MakerBot Industries

When putting together my Thing-O-Matic, first with a Mk5 extruder and then a Mk6+, a perpetual source of frustration was attaching the thermocouple to the thermal block. It seemed like tightening the nut would rotate it, and it would come loose. Moreover, more than once the stresses of assembling and disassembling the extruder would loosen it as well.

Fortunately, there is a simple solution:

Mk6+ heater block with groove for thermocouple

I used a Dremel tool with a diamond grinding bit to make a groove in the side of the heater block, where the thermocouple fits. It grips the thermocouple wire as the nut is tightened, and prevents it from rotating loose. With this simple change, I can always get the thermocouple installed on the first try, and it feels really solid.

Another kind of feed rate

I’ve been a MakerBot operator for about 18 months now, and only just now am I really coming to understand Skeinforge’s nomenclature. ¹  Dave Durant did an excellent job of explaining the most important settings and terminology for Skeinforge in his series on configuring Skeinforge.  “Flowrate” is the rate of plastic exiting the extruder’s nozzle.  With a little math and science the flowrate is a function of the extruder motor speed and the volume of filament.  “Feedrate” is the rate of speed of the XY platform.  There are several different methods proposed on Thingiverse for configuring Skeinforge, and most of them involve solving for the proper flowrate or the proper feedrate for a given profile.

When using a MK5 or other DC motor powered extruder, there’s a very narrow range of reasonable flowrate settings.  The MK5 DC motors had a maximum speed of 255 PWM and a functional minimum of around 230-240. ²  Much less below 240 or so and the motor just wouldn’t turn.  The MK6 stepper motor extruder, on the other hand, has a much wider range of speeds – from around about 0.0 to up to about 3.0 RPM or so. ³

What values, then, is a 3D printer operator supposed to modify to get the best profile?

If you’re using a MK5 or DC motor extruder, I would recommend calibrating Skeinforge as you hold the DC motor speed at 255 and the maximum reliable flowrate you are able to obtain.  You’re going to want to adjust the other variables that are easier to control – the feedrate⁴ , the layer thickness⁵ , and thread width⁶ .  Since I like choosing my resolution and there’s not much choice in flowrate, I like to vary the feedrate since it’s easier to control.

If you’re using a MK6 or other stepper extruder the flowrate, feedrate, layer thickness, and thread width are all easily adjustable.  I prefer to choose the layer thickness and thread width, hold the feedrate at a level I know my robot can handle without problems, and vary the flowrate.

Here’s the interesting thing about solving for flowrate: the nozzle size ends up not being super critical.  As long as you’re choosing sane numbers for layer thickness and thread width, the amount of plastic going in will always equal the amount of plastic coming out.  While a 0.5mm nozzle would be capable of up to about 0.36mm thick layers and 0.75mm wide threads⁷ , these numbers might not make sense for a 0.4mm nozzle⁸ but could still work for a 1.0mm nozzle.

What settings do you like to vary when you’re configuring Skeinforge?

1. Photo courtesy of MadBuster75 [↩]
2. My understanding is that a PWM of 255 equates to about 2 RPM [↩]
3. A few disclaimers are in order.  You can set your RPM in the control panel to whatever you want and try to extrude – with no guarrantee it’s actually going to extrude at the mathematically calculated flowrate.  Additionally, the maximum effective RPM will be a function of the nozzle aperture. [↩]
4. XY platform speed [↩]
5. The thickness of each layer as it is laid down.  0.25mm/layer to 0.36mm/layer are reasonable ranges [↩]
6. This is the width of the filament as it comes out of the extruder nozzle. [↩]
7. Since they’re getting squished as they’re laid down [↩]
8. Since it would be tough to squeeze a 0.75mm thick layer out of a 0.4mm filament [↩]

Plastruder base top plate - with ooze!

Yesterday I tried to print in pink plastic (a bath hook for my daughter) only to discover I was unable to back the black ABS out of my MK5 plastruder.  I could extrude, but I just couldn’t back it out with the motor or pull it out after loosening the Delrin plug.

