Saturday, 3 February 2018

Household 3D Prints Roundup 2

LED candle shrouds
     Been a while since I covered prints that I've put around the house, so here's some of the stuff that's been made over the past year. Starting with the above, those are simple spiral vase prints that have a small LED tea light inside for decoration. Design wise, they're really simple, just a faceted cylinder with lots of extra horizontal rings twisted along it's vertical axis, I used Wings 3D 2.1.5 as my design program of choice here. Filament is MG Chemicals Gold PLA, prints look like natural beeswax in colour, really nice for making decorative shades and such.


Toothbrush Holders
     Not the kind of thing you'd expect to use transparent PLA vase-mode prints for, but the faceted spiraling makes them surprisingly sturdy for this kind of use. Again, custom designs done in Wings 3D, using 3D Solutech Natural PLA this time, all printed on the Mega Kossel.

Birdseed filter in use
Birdseed filter clean
Birdseed filtering funnel
     Next up is something bird related, I've got a couple of budgies that are rather picky eaters, they don't like large sunflower seeds and keep emptying the food dish onto the floor if the find them. My solution to the problem is to design a funnel with a removable filter grate that strains the largest seeds out but lets all of the smaller stuff through. It's a simple design, just a cone with a 2 cm cylinder stuck on the point and hollowed out to about 3mm thick walls, the real trick is that rather than modeling the final grill in the CAD software, I just modeled the outer dimensions for it and used the infill settings in my slicer software to create the grid. I'm using Slic3r for my slicing program and the filter grid is the result of using 10% rectangular infill with zero top and bottom layers.

Towel hook on bathroom door
    And finally, this is the current iteration of the Customisable U-Hook from Thingiverse that's being used as a towel hook on the bathroom door. It's printed in MG Chemicals PETG for strength, the previous ones were in PLA and both snapped after about 6-9 months of daily use, this one is currently at 5 months and counting, be interesting to see how it holds up after a year of use. And in closing, I redid the oven knob from last time, the original part wasn't thick enough on the walls for the adaptor plug and failed after about a year of use. the current one was made in one piece with a thicker walled plug and is still going strong after about 9 months of regular use.

Oven knob V2

Thursday, 4 January 2018

3D printing and Musical instruments

3D Printed Electric Violin and Cello
    Over the Christmas season my family discovered some of The Piano Guys music videos and I remembered coming across the O'Cello & Mina Violin files, so I got some transparent PLA and started printing the parts. The designers intended for the parts of both projects to be printed on a standard i3-type printer, but I ended up using the Mega Kossel instead. 


Mina Violin neck freshly printed
Printing the Mina Violin body on the Mega Kossel
   As you can see, printing some of the parts needed some creative positioning of the parts, the lower body of the violin was particularly tricky, I ended up standing it on end, rotated by 30° onto one side and tilted back by 10° to get it to fit. After printing, some cleanup was needed, mostly stripping support material and stringing off, but on the violin the fingerboard came out a bit rough, so I sanded it down smooth then removed the heat induced scarring, as sanding PLA tends to mess up the finish.

Sanded PLA fidget cube
     The solution to the scarring issue is to take a heat-gun and run it lightly over the surface, don't pause in one place or you'll melt the print, and after a few passes the colour will reset and be indistinguishable from a freshly printed surface in colour. I'm using a light-duty 'art & craft' gun, but for the more industrial models you'll want to stick to the low setting, the higher one's will just wreck the print.

Heat-gun from Opus Art Supplies


Heat-treating before (bottom of cube) and after (top of cube)
Finished cube, note the distortion on the thin sections from overheating
    With the printing finished, the next step was sourcing the non-printed parts, I ended up using these piezo pickups from Amazon.ca instead of the default ones on both the violin and cello because the original ones are only available on Amazon.com and don't ship to Canada. The tuning pegs were these ones for the violin, and these for the cello. The strings are D'Addario Prelude for the cello and Thomastik Dominant for the violin, both bough from the local music store. All the fasteners were easily found at the local hardware store along with the structural threaded rods, 5/16" for the Violin and 7/16" for the Cello.

