Just Another RepRap

[ BE SURE TO READ THE BIT AT THE END ABOUT MAKERFAIRE 2012 ]

I think I’m actually going to have all of my credit cards paid off before I catch up on my RepRap blog. 🙂 [Edit: I wrote that months ago. It turns out, I actually did! Credit cards all paid off. 🙂 ] Here’s another update – at the end of the last post I was caught up to September 25th, 2011. That means that I’m starting this post out about twelve months behind. (When I first typed that sentence, it said seven months behind. Procrastination = bad)

Here’s where things finally started to get good. My friend Chris was now building a Prusa RepRap, and purchased RAMPS 1.4 and a PCB heatbed. That motivated me to finally get around to upgrading my electronics – I was still using the old Gen 3 Makerbot electronics (which only barely did microstepping, and didn’t have a standard way to drive a heatbed). So I too purchased the RAMPS 1.4 electronics and Prusa’s PCB heatbed.

Here’s the last video of my RepRap before I tore it apart:

For those who aren’t familiar with RAMPS, RAMPS stands for RepRap Arduino Mega Pololu Shield. It’s a shield that plugs onto an Arduino Mega, it has Pololu stepper drivers plug into it, and it can support a heatbed and up to two extruders. Here’s what it looked like when it arrived.

I tore all of the original electronics off (actually I left the stepper drivers underneath on the back unused), cut a new piece of MDF, and mounted RAMPS to that. Then I rewired everything up (which was considerably easier since almost everything is on the same board).

This change thankfully also pushed me to make three others:

1. I decided to try using Sprinter firmware instead of the RepRap 5D firmware I’d been using.
2. I tried out pronterface instead of RepRap Host for control software
3. I tried slicing with SFact (a simplified Skeinforge) instead of using the built-in slicing that RepRap Host used to do for me. I’d used Skeinforge before with the Makerbot to make my original RepRap parts, but I’d read that SFact automated lots of the redundant settings and made calibration easier.

Using Sprinter and pronterface, here I am successfully demonstrating working X Y and Z axes:

(Note: After this point my wife and I finally upgraded our phones from older iPhone 3GSes to iPhone 4Ses, so from this point on my pictures will all be higher res and hopefully the video will be more stable.)

Here are a few pics of the new electronics taken with the new phone:

This next video is where I started to really get excited. The excitement was a sliiiight bit premature because I didn’t realize yet that my Z-axis wasn’t going up enough (I needed to slow down the Z speed to accommodate my RepRap which can’t move that fast in Z), so I ended up canceling this print, but I was happy nevertheless:

It turns out I didn’t have my Z speed limit set properly, so it was trying to go up too fast and wasn’t moving. In setting that I also had to fill in other settings, such as layer height. I decided to print some hollow calibration cubes to try to get that right.

(If I add a fan to my x-carriage that should actually let me print at far smaller layer heights, I think)

I tried going smaller. 0.2 looked great until it didn’t:

The back of 0.2 didn’t come out well though:

But still that layer height looks so small..

Then during this print:

the extruder barrel pushed out of the PTFE:

Even though the heater barrel pushed out of the PTFE insulator (a problem that always caused me headaches), I remember not minding at all this time, because things were going so well.

I realized one thing that helps out a lot when fixing your RepRap is actually putting it ON a lazy susan:

So.. I rebuilt the hot-end (twice, as it failed almost immediately the first time), and printed more of the calibration cubes. Here’s the result:

With that progress made, I went back to trying to print a money clip. Here was my nervous mid-print video:

and even further along:

Behold! Apologies for there being so many pictures of this, but this was a huge moment for me. Check out the quality of this money clip:

Watch this video:

And this was even worth a second summary video:

Can you tell I was excited? 🙂

I have so much more to post, but I need to cut it here. There are a lot of really great quality print pictures coming up in the next blog post!

Today is September 23rd, 2012. I just got you all caught up to October 29th, 2011 (about 11 months ago).

In one week I’m heading down to Makerfaire NY 2012 to show off my RepRap (I’ve attended for the past two years but this is the first time I dragged my RepRap down there to exhibit). I’m booth #8736 in the 3D printing village. If you’re in the area, come on down! It’s September 29th and 30th. Makerfaire NY 2012.

(At this rate, you won’t see pictures of that for a year. 😀 )

I’m catching up! I’m not as far behind as I was before. Hopefully with just a few more blog posts I can catch you all up to my current state of RepRap (which, again, is far ahead of what I’m currently blogging about – April of last year (2011)!