By way of background, and in the interests of experimentation, I had recently tried a few things with my Thing-O-Matic that are kinda contra-indicated by the assembly and usage instructions.  In no particular order, and at different times I had:

• Kept the plastruder warmed up and let it sit for a while
• Let a filament run all the way into the plastruder and shoved another filament in after it

After trying the above and a few different filaments recently, I noticed a  whitish smear along the inside of my plastruder.  The smear was between the toothed pulley and where the filament entered the heater barrel.  I didn’t think anything of it since I had still been able to extrude and print.

Obviously, being unable to remove filament from my plastruder is not really an option.  In order to fix this problem I disassembled my plastruder.  I quickly discovered that the white smear wasn’t just along the path of the filament.  Apparently I had managed to ooze some plastic up out of the heater barrel, around the top of the heater barrel, around the circular hole in the acrylic base to the plastruder, and around the entrance to the heater barrel all along the inside of the plastruder.  This had the effect of “gluing” the acrylic plates of the plastruder together.  Fortunately, there wasn’t a lot of plastic and it came apart rather easily.

The only explanation I have for this is that I must have kept my plastruder hot for too long without running the extruder motor.  This would allow the heat from the extruder to travel up the barrel and essentially liquify the plastic.  Then, once I started up the extruder motor, it must have squished the plastic out of the heater barrel and up into the plastruder.  I also suspect that jamming one filament in after another exacerbated this problem by squishing the filament entering the the barrel with the new filament. ¹

With the plastruder disassembled I found white smear could be scraped off.  It had a gooey residue-like consistency – like old toothpaste.  It was easy enough to scrape it off with a putty knife.  Since I already had the plastruder disassembled, I flossed the extruder toothed pulley.  I also was able to remove the filament with everything apart.  I discovered that the black ABS filament was also covered with the white plastic residue.  The extra width created by the residue on the filament probably contributed to my inability remove the filament in the first place.  I also noticed a few notches ground into the filament from when I was trying to back it out.

Once done, I reassembled the plastruder, reinstalled it, fired up the Thing-O-Matic, heated the plastruder, and did a test extrusion.  All in working order!

1. Perhaps it even squished the first filament up against the sides of the plastruder causing the smear in the first place. [↩]

Completely Printed Chess Set

Not only have I finished printing a full chess set, but the black pieces are in ABS while the clear/white-opaque pieces are in PLA.¹  When I was rocking my MK4 Plastruder² the 5 pound roll of PLA I had purchased was a sore subject for me.  Given the amount of heartache I went through trying to print with clear PLA, it seemed a natural foil to a black ABS side.

Basking in the warmth that follows a nice big Thanksgiving dinner I played chess against my younger brother, whom I haven’t bested in about ten years³ , on a board made by our father, using pieces designed by cbiffle, and printed on my MakerBot.  After a hard fought game⁴ with only nine pieces left on the board, I was up a knight, offered a draw, and we shook on a good game.⁵⁶⁷

1. Printing a full chess set has been on my to-do list for nearly a year, even before I had an operational 3D printer. [↩]
2. If you’ve got a MK4 Plastruder, I highly highly recommend upgrading to the MK5.  I could gush for days about my MK5.  You really deserve an extruder like this. [↩]
3. In chess.  Well, or anything else for that matter… [↩]
4. With much kibitzing from our dad. [↩]
5. Call me a coward, but I’d rather offer the draw while I was up, rather than fumble the endgame and get brow beaten by a pawn I forgot to take.    [↩]
6. See Allan?  I’m not ALL bravado! [↩]
7. Now, where’s did I place that printed thimble I keep all my humility in??? [↩]

Easy path to upgrades

Here’s one of the things I love about open source projects – clear to understand and consistent upgrade paths.¹

To put this in perspective, I have a name brand GPS unit.  Shortly after buying it that company released a new version that was thinner and had some new features.  Alas, no such upgrades were available for my unit.  And, really, that’s just silly.  My unit has a processor, memory, a small hard drive, and a GPS receiver.  Why shouldn’t I be able to just update the firmware to get these new features? ²