Mostly assembled O'Cello
Stringed O'Cello with Viola bow for scale
     After all that, the actual assembly was fairly straightforward, and they both sound surprisingly good, the O'Cello actually producing a decent level of sound even without the amplifier. The Mina violin turned out to produce a decent level of sound when played with a viola bow, I'm guessing the extra weight compared to a violin bow helps produce more volume of sound? Regardless I'm happy with how both turned out.

Completed Mina violin with pickup

Thursday, 30 November 2017

Flying extruders and Deltas, Round 2

Mini Kossel and Mega Kossel with Flying Extruders
     Back in February, I tried installing a flying extruder on my Mini Kossel, didn't work out very well because one of the carriages would randomly slip during the first few layers of a print and suddenly it was air-printing and making a complete mess. So I switched back to the normal extruder configuration and kept it that way until I built the Mega Kossel

Mega Kossel Flying extruder V2
    For this version I decided to use elastic bands to hold everything in place, worked a bit better but still hit the same issue of random carriage slips. After watching it do this a couple of times I realised that the root cause was the extruder assembly's inertia damping the carriage movement and overwhelming the motor on the tower, causing the drive belt to slip on the pulley and produce the effect I'd been seeing. At that point I was just trying to get the new printer working, so I striped the flying extruder components off and set things up for a conventional long Bowden system. Further research into Kossel design was obviously needed to sort out what the solution to the issue was.

34mm (left) & 48mm (right) NEMA 17 stepper motors
    It was while building the Sculptor that I finally figured out the issue, it was the NEMA17/34mm steppers that were originally part of my Mini Kossel kit that were the root cause of the problem. Like most ~$300 kits, it had fairly light stepper motors for powering the motion mechanics, and while they were sufficient for normal use, I'd found some old forum posts that recommended longer NEMA17/48mm steppers for use on delta printers, and indeed I'd followed that advice when building the Micro Kossel in the first place, so I did a motor transplant and installed the 48mm motors on the Mega Kossel. The difference was quite noticeable once I powered it up and homed the effector. With the original motors it was possible to shake the effector about 3mm sideways even with the motors powered up, with the new ones the effector felt like it was glued in place, zero wobble or shake that I could detect. 
Mega Kossel Flying Extruder V2
    I used the Mega in that state until early November, then decided to try the flying configuration again. Surprisingly, it worked perfectly even without the counterweight that I'd used in the previous versions so I decided to stay with this version for the foreseeable future. 
Mega Kossel with flying extruder

Saturday, 28 October 2017

Building a Prusa i3 MK2 from Scratch

Prusa i3MK2 'Sculptor"
    So, time to reveal what the mystery replacement for the Micro Kossel is, I've repurposed the electrical harness into a Prusa i3MK2 clone build. Sourcing the parts was a bit of a nuisance, but most of it had close matches at my local hardware store or was stuff I had on hand.
Imperial equivalents of the M8 (5/16") and M10(3/8")
 with 1/4" threaded rod for the z-axis screws
Idler arm for the extruder with a 625 bearing from OpenBuilds
     Now, some of the parts for this machine do require CNC cutting normally, but following some of the tricks used by Tom's 3D build from march, I printed the frame template on paper, glued it onto and cut it out from some 3/4" plywood for the frame and 1/4" plywood for the y-carriage.
3/4" plywood frame with template attached
Frame and y-carriage ready for painting
     After cutting out the wooden parts, the next step was to paint and seal them, otherwise they'd change size with the ambient humidity. I used some black acrylic for the first coat, then followed up with a coat of Varathane to seal them.


Main frame waiting for paint to dry
Y-axis parts and tools
     While waiting for the paint to dry I moved on to assembling the y-axis/base frame. Assembly is fairly straightforward, only major thing that you need to watch is that the frame is completely square and flat. Squaring the frame is fairly simple with an old woodworking trick, just take a tape measure and check that both diagonals are the same length, if they match the frame is square in that plane. As for making sure it was flat, I used the one surface in the workshop guaranteed to be perfectly flat, the top of the table saw.


Checking the diagonals for squareness
Checking the other diagonal
     Once that was assembled, the next step was adding the vertical frame, then installing the x/z-axis assembly followed by the extruder carriage to complete the primary frame assembly.


