by Chris Labelle
@Ivan - I pointed one of Mosaic's cofounders to your questions so we could get you a more detailed response - answer copied below:
Splicing: Splicing is a complex interplay between the drive systems, splicer machined parts, and the filament you are using. We are still working hard to uncover the nature of the splice issues that are showing up in a handful of Palettes on the production line but we believe it is being caused by out of spec machined parts and out of spec drive gears. We also found that the alignment of the hot-tool can have an affect on splice quality. Based on this finding we have included a hot-tool alignment tool with every Palette to ensure that you can level your hot tool if ever this is required. The good thing is we are able to detect splicing issue very easily (broken splices or splices that will not feed into a printer) and thus we can ensure no Palette leaves our facility before being fixed. We will let you know more as we get to the route of the problem.
Firmware: A pong is a process that is responsible for homing the buffer loop. When the Palette is printing, the buffer Teflon and the Palette Teflon slide along one another forming a “filament buffer” that is required for the printer and Palette to work together. As you may have seen, there is a ring magnet attached to the end of the Teflon tube leading to the ingoing side of the scroll wheel. Every once in a while (10 minutes or so) the Palette executes a pong sequence where it shrinks this buffer all the way so that this ring magnet gets very close to the Palette. The Palette has a magnetic sensor inside of it that detects the presence of the magnet to get the true location of the buffer. This homing sequence account for error that accumulated in the buffer system over time effectively eliminating it.
Scroll Wheel/Pinging: We have learned a lot about why the pings were not being registered, but similar to the isolated splice problems we still have some work to do in getting to the route of this. We found that the number one causes of ping issues were poorly calibrated Palette settings as well as poor print conditions (bed not level, poor adhesion, old dry filament). The pinging process is something that we plan to work on substantially in future firmware releases. We would like to make it more robust and less sensitive to the factors mentioned above.
A ping is a checkpoint in a print that the Palette uses to stay in calibration. The Palette software adds pings into the G-code of the print. For instance, a 2 hour print may have 15-20 pings inserted into it. A ping sequence consists of a few short pauses that the printer will do while printing the transition towers. The Palette is continuously searching for these pings by using the scroll wheel to monitor the filament motion. When the Palette registers a ping, it uses the value it has form the scroll wheel (filament printer) and compares it to the filament that should have been used according to the g-code at that point. If there is a difference between these numbers (i.e. if the scroll wheel thinks the printer used 5cm less then the g-code required) then it will make a corrective action by adjusting future filament generation by about 5cm. Pinging (and ponging) are not something that you need to worry about as a user, they are simply processes that the Palette uses to keep in tune with the printer.
Drive: Lastly, we have continued to invest a lot of time investigating the nature of the drive failures we have seen. You will notice when you get your Palette that the drive system we use is a “constant distance” type. This means that the idler and drive gear are a constant distance from each other (no spring involved to push them together). This system leads to very consistent motion control across different filaments. One of the challenges of this system however is that this drive spacing relies on the tolerances of the steel drive gears, stepper motors, idler bearing, and the acrylic drive casing. We have found that gears coming from our supplier will vary in diameter by a maximum of 0.5mm, which is quite substantial. We have implemented a QC check to measure drive gears to ensure we only use gears in a band of +/-0.15mm. We have also found that the tolerance on laser cut acrylic can lead to significant variance. We have implemented solutions such as optimized cutting patterns, and better location through the implementation of dowels that have helped largely solve this problem. We still need to get the last 10% of the way by further refining our drive gear sorting QC process to ensure that a higher percentage of finished drive assemblies pass our final drive QC process. Similar to the splice issues we have a very robust system in place to detect faulty drives so the concern of them leaving in a finished Palette is very low. In the mean time, failed components are costly in terms of money and time so we hope to solve this problem as soon as possible. Our production goal is to always be increasing the amount of time we spend building new Palette’s rather then fixing faulty components. We will keep you posed!
Hope this answers your questions - let us know if you need anything else
:) What a great team! Love these updates Chris, so rich with design detail, and seeing all the troubleshooting steps really makes one appreciate how much work goes into a good product!
Hi Chris-Mosaic Team,
From your update: "...the four areas we had issues":
Splicing: What are the causes? Out of spec machined parts, or faulty stepper motors/drivers or ?
Firmware: What's a "pong"?
Scroll Wheel/Pinging: Have you all identified why the scroll wheels were unable to recognize pings? What's a "ping"?
Drive: cause identified?
Keep it up guys! Cant wait to get mine! One of the best projects ive been following and getting into since im on crowdfunding!
Woo awesome news! Keep up the good work!