Just Another RepRap

My metal lathe arrives! (back at the end of January 2011)

First, let me say that I’m cursed by the fact that so far I’ve blogged everything I’ve done with my RepRap, and I don’t want to stop now. I’m currently behind by about 9-10 months! My last post was in May of 2011 – today is October 2, 2011. (This feels like debt!) What makes it worse is there’s a whole bunch of pain in those months, but I pride myself on posting my failures as well as my successes, so I’ll post it all – but know that I’m not doing as bad now as these posts are going to show!

Ok.. So.. Back at the end of January 2011, my birthday gift arrived – a metal lathe! Yes, I know that today there are so many people with RepRaps that everyone just buys pre-made hot ends and never experiments anymore, but screw that – I wanted a lathe. I was so sick of trying workarounds that involved doing strange things with my drill press, which were in fact clearly lathe tasks.

So, it arrived:

As you can see it was covered with packing grease, which happened to be a deep crimson red, which felt at least a little unsettling, this being a powerful lathe and all…

After a whole lot of cleaning off of grease, Alicia stood guard over my first test run:

Next I practiced using the lathe with some machinable wax, and made my first thing with the lathe:

And here are two more pictures just for the hell of it, the 2nd one 3D (cross your eyes to get the images to line up):

I then left the lathe for a while to make a heater block for my extruder (so I’d no longer ever have to wrap nichrome wire again).

Then I tapped the hole to M6. Adrian’s original extruder nozzle was threaded (and thus the heater block was threaded too). A later heater block of his (which I eventually make months later) is smooth instead, using just an interference fit. This particular heater block I ended up using on my old M6 heater barrel. Here’s what the tapping looks like for those unfamiliar with it.

Then it was time to assemble it. A large resistor goes through the large untapped hole, and the end of the thermistor goes in the smaller hole.

The resistor needed to be cemented into the heater block, which required me to open the fire cement that I’d bought a year earlier for one of the earliest designs of extruder that I thought I’d make.

I took out the thermistor, packed up the resister hole with some of the cement, put it in the oven at 200°C (around 392°F) for like 15 minutes, and here was the result.

Here’s a picture of it screwed onto my old-style heater barrel (which I’d made on the drill press, but which still works today):

…and then here’s the whole thing in my Wade’s extruder, with the thermistor back in, all wired up with crimped bootlace ferrules, and wrapped in kapton tape:

Now it was time to test it! I hooked it all up, got very excited at the prospect of never having to rewrap a heater barrel with nichrome wire again, and tried heating up the heater barrel. And? And.. nothing. It wouldn’t heat at all. I was trying to bring it to 200°C but it stayed around 16-17°C:

So I double checked everything and eventually found……. ….that I had accidentally ordered the wrong resistors! AAGGHH!! I had accidentally ordered 6.8Kohms resistors instead of 6.8 ohms resistors. I’d have to order new resistors and redo the entire heater block, fire cement and all. Grrrrrr…

Ok, so I ordered the resistors, then the next day – back to lathe work.

Ok, so the main reason I wanted a metal lathe was to be able to make heater barrels better than the ones I’d been making. To drill a 0.5mm hole in the end of the barrel, I’d need to be able to hold onto one of these tiny 0.5mm drill bits somehow in the tailstock of my lathe (so it can be slowly pressed into the spinning barrel, cutting the hole). Before my lathe I’d been able to hold the 0.5mm drill bits in a vise in my drill press but now it had to be held somehow in the tailstock. The thing is, the chuck in my tailstock was way, way too big to hold those drill bits.

I’d wanted to make a heater barrel that day (a Saturday). The smallest thing I could typically find in stores around me were Dremel drill bits (and the collets that hold them) that were way too big (1/32″, around 0.8mm). But then I saw this Dremel 3-jaw chuck online.. It said it only went down to 1/32″ as well, but then I found a European ad for the same product that said it would hold something as small as 0.4mm, which had me believe it would work fine. Here’s what the 3-jaw chuck looked like:

I called a bunch of hardware stores around me and found one that had it in stock. I drove over, wanting to do all of this that day, bought it, and returned home. I figured that I’d hold THIS 3-jaw chuck in the large chuck in the tailstock of my lathe, and that’d be that. But it turned out not to be so easy.

I’m going to show the result video first, because I think it makes a better presentation that way. Watch this video:

Man that made me happy. Yes, I know there are other ways to hold small drill bits (and in fact just this week I finally placed an order for a mini pin vise that can hold a #80 drill bit (0.0135″, or 0.343 mm)), but the cool part was that I had a want to do that THAT day, and I was able to, by making that part with my lathe (including learning to cut threads, which was really cool).

