The M3D Pro Update 2: Technical

Outline:

  • First Tech Update Recap
  • Pro Philosophy Recap
  • The “Claw Carriage” Story
  • Rods & Racks - X System
  • Precision Shots
  • Setbacks of the Week
  • Molding Progress
  • Native Mac Developer Hired
  • Internal Port Function - Warning Highly Technical Content!
  • Detailed Look at the Extruder
  • Native Mac Developer Hired
  • The Pro Shows Signs of Self Awareness

First Tech Update Recap

  • Updates here are being provided for primarily your curiosity/entertainment/education value.  

  • TMI Warning - Details are meant for our most active Backers, as we will share nitty gritty detail which gear-heads, and industry enthusiasts like to learn about, or parts to oggle over.  They’re also wordy, because it’s unrefined.

  • We aren’t checking spelling/grammar so we can provide you all this detail alongside everything else that we need to keep up with.

  • This info presented here isn’t always going to be unique.  Some parts will be reused elsewhere like in mega updates and comments.

  • We’re trying something new here and will see how it goes.  We are open to feedback or suggestions.  But because the level of detail is intense we are not promising to put in any time into answering questions, especially those that are distractingly specific, probing questions, or will be confusing to the majority of non-technical backers, and this would detract from our success and timely progress.  We’re going to have to let some of you speculate at times and stay out of the discourse.

 

 

Pro Philosophy Recap

The goal of the Pro is create a completely different legacy than other printers by

  • Designing as much as possible to be closed loop, making the printer more “self-aware”

  • Using extremely high-end metal parts you usually only see in products priced in the thousands, giving away a lot of value with lower margins in order to reshape the field of 3D Printing towards a better standard - one where quality and reliability are a reasonable expectation and requirement.

The “Claw Carriage” Story

Through a great deal consideration, we spent the last year considering ways to make the X axis more reliable, straight, and especially more rigid.  One of the number one comments we get about the printhead for the Micro to this day is that it seems loose, and gets surprisingly good prints given that.

Rigidity is important for more accurate bed leveling, more accurate parts, better infill, getting more consistency in bonding or layer coloring, and for flattening over extruded layers with consistency.  But make no mistake - a printer has to have a little bit of give, or it will fail for being too rigid.  Most printers have a slightly springy bed or flexible printhead.  This is necessary when a print has enough warp to crash into the print head - if the printhead doesn’t have time to melt through this part of the print and reflow some of it around it could stall and ruin a print.

This was one of the most contentious points for redesign on our team.  The Micro used high quality engineered polymers that have low friction and are resistant to wear.  This reduces cost and noise, especially compared to a linear bearing.  It also doesn’t need lubrication and can’t wear enough over 1000s of hours of use to need replacement.  The concept was a solid but the a problem is that injection molding tolerances and carbon fiber rod tolerances are poor (+- 0.1mm) do not allow for a tight enough gap.  As a result we made the plastic a little flexible, and hand-tune each one in assembly line.  In effect we pay for it indirectly in labor costs.

With the Pro, we looked at an offering of plastic linear bearing sleeve inserts, which could be inserted into the side of a sheet metal piece.  These are made of engineered polymers (i.e. “plastic” but fancier like PTFE infused POM).  The sheet metal would offer a more precise way to screw everything together, and the sleeve inserts would be a neat way to get a higher-precision low-noise system together.  These inserts are then coupled with a very high precision aluminum rod, coated in a thick anodization layer to make it black, shiny, and durable.

An assembly of this system has arrived!

Surprisingly this was our first shot at making a pre-production stainless steel sheet metal parts, and they came in phenomenally on target.  There is nothing left to do at this stage but do final prototyping on these production parts.  We’ve assembled some and can see that it’s on the order of 5-10 times more rigid than a Micro printhead but haven’t quantified it exactly yet.  

2016-09-07 01.41.29.jpg

As a matter of daily painstaking subtlety, even this part almost didn’t make it perfect on the first time.  Our manufacturer spotted us by sighting an issue and making a set of parts that alleviated the problem.  In our original design we had a bend without a relief at the brake point.  As seen on the left, this causes the holes to be misshapen.  Can you see the difference between the parts? Our supplier created a solution for it on the spot and gave us a few samples.  So we have to refine our spec a little.

