We apologize that it's now been a month since you've heard from us. Two things have caused us to be a little "MIA" in terms of community outreach and updates - Dennon got in a fairly serious biking accident, and our motor issue required a bit more problem solving than we were hoping for. We thank you all so much for your patience - it's not easy to balance updates and outreach with progressing forward in fulfilment. It seems for every e-mail we get saying "Hey, where are you guys, I haven't heard anything in weeks?!", we get one saying "Hey, stop wasting your time on updates, and start shipping your product!". It's a difficult situation to be in as we don't want to disappear, but also do need to spend the vast, vast majority of our time on fulfillment and growing the company. That being said, this update is mostly positive, and we'll be making up for lost time on both fronts now that everything is back on track!
Injury / delays
About a month ago I (Dennon) was hit while making a left turn on a bike. While I'm very lucky the injuries weren't much worse, I was in the hospital for a few days, spent a week or so in full time recovery, and then another 2 weeks on crutches. As anyone who's been in a similar situation knows, it can be very difficult to make things when your hands are busy holding you up! I'm now fully back in action - hence the return of updates - and tried my hardest to assist Alex with the motor issue during that time. Nevertheless, when your team is as small as ours, even one player at less than 100% can have some major impacts. As you're all aware, this has meant a relative lack of communication over the past month. I did have some pre-existing engagements to fulfill, and otherwise felt it was best to spend the rest of my time assisting with our development (see below). It will take another week or two to get fully caught up on the updates, etc, but I'll be making up for lost time by posting a little more than usual...particularly on some of the great companies we've gotten in touch with, the progress we've made over the past month, and events we have planned in the near future.
Motor issue...fixed! :D
This is the big one, and boy are we happy it's over! As most of you are likely aware, we ran into some issues nearly two months ago with our DC geared motors. They just weren't holding up, and we don't want to ship you a product that won't last. We were initially hoping that either switching to a more robust DC gear motor, or substituting a stepper, would solve the problem quickly and easily. As we looked more and more into the issue, however, two things became apparent: 1) We would need a *much* bigger motor, including changes to our circuitry and control to power it, if we stuck with the geared DC route. The cost and time implications would be significant, and ultimately we'd just be delaying motor failure...not eliminating it completely. Unless we went for an industrial grade and strength motor - which by itself costs half of our product - there was going to be a cost / reliability tradeoff to make, and we didn't want to make it! 2) Choosing to use steppers would require more mechanical re-design than we were hoping for. Instead of just changing our control algorithm and adding a sensor or two to make up for the lack of current feedback, we would actually have to slightly re-design our melting and metering section to work with a stepper motor's constant speed (as opposed to a DC motor's constant torque).
We ultimately to chose to go with the stepper route. We'd started experimenting with it first, while we researched further into just why our DC motors had failed, and what it would take to overcome that. While the DC motor would have been a likely quicker fix, our research showed it would have resulted in an inferior product going forward. While we know how anxious you all are to get your units, we've never been about a quick and dirty approach. The very concept of ProtoCycler was founded on the fact that others had rushed extruder units to market, that didn't always meet the demands required of them. We didn't want to take that same path, and so made the tough decision that we wouldn't mind being a little late, if the wait was really worth it.
Plunging head first into continuing the stepper decision meant full time algorithm modification and optimization - software is *always* quicker than hardware - and when we couldn't quite make it work with just some code changes, we added a second sensor. While this got us even closer to our goal, it ultimately made it even more clear that all of the code changes in the world weren't going to fix things on their own. Instead of fixing the actual problem - inconsistent flow - we were trying to compensate for it by adjusting everything super quickly, to try and cancel out the flow problems. We realized we'd have to do some mechanical design changes to bring the flow back to where it was, so that the sensors could augment our already great extrusion, instead of trying to create it on their own.
To give some detail on the specific mechanical problem we were faced with requires some background on extrusion and motor types. Extruder units, particularly our MixFlow tech, work best with constant pressure - think of squeezing a ketchup bottle softly, or firmly - as opposed to constant speed. With a DC motor, the torque the motor outputs (and therefore the pressure on the plastic), is directly related to the electrical current it consumes...and so keeping the current constant through control algorithms meant that the pressure of extrusion was also constant. As an example, pretend that the feed auger slipped slightly for a second or two. If it was spinning the same speed, this slippage would result in a drop in pressure, and the flow would reduce, and your filament would get too skinny. However if we could sense the slippage, we could speed the auger up temporarily to maintain the pressure until it stopped slipping.
With a stepper, you can't measure pressure. Torque is held relatively constant, as a maximum threshold. The speed is then controlled exactly. So the stepper will always move the same speed, as long as the load on it is less than the maximum torque available. There is no current feedback, other than knowing if you've exceeded the maximum (when the stepper starts skipping steps). In the above example, if the auger slipped, the stepper would have no idea - it would dutifully keep stepping at maximum strength and the same speed. This causes issues because, while auger slip is rare, in the imperfect world we live in there are many factors that influence pressure and flow rate. The stepper was blind to all of them! We tried to "give it eyes" by adding a second sensor so that it could see flow variances immediately, instead of after the cooling, but no control algorithm is perfect and we just couldn't account for things adequetely enough.
Ultimately, we had to find a way to convert the steppers constant flow into constant pressure...and as of earlier this week, we've done so. Without giving too much away, we've added a simple metal plate which ensures that every X degrees of rotation builds Y lbs of force, no matter what. For the technically inclined crowd, it's sort of a (mechanical) stator. We've done quite a bit of testing with this new setup, and it has held up through our initial stress testing very well. And, since there are no gears to wear out on a stepper, there's no danger of it failing over extended usage. There are other benefits, too - we're actually using a little bit less energy, extruding a bit faster, and are less sensitive to variations in particle size (i.e. powders vs pellets vs chunks). As a final, awesome added benefit, the cost hasn't increased - it's actually gone down by about 5$! While this isn't much, we're trying to think of how we can pass this savings onto you folks for being so patient as we work through this - some free pellets to start maybe? We're open to suggestions :)
So, we're now back on track with our endurance testing, have our motor issue solved, and are extruding awesome filament again. Our estimated ship date will likely slip slightly more, i.e. fully into October, as between the injury and the longer process to resolve things, we're a little further behind than we'd hoped. Rest assured, we are working around the clock - at times literally - to try and catch up, and there is no one who wants to see ProtoCycler delivered to all of you more than us. And, as always, we welcome your feedback and comments at firstname.lastname@example.org - with the caveat that it may take us a little longer than you'd like to respond, given everything else going on right now.
Thanks again everyone, and stay tuned for multiple updates next week!
- The ReDeTec team.