Going In-Depth with The Glowforge Laser Cutter

By Norman Chan

We sit down with Dan Shapiro, CEO of Glowforge, to ask our (and your) pressing questions about his company's laser cutter.

A lot of you had interest and questions about the Glowforge laser cutter we shared earlier this month. We had many of those same questions, and were invited back to Glowforge's headquarters in Seattle to chat with CEO Dan Shapiro about how the printer works and what potential buyers can expect. Our full conversation is below, and I'd love to hear what other questions you may have that we can relay to Dan and his team!

Tested: First of all, congratulations on how successful the pre-orders have been on the Glowforge. We had a lot of responses to the first video that we did with you about the Glowforge. And I have to start by asking: why are you calling it a "3D laser printer"?

Dan Shapiro: You know, I had this giant industrial China-imported laser in my garage for a year, and I brought people through to show them what it was. I told them that it was my CNC laser cutter and engraver, but they'd say "I love your laser printer." And then later they'd say "your 3D printer is so cool." And I was like "OK, that's not what it's called" and I explained and explained. And at some point, I thought--you know we, want to make this accessible to a huge range of people. Designers and creators, as well as engineers and makers know what it is already. What can we call it that isn't already taken, like laser cutter and 3D printer, that's going to help convey what this is about.

So, it's a CNC laser cutter/engraver. All your readers are going to recognize it as such, but we think it's a little bit more than that. So we coined that term that we thought would be a little more accessible and convey the superpowers we built into it. Like the real-time full 3D autofocus and the flying lens--the ability to actually measure the depth of the material before you engrave and cut over curved surfaces, like warped wood or a MacBook. The name captures that in a way people will remember.

I think that when people look at the price and functionality, it feels a little too good to be true. There's a lot that you're doing in the cloud, and technologies you're able to leverage like cheap cameras. We know what the Glowforge can do that other laser cutters maybe can't do, but what can Glowforge not do that other laser cutters can? For example, you have to be connected to the internet to use the Glowforge.

Yep, you'll need an internet connection, and you'll need that at the start of the print. For medium-sized prints, you can cache everything and disconnect mid-print. For large prints it'll pause if it loses connectivity.

There's a couple things we sacrifice to make this possible. The biggest and most obvious one is Z-depth. So, I carved a pumpkin on my old chinese laser, because it was actually three and a half feet tall. You could load a whole pumpkin into it. It has this huge elevator bed that lowers everything down. What I'm going to do this year is take a pumpkin that has a relatively flat face, slice it off, load it in the Glowforge, engrave the face onto it, and stick it back onto the pumpkin.

So what is the depth of the printer? How thick of an object can you put into it?

You can use the tray that's included. The tray's really convenient because it has cutouts for little pieces that fall into it--shake them out like a toaster. If you use that tray then you've got a half inch of focusable space. So anything from paper that's basically zero height all the way to half an inch you can focus. If you remove that [tray], you get an extra inch for a total of one and a half inches. So, spice jar lids we've done work great, we've done flask bottles that are flat and wide. But you can't do like a wine bottle or a wine glass. That's one of the things we get a lot of questions about, that we say "for that, get something like an Epilog or a Trotec". They're awesome machines, they'll do right by you, and some have big rotary things that can rotate your object so you can etch all sides, and they can handle all that z-depth. So that was one of the big things we gave up.

One of the tradeoffs we made, and I think it worked out for the better, is that we're using an injection molded case. We're using a very particular plastic that's stronger and more robust against heat than a typical plastic. That lowers our cost, and it means you get a much more beautiful machine that's able to reduce cost a bunch of other places. Like, if you use sheet metal, like almost every other laser out there, you end up having to separately route wires, whereas we can actually mold routing into the case. Assembly of laser cutters is usually a matter of people using screws, whereas we have parts that snap together and make assembly much quicker. It's a huge upfront investment--six digits for the mold. But once we've done that, we can make Glowforges that are lighter and easier to ship. Even though shipping's really difficult, and apologies to all the international readers--we know international shipping's really terrible right now--it's way more affordable to ship that size box via FedEx and UPS than if we had made the same thing out of sheet metal.

