November News - Calibration, New Closed Loop System, and Mosaic Update

The whole team here is pretty excited to give everyone a look at what’s been going on here the last month. We’ve seen some really promising results, as always, we want to give you an in-depth view into the behind the scenes technical progress. We’ve also got some pretty awesome company updates we’d like to let you in on.

So like always…. let’s dive right into the tech!

To start off, we’ll give you a quick recap as to where we left off.

Last month, we decided we needed to hit a target of 100-120 meter long prints before we moved into production. We mentioned a few systems we were considering putting into The Palette to close the loop on our drive systems. Our hypothesis is that by closing the loop on our own drives we would be able to occasionally “reset” calibration back to Time 0, never letting anything slip too far in the print.

As with any solution, it might sound good in theory, but it needs to be investigated and validated. In order to do this, we built ourselves a quick test rig with a limit switch and some 3D printed parts. Check out the rig below:

So, what are you looking at? (Warning – technical content)

We’re going to provide a bit of background context, as well as an analogy that we believe will help illustrate the entire situation better. It’s also going to help you understand where some of the initial inaccuracies were coming from, and how our new system has fixed the calibration issue we were having.

When we were first starting to make The Palette we realized we needed to be able to make filament in a non-continuous manner (stop/start), while keeping up with a largely continuous process, 3D printing. To do this, we came up with the solution of the “Buffer Loop,” or male/female Teflon overlapping sections.

This telescopic mechanism expands when The Palette makes filament, and contracts when the printer uses filament. This lets the printer continuously print filament while The Palette is splicing more filaments together.  

Expanded Buffer Loop

Contracted Buffer Loop

We built the new sensor system to measure the length of the buffer from time to time. Think of it like homing the axes on your 3D printer. When your printer starts up, it hits a limit switch for the X, Y and Z axes. This homing sequence tells the printer where the extruder is, in an absolute, rather than a relative sense. This is a very important concept to grasp (absolute vs relative), as it is the basis of where our inaccuracies were coming from.

Something you may, or may not know, it that during a print your printer does not know for certain where the extruder is. Your printer knows where the extruder is relative to where it used to be.

Printers track the amount of steps it sends to each stepper motor to move in a certain direction. So what it understands it that the extruder is “n” steps away from its homing point on the Y-axis, and “m” steps away from its homing point on the X axis. It does not, however, know in an absolute sense, where it is. It believes it knows where the extruder is located, but this location isn’t always accurate.

Have you ever had your extruder catch on something, motors skip steps, and the printer starts printing somewhere else on the bed? This is because the printer doesn’t know the skipped steps occurred, and believes it is somewhere it isn’t.

The filament got tied up on an unattended print, and caused the motor to skip steps. These skipped steps have no way of being measured without re-homing the printer, something current machines don’t do. The result? A ruined print

The same was true with The Palette. We always talk about how drive systems aren’t perfect, how filament skips steps and isn’t driven the true amount. Over time, these tiny miss-steps add up, causing differences in how much filament was driven, compared to how much filament The Palette believes was driven. The Palette used to operate on a relative basis, but that wasn’t good enough.

So we fixed the slipping problem by “homing” the buffer loop at specific times in the print. This homing lets us create a check in point, and allows The Palette to understand the absolute (homed) point of where it is, versus where it should be. When The Palette sees this homed point, it compares where the filament is to where it thought the filament was. Using this comparison, The Palette is able to figure out how much of an adjustment needs to be made to bring the print back into perfect calibration. The Palette is no longer comparing to relative lengths, it is comparing to an absolute point, a point of theoretically perfect calibration.

Now that we’ve proven the concept with the quick test rig shown above, we need to turn the solution into a finished system. One of the design challenges we have is how integrated each part of The Palette system is. Because of this, when we design, we must consider a holistic view of the system to ensure that each new part compliments the rest of the system and leads to the reliable, repeatable experience we are committed to.

So where are we now?

Long-winded technical explanations aside, we’re a lot further along than we were 30 days ago. But this doesn’t quite mean things are perfect. As with any new features and systems, this one has introduced one more challenge that we weren’t experiencing before. Extra strain on splices.

Because of the way the buffer measurement works (tube hitting a limit switch), we have seen some extra pressure being put on splices, occasionally causing them to break. This is the last hurdle we need to get over before freezing design, and moving into full production.

But don’t worry, we were successful with our first test regarding calibration, we’re still on track to deliver on the timeline we announced in our last update!

We will have the production ready version of The Palette frozen (ready) by January. But here’s the thing, just because the design isn’t frozen doesn’t mean we aren’t working on the best way to produce The Palette today.

That leads us to the next portion of our update – production preparation.

This isn't even our final form!

After we finished our Kickstarter campaign, we went out and expanded our team to include a mechatronics designer, software developer, and production designer/engineer. This means we’ve had an experienced individual working on production preparation since July. We actually hired one of you (our backers) who contacted us after we put out a call to hire in our May Update.

So what progress has been made with our manufacturing setup?

There’s a lot that goes into manufacturing a product – from certification, to sourcing, to assembly and quality control. These things have been going on in parallel over the past four months.

To start off, we’ve broken The Palette down into a series of sub-systems. By doing this, we’ve been able to start documenting the assembly instructions that (one day soon) we’ll train our assembly team on. We also have a much better idea of how many Palettes we can make each month, and the time commitment that goes into each unit.

We’ve been ordering more test runs of some of the critical components needed for certification. When creating a product we need to ensure that we’re able to hit certain certification guidelines, which means using the right power supplies, wires, motors, etc. By ordering test runs now, we’re able to start to cement our supply chain for things we know are not going to change.

Production’s going to be a huge challenge, but we’ve been prepping for months. When its time to pull the trigger on the frozen design, we’re going to be ready to go and get you the best possible product in the shortest period of time.

There’s one more thing we wanted to talk about on the tech front – The Scroll Wheel.

Its coming along super well - Our goal is to have the production design finalized early next week. We didn’t want to take away from the importance of The Scroll Wheel being finalized, so we’re going to save an in-depth dive for our next update. Seriously exciting things happening on this front, keep your eyes peeled next month.

That wraps up the technical side of things – now we want to take a minute to give you an update on the rest of the happenings at Mosaic.

We finally moved into our new office, after months of contract work to get things complete, it finally feels like home. We wanted to throw up a few pictures for everyone so you could take a second and understand what you really helped bring to life.

View from the entry way to our office – production space on the right, kitchen on the left. Open work space straight ahead

View of our open work space – complete with 10 works stations, a printer wall, and a couch/Wii for when we need to relax.

We also went out and hired two more people to add to our team. We realized we never took the time to show you the people behind Mosaic, so we all got together earlier this week to take a family photo. Below is the team of really smart engineers (plus Chris) working to ship you an amazing product.

As always, if there’s any information you would like feel free to drop us a comment below!