Such a sense of triumph when tiny strings of fire-engine red plastic appeared from the extruder nozzle. Not under any sort of control yet, as I still have lots of setup to do, but all the motors work, the heater and temperature sensors do their thing, and I can compile and upload the printer firmware without problems.
It’s been a fun and head-scratching experience, watching a box of bits turn into a physical object, and then having it come to life. Anyone who has watched much science fiction will be familiar with the distinctive noise of electric motor driven robots. To build it myself and hear those iconic sounds felt great, though my anxiety levels peak when my end stops don’t work and the motors try to drive themselves off the rods with great rattling noises.
Below I’ll describe step by step the notes I wrote on using the two essential references (the ’Prusa Mendel Visual Instructions’ and the ’MakerGear Prusa Mendel differences from stock Prusa’) to assemble my MakerGear Mendel Prusa 3D printer kit.
Building My Printer
First I’ll apologise to my friend Paul for the lack of images. Really this post is pretty useless to anyone who isn’t building a MakerGear Prusa kit right now. So buy the kit, then read this.
Printing out both my major build references and punching them for ring-binders turned out to be a really good idea. I thought about plastic display-book things, but you can’t easily write notes on the pages that way, and I wrote lots of notes. In pencil. So I could erase them again when I found I was wrong, or finally worked out what was right. I don’t see an easy way of adding my comments onto the original webpages, so that other people could see them, but I’ve typed them up here.
The first thing I did was to go through the ‘Differences’ printout, and note any changes directly on the ‘Visual Instructions’. That way I could just follow along on the visual instructions until I came to a pencil mark telling me which page of the ‘differences’ printout to refer to.
The ‘Visual Instructions’ are excellent. Very clear and well explained. The ‘differences’ comments vary from well explained and well illustrated with photographs, to, um, not so clearly understandable. The ‘differences’ file is only necessary because the MakerGear kit is so much better than the standard Prusa. Remember that each time you get confused by the gap between the two references.
Most of my problems came from the not-so-clear bits. I can tell this, because when I had a problem, and traced it back to its origins, there was usually a pencilled ‘what?!’ or ‘this is instruction is c**p’ comment on my printout.
An example. Under ‘Part 7, Steps 1-7’ is the comment “The parts look slightly different here but it shouldn’t be too hard to figure out.” Then a photograph of the supplied printed plastic parts next to the original ‘Visual Instructions’ diagram. Followed by a build tip saying “Don’t forget the two fine rod adjustment screws”. I still don’t know what or where those screws are supposed to be, or what to do with them if I did know where they were. There are a bunch (more than two) of screws that fit into the underside of the x-carriage that look like they should fix the x-rods in place, but every time I tighten them the x-carriage binds solid on the rods at either end of its travel. If I leave the screws loose they fall out when the printer vibrates (Marlin doesn’t like my end-stops, Sprinter is more accommodating…). I removed the screws altogether. I hope they aren’t vital…
[Update: I just asked a related question on the makergear forum, and now I understand. I’ll explain in Part 7, below]
And while I was tearing my hair out over that, I misunderstood the very important line in the ‘Visual guide’ that says “Drill out the centre hole in the hexagonal sections of the x-end-idler and x-end-motor parts to 8mm”. If you don’t do this when the visual guide tells you to, you’ll find (much later) that your z-rods won’t turn because of friction in those holes. Fixing it then will require a lot of disassembly. I found a solution when I remembered I had a special drill bit with a stepped cutting end that widened from 4mm to 12mm in 1mm increments. And it was short enough that I could turn it by hand without taking everything apart. That showed me that 8mm wasn’t wide enough, and I enlarged it to 9 or 10mm.
Step by Step Comments
All these comments only make any sense in combination with the two essential references (the Visual Guide and the Differences) linked to above.
Part 2 Section 1 – Which rod do I use?
The unspecified bottom rod is 305mm long. So are all three other rods nearest the table top at the front and rear of the printer. The two longer, 455mm, rods go right at the top of the printer (in Part 4 step 1).