I have to imagine that for a lot of people looking at the Glowforge, this will be their first laser cutter. They're going to look at factors like the laser power, software interface, and how big the bed is. How did you guys end up that 12x20" bed size, and 40 watt laser? And how does that compare to something you would otherwise spend 10 thousand dollars on?

That's another great place where we thought about tradeoffs. We've seen really high quality lasers which are as small as 12x16" in terms of the cutting area, and it's great when you can get something like 20x24". There, we just came up against logistics. Again, as expensive as it is to ship, if we added just two more inches of X, it would have ballooned in terms of the shipping cost. So it's as big as we could make it while still maintaining shipability.

We also decided that we wanted it a little wider and narrower, which is why it's not more square. Because that means with the Pro passthrough, you have 20 inches by infinity, instead of having like, 16 inches by infinity.

There are basically two choices of laser--metal tube or glass tube. Meta tube lasers are great...the problem is they're ridiculously expensive.

For the laser tube, we played around with a lot of different designs, and it came down to how much power can we pack into 800 millimeters of space. Because once we've established the size of the shipping box, we knew how big the laser had to be. So, the first thing was, what's the most power we could pack in an 850mm tube. We actually had one custom made because it's an unusual size, and because with the custom treatment we can actually get better beam quality which gives us higher power density. Basically the same as turning up the power.

There are basically two choices of laser--metal tube or glass tube. Metal tube lasers are great; they're what Epilog and I believe Trotec use. You can hit them with a hammer, I'm told, and they'll run fine. The problem is they're ridiculously expensive. The laser alone would have cost us the price of the entire Glowforge. The RF transistors that are necessary to drive the power supply are on the order of a hundred dollars each. So there's just no way we could get something to the price point we wanted for that.

Using a glass laser tube means we have to be more careful about shipping, more careful about transport. But it means we can get down to our price point, and there's an extra bonus. With RF lasers, if you cut acrylic, you get not quite as nice of an edge finish, because RF lasers are pulsed. But with CO2 lasers, you get a continuous beam, so you basically get a flame-polished edge.

Right, and some people might not know that--the laser is actually in the back of the cutter, not in the moving head above your material.

Yeah, so one of the things that's unique about the Glowforge is that we actually have the tube moving. It sounded kind of insane when my CTO suggested it. It seemed really weird to go take this giant glass tube and be moving it back and forth while you're cutting. But it let us do some really cool engineering stuff that's going to make the product much easier to use. I think the biggest one of those benefits is that we completely enclose both of the turning mirrors, and the output coupler of the tube. So from the mirror that's the output of the tube to the two turning mirrors to the output, there's just one window. It's flat, it's inexpensive--you can go in and wipe it, and you don't risk knocking the mirrors out of alignment when you do it. I know some of your viewers noticed we weren't showing that at New York Maker Faire--that's one of the things we're putting into our most recent builds.

With regard to materials, you have listed on the website that you can cut wood, acrylic, fabric, and etch on metal. Those are things determined by the type and power of laser you're using. How do people configure and calibrate what material they're going to be using in the Glowforge?

You have two choices for calibration as far as materials go. What we've been doing up until now, and what anyone's welcome to do, is buy any random material they want, research to make sure it's safe. So organic materials like wood are fine, but for plastics you're going to want to read into the materials safety sheet to make your own determination. Once you've done that, you can put it into the machine and start cutting squares [to calibrate]. We'll explain to users how to do this, but basically, you do a cut, you adjust the speed or power, and you do another cut. If it goes through, you try faster, if it doesn't, you try slower or with more power. You play around until you dial that in. Then you can save that setting and use it forever, as long as Home Depot doesn't change what the plywood formulation is.

The other thing is that we're going to be selling material that comes with a barcode on it that we can read in the cutter. Then all the settings just get sucked in automatically and we can guarantee it'll work.

Are you saving calibration settings locally and can users share their settings?

It's easy to share because it's just three numbers: speed, power, and focus. We haven't thought about a feature to share those, but it'd be interesting. The problem is that there's so much variability. We'll buy wood from Home Depot on two different days, and it'll require different settings, because of small differences in the material, wood, and thickness.