Last month brought lots of cool new upgrades for MakerBot printers as well as a brand new printer, the Thing-O-Matic.  Since MakerBot Cupcake CNC’s are an open source project, you have an opportunity with your printer I don’t have with my GPS.  The parts are designed to be modular, updated, swapped out, and hacked.  Unlike with my GPS, a 3D printer owner need not ever worry about being left behind by upgrades.  With the designs online, you can even source (or print!) all the parts you need.

Now that I’ve picked up all the upgrades for my own Cupcake, I can highly recommend them.  The MK5 Plastruder gives an extremely smooth extrusion and the Deluxe Filament Spindle Box box gives you tangle and hassle free filament feeding.  The Automated Build Platform has worked for me admirably as a heated build platform – but I expect it will work better as I get the hang of it.

1. Photo courtesy of eggman [↩]
2. Also, if it were open source I could have modified the on-screen keyboard to be QWERTY rather than have keys in alphabetical order. [↩]

PLA Pawn temperature tests

Above are twelve nearly-consecutive print-tests of Cliff Biffle’s pawn.  Once I pulled the black ABS filament out, I inserted the clear PLA, and ran it until it looked reasonably clear.  However, you can see how the plastruder was still clearing out the remaining black ABS for the first three prints.  Here are the relevant Skeinforge settings for the above pawn prints:

• The entire top row was printed way too hot – at 230 degrees Celsius.  Each print looks essentially identical, except for the variation in their coloring.  The bases are very well defined – but the “stem” of  each is slightly warped.  This is likely the result of the layers not having sufficient time too cool before the next layer is deposited.  Molten PLA holds its heat longer and stays gooier longer than ABS, so it is more prone to having layers pushed around slightly when a layer is put on top of it.  The little globe at the top of the pawn is droopy and doesn’t look very spherical.  Overall, these pieces could be acceptable as playing pieces, but I’ve seen some amazing prints in PLA and I want better for my 3D printer.
• For the middle row I tried lowering the temperature and experimented with Skeinforge’s “Cool” feature.  Notes on these prints, left to right, follow.
  • The far left pawn printed at 190 degrees Celsius and with the “Cool” feature set to 20 seconds.  The “Cool” setting made the extruder trace over the perimeter of the layer until it had spent a total of 20 seconds at that layer.  If you look closely at the picture, you’ll notice there appears to be a slight gouge out of the spherical part of the pawn.  I believe the gouge was created as the extruder essentially tore bits of plastic off the layer.  That said, this print has two other interesting qualities.  First, this pawn is significantly whiter than all other prints.  I believe this was caused by the “Cool” feature being set so high.  Secondly, aside from the gouge this pawn turned out better than any other print.
  • The next pawn over was printed at 190 degrees Celsius with “Cool” set to 10 seconds.  This print is slightly less opaque than the prior and has much larger gouges throughout the model.  This was one of the most unsuccessful prints.
  • The next pawn over to the right was printed at 190 degrees Celsius with “Cool” set to 5 seconds.  I canceled this print part way through because it was evident the stem of the print was badly mangled.  However, it was slightly more clear than the prior prints.
  • The last pawn to the right in the middle row was printed at 195 degrees Celsius with “Cool” set to 5 seconds.  This pawn’s stem was also mangled and the entire model deformed.
• The bottom row contains three prints with the “Cool” feature set to 1 second.  I’m not sure this helped at all.
  • The far left pawn was printed at 195 degrees Celsius.  I didn’t see much of an appreciable difference between this model and the ones printed at 220 degrees Celsius.
  • The middle pawn was printed at 187 degrees Celsius.  I didn’t see an appreciable difference from the prior print.
  • The far right pawn was printed at 210 degrees Celsius.  I didn’t see any appreciable difference from the prior two prints or any of the prints at 230 degrees.

