the inductor). On April 13, last Thursday, I said yes! Making the plastic parts takes about 50% of the total time required for production. So presumably we'll be able to get a long way before the lockdown impedes progress further.

At this point, we're in the process of finishing up translations and tying up loose ends. I find this part is one of the hardest for me because everything is out of our hands. There's nothing we can do to help or speed things up. We just...wait. I offered to help by communicating directly with the spring suppliers, but Christina told me no because I don't speak Chinese. *sigh*

Regardless, progress is being made. I expect that by the next update, we'll have all the remaining issues sorted out, the plastic parts will be produced, and we'll be assembling games. We can't tell you how excited we are to get this into your hands. Thanks again for everything!

Paul and the Team at Upper Story

Spintronics Progress Update #8: CNY Is Over and Things Are Moving Again
about 2 years ago – Sat, Mar 05, 2022 at 12:10:45 AM

Friends,

It's been quite a week. In the midst of an invasion that continues to uproot and kill people, Spintronics feels somewhat trivial. What's happening in Ukraine weighs heavy on our thoughts.

Nonetheless, Chinese New Year (CNY) is over and everyone at LongPack has been back to work for a couple of weeks now. We got some new samples of the ammeter and the chain (I'll show you those later), but generally, progress has been unexpectedly slow on the manufacturing end. Part of it is my fault for not getting the box art to LongPack as fast as I thought I could, part of it is LongPack taking significantly longer than they expected, and part of it is a COVID-related delay in Shenzhen, where the steel bearings are manufactured. We were planning to have pre-production samples ready in January (before CNY even began) and then we would begin production right after CNY, but as it is, we still don't have the pre-production sample. It should arrive in about a week.

Unfortunately, that sets our schedule back quite a bit. With production beginning in April and freight shipping beginning in May, it will be June, or to be safe, July before everything ships. I'm really sorry to you all - we initially estimated a January ship date. July is no small delay.

Hints and Solutions

The silver lining of a delay like this is that it gives us time to polish things even more. For the last month and a half, one thing we've been working hard on is a website for Spintronics, and a key part of it is the Hints and Solutions page.

With Turing Tumble, we put all the solutions in the book, but it took up a lot of space and there wasn't room to really explain the solutions, much less to add hints. So in Spintronics, we put only selected solutions in the book where they are needed. The rest of the solutions are online.

Spintronics Simulator

I made some major updates to the Spintronics Simulator. The biggest update is that it now allows you to save and load links to circuits. Before, you had to share files, but now you can build a circuit, create a link, and send it to someone else for them to try out! Like this:

A link to a spintronic circuit!

Here's the one that's up above.

A ring oscillator...in a ring!

Each of those links should load the Spintronics Simulator and show a circuit, but it's the first real server-sided application I've written and I haven't done any sort of stress testing to see what happens when more than 1 person tries to use it at a time. This update might just crash it or bring it to a crawl, sooooo...it'll be a nice little test! Thanks for being a Guinea pig. :)

Ammeter and Chain

There has been some key progress on the manufacturing side. If you remember in the last update, LongPack made a new mold for the chain links, but they only put 5 cavities in it so I could test them before they made the other 75. It was a great idea, and the 5 cavities worked well, but it was unclear whether it would work with all 80. It takes a lot more pressure to squeeze the mold together when you're injecting plastic into 80 cavities than when you're injecting into 5. Would the quality degrade? Well, it didn't! The chain mold is complete!

Additionally, the ammeter is good to go. LongPack made the grooves on the "record" deeper to make the sound louder and they fixed a number of other small issues. It works great! Here's the ammeter in one of my favorite circuits. This is called an "active current source". It tries to keep the chain moving through the ammeter at a constant speed. It constantly measures the speed of the chain and then adjusts the transistor (the magenta part) to make it faster or slower. 

Spintronic transistors aren't quite as good as electronic ones, so you'll see it doesn't keep the current as constant as the equivalent electronic circuit, but gosh darn it, it sure does try. Watch the magenta transistor make small adjustments to try to keep the current through the ammeter constant.

So at this point, the injection molds are complete! That's right. That huge, time-consuming step in the process is done! And I couldn't be happier with the results.

Preparing for Production

LongPack has started preparing for production. They ran 1,000 shots through each mold as a final test and they've created assembly guides for each part that look like this:

Once the bearings from Shenzhen arrive, they'll start practicing the assembly, making hundreds of each part to figure out the most efficient configuration for an assembly line. Here's a picture of a table they have set up for that purpose:

I'm pretty sure this will be the most complicated assembly LongPack has ever completed. I'm sure they can do it, though. Over the years that they've been making Turing Tumble, the quality of the manufacturing and assembly has improved substantially. Production samples always arrive in perfect condition and customer support requests for missing parts are far less frequent.