Here’s how I built that.

First, I knew I had to cut threads with the lathe (my tap&die kit is metric, and probably wouldn’t cut threads that small anyway). I determined what the threading was on my Dremel shaft by inspection first, then I confirmed it by Googling it and finding out that it was 40 threads per inch. So the first thing I had to do was open up the panel on the left hand side of my lathe and change some gears to make the leadscrew turn at the correct rate to cut 40 threads per inch. Here’s what the original gear configuration looked like before I changed it:

(That’s a 20 toothed gear for A, 80 toothed gear for B, 20 toothed gear for C, and an 80 toothed gear for D).

I followed the lathe manual’s instructions on changing up the gears for 40 threads per inch, which involved removing one of the gears and changing some.

Then I put a piece of aluminum rod in and turned it down to the correct diameter. I marked where the threads needed to go with black marker.

I put in the right cutting tool for cutting threads and lined it up perpendicular to the piece. I started cutting the threads, one pass at a time, and frequently went back to check whether a normal Dremel cover (which has no center axle) would screw on or not yet.

Thread cutting is pretty fun – it’s an interesting experience. I wanted to document it and show exactly what steps I go through to do it, so here’s a pretty detailed video showing a few passes of cutting threads and what I do to cut them. I like this next video:

Then after a while I ran into a problem where it seems the diameter of the cut threads was actually larger than it was before cutting, as if the metal was swelling or something (maybe it’s because it’s aluminum?). I’m all ears if any experienced lathe operators want to explain this:

Well I think what I did after that was actually file down the threads a bit, and eventually, I got it just right:

That allowed a standard Dremel cover to be screwed on, but since it wasn’t a hollow shaft yet, the Dremel chuck couldn’t screw on (because if its inner axle). So here I used a center drill to start the hole that I’d ultimately cut with regular drill bits:

And then, success!

Here I’ll quickly mention one non-RepRap thing that I did with my lathe right after that. The handle on my toilet broke, leaving it looking like this:

Well I decided I didn’t need it to be pretty, but I wanted to figure out a way to fit those two back together. I didn’t even need to thread one side and tap the other – it’d be fine if I just had a section of smooth rod that I could hammer into a slightly smaller hole on the base for an interference fit. It was, of course, just an excuse to cut something else on my lathe.

Fixed!

One week passed, and the new (correct) resistors that I ordered for my heater block came in. The first thing I had to was drill out the old resistor that I’d cemented in the block.

Then, while I was downstairs, I made a perfect PTFE insulator on my lathe (feel free to go back and compare it to the off-center horrible insulators I’d made in the past on my drill press):

I then drilled holes for mounting it in Wade’s extruder with the jig that I’d made previously:

It was at this point that I made the CRITICALLY WRONG HORRIBLE MISTAKE OF NOT USING THIS PTFE INSULATOR! I got lazy, and decided I’d use the previous one, since it still seemed to have pretty good tension when I screwed it in. (Jeff contains his anger as he types this). This turned out to be a monumental mistake that cost me dearly.. but you won’t read about that until my next blog post.

Not realizing the mistake I was making (the affects of which I wouldn’t feel for another whole day), I hooked up everything again (the heater block with the new resistor, but mounted in my OLD PTFE insulator), and tried testing the heater block.

The next day, I hooked it up, and did my first print with my new heater block.

And I’ll leave you with one peaceful, uneventful video I took of it printing that afternoon, and end this post on that happy note. My printer was printing nicely, everything was going well. What a beautiful machine.

Stay tuned for part 4 of ?

This is part 2 of my recap of what I’ve been up to since January. My last post “Four months.. part 1 of ?” brought us up to speed with everything I skipped over when posting my minimug post in January. Now let’s continue with what I did after that..

First I tried printing the small gear for Wade’s extruder. I tried with an infill width of 0.6, which was too close together (should have stuck with 0.7).

Here’s more video of it printing (no narration):

And here’s video comparing the result to previous failed attempts:

The result:

Given the globs of PLA inside that previous gear, I thought that maybe one of the issues was that the previous layers weren’t cooling in time, resulting in a messy goopy print. So instead I tried printing all of Bed 3 of the Mendel RepRap (an entire bed of parts), so that each layer was separated from its previous layer by a huge chunk of time.

