Product Development Series: Lens Development
The following is a transcript of a phone interview between LAFORGE Optical's CEO Corey Mack, and LAFORGE XD's Forum Manager Joshua Quillin. This is the first in our five-part Product Development series of blog posts featuring more in-depth discussions on the Shima project. These will be released in conjunction with project updates as we move forward.
JQ: This is Josh Quillin, and I am here with LAFORGE Optical's CEO and Head of Design, Corey Mack talking about the LAFORGE Shima Lens Technology: Design and Production. How are you today, Corey?
CM: I'm doing pretty good, how are you?
JQ: I'm just fine, thanks.
So, what was your inspiration for developing Shima?
CM: The genesis of Shima actually came out of a failed college project on disaster relief smart homes. The 'issue' that I had, in a sense, was the U.S. didn't experience enough major disasters around 2010-2011 to continue the research project in that vein.
So I decided to investigate just the smart home concept and discarded the disaster relief element. From that came the concept of a disk shaped device that goes on your wall, featuring a twisting frosted glass mechanism for adjusting its settings and an app to run it with your phone. But then on day one of prototyping, Nest comes out! As Nest was a very similar device, I quit because I thought that investors would be confused.
A few months after that I was at a bar, and I realized that most people only had their smart phones on them while in a restaurant, or at a party, in the office, generally when they're in an unfamiliar place. While at home, their phones are usually not being used and are either in a bag or charging in another room.
Through this I realized, “Oh, you don’t want to interface with your phone to control your house, since you’d have to get up anyway. Let’s make it on something you can wear.”
That's when I started looking at the eyewear industry and how could I make a smart home controlled by glasses, and it became clear that the current smart glasses that are out there don't look like things that people actually want to wear.
It was basically a series of bad timing and novel observation.
JQ: Why don't you give us a breakdown of the components that Shima is comprised of?
CM: At its core, Shima is a digital eyewear. The difference between what we're doing at LAFORGE and other companies is that we are an eyewear company. We're not a software development or optics company, though we deal with both optics and software developers.
The key components are:
A prescription lens with a curved mirror inside.
A frame with cutouts to accommodate the battery and other various electronics.
Skins that cover a metal frame.
That's the core of Shima. Whereas most eyewear has maybe five major components, ours has 12 - 15 depending on the design. And then in regard to total materials, ours has about 100 different individual [unique] components where a normal pair of glasses is about... ten.
JQ: How much of the Shima hardware had to be custom designed/molded/fitted to the project as opposed to being able to fit existing technology and materials to your needs?
CM: Let's look at it this way: I'll tell you what you would normally find on a pair of prescription eyewear and what we would have in common with this.
The nose pads, nose pad arms, and hinges. That’s it. That shows you how little is off-the-shelf from the eyewear industry. In terms of normal off-the-shelf from the electronics industry, the only things are the USB-C port, the auxiliary port, and the micro display. Our battery had to be specially manufactured to size, and our circuit board is completely custom from scratch; that is, basically all the major sensors you'd have on a smart phone and enough horsepower to power them fitted onto about a 10mm x 45mm footprint, or just larger than a piece of Trident chewing gum.
JQ: What does it take to make them custom like that?
CM: On the circuit board, we contract with an off-the-grid genius; we privately joke that he’s some retired secret agent or covert ops government contractor as he’s so good it’s almost scary, but he can make extremely small BGAs and integrated circuits.
On the frame side, it's more so been me being aggressive with frame designers, or, assertive about their expectations. For example, we don’t want them to try and reinvent what glasses are, we want frames just like what people prefer today but with room made for electronic components to fit within the same footprint seamlessly. We don’t want added bulk, it must fit the existing footprint.
It also took several discussions on what can and can't work with a number of our potential and existing electronics partners. For instance Intel, who was really trying to get a particular camera in Shima. We had to keep reiterating that what they were trying to do with that camera literally would not fit. It’s just one of those things where our device already weighs this much, and everything they’d need to drive that camera would double the weight just for the one component. Their camera was also 5-8x larger in volume than one we were already planning on using for a future device, so that was a really hard sell for us.
JQ: Okay. So, you were talking about the lenses and how they have a mirror inside of them, what does it take to make that?
CM: There's a few ways we make these lenses. I'll share one of the ways as the patent was just granted on it, so it's not a secret.
Right now we're casting these lenses, which means we take molds and put a clear resin or 'goo' in them, then we bake them for 2-3 days. This is actually a typical way to make finished lenses, but in our case instead of making the entire lens in one go, we actually make the back half of the lens, and then we make the front half of the lens. So it takes about twice as long as a typical prescription lens to manufacture. Between those two processes, we apply a thin mirror coating before moving the lens to the second mold. After pouring more molten material over top and baking it for a few more days, you now have a finished Shima lens.
Currently we're using Trivex, and we're also experimenting with several Mitsui products (MR6, MR8, MR10), as well as polycarbonate. Polycarbonate has been giving us some pleasant surprises regarding flexibility in what we can do with different coatings, and reducing the lens production time from four days to a matter of minutes, but further testing is required.
JQ: Working with companies overseas must be quite the logistics chain. What are some of the pitfalls and delays you've experienced due to global collaboration and production?
CM: Basically two things. One is more obvious, that being language barriers and time zone differences. We've basically gotten over that by increasing documentation, communication, and physical meetings. Back when we started we'd have communication maybe every two months and a meeting every three months. Now we've increased communication to about once a week and a meeting roughly once a month. That really helped surmount a number of delays, as well as introducing part numbers and specification sheets for prototypes.