Part 2 Section 5 diagram 3 – Getting the bearing placed correctly
This caused the largest amount of disassembly for me. It looks from the photos in the ‘differences’ file as if the 688 bearing (one of the small all-metal ones) is not fully pushed into the centre of the flanged wheel. I think that’s wrong. If you look inside the wheel you can see a ridge which to me looks like it is supposed to centre the bearing. So push the bearing all the way, from the slightly wider side, til it settles against that ridge. (I used a 12mm hex socket to keep the insertion pressure aligned and away from the ball race). That way the bearing is at the centre of effort of the wheel, not out on one side leaving the wheel to experience a bending force.
Doing this my way means that instead of two washers side by side, you need three washers. This is because the nut next to the three washers moves inside the flanged wheel, leaving the wheel to move left and rub against the hex-headed bolt securing the y-motor to the plastic motor mount. To test that you won’t have problems later, put a hex-headed bolt (with a washer on it) into the hole in the y-motor mount nearest the flanged wheeled and ensure there’s enough clearance.
Part 3 Section 2 – Lock nuts to stop the idler wheel wandering
The bearing inside the flanged wheel on the rear upper rod has nothing to hold it in place but the nuts on either side of it. I was concerned that the turning of the wheel might turn the nuts up against it, leaving a gap and allowing wobbling. So I put two nuts instead of one on each side of the bearing, so they could act as lock-nuts. Though don’t tighten them too much at this stage since you’re going to have to align the front and rear flanged wheels and the y-motor gear with each other later on. (Having just checked my printer I see I only put doubled nuts on one side of the bearing. Oops. I’m not so sure that this will cause a problem that I’m willing to take things apart again.)
Part 4 Section 1 – Use the long rods this time
The unspecified rods in sections 1 and 2 are the 460mm ones.
Part 4 Section 5 – Watch your orientation
Here is one of the places where I got confused about orientation. Pay close attention to which side has the y-motor on it, and which side does not. Nobody warns you when a photo or diagram is suddenly showing the other side of the printer from what you thought you were looking at. Different people seem to set their printers up with different orientations. Mine has the y-motor at the front, and the X motor on the right. So on my printer the gap between the Z-motor mount and the top of the frame is on the left – the end away from the X-motor. Check the photos on the ‘differences’ file for Part 4.
[Updated 2011/12/26:- Spacing of the z-motor mounts]
The Visual guide (Part 4, step 5 onward) shows both the z-motor mounts separated from the frame apex printed pieces by a nut and washer . This is wrong.
The differences file (Part 4: Top threaded rods) shows one motor mount separated from the apex by a nut and washer, and the other by two nuts, two washers and some space. This is also wrong.
The differences file (Part 5, Step 6) then shows how it really should be!
- The z-motor mount above the x-motor has no nut and washer between the mount and the frame apex.
- The other z-motor mount (above the x-idler wheel and electronics) has two nuts and washers. The space between the z-motor mount and the frame apex should be almost exactly 40mm. This gives the greatest amount of x-axis movement, for the greatest build area, and gives you plenty of space for mounting the electronics support panel.
Part 5 Step 5 – We don’t need no steekin’ PTFE tube
I don’t think you need to use the PTFE tubing any more, because of the change I’m about to describe for Part 5 Step 6.
Part 5 Step 6 – Run the z-support rod underneath
Turn the U-shaped bar clamps upside down from the way the ‘Visual Guide’ shows them (so they look like a u, not an n) which puts the z-support rod under the frame rods instead of over them. This keeps the y-belt away from the rod that it otherwise would have rubbed against.
Also, note the alignment of the printed plastic pieces that will hold the z-rod bearings. The u-shaped openings point away from the side with the y-motor on it. You can see it if you squint at the photo in the ‘differences’ file, but it took me a while to realise that it mattered.