My first attempt failed – it ALMOST completed the first layer, then jammed the head into the bed, resulting in a video I won’t include here because of the anger it contains. Here are pictures after the head crash (which was causing the heater barrel to be dragged from side to side):

A week later, I used my new chop saw to cut a new piece of PTFE:

I then had a worse attempt at printing Bed 3:

To the basement! I went downstairs to make (another? can’t remember) PTFE insulator, and hob a new bolt (with the hobbing jig I’d printed previously, which I showed a few posts ago).

Here was a wood jig I made to cut the center hole in the PTFE, and the two side holes to mount it in Wade’s extruder (which I’ve now used successfully multiple times):

Next I wanted to hob another bolt. Last time I’d done this, I held two bearings together between two pieces of wood clamped in a vice, which wasn’t the most ideal solution. This time I used the Pinchwheel Worm Drill Jigs that I posted about previously.

Then I assembled the extruder. Here you can see the newly hobbed bolt in place:

Next, I tried printing a bed of parts again.

At this point I started to get pretty hopeful.. The shapes were the right proportion/size, I’d made it past the initial layers, and everything seemed to be going well..

So I went upstairs to go to sleep, and allow my RepRap to continue printing its own spare parts into the night. Of course I brought my iPhone with me, and occasionally woke up just enough to take a peak at my iPhone to see my webcam feed of the RepRap to see how it was going. I saw these:

Fail! It was fine in the first two of those pics (at night), but the last two show a mess.

The reason? The reason was this:

Aaaaggggh!

The pictures taken at night showed everything going well, so the problem only occurred after the knot held the spool from turning, and plastic must have poured all over the place..

Since then I’ve found myself pretty paranoid about the filament spindle and whether it has enough slack in it, and I check on prints every 15 minutes or so to make sure they’re not having problems. PLA is pretty strong, so a knot isn’t pretty.

Anyway, two days later, I tried printing again, and managed to print this entirely usable alternate X-axis motor bracket, which takes the X-axis belt on a far less torturous path than the old one. (I still haven’t used this piece)

At some point I tried printing out this gear for the Z-axis, which didn’t come out too great:

It was around this point where I started trying to print parts for Adrian’s Geared Extruder. With the temperature problems I was having I didn’t have the best quality prints, but at least they were close:

And then this piece..

Here was a picture of a bunch of the PLA pieces I’d recently printed, all arranged together:

Finally, I tried printing the big gear for Adrian’s Geared Extruder:

Here’s another arrangement of pieces (this time just for Adrian’s Extruder).

And finally, here are two close-up pictures of that gear. The temperature problems created this possibly usable, but certainly strange looking, gear. It’s more interesting to look at than to use at this point:

Well even though this is a pretty short post compared to most of mine, I feel like this is a good point to stop. Coming up in part 3 – my metal lathe arrives! (remember, I’m still catching you all up on things from around January/February).

On January 3rd I posted my celebration about finally creating an alcohol-tight minimug, and I haven’t posted anything since then. That leaves four months worth of stuff to talk about. It’s a huge amount to blog, so I’ll be breaking it up into multiple parts, and I’ll post the parts on different days.

Stuff From Before My MiniMug Post

For my January 3rd post I jumped ahead and immediately posted the pic of me celebrating my minimug. I still have stuff from before then to post.. Here’s the best I can do from memory and piecing it all together from pictures (as always, there are a lot).

First of all, I’d fixed my X-axis problem as I said two posts ago, but little did I know that I had a subtle Y-axis problem too (that I just hadn’t noticed, because of my large X-axis problem). But I hadn’t figured that out yet. All I knew was that my horrible X-axis problem was solved, and parts looked way better.

So, I printed some parts (which, again, still weren’t perfect, but I didn’t fully realize that yet).

A calibration piece (lit by a flashlight I loosely attached to the frame):

There’s a subtle y-axis problem with that print (which I’ll get to below), but I wasn’t focused on that problem yet – I considered this print a success.
Encouraged by that relative success, I decided to try printing better quality versions of the filament spindle that I’d created before.

Compare that to this previous print from two posts ago, before I’d fixed my X-axis:

Here was video of it printing:

Then on December 30th the kids were playing one of the versions of PacMan we have on the xbox 360. I had an idea, pulled out my laptop, quickly designed a PacMan model in OpenSCAD, and started printing six of them (one for each of us).

I gave them to Laurie and the kids, all of whom loved them, and asked if they wanted to see how I designed it. (At the time Emily was 7, Alicia 5, Cara 3, and David 1). They said yes and I showed them how we could make a cylinder, how we could extrude a triangle, how we could add those two with a union, and how we could subtract those with a difference. They immediately asked if we could design the ghosts.