What's Next

In about a week, we should have complete pre-production samples of Act One, Act Two, and the Power Pack - the whole thing. I am sure there will be small adjustments to make, but there won't be anything substantial. It will be fast and easy things to fix like, "this spring needs to be slightly shorter", or "the battery spring needs to be wound three full turns before you attach the string", or "we found an error in the puzzle book and we sent you updated files".

Again, I'm so sorry for the delay, and thanks again for the patience you've shown already. It's turning out beautifully and I can't wait to get it into your hands!

Yours,

Paul and the Team at Upper Story

Spintronics Progress Update #7: The Chain Works!
over 2 years ago – Sun, Jan 16, 2022 at 11:03:32 PM

Friends,

I didn't run over any more parts with my car since the last update, but some important progress has been made! In terms of timeline, LongPack is shooting to get a complete, fully functional sample to us before February (when Chinese New Year begins). Then, after Chinese New Year ends, they plan to go into production. April is still looking reasonable for a ship date, assuming no major problems along the way.

Chain

The biggest news is related to the chain. You might remember that the first attempt at making an injection mold for the chain didn't work out. It's a big mold with 80 cavities in it, meaning each time the mold opens and closes, it makes 80 individual links. There was one big error in all 80 of the cavities, so rather than try to modify the mold, LongPack decided to scrap it and start over. The chain is an extremely important part! Nothing will work without it.

This time, they had the fantastic idea to just cut out 5 of the 80 cavities to start. If they look acceptable, then they'd cut out the remaining 75. They did that, sent me a sample, and...*sigh*...they were not good. No, they were great! They look fantastic and they work well!

The only improvement I asked for is that they make them snap together with just a little more force. Once in a while, as we were fiddling with one part of the chain, a connection would break somewhere else on the chain. Not a big deal, though. The chain problem is solved! That was probably the biggest remaining risk.

Tiles, Tile Connectors, and the Tile Connector Tool

Along with the chain, we also received samples of the injection molded tiles, the tile connectors, and the tile connector tool. As you can see from the picture, these weren't good, either. Nope, they were great, too!

They fit together perfectly, the tile connectors held them together snugly, and the tile connector tool snapped onto the tile connectors like a dream. There were only a couple of improvements I requested from LongPack:

1. The metal surface is a little slipperier than in my prototypes. In some circuits, like those with relatively high voltage, the parts slip too easily. To fix this, I asked LongPack to make the rubber-like bases of the parts slightly softer. They're pretty hard right now.

2. The tile connectors work very well, but they don't snap as much as I'd like when you've turned them all the way.

Silicone Oil Sourcing

You may remember that we need silicone oil to make resistance inside our spintronic resistors. Well, our engineering manager at LongPack flew to a silicone oil factory a couple weeks ago to check them out before they buy from them. He was nice enough to take some pictures and video so we could share them with you. I'm not sure if this factory actually synthesizes the silicone oil or if they just process it and make mixtures. In any case, I've worked in a few chemistry labs in my life, but I've never seen a lot of this equipment before. I guess working with extremely viscous materials all the time is sort of exotic, and you need special equipment to do it.

To start, as you'd expect, there are big drums full of silicone oil all over the place. I love the way the light refracts off the surface imperfections of the silicone oil. It just looks...off. Almost like water, but you can tell something's not quite right.

And here's a heavy-duty mixer. If the diameter of the mixing arm says anything, it looks like it could provide an awful lot of torque.

And then there are these things:

What are they? They're round, so I'm guessing they have pressure inside...or a vacuum? Maybe they pull out unwanted, volatile chemicals from the silicone oil? If you know, please comment! I'd love to learn what they are.

And then there are these:

What are they? I'd love to know.

He also sent pictures of some of the labs. Here's one of them. And again, I'm not sure what this apparatus is for:

Notice the green floor? I've always wondered about that. Why is it green? It seems like something you often see in Chinese and Indian factories, but you don't see green floors in many other countries. You certainly never see them in the US. Apparently it's there for two main reasons:

1. It's supposed to be calming, like green grass beneath your feet.

2. It distinguishes safe space from factory space. Green areas are safe to stand and relax. Notice the factory floor isn't green.

More Improvements to the Puzzle Books

The puzzle books are getting better and better as we work out unclear parts and smooth out puzzles that are too difficult. I've added even more tutorials to help with difficult concepts and I'm really happy with where things are now. These books were a LOT of work. As they stand now, the Act One puzzle book is 148 pages, the Act Two puzzle book is 158 pages, and the Power Pack puzzle book is 18 pages. That's a total of 324 pages in all! This is the biggest, baddest thing I've ever made.

