A Scaled-up Form 2 Print Engine, Using Giant Galvanometers

This is an ongoing project! This, and future posts will all be linked on the Project Page (once I create it)

Some time after finding the camera from my last post, I found myself again at the MIT Swapfest, amongst yet more relics from the labs and basements of the greater Boston area. I wasn’t looking for cameras this time, just…seeing if anything caught my eye. Well, something did – it wasn’t useful, cheap, or practical to carry home by hand – but I bartered the seller down and picked it up anyways, because right in that moment, I’d realized this odd chunk of metal would solve a problem I’ve been having for some time now.

source: Formlabs

The Form 2 and Form 3 printers use the careful manipulation of a UV laser to produce solid objets out of a photopolymer goo. We can turn the laser on and off quickly, and move that laser by rotating some mirrors (via some carefully-tuned Galvanometers) to redirect the beam. That seems pretty simple to understand, at least at a high level. To a user though, the underlying technology appears to be shrouded in mystery, both by our industry-focused marketing videos and the slick enclosures nicely fastened around our products. Even with the enclosures and safety covers removed (Author’s note: don’t take apart your printer, I suggest instead checking out Bunny’s Form 2 teardown), the components which perform all the magic are both tiny and swift, moving our laser beam at highway speeds during a typical print. Add to that Stereolithography, the name Chuck Hull chose back in the ’80s to describe this method of printing, and you’ve got a pretty difficult context around which to answer the question “Hey, so how does one of these work?”

That’s where my flea-market find comes into play. This is a galvo-based XY scanner, blown up to gargantuan proportions. The mirrors vary in size, but the smaller one is a 150x90mm oval – more than an order of magnitude bigger than those on the Form 2! The actual galvanometers themselves are also massive, and are designed with a rotating coil (as opposed to our design, where the permanent magnet is on the rotor ). The origins of the block are a mystery, but the galvos were made by Cambridge Technology, model #6550. This is an excellent platform to build from: it’s very open, bomb-proof, and big enough that users can interact with the components without the fear of smudging mirrors/breaking anything.

Galvanometers are only effective when paired with an effective controller; they are positive/negative current-driven devices, and typically require a closed-loop controller for servo control. I had a few potential paths forwards and ultimately decided to weigh the following options:

  1. locate a controller compatible with the Cambridge Tech. Galvos
    • Likely very pricey, and hard to acquire (these galvos are effectively discontinued)
  2. fabricate a custom controller
    • This would also require coming up with a new means of tuning the scanners, and feels like I’m re-inventing the wheel a bit…
  3. FrankenGalvos – splice together the giant Galvos and Formlabs Galvos, and control them from our own hardware!
    • We’ve already created a high-performance galvo controller, and written software for tuning/calibration – leveraging that should allow us to get up and running quickly

The position detector on these scanners appears to be capacitive, which isn’t compatible with our controller. I would need to invest some serious reverse engineering to bridge the interfaces. unless I don’t even bother with that…

Our galvanometers utilize a clever optical position sensor, and the only moving part needed is this set of light-blocking fins on the back of the rotating shaft. I decided to retro-fit this onto the giant galvos, as well as our bespoke position detector PCB. These don’t have a back-shaft unfortunately, but their mirrors do have a nicely threaded hole tapped into the ends of their drive shafts – an incredibly convenient addition. I devised an adapter, turned out of an aluminum standoff, which the fin piece was adhered to with epoxy. A printed adapter was then used to affix the position detector PCB in the center of these fins. the printed adapters allow the detector to be rotated freely for centering, then clamped into place.

A plot of impulse responses after tuning the giant galvo. The error’s quite small, despite the massive inertia of the mirrors!

With the giant galvo coils and feedback detector connected up to the driver, we tuned our controller to be stable with well over 10x as much inertia, and implemented some already-existing tools to move things to arbitrary locations. This involves some pretty standard control theory – but it’s secret sauce stuff – so I’ll suffice to say that we can now move a laser around pretty well, but in plain sight!

A janky stand, some lasercut acrylic panels and we’re ready to go!

This project was timed nicely to be showcased at an event I’ve assisted with a few times now with the Boston Museum of Science. We strapped a UV laser pointer to a hacked-together fixture, and created a board of Glow-in-the-dark plastic sheeting. despite being a very low-power 5mW laser, this material is rapidly energized, and glows brightly for a few seconds after being exposed – great for “drawing” shapes out!

The demo was very well-received, and we had quite a few “a-ha” moments from visitors when we showed them how you can turn this light painting into 3D objects. Thanks to the MoS for hosting us at their event, and to several of my co-workers who contributed to the project!

I’m hoping to continue to work on this over the coming months – I’ve wanted to collaborate with a few education connections I have to host demos of this thing at schools in the area. I’d also like to tie our controller to our more powerful tools, which can draw just about any shape! The future for this big guy is bright – hopefully not blindingly so.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s