(For later ghost prints I made the “eye” part an optional OpenSCAD parameter, because it didn’t look as good with the pupil in it).

Next, I wanted to fix the subtle Y-axis problem I was having. My Y-axis was still using the old McMaster pulleys from before, and although it hadn’t caused any huge noticeable problems like my X-axis had, it did cause problems. The calibration piece at the top of this post shows it best – the infill of the object was crossing over the perimeter that had been drawn – I was accumulating Y-error. The fix involved removing my bed, so that I could loosen the belt, so that I could remove the Y-axis motor, so that I could get off the pulley, and replace it.

I replaced it with a printed pulley, put it all back together, and tried printing another calibration piece. Even though I had to abort the piece mid-print (because a loose connection caused the thermistor input to glitch, causing the temperature to drop and the plastic to stop flowing), the results before I aborted showed that I had indeed fixed my Y-axis problem. Pay particular attention to the front of the right-most part of the piece, and see how it’s smooth instead of bumpy.

Y-axis problem solved! X and Y now print beautifully!

That confidence led me to go back to trying to print a minimug again, which I did (and blogged about out of turn last time).

Even though I had these great smooth perimeters in the calibration piece above, that wasn’t what was needed for me to get an alcohol-tight minimug. I’d determined previously that in order for it to be alcohol-tight, I needed to cram plastic together. I changed the ExtrusionInfillWidth from 0.7mm to 0.5mm, and plastic globbed everywhere, making a nice alcohol-tight minimug (which looked much more bumpy on the outside).

Here was the first of the recent attempts, which was close but not even water-tight:

…but, as you can see…

The next one started out much better, and had a promising base.

It seemed water tight at first:

but then it leaked:

That made me try an infill width of 0.3mm, which was WAY too small. Here’s a horrible video of the extruder just plowing through plastic (not a good idea – don’t try this at home):

I aborted that one (in the foreground of the next pic), then started printing what would be my first alcohol-tight minimug (the star of my last post from January) in the background on the bed:

The successful minimug (0.5mm infill width) was born. If you missed it, you can see that post here, with its pictures and video.

Now you’re caught up on everything that happened before the last post. And this seems like a good place to end this one. Coming up next, more of what’s happened since January – I start trying to print parts, my new lathe arrives (!), etc. Woohoo!

DONE!!!

Here’s video of my official toast!

More pics:

Here’s what it looked like before putting any alcohol in it:

The alcohol is Rīgas Melnais Balzams (english: Riga Black Balsam). My mother was born in Riga, Latvia, and in 1999 we (my parents, my brother, and I) backpacked through Europe to get to Latvia. I’ve had that Balzams sitting around for a long time waiting for an occasion – this was it!

And this minimug was no fickle just-barely-works-once-well-enough-for-a-photo piece – no! We went out for dinner tonight and I brought it with me, and filled it again! Here is is filled with Merlot:

I’d had many water-tight minimugs before this, but none of them alcohol-tight. Then for a long while I put the goal of trying to get an alcohol-tight minimug on hold until I’d made it further with calibrating my machine. After recent successes, I tried printing a MiniMug again, and a few prints later, I had it.

I’d say the main differences between now and previous attempts were:

• My X and Y axes used printed 8-toothed pulleys instead of the 10-toothed McMaster pulleys I was using before. (The 3rd generation boards I have only do half-stepping, not the 16-step microstepping that newer boards can do – and without that I wasn’t able to get great resolution with the 10-toothed pulleys)
• The filament spindle I made prevents drag on the X carriage, which was severely hurting my prints
• A geared extruder (in my case, Wade’s Extruder) with a strong stepper was necessary to get a constant, reliable flow of plastic

This minimug (“try39″) was printed with Ultimachine 4042 PLA, printed at 200ºC, using the RepRap Host software. The firmware is from the 20101211 debian package. The RepRap Host preferences were mostly the same as the standard preferences with the following key exceptions:

• For the minimug, I changed the infill profile to be the same profile as the perimiter (“PLA”) since I wanted a ton of plastic
• After trying just the above change and still having the bottom leak, I tried changing “Extruder0_ExtrusionInfillWidth(mm)” from its previous value (0.7) to something real tight – 0.3. That was waaaaaay too close – a TON of plastic printed out, and the extruder head was basically swimming in plastic and it wasn’t going well. I aborted that one, split the difference, and tried 0.5, which made try39.