Final Box Art

Now that I had injection molded samples of all the parts, I could take the final pictures of the game that needed to go on the back of the boxes. Once I took the pictures, I sent them to Mackey Creative (the amazing design company we've been working with for our packaging) and they finalized the box art. They did an outstanding job through this entire process. I thought you might like to see the progression - all the phases it went through before we got to the final artwork.

Remember, too, that even though Mackey was working specifically on designing the box art, they were setting the look and feel for everything related to Spintronics. The packaging, the puzzle books, the website...everything.

Phase 1

We started the process almost two years ago! Alyssa and I went to their office and brought the prototype we had at the time. We explained, as best we could, the general look and feel we were shooting for: the "mood." We gave them a copy of the story script (there were no illustrations, yet) and 30 or so puzzles. They spent quite a while with us, asking questions to understand as much as they could. If you're ever looking for something to be designed, I think this may be one of the most important metrics you can use to gauge the quality of the designer: how many questions do they ask?

We'd already worked with them in the past with the box for Turing Tumble, so we weren't as surprised this time. A couple weeks later, they shared mood boards with us. The idea was that we could pick and choose the parts we liked and zero in on the general look. Here are the three mood boards they created:

There's an awful lot there: fonts, color schemes, art styles, compositions. In some cases, we chose pieces of it that fit the image in our heads, and in others, we changed the image in our heads because it was so dang cool. Like the Spintronics logo in 1A. We adored it. In fact, we leaned hard toward a lot of the parts of 1A.

Phase 2

We realized we'd need to have our illustrator picked out before we could continue much further, since illustrations would go on the box cover. Several months later, we had rough illustrations, a more polished look for the spintronic parts, and we continued on to phase 2. Based on our feedback from phase 1, they sent us these four box covers:

Now it was starting to look like a real box! We knew right away that we liked 1a the best, but we liked the color of 1b better.

Phase 3

Once the illustrations from Alexandar were colored, Mackey continued on to the next phase. They took our feedback into account and made a more polished version of the Act One box as well as a first pass at the Act Two box. We also replaced my cruddy photographs with nicer images that I rendered. I probably could have improved the photography, but at that point, I didn't have final versions of all the parts on hand. Really a rendering was the only way to go:

Phase 4

At this point, we had just about everything we needed to finish up the box art, so we made some quick iterations. Here, they asked me to improve a number of things in the renderings. I created new renderings and they polished the art further.

Phase 5

Side panels! Mackey created 4 versions of side panels:

We leaned toward #3 without any changes. They also spent time thinking about the illustrated pictures for the front of the boxes. Up until now they had just pulled art from pages of the comic story. They created a document that told Aleksandar just what they wanted him to draw for the images on the front of the boxes.

Phase 6

Once Aleksandar created the custom art for the front of the boxes, Mackey incorporated it. They also toned down the color of the illustrations to make the box look a little less busy.

Phase 7

Now it was really close! Only minor changes from here on out. Here we made small changes to the renderings. The transistor in the Act Two is turned to more closely match the angle of the junction in Act One.

Phase 8

At this point, we took another long break. During this break, we ran the Kickstarter. (Notice the pictures of the boxes are all from the front - we didn't have art for the back of the box yet!) And after the Kickstarter was over, it was time to start work on the back of the boxes. I put together a document with all the possible information that could go on the back. They did a wonderful job organizing it and making it look good. They gave us three versions:

We leaned toward #3. It felt less cluttered than the other two because the dense information was held in the dark box at the bottom. It seemed to us like there was a clearer hierarchy of information.

Phase 9

We requested some small changes to the text and the art, and they requested some better renderings. At the end of phase 9, we had a more polished back of the box, but still without photographs that go in the holes. 

Phase 10

Some small adjustments: larger text, a better rendering of the chain. It was about done now.

We also got a first pass at the back of the Act Two box:

Phase 11

This is the last phase! When I received samples of the final outstanding parts from LongPack, I took some photographs and sent them to Mackey to incorporate in the back of the boxes. And without further ado, here are the final keylines. I just received these a few hours ago:

Act One

Act Two

That's it! I'm sending these on to LongPack today and they'll create complete samples of the entire Act One and Act Two kits: the boxes, the puzzle books, and the parts. The Power Pack will follow shortly after. That should be easy, though. It doesn't have any parts that aren't already in Act One.

Fingers crossed that everything is 100% perfect in the next sample from LongPack. Thanks again, everyone, for your support and kind words. It means a lot to us. We can't wait to get this into your hands. :)

Paul and the team at Upper Story