Part 6 – Assembling the Y-Axis
Follow the ‘differences’ file to build the Y-axis carriage. Do as the build tip says, and use the y-stage base plate (the biggest of the laser-cut pieces) to mark the drilling holes for the big square plywood build plate. Do it now, because access gets impossible later. The y-stage base plate has its long axis running from left to right when installed. The M3x12 bolts are in the miscellaneous hardware bag, they don’t have a labelled bag to themselves. (You do have a digital calliper, don’t you? My calliper was an essential purchase for building this thing. Checking lengths and widths was a common task.)
My ‘bushings’ were bronze-coloured cylinders encased in black plastic. I didn’t need to hammer them in, they were a perfect friction fit. There’s a bit of movement in them, once the rods are inserted, and the alignment gets better over time, as you move the carriage back and forth.
Part 6 Step 20 – Cutting the belt
According to the bill of materials, there is a 7 foot length of belt. The y-belt needs to be 3 feet long, and the x-belt is 3.5 feet long. That is 914mm and 1066mm. Measure twice, cut once! (and don’t just trust me, do the maths for yourself!) I cut 3 feet for the y-belt, and left the x-belt untrimmed until both were fitted and working.
Part 6 Step 24 – Installing the belt clamps
As the ‘differences’ file states, the belt clamps have two pieces each. The bottom pieces (the ones with the indents for the M3 nuts), go under the y-carriage, with the nuts at the bottom. This is a big improvement on the original design, because you can loosen and tighten the belt with a single hex key, without having to put a spanner on the nuts under the y-carriage.
Part 7 Step 1 – Make big holes bigger
I mentioned this at the beginning. Drill out the big holes in the bottom of the hexagonal tubes. I don’t think 8mm is enough, I’d recommend 9 or 10mm. You aren’t looking for precision, you need them to be large enough that the threaded rod fits through without touching the sides, but small enough that the springs don’t catch or fall through.
[Updated: New information about installing nut traps]
Part 7 Step 3 – Install nut traps
This is horribly fiddly, and most of my bolts vibrated out later (but the nuts are still there if I need them). If your plastic pieces are at all warped you might need these nuts and bolts to adjust the alignment of the X smooth rods, to get them parallel.
The trick I developed was to hold the nut roughly in place over its captive hole with a pencil through the smooth rod channel. Then screw a screw through the set-screw hole til it pinned the nut against the other side of the channel. The bolts are just long enough to generate enough friction to wind the nuts onto the bolt heads. Once the nuts are threaded on, you can drag them back into their captive slots by tightening the bolt up. Then undo the bolt til it’s not sticking out into the smooth rod channel, or you won’t be able to insert the smooth rods.
Important: Ian’s tip about the “two fine rod adjustment screws” does matter. Set the x-end motor mount and x-end idler up on your bench following the photo on the ‘differences’ file, so the hexagonal columns are nearest each other. The flat rectangular sections at the top of the photo have big holes through them running left to right. They also have tiny holes for set screws on either side of the 8mm opening. These are the only screws that directly oppose each other, which means you can tighten them both up the same amount to lock the smooth rod in place without pushing it sideways. Put captive nuts in both holes on each piece and leave a screw loosely threaded into each of them, as above. You can tighten them up much later when you know the stepper motor can drive the x-carriage smoothly.
Part 7 Step 6
My smooth x-rods were 458mm.
Part 7 Step 10 – X-idler wheel assembly
To get my X-motor geared wheel aligned with the idler wheel, I used the following:
M8x30mm hex-headed bolt –> 608 bearing –> flanged wheel (sloping face nearest bolt head) –> nut –> washer –> x-idler –>washer –> nut.
Part 8 Step 1 – Levelling the whole machine, temporarily
I had made an attempt to level my workbench top before starting the build. I didn’t get it perfect, but it made this step a little easier. I found an unused pack of playing cards in my desk drawer, and used them instead of scraps of paper for getting the top rods of the printer level. If you can, check your level against some other reference. I had three bubble levels, and one of them was way out compared to the other two.
Part 8 Step 13 – Use the metal z-motor couplings
After much cursing of the plastic z-motor couplings (they simply wouldn’t grip the motor shafts, and kept slipping down, throwing out my z-alignment. Grr) I realised that a zip-locked baggie with two odd metal cylinders in it was a pair of metal z-couplers. Yay!