It's easy to share calibration settings for materials because it's just three numbers: speed, power, and focus.

Most of the time, for focus, you just say "make it focused". And we actually have the ability to do depth measurements, so it doesn't just focus once--we can track the focus over the material. If you want to get crazy, we're going to give you the ability to override autofocus, and heat the material without vaporizing it. And we're excited to see how people use that--there are ideas like cooking bacon or things like the laser origami project.

Can you explain how the laser cutter uses its two cameras to calibrate for uneven shaped materials?

I don't think I've talked about this before, but the way it actually works is on the head, there's a macro camera that can look very closely. And there's an extremely low-powered laser diode that fires down at an angle at the material. And based on how thick the material is, it'll intersect it at a different point. This is also one of the reasons we use paper on top of all of our materials, which helps with that focusing. That allows us to measure at various points of the material how thick it is so that the focusing lens can track that curve. And it happens in real time as it's moving.

For file compatibility, what types of files can the Glowforge cloud service handle?

You can send us almost any bitmap format, almost any vector format. That includes things like SVG, Adobe illustrator, PDF, TIFFs, PNGS, and JPEGs and anything else. If you're using CAD software, you're going to want to export it to something 2D, like DXF. Or, one of my favorite tricks is to use Autodesk 123D Make, which is slicing software that and take a 3D model and spit it out into layers of SVG. That's probably the most complicated workflow of anything we've done so far. It's not totally automatic to take a 3D shape and decompose it into 2D--there are design decisions to be made there--but it's not too painful.

So on to maintenance. You're selling the cutter with an optional purifier so you don't have to vent it. How long can you use that and what's the replacement process like?

First thing I'll say, is that if you're running the Glowforge non-stop, you'll want the Pro model. The hobby model is designed for someone messing around in design and then printing from time to time, maybe a few in a row. If you're cranking things out for your Etsy store, you'll want the Pro model, which has upgrade optics, cooling, and more powerful tube so it runs faster, and a longer warranty (12 months instead of 6).

That said, in terms of stuff you're going to want to replace, the air filter unit basically lives underneath the Glowforge. And it has a HEPA filter and charcoal. The HEPA filter pulls out the big stuff, and the charcoal absorbs the gases and everything else. There are a bunch of people who sell these, and we we looked to try and figure out how they rated it, because we couldn't figure out how to answer the simple question of how frequently would you need to change one. We couldn't find anyone explaining how often you would need to change those, because it depends so much on how you're using it. Acrylic clogs things up faster than wood. Wood clogs things up faster than engraving something like stone or glass. It just depends so much on usage. It may be less than a month if you're cranking a Pro model non stop and it may be more than a year if you're a hobbyist who's using it a few times on the weekend.

Is that a process where you'd have to buy a whole new $500 filter?

So you're going to lift the Glowforge off of it, pull out the filter that's in there, and drop in a new one. We haven't announced pricing for that, but we're going to keep it under $250. We're going to go as low as we can, and $250 is what we're targeting as our price cap for that.

What about the lifespan of the laser?

The tube we get from the glassblower has a rated lifetime of two years, but that's going to vary with usage. We're still running tests to see what lifespan you should expect with different uses of that. But that's something we're going to allow the user to replace as well. We priced it out and think we'll be able to get it under $500, and it comes in a kit that makes it really easy to replace without you having to do alignment, which is terrible and scary. We're going to make sure Glowforge owners don't have to do that.

You have a Basic model, and a Pro model. You said that the Pro model is really for someone who's going to be making things for production. What are the other differences?

The Pro is the more powerful laser with better optics. That's going to get you a faster cut--about 20% faster cutting times.

The Pro is the more powerful laser with better optics. That's going to get you a faster churn through--about 20% faster cutting times. It's got an enhanced cooling system, too. The Basic is designed for room temperature--70-72 degrees Fahrenheit, because it's cooled with ambient air. If you're in a shop, where the air is 80 degrees or higher, you're going to blowing hot air over the cooling fins, and it's going to pause. Pro is designed for warmer temperatures--75 to 80 degrees. You're going to get a higher duty cycle. It's not going to pause at all or nearly as much, running nonstop.

How does the material passthrough in the Pro model work?

