It’s the update we’ve been waiting all year for – the first Palettes have finally shipped!
This didn’t happen without a huge amount of challenges in front of our team. We swear every problem that could have happened did – from splicing issues, to Scroll Wheel problems, to drives slipping….. it was not an easy week. Murphy’s law at it again…
BUT, our team was able to relentlessly work through these issues and ship units. This shipping is the biggest milestone our company has ever hit, thank you all for playing a part.
Now, on to some of the details:
Unfortunately, we only got two units out the door. We have another eight that are assembled, and in various stages of the QC process. As always, we’d like to take some time to help you understand what challenges are facing our team, and be transparent about how things are going.
Now, the hard truth – 20% of our units passed our QC processes. We built a total of ten units, and two of them were cleared at the end of the week. Now, this isn’t exactly bad news….it means two things: 1) We learned a TON about what can go wrong with these machines during assembly and testing. 2) Our QC processes caught issues, and stopped faulty units from leaving our door.
So what were some of the issues?
We’ve mentioned this before but it’s important to re-iterate here. The Palette is a very systems heavy product. When one component doesn’t work, the entire system can fail.
Think of your car – if for some reason your air conditioning malfunctions, your engine will still start. Every system in a car is not completely reliant on all the others.
Another hurdle we face is identifying which component/subsystem is causing the failure. If The Palette stops making filament when it should, is that a Scroll Wheel problem, fried electronics, or a firmware issue?
With The Palette, if the splicer isn’t working correctly – the system fails. If the drives are not working correctly, the system fail. If there is a bug in the firmware, the system fails. If the Scroll Wheel won’t ping, the system fails.
Now…with that being said. Can you guess the four areas we had issues?
Splicing: One of the Palettes was producing splices that were too large to be fed through our QC printer. We’re investigating this issue right now, and think it may have to do with either out of spec machined parts, or faulty stepper motors/drivers.
Firmware: There was an issue with The Palette where it wasn’t making filament after you hit “Enter” to begin a print. We originally thought it was a Scroll Wheel problem, but identified that it was happening in exactly the same way, across a couple Palettes. This was identified as a Firmware bug – the firmware was forcing a “Pong,” but The Palette may refuse to recognize said Pong, if it was too close to a splice. And just like that…. Once it was identified it was fixed.
Scroll Wheel/Pinging: Three of our Palettes decided they didn’t want to recognize pings. They would make filament, the Scroll Wheel would read this filament, splices were great, prints were happening. But, without Pinging, these Palettes would slowly lose calibration over time, causing ruined QC prints. This issue is being investigated right now, and we expect to have it fixed relatively soon.
And finally…. Drives! One of our Palette’s drive gears was slipping on filament. If The Palette cannot drive filament effectively, it cannot do anything. Luckily, this was only one Palette, and it’s being investigated by our mechanical engineers.
We were having a conversation on Friday about how beautiful it’s been to develop the manufacturing processes, assembly process, and QC processes. As a team, we’ve been able to get our finger on the heartbeat of these machines, and understand what’s really going on inside them. Discovering, investigating and solving these issues today are going to allow us to ramp up our manufacturing numbers over the coming months. This also allows us to harden our troubleshooting understanding so that you can have more information at your disposal when working on your Palette.
Manufacturing is a process that is front loaded with a ton of pains. A ton of learning. And a ton of issues. These issues are things you’re always trying to account for in the sourcing process, and the assembly process – but until you jump in to build you’re never going to be able to figure every issue out.
We believe that this week we will be able to catch up with our backlog from last week (3 units), plus will aim to ship out seven new units. The reason we believe this is because the units from last week are being worked on by our engineering team, while our next batch of seven is almost complete from our assembly team. Because these products are at a different phase in their manufacturing process we’re able to work on them in parallel.
Something really important that is happening as well is that we have looked at all the failures we found in production week 1 and have already implemented further QC checks into other areas of the process. When we discover a problem, we do our best to uncover the nature of the issue to ensure that we can avoid it, or at worst detect it as early as possible next time. These additional QC checks have already caught several component defects early on that would have wasted significantly more time on solving them if they had made it into the Palette. While it is frustrating that this process of discovering problems and implementing corrective actions is so front loaded, we ultimately know that this investment will benefit everyone.
From there, we will be aiming to get ten more out, then runs of 20 per week. We’re in the process of hiring more assemblers to come on, and expanding night and weekend shifts. We’re going to be ramping up to hit our delivery goals over the next two months, and are still pushing really hard to get all of the units out by the end of May. Please understand, this may change as more issues are discovered.
We took a breather on Saturday to catch up on sleep, but our team is back in the office pushing towards these deadlines.
We’ll be posting in the comments section at the end of the week with a mini update.
And now – it’s picture time:
Close up of the splicing area sub-assemblies. Opto Sensor, machined Teflon, aluminum plate, and FQMM (Filament Quality Management Module).
How the outgoing sheet metal plates show up at our office in individually wrapped bubble packages. We love seeing how our manufacturers put as much care into the production of each part as we do.
Small sampling of mid-assembled Palettes. Set to have ten of these complete by Wednesday for QC testing/Shipping on Friday.
Proud Parents – Right/Front: Derek, Jared, Jon. Middle Row: Manmeet, Bobby, Chris. Back Row: Mitch, Firas, Justin, Brendan. Missing from photo (but always present in spirit) – Alireza
One last thing – please understand that The Palette is a first generation product. As with any first generation product, there will be limitations out the gate. The Palettes that left last week perform well, but they are not perfect. We will be releasing firmware, and software updates very frequently.
These updates will improve The Palette’s performance substantially over time. Please keep your eyes out for these updates, we will be sure to share them on our email list and our Kickstarter comments. We will be including more details on these limitations with the welcome document included in each Palette to ensure we set you up for success on the prints you are tackling.
We ask that you be patient, and share your experience with us. Your feedback will help us continue to improve the Palette . We also ask that you are around to help each other, as you will gain knowledge about The Palette that will become useful to other Palette users. We will do our best to compile and distribute useful information but ultimately community collaboration will lead to the best support system.
And with that, we need to wrap up this post so I (Chris) can get back to focusing on Palette work. Keep your eyes on our Kickstarter comment section around the end of this week for more info!