As you all know it's been a very busy time for us as we try to get everything back on track for UL. In this update we'll go over just what changes we've made on some components to re-adhere to the "no 3D printing" standards on UL, and the work involved. But first, a quick progress update:
As stated, all of our parts have been spec'd and samples ordered. We're still waiting for a few that are ~a week late, but we've decided to withdraw from Maker Faire SF this year in an attempt to use that time to catch back up, so we should still be on track. While we absolutely love going to Maker Faire and were very excited to show off the final version of ProtoCycler, our current backers and customers are priority #1 and so we'll be staying home and working through the weekend to keep on top of production. Once all the samples are in we'll take some more pics of them and of the finished unit, ready for final UL.
So what are the actual changes that we've made? As we mentioned earlier, we had to remove all of the 3D printed parts - a number of you have e-mailed us expressing sympathetic annoyance at this fact, as it does seem ridiculous. Sadly, there's no way around it, and so a partial re-design was required. The two main factors at work are that everything needs to meet UL fire safety regulations (which 3D printed things that we print ourselves do not, unless we become certified as a 3D printer production facility, which adds more time and cost) and that we need to keep the cost of ProtoCycler within the realm of reason. For instance, while outsourcing to shape-ways and asking them to simply 3D print our parts in metal so that no design changes were needed drives the cost up drastically, especially considering how many 3D printed parts were involved in the previous design. So, we've had to do some re-design both to remove any 3D printed parts, as well as keep the costs down.
Two parts of ProtoCycler where this is very noticeable are the spreader and spooler. Together, they ensure the filament you produce gets nicely wound on the spool without any bunching or kinking, so they're pretty important. Some of the design constraints are below:
- Filament can't bind or get caught on the spreader, ever.
- The spreader needs to have sufficient resolution to allow even spreading of the filament...not just a few spots to spread the bunching.
- It should be easy to feed filament through, and integrate seamlessly with the rest of the device.
- Servo driven.
- Be able to spread across the full width of all the spools we've seen.
- Fit any size spool that we've come across or know of.
- Ensure the filament is tightly wound on the spool.
- Never tug on the filament more than the puller is able, to ensure that it doesn't stretch the filament overly thin before it is cooled.
- Operate at the correct speed at all times to ensure there is minimal difference between the puller and spooler speeds, and account for any difference in speeds automatically and without harm to the filament.
- Do all of the above for various polymers, all of which require different forces to wind without stretching.
- Easy loading and unloading of the spool.
- Minimal adjustment required between spools or filament types - having to adjust things everytime you wanted to extrude would really suck.
Of course, in addition to all of the above, cost is a serious factor, as is manufacturability and ease of assembly. Thankfully, we've been able to redesign these two parts meeting all of the above goals (and some additional ones) without increasing the cost beyond a very marginal amount that we've been able to account for elsewhere. Below, the old and new spreaders and spoolers:
Spreader - old vs new - front
Spreader - new vs old - back
As you can see, the entire thing is now metal! The filament guide has been switched from a 3D printed guide with teflon O-ring to a standard flanged bushing, and the servo link has been switched to a simple CNC bent wire from the 3D printed link employed before. As the major plate formed part of the fire enclosure, and required four standoff mounts, it became more cost effective to switch to a steel bracket with integrated standoffs than to assemble the standoffs into a UL fire rated plastic part. All in all, the plastic part count falls from 3 to 0, and the cost is almost identical.
Spooler - old
Spooler - new
Like the spreader, the spooler's seen some changes that reduce the plastic part count significantly (somewhat obviously, the motor is only mounted in the new version). The major issue we had with the spooler is that we need to accommodate any spool. We don't want to lock you into the spool we think is best, we want you to use whatever you want! Our previous solution was to include a few "standard" 3D printed hubs, and allow you to 3D print your own in case they didn't fit your particular spool. However, even though the spooler sits outside the fire enclosure, if the fire did escape, the spooler could fuel it - so it still needs to be rated (though not nearly as strictly as the insides). This meant no 3D printed spooler hubs - so we made a number of changes to eliminate almost all of the plastic (only two unique parts remain) and to ensure all of the user convenience was maintained.
The first big change was to mount the motor directly to the plate and create a new, all metal standoff system to brace the shaft along it's length. Together, these replaced two large pieces of plastic previously assigned to those jobs. Next, the custom hub/clutch system was replaced with standard timing belt pulleys with more flanged bushings (the same ones as used in the spreader, to save cost). As an added benefit the spool is less jerky in it's rotation now as the spring only contacts the bushings themselves, meaning it doesn't twist at all upon engagement / disengagement of the clutch.
Finally, the spooler hubs. The original design incorporated a nut in a 3D printed mount, then the first hub, then the second hub, then a plastic stopper to hold them at the right place on the shaft. After that came the mount, the custom hub/clutch, and another different nut in a 3D printed mount. We got rid of almost all of this, keeping only the two hubs (which are now coned to universally fit almost any spool), and the rear nut with 3D printed mount. The nut to tighten the spool is now built into the first hub, and the second hub rests on a pin to hold everything in place - and both hubs are molded with a combined profile so they're actually the same part, again to save costs.
The rear of the hubs - the 4 reliefs both save cost and weight, and offer a nice bonus - you can quickly tighten any slack by holding your finger against one of the internal edges and quickly rotating. The combined profile for both a pin and a nut is also clearly visible.
As you can see, the hubs are currently still 3D printed to test the fit - these are one of the samples we're still waiting on. But they work great, and allow us to eliminate almost all of the plastic parts on the spooler without really increasing the cost again! There's also less parts for you to fiddle with as you mount and un-mount spools, and you won't need to print your own spool hub if you have a non standard spool hub.
We're not generally ones to brag, but we're pretty happy with the spooler and spreader combo. ProtoCycler is the only unit on the market with spooler and spreader built in (meaning they can exactly match the speed of extrusion, giving better spool quality than anyone else). And, our competitors offer their "add on" spooler/spreader systems with no safety certification, and some serious performance issues on some models...and yet they still cost hundreds of dollars!
We hope the quick overview of some changes above gives you an idea of what we've been doing over the past few weeks to try and hit UL requirements within the cost constraints we have. It's upsetting that we've had to make these changes so late down the road when things are increasingly far behind, but ultimately they combine to make a safer machine and allow us to create a better overall user experience while we're at it.
-The ReDeTec team