There's a silicone skirt in the front, and a skirt in the back, so it's closed most of the time. but you can put really big material in it. Really big material, like 20-inches by as long as you like. It's human powered, so you're going to push this material through. You put one end in, and you'll cut and engrave, then you'll push it in partway, not all the way through, so the camera can still see the end of what it did last pass, register that, and pick up and continue. So that's going to let you do enormous stuff like furniture and macro scale projects.

The lid [camera] picks up what's there and sends the head over to pick up the details. It's the same technique we use if you pause in the middle of a print, pull the material out, look at it, and put it back in the bed. This is something we've got on the prototypes working, but not production ready, which is why we haven't put any videos up about it.

The Glowforge is connected, and there are the obvious benefits to that, like the database of designs that you can download. Now, if you don't want to stay connected to Glowforge, do your designs have to stay in the system?

There's no offline, there's no USB or Ethernet. So, Glowforge pulls down from the cloud. You can upload directly to the cloud service, and that can be either through a plug-in in Illustrator or by just dragging and dropping. You send it to Glowforge, and those designs are of course private. You don't have to share them with anybody if you don't want.

But we've talked to some folks, like in the military, who say they can't let their designs off of their servers. And my answer to them is that Epilog, Trotec, those guys make awesome lasers, they should get one of those. We're not going to be everything to everybody. And this is what we needed to do to make it as easy to use as we wanted to, and to not have our software team re-implementing everything three times for Mac and PC and cloud.

It's [platform agnostic]. Right now, we have plug-ins for Illustrator and we're planning them for Photoshop, Inkscape, and other common design software as well.

A big fear people may have, is what happens if they can't connect to your Glowforge servers, or if the Glowforge service goes away? What happens if you guys aren't around in five years?

This is something that came up on the first day, loud and clear. We sat down and thought about this. On the one hand, we don't want to split up our development resources to make local clients and standalone server packages. On the other hand, if you bought it, you own it, and you should be able to do whatever you want with it. And if we disappear there should be some fallback plan. So what we decided to do is open-source, basically publish one version of the firmware under the open source license so people can go mess around with it. If we disappear, they can go modify that and flash it back. Now, the unfortunate thing about that is we have to void the warranty if you do that, because we destroy Glowforges here at an amazing rate when we're tweaking firmware parameters. It's really easy to damage your Glowforge if you're playing around with firmware, but it's not designed as a primary use case, but as a backup. And there are some good solid jumping off points to make it a good standalone G-code driven laser if that's what people need to do.

So that won't include any of the cloud software, just the firmware to make it compatible with G-code. Is that open source firmware you'll be updating over time?

We haven't decided yet. We'll publish a reference version because the firmware on there honestly doesn't even understand G-code. It's very low level. It's not something where you flip a switch in and it'll work offline. But we want to make sure people can get access to configure the Wi-Fi on board and other parameters. It'll be a framework to get started. We actually do the motion plan in the cloud, so what comes down isn't even G-code.

Let's talk availability. You've said you'll start shipping in December. If someone pre-ordered on Day 1, is that when they'll be getting it?

We're shipping out the first units in December, and our goal is to make sure nobody ends up with something they're upset about. So we'll trickle out a few units, and we'll see how those go. We'll trickle out more units, and it's really going to depend on the feedback of those first units. Our very earliest ones will go to makerspaces and reviewers where they can get a ton of use. Then if everything's good to go, we'll crank up the volumes and ship on a first-ordered first serve basis. If we find out there are problems, we'll push it to early 2016 or however long it takes to get those solved. Quality is our most important consideration.

We also don't know how the factory is going to decide to manufacture. Basic and Pro, and which first. As soon as we know that, we'll let everybody know, and give everybody a chance to change their order. And I should mention that right before we ship people their Glowforge, we're going to send something that asks for their address, because we haven't asked for that yet. People move. Up until the moment you fill out that form, you can cancel with no penalties and no questions asked. We're working really hard to make something our backers are going to love, but I would rather have someone be a happy customer of someone else's company than a sad customer of our company.

Thanks to Dan for spending the time to answer our questions. There's still a week left on the Glowforge laser cutter pre-order period. You can check it out here.