The sides are very thick, the minimug feels far more sturdy and solid than the others did, and it feels a bit heavier. I used to try printing with the number of shells set to 2 – I think that was a mistake.. Here I went with the default (1) and it still did an outer and inner wall, and filled in the solid part of the wall.. Perhaps the extra shells left no room for infill before.

The outside looks all bumpy because I had “Extruder0_ExtrusionInfillWidth(mm)” set so tight that plastic was bunching up everywhere, which is what I wanted. I wanted a molten pile of plastic holding in alcohol – not a fine mesh air filter like I was making before. I won’t be using that set of RepRap properties for everything – that’s for sure. I do think that if I spent more time tweaking the value I could probably get a smooth-outer-edge minimug that was still as solid and alcohol-tight, but that seems less rewarding than printing out a set of spare parts.

So – WOOHOOOOO!!!!! IT’S DONE!!! I did my toast. That was a long wait. Now my buddies can’t use that “but it can’t even print an alcohol-tight minimug….” joke anymore.

I’d done other things before this minimug, which I’ll blog about later – I wanted to get this out first.

So much to cover. A lot has happened in the past month (aside from my virtual tangents). And I REALLY should break this up into multiple posts, but I can’t – so it’s going to be a huge long post of LOTS of pictures and video (yes, I say that all the time). Let’s get to it.

Printing again

One shipment of gears later, and I’m printing again.

I took the opportunity to do a few things, each of which had benefits and consequences:

1) I finally corrected my mistake and set up the bolts in Wade’s Extruder correctly. (I bought 4 M4 bolts with hex heads (70mm long I think) at Lowes – it was nice being able to go to a local hardware store and find metric anything.. There’s nothing metric at Home Depot that I’ve seen (at least not near my house). That did make it much easier to mount (no worrying about trapped nuts falling out, because there are no trapped nuts – those holes are for hex bolt heads).

Ok, that one didn’t have many consequences. I’d just been putting it off.

2) I mounted my Wade’s Extruder sideways with Tony Buser’s sideways mount piece.

When I ordered some gears from Tony, I also ordered a print of his sideways mount for Wade’s extruder. I’d expected this but he sent me two versions; the version I used was this:

I’m not sure exactly where this particular object is on thingiverse, if it is at all. It looks a lot like this part of thing:4307 but it’s different enough that that’s not it. I haven’t gotten around to asking Tony where the model is for this. Hey Tony, upload this version, it works great!

The immediate benefit is more build height. When printing the bag holder that I use to take the trash out to the curb, I wasn’t able to print all of it because my extruder’s stepper motor would hit the top of the frame. With Wade’s Extruder rotated 90 degrees, the X axis can raise much higher.

The second huge benefit is now you can mount and unmount Wade’s extruder without having to disassemble the whole thing. The idler bracket can stay exactly where it is.

The third benefit was that the heater barrel is now centered in my X carriage. This had been a concern for me before, as it was previously off-center, with the heater barrel right up against my easily-meltable PLA x-carriage.

And the fourth benefit is that the bolt hole in front of the large gear has a recessed area for the head of the bolt, so it’s well out of the way of the large gear. Previously, taking the large gear off of Wade’s Extruder meant straining everything to bend back the extruder to pull the gear out. Now, it comes right out.

Ok, but issues arose. First, to get my PTFE insulator through the hole, I had to remove it from my Wade’s extruder (because of the wires for the heater and thermistor). Then, some of the kapton tape fell off (I’m looking forward to someday leaving kapton tape behind), and the thermistor floated free in the air instead of being held against the nozzle. I agree with what I read someone say a week or so ago – kapton tape is not a real solution for holding thermistors against barrels. (or maybe my roll of kapton tape is just getting old – it doesn’t seem to stick like it used to).

So I had to at least slightly rewrap. But I had to do that anyway, for a different reason. Tony’s mount is at least 6-7mm tall if I remember, and with the short heater barrel I was using, that made my heater barrel not go down as far as the bolts on the bottom of my X axis.

It also meant that the inside of my X carrriage was exposed to more heater barrel. That’s one of the reasons I wanted to rotate the extruder in the first place – concern for heat in my x carriage.

I dreaded having to make another (longer) heater barrel right now. I just made one recently, and it was the one I made on my own without any external nozzle (the original Mendel heater barrel design). Plus, I’ve wanted to keep the heat zone on my heater barrel very short, so I don’t use longer purchasable heater barrels (even though I have some lying around). Plus, that wouldn’t solve the heat-in-my-x-carriage issue.