That may seem obvious to some people, but in the eyewear space where it's more or less the same lens with different materials, there's been no real need to have super in-depth tracking in the past. Like if you break a lens or it comes out wrong, okay, you lost $0.02. It's not really worth anyone's time to track it. But with ours, where several lenses look the same but they've been made in different ways or through different processes, we've implemented ways to track them down.
In regard to delays, the biggest issue is, and not many are aware of this, we aren’t just the customer from the perspective of the supplier. We are development partners in certain cases. For example, we have to buy the lenses from our strategic investors straight from the factory. Now, we are responsible for coming up with the design; most companies you’d approach about having a lens made have their own design, but in this case our suppliers outsource parts of R&D to us, which makes for an interesting loop of information. Then, on top of that, we are responsible for dealing with Tier Two suppliers who design the molds, because as this is a different way to mold glasses for them, we also have to be their R&D team and design the molds for them. Getting even deeper into it, both our Tier One suppliers and us are acting as R&D to the Tier Three suppliers which means making the tools to make the molds that make the lenses that go into our glasses. We go several tiers down, and it's quite the process.
This is different from a lot of other companies in the “wearables” space in the sense they're just just taking a printed circuit board; something that's been mastered for about 40 years, and putting it into some sort of case; something that humans have been doing for centuries. So the downside is that due to the cutting-edge nature of what we are doing, every time we get a new supplier/partner involved there's the whole buy-in. “Who is this tiny company?” “Does that really work?” So we show them the prototypes. “Well that's really cool, but how in the hell do we make this thing?” We then go over our theories, and maybe a ten-step process. Their immediate reply? “Well, three of those you can't do here. And with this fourth one, if you did it differently, you could cut your costs down by a factor of 10.”
To show you that such a difference in cost isn't being exaggerated, I'll talk about our prototypes. In the early days at LAFORGE, in 2013 our very first prototype didn't cost that much, but it looked horrible. It cost about $250; we used parts from the drug store and some stuff we got off of SparkFun just to say, “Hey, here's something that communicates with your phone, and it shows a very blurry image, and these glasses look like hell.” That was what really drove me to start the company, which at the time was going to be called MirageTek but my friends were trolling me by saying it sounded like something from Office Space.
Then we made several other "look-like prototypes" in 2014 under LAFORGE Optical Inc., and our first real working prototype was a $15,000 thing you would put on an aluminum stick and look into the prism it held. You couldn't even attach it to your face. It came from our first lens investor, and at that point I was pretty much flat broke and just said, “Just try and make it.” We had been in negotiations for about four months, and it was to the point that I was thinking, 'If you don't believe I can make this, don't talk to me anymore.' Because I was trying to get this done. “If you think you can make it, just have your team try to make one of these lenses.” (Because that was the only unknown part then.) “Make ten. If you get one that's close, or if you find some problem, let me know.” 'And I will perhaps quit the entire project.' As is typical of when you ask an investor to do something (they don't tend to take “Orders” well from prospective investments), he didn't make 10. He made 100. And he said, “Corey, three of them came out perfect. I'm in.” So that's where we spent the first $15k on the one that wasn't wearable.
The following prototype was about six months later and $10,000, and that was the first one that you could wear. It looked like a pair of glasses with a huge beetle hanging off the side. That one really showed the concept that you could see something through this tiny little mirror that's both legible and in full color. From there the prototypes got smaller and started looking more like a pair of glasses and less like a very expensive science fair project. They actually went through our production process, the frames were lightweight, the lens wasn't a collage of plastic that was carefully glued together, and it had both the custom circuit board and battery.
The latest ones that are being printed as of today (ugh, I think we're about three months late off of our latest timeline), those are going to be very much elegant, lightweight, and should only cost us about $250 to make apiece. So, in three years it went from $15k/prototype to ~$250. Custom is expensive upfront, but the cost structure drops dramatically with each fine tuning. We think we can drop the cost by another 66% over the course of this year; our frames now cost ~5x more than a normal eyewear frame, and I know how to get cost down closer to industry average without going so low that sacrifices become noticeable. We are going for something that remains tasteful.
JQ: And for the last thing I'm sure everyone wants to know, what is the current status of Shima Alpha?
CM: Alpha is across-the-board in production and going into assembly. We should have a few dozen shortly to weigh the quality levels and will produce rev0 until we are comfortable enough with quality to progress to rev1, which has more expensive electronics. The purpose of rev0 is to confirm assumptions we have of the public, such as, “What is visible when you turn it on?”, “Is the image in the correct location, or off in any direction?”, and other questions, via survey.
Why a survey? We're using a survey because it’s a different type of eye experiment; all eye exams are subjective and are all surveys. “Better 1, or better 2?” is literally a survey. For the prescription that your doctor or his machine sees you need, what it's predicting in what it sees and the error that your brain has adapted to over your life are two different things. So the 'software' in your brain, if that makes sense, can make your eyes work a little bit better or a little bit worse than the mechanical limitations of your eye. We want to confirm our assumptions across a pool of different people before then moving to Rev2.
Beta is already in design; we have a pretty good conceptual idea of what it will look like and we are working with a connector company to figure out if there's a way to make the electronics even thinner and smaller than they are now. We're trying to get it [the circuit board] from 3mm thick to 2mm which will allow us to eliminate 50% of the components on the frame to where there will be no difference structurally between the frame, its touch sensor, its electronics, and the skin. It was a technology that was invented 5-7 years ago by a 60+ year old company that makes electronics at massive scales, and they want us to be one of the first to try this out; some of this technology is already in the Apple Watch.
So to sum up, Alphas are being made slowly we are late on them just like any other company that make tangible things: Tesla, Google, Apple, Pebble, all of us are late but we are just trying to avoid problems. Its getting to be an exciting time, and this is finally going to be the year for us to push these across the board. I’m very anxious to put these on myself. I’m looking forward to it.