Ok, so where was I.. At the end of the last post, I had just recovered from a catastrophic failure – I replaced my old melted X-carriage with a brand new OpenX carriage and had started printing. I was having extrusion problems again (resulting in rice-krispie consistency), but at least I was printing again.

This seems like a good place to post something I often talk about regarding having printing problems and not knowing what to do about it. When doing RepRap printing, there seem to be times where something isn’t working and you have no idea what exactly is causing it. The problem is that there are so many things that potentially could be a problem that you have no idea which one needs fixing. I liken it to being lost inside a cave or mountain in the game Minecraft. For those of you who aren’t familiar with Minecraft, all you need to know right now is that it’s a blocky cube-based video game with a custom world for you, with hills, valleys, caves, mines, and you can dig and climb. Sometimes with RepRap it feels like you’re inside a mountain in Minecraft, and you’re lost, and you just want to dig your way out. When you’re lost in a mountain in Minecraft, for all you know, you’re right up against an outside wall, and if you just dug West by one block you’d be outside, but instead you arbitrarily decide to dig East instead, and dig yourself deeper into the mountain, still lost. (leave aside Minecraft tips like digging upwards in this scenario!).

It feels like that with RepRap sometimes. Hey, this print is awful.. should I adjust the tension on this belt? Or maybe a variable in the firmware about my rate of extrusion is wrong? What if I got a heated bed? The extruder! Maybe I need a new extruder!

That. That problem where there are so many dimensions of variables that it’s tough to know what to do to improve your situation. Maybe all you needed to get back to working was adjust the tension on that one screw on the extruder idler (just half a turn!), but you instead decide to take apart your extruder to floss the hobbed bolt. That’s how I felt for a long time. It’s not entirely how I feel now, which is good, but I’m blogging about back then. So if you’re reading this and you’re like “YES! THAT’S HOW I FEEL!”, you’re not alone – you at least have the last-year-version-of-me in your camp too. 🙂

Looking back on my pictures, the next dimension that I tried tackling was the extruder. I arbitrarily picked East.

One extruder that I wanted to build (which, spoiler alert, I still haven’t yet built) is the RepRap Universal Mini Extruder (also here on Thingiverse). That extruder uses 1.75mm filament.

So I started with its heater barrel. I needed to make this on my lathe.

First I figured out the angle for the cone at the tip.

After turning the cone, I had something that looked like this:

I then cut the brass for the heater block for it:

Here’s a better view of the barrel at this point:

Then I drilled out the back, with progressively larger drill bits.

The way this barrel will work is that a PEEK section will actually rest inside the brass, and inside both of those will be a PTFE tube. Here’s the PTFE tube inside just the heater barrel for show:

To have the PEEK rest inside the barrel nicely, the big hole in the heater barrel is a flat counterbore (as seen in the hot end drawings), not drilled with a standard drill bit. I purchased a 6mm end mill from mcmaster to flaten out that hole:

A few weeks later (after unrelated quadcopter, telescope, and pool opening fun) I then got on to the PEEK part:

..and for fun, here’s a 3D view of that (let your eyes drift until both images overlap on top of each other):

That was all around the May/June timeframe. That’s the last RepRap photo I took until August 21st (now I’m only around 7 months behind in blogging again!). Then I seemed to lose track for a while of the new extruder barrel, and instead just tried printing again. The next pictures I have of are a hobbed bolt:

… a horrible octopus attempt:

(which wasn’t quite as bad after cleanup):

… and then the Minecraft pickaxe I printed. 😀 :

The Minecraft pickaxe print did not go without incident. At one point I paused, triggering a bug I’d seen in either the RepRap Host software or the 5D firmware where it would print at incredibly slow speeds:

Despite that, I ended up with my deliberately-heavily-pixelated pickaxe:

Maybe I could use this pickaxe to dig West. 🙂

Then sometime in August, I saw a Google+ post by Forrest Higgs saying how he finally bought a proper crimping tool and was now making harness cables about ten times faster. Ever since the time that I’d initially found header housings/pins on mouser and shuddered at the crazy crimper price I saw there ($267.45 on mouser.com! AAGH!), I’d dreamed of being able to crimp correctly (but I wasn’t spending that crazy money on some professional crimper). It turns out an excellent crimper can be found for about a tenth of the price. He pointed me at these URLs:

Crimping Tool: 0.08-0.5 mm² Capacity, 20-28 AWG
http://www.pololu.com/catalog/product/1929
($29.95)

Female Crimp Pins for 0.1″ Housings 100-Pack
http://www.pololu.com/catalog/product/1930
($5.95 for 100)

0.1″ (2.54mm) Crimp Connector Housing: 1×2-Pin 25-Pack
http://www.pololu.com/catalog/product/1901
($0.69 for 25)

(and there are 1×3, 1×4, and 1×1 housings too).