So instead I realized the solution was to make a longer PTFE insulator. That would lower the heater barrel, and insulate the carriage from heat. I’ve made several, but they’ve all been a pain. This is another area where a lathe would be nice. I couldn’t do my spin-the-piece-in-my-drill-press trick because the 16mm PTFE was too wide for the chuck of my drill press (I did wonder if people sell some sort of chuck extender to fit into an existing chuck and provide a much wider one.. but then I realized they probably don’t, because really that’s all the stuff of lathes). So instead I just drilled it, it wasn’t centered perfectly, but it worked. I decided to make it 13 mm longer, so I made it 48mm long. 3mm hole all the way through, then I think I used a 5mm bit on the other side (15 mm deep) before tapping it M6.

The next issue I ran into (that might have been a problem I already had?) was that with 3 washers, the large gear had at least one time on its full rotation where it’d push up against the small gear too close and catch/hang. Spacing it out with an extra washer fixed that, but moved the knurled part of the screw even further away from center (but again, it extrudes and prints, so..)

Battling with my X-axis: Part 1 (The Problem)

So now that I was printing again, and that I’d tested printing that taller bag handle above, I printed a test “K” block. Despite my attempts to ignore my x-axis problem for the short term, this was un-ignorable. Look how horribly all-over-the-place this block is in along the X-axis (left to right):

The Y axis is fine – looking at that from the side shows the edges line up. But looking at that from the front, you can see that layers go left and right from where they should be. As it turns out, this was caused by one of two overlapping problems I had with the X-axis – but I didn’t know this yet.

I had several theories about what might be wrong with my X-axis. One theory – the one that I started working on first, was that there might be some drag on the X-axis from the filament supply I had resting horizontally on a lazy susan behind the RepRap.

I’d used this lazy susan with the Makerbot successfully before, but I think the reason it worked in that case was because the Makerbot’s extruder doesn’t move – it just stays in the same location. As a result, the only issue for the Makerbot was whether it grabbed enough plastic. On Mendel, however, the carriage moves left and right, so extra drag on the filament might cause problems with the carriage moving. Plus, there’s a horrible coiling affect that happens when plastic comes off of the top of the spool, converting from its horizontal resting position to a vertical pull into the extruder:

Battling with my X-axis: Part 2 (Attempt to Test a Theory)

Upon describing this to my Dad during a long drive to a football game (Go Pats!), he asked if there was a way to remove the extruder and instead mount something that could hold a pen. That way, the filament-drag idea could be tested, by removing filament from the equation.

I immediately replied that yes there was, and that I’d planned to download and print a pen holder one day, but probably after I had everything working. Still, I figured it was worth a try. So I downloaded the Mendel Pen Holder, put the parts in an RFO file, and tried printing all three parts at once. (That was a knowingly foolish move, but it turns out to have yielded a good example of my X-axis problems.

It didn’t turn out well. Here you can see the three parts. Because I forgot to set the print order of the parts correctly, the back part printed first, then the part on the right, then finally the part on the left. As you can see in these pictures, my X-axis was accumulating error as it progressed within each layer, such that the third object printed was in a far different location than where it should be.

Specifically, look at the tall portion of the third part printed (on the left). It seems shifted way to the left of the base of the piece. What’s actually happening is that the base was shifted way to the right of where it should be (because of all of the error accumulated during the lower layers, from printing out the other pieces). The carriage was drifting more and more to the right of where it should be. But once the Z-axis had raised enough that the other two parts were complete, you see the higher portions of that piece were now printing where they were supposed to be (but since the parts below it were to the right, the top appears to drag left).

This was actually an illustration of my second overlapping problem with the X-axis, but again, I didn’t really know for sure yet that there were two problems.

I believe I made one more attempt to print that tall piece on its own, but I then put on hold the pen mount until later, instead deciding I’d just plow ahead with one of the solutions I’d already planned. (McMaster parts arriving in the mail helped that decision along, as you’ll see).

Battling with my X-axis: Part 3 (The Filament Spool/Spindle)

I’d determined that even if it wasn’t the cause of my X-axis problems, I wanted to build a vertically mounted filament spindle (or filament spool – both terms are out there), to reduce any drag whatsoever on the X-axis.