Rods & Racks - X System

The rods themselves came in as well.  This is the 3rd time we’ve ordered them because they came in with turned cuts on them.  This is a subtle feature, perhaps only 2-3 microns, barely something you could feel by hand, but the ridges created noise and could add friction to the system.  We asked them to polish it better and refined our spec to match.  

Racks came in.  This one was actually on the Kickstarter risks section because it is difficult to source a straight and accurately toleranced one.  The challenge isn’t just to make a rack, but also a circular concentric one.  We started with aluminum, and it can bend when internal stresses are removed from the cutting process on only one side of an extruded rod.  So we made some more expensive stainless steel ones as a backup, to speed things along.  This rod is coated by vapor deposition with TiAlCN to make it black (Titanium-Aluminum-Carbon-Nitride coating as used on tool bits). It’s functionally pretty good, with very mild tolerancing issues, and is a little sticky.  This is our 4th iteration of getting racks and we finally have two companies that can make them, the second in aluminum (Al7075-T6), which is more likely the one we will use.  We still want to refine them slightly to eliminate any chance of interference between the teeth and bearing surface.  We’re also not thrilled about a half-circle shape as a press fit but it is a useable design.

Precision Shots

Setbacks of the Week

One of our typical challenges for the week is the heater block holders.  It’s mild setback just requiring much more careful tolerancing.  The inner hole came in as 7.8mm but should have been 8.1mm.  

Another hiccup is dimensioning on a gantry holder sample we got.  The fit is far too tight to fit the idler bearing (shown in brass).  It can be pressed in with effort but ends up with far too much friction, as it goes beyond the bearing clearances.  This test was using a new supplier so it isn’t a big surprise and isn’t a hard fix.  

Changing this week is our spool mold design.  After a great deal of design we found the Pro spool flange cost almost $2 to make.  That’s unacceptably expensive for a consumable, and would greatly raise the cost of the final product.  So we are redesigning that part.

An interesting development is a cooling design issue.  We did complex CFD analysis on our first fan designs to make sure air would flow over the right areas and keep the motors and nozzle transition areas cool.  Initial tests showed great results but our latest system isn’t cooling as well as expected.  This seems to be due to excess openings in the Pro cover.  We’re going to close them up and redo tests to get the pressure in the right places to create the right airflow.

The last issue we’re having is with power supplies, which ought to have been done by now.  Using our preferred supplier, we’ve shipped over 50,000 units for the Pro and rarely if ever seen an issue.  But their entire first batch of 100 units for the Pro were failing under stress testing.  We have a backup supplier that we are scaling up and seems to be working perfectly.  This one was unexpected.

We see things like this constantly on our path to refining the parts to perfection.  With several 1000 dimensions and features, it’s expected that about 50-100 come in wrong the first time, and that’s not considering the many additional permutations caused by multiple revisions and different suppliers.

Molding Progress

Word is the biggest two and also riskiest parts mentioned in the Kickstarter risks section are going to be ready for T1 shots on the 18th!  As this is a play-by-play, it seems optimistic, so expect delays before we see them in person ourselves!

Internal Port Function - Warning Highly Technical Content!

This is a complex topic involving a lot of moving parts and optimizations but it is worth going into.  For those who don't know what means, it’s just another way to use the micro or Pro, but not required.

We left the internal port functionality intact with the Pro as a legacy feature. It's beautiful to see a printer make prints out of seemingly nowhere. However it had proved unreliable on the Micro because the force needed to extrude was right around the safety factor needed (not enough margin for error deviation).  We stopped advertising the feature some time ago, so that there would be no expectations about it.  Also for rapid filament changes, it took more effort; with tough filament, it isn't feasible to feed it internally (you can't push a rope), which added confusion.

In order to make the Pro have a much more reliable internal port system, we're planning to make it bullet-proof for ABS-R, which is soon to be renamed M3D Professional Filament (ABS-R).  This material is a workhorse, and we call it "the gift that keeps on giving."  Compared to PLA it has much less "snap" and doesn't have sudden brittle failure like PLA can have.  It can be wound much more consistently on a spool and has a very low likelihood of catching itself and causing a jam.  In addition, its much higher glass transition temperature will make sure it can work under a heated printbed.  It requires less force for the extruder motor to crush, and also has less friction itself on the internal port tube.  