Main frame assembled
Extruder ready for installation
    With the Extruder installed, it was time to sort out where to mount the main control boards. The normal place for them is on the back left of the main frame, but the cables on the LCD module I'm using are only 10cm long, so I've mounted the Ramps package on the front left side with the power block on the back.

Installing the Ramps casing (Thing:761806)
Power brick with custom brackets
Raspberry Pi Zero W for Octoprint host
    I designed a couple custom L-brackets for mounting the power supply and Pi camera, with those installed, it was time to sort out the firmware and calibrate the z-height. For the print-bed, I'm just using a sheet of Buildtak stuck to some 3/8" plywood, basic but workable.

Completed i3MK2

Thursday, 28 September 2017

Decommissioning the Micro Kossel

Micro Kossel
    After almost 18 months of heavy use the Micro Kossel was long overdue for maintenance and wear checking, so I decided to partly disassemble it last week, this unfortunately lead to the discovery of several fatal flaws that would have required basically replacing the frame to correct, so I decided that it was time to retire it.

Micro Kossel parts after disassembly
    Most of the flaws were wear and tear issues that some of the PLA parts hadn't stood up to well at all, specifically the AMZ3D transparent red carriages and SEACANS.com green rails and corner/motor brackets. I don't have photos but the motor corners were warped badly enough that the motors were stuck pointing about 20 degrees up from the normal horizontal position, making any attempt to tighten the drive belts pointless, they'd just slip off the mounts or the plastic would just melt further and worsen the problem over time. 

Micro Kossel frame rail with cracked section

Micro Kossel frame rails with cracked sections
     As you can see, the Seacans green PLA fractured across the layers on these sections, a highly atypical failure for a printed part, normally you'd see fracturing along the layer lines with PLA, not this glass-like fracturing. It's worth noting that the Tinkerine Lavender PLA shows only minor damage where stuck nuts and bolts were removed, and that's mostly cosmetic, so I'll probably be reusing those parts in a future project. 

Carriage plate showing stress fractures
    The AMZ3D PLA carriages are by far the most alarming of the parts, two of them cracked at the hinge point (lower left of the image above), and all three where showing some alarming stress fractures that you can just see in the picture above, I've warped the contrast to make it a bit clearer, my camera has a hard time seeing them but they're very clear to the naked eye. Taken together, these issues mean that it was just a matter of when, not if, that the hot-end would come off axis during a print and make a complete mess at minimum, start a fire at worst, so I've dismantled the frame and salvaged the electronics for a new project, here's a preview, but more on that next time.

Mystery parts for next project

Monday, 18 September 2017

Stress Testing the Mega Kossel: Building C. Laimer's Tourbillon Watch

Tourbillon Clock
     After the painting project from last time, I decided to try building a more mechanically complex project, Christoph Laimer's Tourbillon Watch (thing:1249221). It's a very complicated model with very tight printing tolerances, all the parts need to be as close to perfect as possible or the clock won't work properly. 


Printing the mainspring pinion gear on the Mega Kossel
     Obviously the first step in this project was lots of printing, about 100-120 hours in total, perfect for shaking out any glitches in the printer. Aside from a couple of spaghetti incidents caused by the print bed shifting on it's mounts, actually printing the parts was the easiest part of the project, tracking down compatible screws was the hardest part, I ended up reaming most of the screw holes to accept M2 screws since those were the closest that I could find.


Clock parts organized for assembly with mainspring being assembled
Mainspring clamped for assembly with printed machine vise
    Actually building the clock took a few days, the virtual walkthrough and assembly videos were invaluable for this stage, always making it clear what part was installed next at every step of the process.


Mainspring fully installed with outer casing in place
Clock parts set out for assembly
   I ended up using some 14-gauge wire cut to length for the gearing axles, long needle nose pliers are crucial for installing them in the upper center plate. Other tools I used were a 3/64-inch drill bit for reaming the gears mounting holes and a Philips screwdriver for the M2 screws. After a couple of days work I had it assembled and mostly working, one of the gears is a bit sticky on its axis, but it's mostly complete.


Test fitting Clock face components
Completed Tourbillon Clock