Preassembled printer test plan, logistics and documentation

Terence here.  Fatherhood certainly brought new logistics challenges, but I am happy to report that we are finally settling down into a routine with our newest addition to the family (ZT-KID-00101...) and we've been making steady progress at the shop.   You can see some of the progress on the shop here.

Early this week, I've retrieved 9 Kossel Pros in various stages of assembly from Nick, our tech.  (7 of these are fully assembled, the other 2 just needs a few parts).  We are losing Nick to a new job on the south side of the lake; we certainly wish Nick all the best in his new endeavors and we're happy to also see him grow professionally and all the cool things he's been able to do with his personal Kossel Pro.  Additionally, I have 2 machines that I've been building / rebuilding, and I will also be building 2 more machines as part of our documentation upgrade as well as training process for our new tech (more on that below).  This is enough assembled Kossel Pros to fulfill our existing Kickstarter liability (8 KSPs, 3 Orange KSPs) after refunds that have already been issued.

Midnight run to the shop to drop off the Kossel Pros, so that the car seat can be reinstalled for next morning's day care dropoff. :-P

Midnight run to the shop to drop off the Kossel Pros, so that the car seat can be reinstalled for next morning's day care dropoff. :-P

So much for trying to keep the shipping area clear.

So much for trying to keep the shipping area clear.

The first shipping crates are going out this week for laser cutting as well.  Once we receive our crates back in, there's some additional prep work to get the printers ready for shipping.

One of the most frustrating things about this project, is that comments that make it back to us on the printer falls onto two extremes:

1)  There are a small subset (<10%) of (very vocal) people for whom this printer doesn't work.

2)  This printer, especially with the K-Head upgrade, is one of the most reliable / accurate machines in this price range if reasonable care goes into the construction of the machine, and that it is the most commonly used tool in their arsenal.  Google, for instance, is about to buy their 3rd machine for their Seattle office.  

Now, obviously our shitty instructions are to blame for 1), but we don't have a very good picture of what the common pitfalls are in assembly and how it affects accuracy and user experience.  We are hoping to  unearth some data though to hopefully get a better understanding to where the pitfalls are.

Throughout this project's  history, printers have always been in short supply.  This is our biggest sample size, and this is also the one time where we get to test the printers side by side (and finally have the *space* to set up 12 printers side by side - where we aren't finally tripping over boxes of parts in a living room).  At the same time, we need to devise a way to test the printers and to test it in such a way that our backers can receive them, run the same test and compare results to see if something came out of calibration along the way.  These are the following tests that we are considering for our test plan:

A)  "Normal" tuning enchancement.  PID tuning, extruder amount tuning, logging all the custom values for the firmware.  Part of this includes writing up the procedures we use to tune our printers so that others can repeat it for themselves to dial prints in.

B)  G29 repeatability.  At this time, knowing the limitations of an offset probe, I do not consider a slight tilt in the numerical results a defect unless it affects the ability for the first layer to stick across the plate.  However, successive G29s should result in similar data (showing a high degree of consistency in the returned results).  I'll talk about what our plans are to improve G29 accuracy in a future update.  

The plan is to run G29 on each printer multiple times and record the G29's data plots.  Successive runs should all agree within 0.1mm or so.  If we see a large deviation in successive runs we'll fail the test and look at what options we have (maybe replacing the probe, for example).

From a group perspective, being able to run G29 probing on a bigger statistical sample may also give us insight as to why certain beds are tilted in a certain way.  I have designed and am building a specialty fixture to measure very accurately the center-to-center distance on the ball joint arms with a micrometer - and certainly this is part of the characterization that we'll be using when we are evaluating these printers.  One theory is that uneven gluing of ball joint arms leads to a tilted bed.  

C)  Long Print - We are going to print the 75mm OpenBeam Reprap Kossel Vertex - this is approximately a 12 hour print - from the SD card of each machine, to ensure machine stability on long prints.  We'd figure that we'd donate these parts to local schools trying to build 3D printers afterwards.  We'll be using MatterHackers Pro PLA for this test.

D)  Calibration Objects - As part of the test, we plan on printing calibration objects (and shipping these with the printers) as part of the calibration process.  Likely, this will be printed with MatterHackers Pro PLA as well.  

E)  Shipping Restraints - I've designed a 3D Printable shipping restraint that locks out motion on the arms to prevent the end effector from flying around and damaging things.  These will be printed on PolyMakr PolyMax - I've been using these for my fixtures and such and it's been working out great.  It's a good "fall back" filament if an extruder becomes fussy - I've never had a jam on PolyMaxx.  This test will also validate proper retraction settings on the printer / slic3r toolchain, as we'll be printing 3 copies of the shipping restraints simultaneously per printer.

Assuming each printer's power supply draws 400W, I can safely power and test 3 printers on our shop's circuit, and the tests and calibration should take about a week to cycle through on each batch of 3 printers.  (assuming nothing major goes wrong).  Here's our new 3D Printer test area in our shop:

Next trip:  Install power strips, laptop, networking cables for the Raspberry Pis, MatterControl Touch units, etc...

Next trip:  Install power strips, laptop, networking cables for the Raspberry Pis, MatterControl Touch units, etc...

As for the rest of the Kickstarter "plastic parts only" rewards, etc, we are waiting for our next shipment of plastics to come in.  They are due in a little bit over two weeks.  Good thing we got most of the shop cleaned up before the new shipment arrives.  Unfortunately, we are still f*cked with my other CM- my stuff still have not left on a boat yet.  I made an expensive call to ship the rest of my merchandise across on an ocean shipment first - and we should have enough stuff to hold us over for a bit.  

Finally, as these printers enter V&V testing, we will be finishing up some assemblies and shooting new video and pictures to finish up the documentation on the Dozuki site.  We'll also be documenting our printer tuning tricks as well.

That's all for this update,

-=- Terence