Well, it took us about a week longer than we’d hoped to solve our auger issue…but it has finally been solved, through a combination of careful material selection, extra reinforcement, and careful machining tolerances. We’ve got some exciting news at the end for production and pellets, but first we thought we’d share the story of how we managed to solve our auger problem, i.e. what we’ve been doing the past two weeks.
Auger Issue – recap
Our aluminium augers were rubbing, which we found out and posted in our update about a month ago, when tons of bad things happened at once (“The good, the bad, the ugly”). We were pretty sure it was due to misalignment of the stators, so we added dowel pins. Nevertheless, things started to wear over time. We switched to brass, a stronger material that should have shown less wear and been more resistant to the scrape of the stator – and as of our last update two weeks ago, things were looking promising – no brass had started to wear yet.
Unfortunately, this is where things took a turn…as we wrung the brass auger through it’s tests, it too started (albeit much more slowly) to show signs of wear. We tossed around the idea of moving to an even stronger, more wear resistant material, but made the decision that continuing to prolong the inevitable wasn’t going to cut it – it was time to re-engineer the auger to eliminate the wear completely. What followed were 3 stages of work that led us to our final solution – and kept us very busy over the past two weeks!
Stage 1 – re-sizing the auger.
Rather than get too involved with re-design, we attempted to selectively cut down on the auger’s diameter. This went through three iterations, each with their own issues. In each case, we reduced the diameter of the auger in a certain area to try and reduce rubbing…first we cut down just the stators, then we cut down just the feed area, and finally we cut down the entire auger. Here are the three attempts we made, from first to last:
The first and perhaps most obvious solution was to remove lots of material off the auger where the stators were, so it couldn’t possibly contact them and start to wear. This, sadly, did not work at all – particularly with recycled material, material could bunch between the stators and the outer flute edge of the auger, and all of the pressure was now concentrated on the end point of the “large” diameter just before the start of the stators, where the auger would rub quite badly against the end of the cold tube:
After a little more investigation, we realized that in general all of this bending over time we’d been seeing was due to pellets getting caught between the auger and the feedthroat, right at the feed throat. For a while, the supporting edge of the cold tube and stators would fight back…eventually, the bending would start to dig into the stators too much and remove material. So we thought it might be worth it to try cutting down the auger just at the feed throat. While this was a lot better than attempt #1, we still had some wear just after the feed through, where the flutes would suddenly grow in size, and could bind on pellets that had been happily floating between the shorter flutes and the cold tube just moments before.
Attempt number two – much better, but a keen eye will still see some where just after the shortened flutes, as material would get pinched by the sudden “growth” of the flutes in the now constrained cold section.
At this stage, we decided to make the whole auger smaller. That way, no matter what happened, it couldn’t get pinched. We assumed based on our previous tests that performance might suffer slightly -5-10% – but it was worth it to have a durable system that we could start shipping. Unfortunately, that’s not what happened. With large pellets, everything worked fine…but the auger was too small to push anything smaller than pure virgin PLA (whose pellets are larger than ABS), and recycled material failed to extrude at all. It would just flow around the outer diameter of the cold tube and back up out the feed throat, as there was nothing stopping it from doing so!
Stage 2 – Supporting the auger where it needed it most
While stage 1 didn’t solve any problems, we now knew exactly what was happening, why it happened, and what we had to do to solve it. The auger was going to be pushed around by the pellets, and if it wasn’t supported sufficiently by the two bearings holding it at the rear, it would eventually start to dig into the stators. The solution, of course, is to also support the auger at the front, so it can’t deflect along it’s length. We knew from testing our first stator iteration over a year ago that nothing could protrude into the cold section to hold a bearing…so we made the auger out of bearing bronze, and surrounded it with a bearing bronze bushing.