I’m not sure how to proceed with testing.  The one print that looked the best, except for gouges, was the one printed at 190 degrees Celsius with “Cool” set to 20 degrees.  One potential issue with Cool set so high is that this print took more than 27 minutes versus about 7 minutes when Cool was set to 1 second.  I suspect any of these print settings might be acceptable for larger objects, allowing the rest of the layer to cool sufficiently to support the next layer.  However, I’m more concerned with dialing in the quality for smaller objects – if I can print those I should be able to print nearly anything.

Do you print successfully with PLA on your Cupcake?  What settings do you think are the most important?  What suggestions could you offer to help others?

Also, one additional photo of the same pawns as above, but with the flash on my phone turned off.  This might help seeing some of the details.

PLA pawn tests - no flash

PLA jam!!!

My MK4 Plastruder is made of entirely stock parts, except for a printed insulator retainer, and worked great with ABS.  My first attempts with PLA about six months ago were alternately wondrous and horrific.  While trying to dial in the proper temperatures for my Plastruder for use with PLA I cranked it too high, the PLA tried to expand inside the insulator retainer, and liquid PLA was forced around the threads of the extruder barrel.  As I said, it was a horrific mess.

How horrific you ask?  Just take a look at this picture of my MK4 plastruder after I tried to extrude PLA:

BLOCKED

Yeah.  It was that bad.  Well, this weekend I gave PLA another shot with my MK5 Plastruder.  At one point I realized I was trying to print while still using the ABS profiles – and heating the PLA up to 230 degrees Celsius.  This was also happening while I left the area to put my daughter to bed.  The result was the extruder was operating way way too hot for way way too long.  These are not good things.

The result is the extruder kept the heat at 230 degrees Celsius, the PLA formed a blockage, and extruder motor kept forcing new PLA down into the extruder barrel.  However, due to the way in which the MK5 is designed, there was no place for the PLA to create a mess.

When I realized what was going on I shut down the extruder and pulled out the filament, pictured at the top of this post.  What you’ll notice is the area of the filament to the right is still clear/translucent while the filament to the left, which was down inside the extruder barrel, is a milky color.  Interestingly, the clear area of filament is still just as flexible as a normal piece of filament while the milky white section is extremely rigid.

This situation, running the extruder too hot for too long, would have been catastrophic for my MK4 plastruder and was basically shrugged off by my MK5.  All I had to do was loosen the thumbscrew for the Delrin plug, yank out the filament and blockage, cut that section out, shove newly cut piece of filament back in, and tighten the thumbscrew again.  This was the difference between a 30 second fix with my MK5 as compared to a problem that actually ruined my MK4 insulating barrier and kept me from printing until I got a replacement.

I’m still dialing in my temperature settings for PLA and will be posting some of my results later today.

On Wednesday Bre revealed MakerBot has a new printer in the works. ¹  Frankly, I have a tough time imagining what a new printer would be like. ²  With an Automated Build Platform, which incorporates a heated build surface, and a MK5 Plastruder you’ve got just about everything you could want out of a 3D printer.

I’ve been really impressed with my MK5 Plastruder.  I’ve been pretty lucky and rarely have filament jams, so I haven’t seen any improvements going from a MK4 to the MK5.  However, it’s given me a really smooth, even extrusion.  With my MK4 I noticed very slight imperfections in the ABS rafts and layers – like small bubbles and imperfections in the extrusion every few millimeters.  With my MK5, these have been almost eliminated.  I’ve also noticed that the minor blobs that used to occur when the Cupcake raised the Z platform have gotten smaller.  I haven’t made many changes to the Skeinforge profile, so I tempted to attribute this improvement to the new plastruder.  Perhaps one of the coolest improvements is the ease with which I can now swap filament.  The MK4 system for pulling out a filament required me to slowly back the filament out using ReplicatorG.  The Paxtruder-inspired Delrin plug can be loosened, filament yanked out of a warm plastruder, new filament jammed in, and the plug tightened back up in 10-15 seconds.  I can definitely tell you this makes me far more likely to swap out colors.  In fact, I’m tempted to try swapping out colors part way through a build, just to see what happens.