I looked at the existing spools on thingiverse and finally settled on a combination of three things:

1) Simple spool system for a Mendel (or other 3d printer) by camiel
I didn’t actually print any parts from this entry – I just followed instructions on it.
2) Simple spool system for a Mendel (or other 3d printer) by brokentoaster
This was a modified spool bearing, where you put the 608 bearings in the spinning part itself. I also studied the pictures in that thing to figure out how I wanted to mount it, including printing out 8 y-bar-clamp_10off pieces and buying the appropriate bars/nuts/bolts/etc.
3) Extensions for simple spool system for a Mendel (or other 3d printer) by Gavilan
These extensions bolt onto the 3 support bars, to keep the filament from spilling over the left or right side of the spool.

What I ended up with was uuuuuuuggggggggggggggggggggglllllllllyyyyyyyyyyyyyyyyyyyy, but it worked, and solved one of my X-axis problems.

(And actually those above weren’t that bad.. what was ugly, was to come). I realized that my X-axis problems would affect this piece, but I figured I’d print it, and if it was even SLIGHTLY usable I’d make it work (however I needed to), and if it solved my X-axis problems, I could print a better quality one later). Here’s the ugly:

This picture actually makes me feel dizzy looking at it, like I’m seeing double, but I’m not:

More ugly:

And if you don’t feel drunk yet, three more:

Then I printed out the extensions.

In this one, you can see the drastic difference that having the Z-axis optoendstop correct has; the piece on the left had the nozzle home just a paper’s thickness over the bed – the others were printed previously, when the endflag was too low (causing the nozzle to start too high).

Undeterred by the fact that there was no way a 608 bearing would smoothly fit into that, I bought a nice set of forstner bits and drilled out an appropriately sized hole. I put in the bearings, used a hot glue gun to make it fit nice (and fill in some of the holes in the plastic), and eventually ended up with this:

Then I put filament on it, and mounted the extension pieces.

And then, I mounted it on my Mendel!

Brace yourself: extremely exciting video ahead:

Then I tried printing another K block. Here was the glorious result of that print (ignore the brown spots):

The filament spindle had indeed fixed the first of my two overlapping problems with the X-axis.

Did you hear my foolish optimism thinking that that might have solved everything?

Ok so once I came down off that high (double rainbow, all the way!), I cautiously tried printing what had failed spectacularly for me before – a large Wade’s Extruder gear. In this next picture, the new print is on the bottom left, the previous attempt from the summer is above it, and an ABS gear printed by someone else (not sure if that was the one from Neil that broke or a new one from Tony) is on the right.

I’d actually let myself think that maybe the filament drag (fixed by the spool) just miiiight have actually been the entire problem. Nope. So this was the first time that I definitively knew that I’d had (at least) two different overlapping problems with the X-axis, and that I’d only solved one of them.

Battling with my X-axis: Part 4 (The Belt and Pulley)

So I moved onto other theories that I’d had. The belt could be too loose, maybe I hadn’t filed the shaft down on the X-axis stepper and the pulley was slipping, and here was a big one – maybe the McMaster pulley (a 10-toothed pulley) required microstepping to get adequate resolution (vs the 8-toothed printed pulley that’s supposed to be standard).

Worse theories hovered in the air like a storm cloud – maybe my X-axis bars were corroded/dirty enough that they were causing problems, or perhaps the plastic had shifted and now the X-axis bars weren’t parallel. Maybe the bearings in my X-carriage were gunked up, and it was just broken. Replacing the X-axis would be a major pain in the ass – I’d have to take off the Z belt, I think I’d have to take off the bed, etc. No fun.

Here’s a video from earlier when I was printing that pen mount.. Once I saw that it was making mistakes with layers, I filmed about a minute or so of the underside of the X-motor-bracket, where the pulley goes through this torturous path around bearings and the pulley. Note that the pulley was the McMaster pulley I purchased last April, which again has 10 teeth:

The belt certainly looked suspicious.. Maybe loose. Just barely hanging on to those turns. Plus, again, there’s the potential resolution problem that a larger pulley could be contributing.. One full turn moves 10 teeth instead of 8, and the number of steps for the motor stays constant, so there’s at least some less precision with this pulley than with a printed one I have lying around waiting to be installed.

So I decided to start down the road of replacing that, and that meant removing the X-belt from my Mendel. That was something I’d wanted to avoid if possible, since it meant changing a key variable in the equation (the tension of that belt), and who knew how much worse that might make things. Worse, if I removed the belt and the carriage still didn’t move smoothly, several of those worst-case scenarios might suddenly emerge as likely.

I dove in and removed the X-belt.

Beautiful. Very smooth. So that left the pulley, motor, and belt tension.