Combined with extruder force improvements, heated zone length, and other nozzle improvements in the Pro, we're getting at least 7Lbs pull force on the filament.  In order to complete the system, we plan to build a completely separate bearing assembly for the internal port, to be purchased as a low cost accessory later on;  Although this can be 3D printed and/or likely made to work without bearings, because the Extruder can detect any reduction in extrusion force and recover on the spot by slowing the print down slightly.

Lastly, we're working on yet another extruder which should increase the max force to 10lb, which may be more than the largest 3D printers out there, and giving a huge safety factor for the system to work extremely well.

Detailed Look at the Extruder

"There seems to be quite a gap between the extruder drive wheels and the entry into the stainless steel tube."

You are all very observant! This and a few other parts that make the complete assembly were not shown in the picture.  One of the key things that makes the PRO extruder is having the most reliable feed path out there - the PTFE extends up to about 0.5 mm away from the gear, and special grooving helps keep filament straight even though it is crushed - so there is no chance the filament misses the tube.  We'll show a more complete assembly as more finalized parts come in.

5tLTLSKRDiEyQ9iXBTIw_blob

"Also, is there a component to the heater missing from that assembly? "

That's right! The heater, heater block, retaining screw, and silicone aren't shown in this picture.  We've drastically improved the flow rates, consistency, and swap-ability in the new system.

"Will the end user of the Pro be able to easily swap the heater element (replace for a new component) if it is problematic or will the end user have to return the printer to M3D (like the Micro)?"

The Pro heater uses the same ceramic heater as the micro because it is extremely reliable and fast, mean lifetimes into the 10,000 hour+ range.  What was not as reliable as hoped was the method of individually calibrating printers, which made them impossible to swap in the field; also that heater moved around easily creating variances.  In the new system the Heater is locked into a brass retaining heat spreader, which increases the heated zone to 15 mm for faster flow rates (technically: better "Residence time" and "latent heat absorption").  The thermistor and retaining ring make the system completely swappable and replaceable and without needing calibration! The system is held firm to the nozzle with a set screw.

aGcSJPMKRTOHNtyC1RIV_Close up heater assembly.png

People don't publish numbers like these often enough:

Overall the designed flow rates push towards continuous printing at 12 mm^3/s with M3D Professional filament (ABS-R) or about 5mm/s which is where most extruders seem to top out.  A 5mm/s filament extrusion rate allows for a 0.5mm width path could be printed at a 0.25mm layer height or 96 mm/s print speed without skipping (continuous with no breaks whatsoever).  Another way to put this is 54 grams an hour.  That may not sound too fast but it's actual quite a lot and it's one thing we optimized for.  Many printers on the market seem to be in the 10-20g/hour range

Native Mac Developer Hired

Welcoming Andrey to the team!  Andrey is a native Mac developer with lots of experience.  With his help we’re going to revamp Mac software so it is quick, streamlined, and has a standard interface.  This is a working contract trial so it is possible things don’t work out, but Andrey has shown great promise at debugging some of our challenges already.  

The Pro Shows Signs of Self Awareness

The following link shows early proof of concept for the Pro’s “Embedded Recovery Mode” Error correction function.  The video shows a printer taking ordinary gcode, simulating printing of a square perimeter.  We got in the way of the printhead quite aggressively 4 times.  When the printer detects the error condition, it pauses, and simulates slowly moving back over the same area to do so with more power.

It may seem simple but this is a BIG milestone and is a huge proof of concept.  It demonstrates robustness, a complete algorithm, correction function, tracking, motion planning, and several layers of processing integration.  About 90% or more of the code is done in order to get this result.  So hats off to the M3D software team for this accomplishment. It was also (and still is) one of the major risks posted on our Kickstarter; the fact that we have results already is great, and even more, we were moving around 115 mm/s in this trial, which was another risk area. It shouldn’t be long before we see a more clear “before and after” style video where everyone can easily appreciate the benefit this offers as related to prints.

Posted in News & Announcements on Sep 08, 2016

  • 3d-printer (1)
  • micropro (1)
  • cheap-3d-printer (0)