Because they have metal grub screws, they grip much better, and one of my biggest fears (z-alignment woes) was fixed at stroke. Don’t use the plastic z-couplers, they’re awful.
Part 8 Steps 19 and 20 – Attach couplers to motors, then motors to frame
Do step 20 before step 19. It makes putting the couplers on much easier.
Part 9 Step 12 – Use bigger bolts for the X belt clamps
I used M4x20 bolts and matching washers. I seem to have run out of longer M3 bolts before finishing the build.
Part 9 step 17
I’m not commenting here on building the extruder. The instructions on the MakerGear website seemed to mostly get me through, but since I was reading them directly off screen I don’t have any useful comments written down. (See – print everything out in advance, so you can become a font of knowledge and experience to others!)
[Update: I’ve written up extrude construction advice]
And I’m still having real trouble getting the filament to feed reliably into the extruder head when I first insert it. The motor and gearhead work, and it feeds fine once properly inserted, it just seems the alignment is not quite right.
[Update: I’ve found a way to feed filament through a bruthead]
Power Supply Strangeness
I will mention a strange problem here that cropped up during the building of the extruder. To make the heater core you need to coat the partially finished heat core with some astonishing heat-resistant ceramic goop that MakerGear supplies. Then you need to heat cure it, using a 12V source of some considerable current. The cold resistance of the nichrome wire was around 6 ohms, which my half-remembered Ham radio experience suggests needs about 2 amps at 12V.
I decided I’d have to use the power supply MakerGear provided with the Prusa kit (if it wouldn’t work to cure the heat-core, it sure wouldn’t power the extruder head!).
[Update: there’s more detailed info on setting up the power supply in my post on building the extruder]
The 600W Eagle ATX power supply had a lot more connections than I needed, and I couldn’t find a connection diagram on the web. It is said (by people who may have had the exact same model as me, or not) to have two separate 12V supplies, one which supplies the little square 4-pin connector, and the other which supplies 12V on all the other yellow wires on all the wiring harnesses.
So I soldered up the heavy black twin-core cable. (Cut it in half, then cut one half in half again. The two short pieces go to the power supply connectors, the longer piece is left for the heat bed which I haven’t built yet). One piece to the square 4-pin connector, the other to the rectangular 4-pin connector with the rounded corners, following the photos in the ‘differences’ file. The bare ends go into the strange green 2-pin connectors that mate with the RAMPS power pins labelled 11A and 5A. Check with a multimeter for connectivity, or undesired shorts between pins, and fix them before plugging anything in.
Don’t plug the green pins into the RAMPS yet. With the ATX power supply switched off, plug the soldered leads into the power supply harness. Switch on the power supply (I held my breath. The last 10 amp power supply I had, back in the late 1970’s, caught fire in my bookcase one evening. Flames and Smoke and Everything! It still worked afterwards.)
Now here’s the weird bit. There was 12V (or in fact 11.65V – I guess PCs aren’t fussy about their 12V line) on both connectors, as expected. Either connector could provide 12V at enough current to cure the heat-core (remember, 2 amps for only 2 minutes at a time). But if one connector was connected to the heat-core and the other was left unconnected, the power supply’s protection circuitry kicked in so fast there was only the slightest twitch of the fan blade, then nothing.
I can only guess what the problem might be, but it seems to have disappeared now that I’ve got things properly terminated on the RAMPS board. Though, as I said, I haven’t built the heated bed yet.
All Done? Not by a long shot!
That’s about all for my comments on the mechanical build. I have a couple of pages of scribbled notes on the electronic side of things, but this is way too long as it is. Hopefully it’ll be useful to someone.
And I’ve yet to get the beastie even basically calibrated. That’s a whole ‘nother can of worms. Thanks from the bottom of my heart to all the folks who’ve posted their own experiences and explanations. I couldn’t have done it without you.
Corrections and comments greatly desired!