Unfortunately, I haven’t had a chance to fire up my heated or automated build platforms yet, so I can’t comment on their utility.  However, I’m keenly aware of just how much even a minor temperature increase can improve a build’s quality.  I’ve noticed that after about two print jobs my acrylic build platform is slightly warm to the touch.  This might be due to having a few warm layers of plastic on it – or might just be due to the slightly increased ambient temperature inside the Cupcake caused by the warm plastic, power supply, and from the plastruder.  In any case, even this slight temperature increase has been enough to cause a second and especially a third print job to noticeably more flat than preceding print jobs.  Having a print bed that gets more than just slightly warm must be pretty great.

I’ve never wanted or needed to print anything larger than my current build size.  Besides, that would also increase the build time.  I’m just not that patient. 

So, just what the heck could a new printer possibly bring to the table?  Or, more importantly, what would you want to see in a new printer?

1. Photo courtesy of jmtimages [↩]
2. I guess that’s why I just blog about their developments and they’re the ones who actually dream up new robots! [↩]

Are you rocking the Five?

When the Plastruder MK5 I ordered hadn’t arrived this last Friday I was a little disappointed because I was hoping to build it over the long weekend.  Luckily it arrived late Saturday!  Huzzah!  My weekend was saved!  I had put off soldering the Relay Board I got with my heated build platform since I haven’t gotten around to picking up a hotplate yet.  However, since the Relay Board was needed to run the Plastruder MK5, I busted out my soldering kit and got to work putting together the Relay Board and Plastruder.¹  Without further ado, here are the things I learned assembling and building my Plastruder MK5:

• When attaching the thumb wheels to the bolts, definitely use a vise.  I tapped the bolts into the thumb wheels with a hammer – and smashed the threads just enough to make the bolts useless.  :/ ²  If I did this again, I would definitely put the nut partway onto the bolt and tap the NUT rather than the bolt itself.  This would keep the force away from the threads and probably keep them from being damaged.
• When threading the nuts onto the retainer mounting plate, I found it easier to put the nuts about halfway onto the bolts – rather than almost all the way to the plate.  If you put them halfway on, it’s easier to tighten each bolt individually.  If the nut is nearly to the plate, you have to simultaneously tighten all screws or tighten them each a little at a time.
• The MK5 thermistor preparation instructions direct you to the MK4 directions for wrapping the thermistor.  The MK4 directions suggest you essentially sandwich the thermistor leads in between two long pieces of Kapton tape.  This is exactly how I built my MK4 plastruder – and it’s just not easy to trap these two thin leads in between pieces of thin sticky tape.  An easier way, and what appears to have been done in the MK5 pictures, is to separate the thermistor leads and wrap the Kapton around the leads – making sure they never touch.
• I would test the thermistor’s resistance and temperature measurement before attaching it to the plastruder.  If there’s something wrong with the thermistor, I would want to know long before I taped it to the plastruder, insulated it, and mounted everything inside the ‘bot.  Both tests are quick and easy to do before you assemble everything – just check the resistance with your multimeter then plug the unattached thermistor leads into the extruder board.  This quick sanity check could save you a lot of time and supplies by not having to undo a lot of work.  This would be a good time to test resistance on the power resistors too.
  With all the wires running from the plastruder and relay board, it can get confusing.  I picked up some heat shrink tubing the last time I was at an electronics store and am using them to color code each of the sets of wires – one small band at each end.  This helps me keep everything straight.  By just looking at the end of the wires and doing a sanity check against the bands on the motor, thermistor, or heating element I can quickly verify I’m wiring everything up correctly.
• I braid the wires together to keep them all in check.  There’s no particular reason for braiding over twist ties other than it just seems more elegant.

What did you learn while building your Plastruder MK5?

1. Photo courtesy of WoodleyWonderWorks [↩]
2. I’m just that strong. [↩]