This meant that the auger was supported at both ends, and couldn’t be bent or deflected over time. We tested it…and it worked! We tested it some more. It still worked. We kept testing it, and it still kept working. There was only one issue…We couldn’t remove stators as we needed a minimum of a “full flute’s worth”, which we knew from the stator design was critical to ensuring consistent performance. But we also needed a full flute of support on the auger…and having both meant we had to extended our extruder. We also had to add stators before the existing ones as well, as moving them back was causing significant pressure loss once the pellets hit the smooth bearing after the stators (those who are familiar with industrial extrusion will know about how the “metering” zone is supposed to come last – and that’s for good reason!).
The grand daddy of all drive sections…stator’d feed throat, triple stator pack, and bearing bronze support at the end of the auger. It is beautiful, it is functional, it is too long, and too expensive.
This solution also required considerably more precision machining, out of much more expensive materials (bronze vs. aluminium), with more room for error in alignment, etc etc. Instead of a cold tube and two laser cut stators, we needed a cold tube, 2 laser cut stators, two machined stator “U”s, a machined stator junction, a precision press fit bronze bushing, and a precision machined bushing housing. Not only did we have way more parts, but they all cost more to make, took longer to assemble, were more difficult to align even with dowel pins, etc etc. And we’d have to modify other parts to get the longer cold section to fit…which would take more time, and raise our costs further. So while making this solution work once was one thing – making it affordably, repeatably, and in a timely fashion was quite another.
Stage 3 – Mass manufacturing
At this point, we knew exactly what the auger needed, where it needed it, and how to do it – we just had to tweak it to be easy to make, quickly and repeatably. We met with some experts at IGUS bearings and spent the morning going over some calculations with them. We examined many, many different materials to replace the bronze press fit bushing…and found a few promising candidates.
When it was all said and done, we could replace the “stage 2″ array with our original cold tube -> stator alignment, but press one of IGUS’ plastic bearings into the end of our cold tube. It would suffer roughly twice the load of the bearing bronze, but lifetime calculations still worked out to around 70,000 hours using their “J” material – and we’re getting 20 samples ASAP to test abusively through the week and ensure they’ll hold up. Keep in mind that 70,000 hours is much longer than ProtoCycler’s been designed to last, so the factor of safety here is through the roof.
The final changes required for all of our units, therefore, are to machine new cold tubes to mate with the IGUS bushing, re-machine the auger out of brass (which, due to stiffness, wear resistance, and lubricity, is *much* better than Aluminium for the IGUS bushing and contributes heavily to the long lifetime), and re-build all of our extruders with the changes. We’re going to start doing this as soon as we’ve done a preliminary test with the IGUS equipment once it shows up tomorrow, and will post another update by the end of the week confirming progress in this regard.
In addition to solving the auger issue, the rest of the team has been diligently moving forward with production. We’ve now got nearly all of our accessory modules ready – like the spreader, spooler, and UI panel – which will help make up some of the time lost to the auger (and other) issue(s). Another hundred or so enclosures are off at the powder coater for those who ordered black, and everything is generally progressing smoothly!
While this has certainly been a long time coming, production is a much higher priority. Now that it’s all settled out, we can confirm…pellets should be live and on sale by the end of the week! We’ll be sure to either include more details in the next update, or have a separate update depending on timing, once everything is ready to go. As a friendly reminder, from our last post on the subject…
“We’ll be selling pellets in 250g and 500g amounts – this makes it easy to make as little or as much as you want, without worrying about moisture or waste. It will also let you pick different colours so you can really experiment with things. We’ll be selling PLA in both 4043D and 3D850, and ABS in 250-X10. As mentioned in a previous post, all pellets will come individually packaged in Mylar bags for 100% moisture protection, and pre blended with the colour of your choice. We’ll be selling them through Amazon to minimize shipping and fulfillment costs, and making it easy for you to buy them. The final cost will be 3$ / 250g, and 5$/500g. There’s also a $3 fulfillment fee for each order (i.e. if you bought 2kg in 500 bags, it would be $23)…shipping should be free within the USA, Canada, and Europe, but we’re still working with Amazon to confirm that. Once we’ve got everything completely finalized, we’ll put up a link and you can start purchasing pellets!”
That’s it for today – thank you all for your endless patience, and stay tuned for a follow up very shortly!
-The ReDeTec team