I removed the X stepper motor and took off the McMaster pulley. Sure enough, I’d been lazy and hadn’t filed down the stepper motor, but if I remember correctly I think the set screw had dug in pretty firmly even on the round shaft. So the first thing I did is file the shaft. When I was filing, I wasn’t sure which way nophead’s pulley (sent to me by Neil Underwood with my Wade’s Extruder) needed to go on, so I filed two parts (rather than filing the whole thing).

It turns out, it needs to go with the fat part on the outside:

You can see why here:

I put the pulley on, and tightened the belt much tighter than it had been (but still not so tight that nothing moved).

I recompiled the firmware to change X_STEPS_PER_MM to be the default setting for the 8-toothed pulley, uploaded it, and printed a calibration object. It came out beautiful.

That was good to make sure I had the new settings right for the new pulley, but it didn’t test the real problem. The best test for that would be the Wade’s Extruder gear, since that had always failed spectacularly.

So I printed that.

So good! It might be just a tad shy of usable because it might be slightly oval (because I still can tighten up the calibration from the last calibration object I printed), but that might be just being overly cautious. But anyway, that’s not the point. The point is – that solved my second overlapping X-axis problem!!!

Need more pictures of the successful gear.. Here it is compared to the most recent failure:

And here it is next to the functional one made out of ABS. Again, the color/transparency issue is because the gear I printed was printed with PLA (which is clear), vs ABS which is opaque and beige colored.

Horray! So I then remembered that I’d wanted to print out two Pinchwheel Worm Drill Jigs in case I needed them in the future – so with my new functioning printer I printed two out:

They came out great. Then even though my X-axis problem was solved, I went to address one of the last items on my list – I figured I’d at least try printing an alternate X-motor bracket that eases the path for the X-belt. I started a build, and headed up for bed. Luckily once I got to bed I pulled out my iPhone and looked at the streaming video of my RepRap and I saw it failing miserably. The barrel had apparently hit the bed and had pushed out of the barrel. For some reason the motor was stuck near the dump location on the bed. Anyway, fail:

A week went by and I rebuilt things, then tried printing (using the new (working) debian package for RepRap Host and firmware), and eventually had it crash and push out AGAIN (in fairness, I was trying to re-use the longer PTFE insulator I’d made). I did witness one issue though.

With the new software, I’d let it specify all new default settings for Preferences (except ones I specifically wanted to set). I noticed that after a layer, Z sounded different.. it didn’t accelerate and decelerate when moving up – it just made a little “beeeeeeep” like sound (presumably moving faster at a constant speed). But later after the crash, I found that it was in a state where even when I used the jog controls to move Z up 10mm, it failed to do so, and instead just made the same beeeeeep sound from the stepper. I forget exactly what I did to reset it, but I think it was strictly jog controls, and then it started responding normally.

I’m planning on trying that new gcode with the old 0806 firmware to see if Z makes the same new sound – I’m thinking it’s just new faster default settings that I can dial back down to speeds that I work OK with.

I did one more thing before finishing this post (I’ve been trying to post this for a while now). I finally took the laptop power supply I’d purchased and soldered on the XLR connector I bought for it – to replace the ATX power supply I’m using.

So there ya go! There’s my progress for the past month or so. Did anyone actually read this far?

Here’s another example of two of my worlds colliding. I built a 300-foot tall Mendel RepRap:

(click image for high resolution)

(keep watching that video until the end, to get a true sense of scale)

I built it with Forge, the map editor in Halo: Reach (entirely on an Xbox 360, using the controller.. no computer involved). I did it all by memory – the only measurement I did beforehand was to compare the width of a frame vertex piece to the lengths of the bars between the left and right triangles (and the length of the bars for the triangles themselves). Here are some more pics (again, click the images for high resolution):

As you can see, it’s printing a minimug!

Of course the best way to view it is to download the video straight to your Xbox 360 so you can see it from within Halo: Reach itself.

I’d always wanted to see what my RepRap Mendel would look like to a mouse or ant crawling onto it – some towering, huge structure. Halo let me do that.

Over the past month I’ve made lots of progress with my actual real-life RepRap Mendel (and it’s print quality) that’s worth posting about, but I’m going to break that up into smaller posts since I have so much to document. First, though, I wanted to post this.

And for anyone who doesn’t know what an actual Mendel RepRap looks like, here are two picture of my actual real-life Mendel RepRap (RepRap Mendel?):

Actual progress updates to come soon – stay tuned!