These changed everything. They took crimping headers from a painful headache experience to something easy and satisfying. You’re insane if you don’t go buy one of those crimpers now. It ratchets, tightens everything exactly as tight as it needs to be, and the pins just slide perfectly into those housings and make that satisfying “click”.

Here is the crimper that arrived at my house:

Here was video that Forrest pointed me to as well that shows how to properly use one of these crimpers:

The only things about that video that aren’t as good as they could be are:
1) The crimper shown is one of the cheap horrible kinds that don’t even ratchet.
2) It seems like he’s skipping a step. I, at least, first crimp the clip around the insulation, then I crimp the inner piece around the wire. To crimp the clip around the insulation, I usually hold the female pin on the wire and use regular pliers to make a temporary crimp against the insulation, then I crimp the insulation part, then I crimp the smaller conductive wire part with another sized notch in the crimper.

More pictures of the crimper:

Ok so I moved my RepRap back down into my basement, and put it on the lazy susan that I used to use for filament. I know, it seems like that would react to the vibrations while printing, but it didn’t. What it DID do is give me easy access to all sides of the printer, by simply unplugging the power and USB. It’s like my RepRap is a car up on the lift at a mechanic’s station!

It certainly felt like it needed to be at a mechanic’s station. Here was a horrible horrible calibration print I did right after moving downstairs:

I’ll end it there. Things are about to get a lot better. A lot better.

In this post I’ll cover the RepRap disaster that happened in February of 2011 (last year) that set me back motivation-wise for several months, which is most of the reason for this Twelve Months part x of x stuff.

(Note: I’m a YEAR BEHIND on blogging now, and I really don’t want to cover these old times where things weren’t working well (since I have such better stuff to get to that came later), but I firmly believe in showing failures as well as success, so I feel compelled to blog it all. Just know – I’m way, way ahead of this now.)

At the end of my last post everything was looking good. I was using my new lathe, I made a new heater block, and I was printing again. Here was the flower I was printing for Laurie:

Well, that print failed because of a bad Y opto-endstop connection – RepRap host homes after every layer, so all later layers had constantly tried homing and then missed seeing the opto-flag, and printed all of the top layers off.. That print looked horrible (but I believe I gave it to Laurie nevertheless).

So, I tried fixing that connection, then started a new build, set up the webcam, and went up to bed. I checked on the print periodically before falling to sleep.

The way I remembered it, I fell asleep and came down the next morning to the following images. Looking at the pictures now, though, it seems everything was fine on February 19th, 6:55pm, and then these pictures were taken starting at 8:42pm. Here’s what I saw:

Near the end of my last post I hinted at what would be the problem – I’d re-used an existing PTFE insulator, because it felt “tight enough” when I screwed back in the heater barrel, even though I’d just created a brand new (perfect) one on my lathe. What clearly happened was the barrel pushed out of the PTFE insulator, the extruder kept pushing, and the heat block stayed powered.

The damage was even worse than it first appeared.. Here’s a picture from above the X carriage, taken two days later:

My X-carriage was destroyed. For comparison, here’s what that X-carriage looked like in January of 2010 – this was one of the coolest parts of the assembly for me – I couldn’t believe I’d made that.

This crushed me. The fact that the X-carriage was broken, before I got to the point where I could print well enough to print replacements, felt like a truck had pressed up against my chest. I remember thinking that I didn’t want to work on it anymore for a while – that I didn’t want to spend months more time working just to have it break again. This affected me way more than it should have.

Here’s video:

Here is a gruesome postmortem I performed later:

The day I filmed the first of those two videos (February 21st, 2011) was the last time I’d touch my RepRap Mendel for over a month.

But the next day, I went into work, and borrowed back our original Makerbot CNC that made the parts for my Mendel RepRap (a Makerbot which I and my friend Chris built in 2009, owned by work). Followers of this blog (are there actually “followers”? cool!) remember that towards the end of my printing parts with the MakerBot, it was performing rather poorly. I’d switched to PLA because all of the large parts would warp under ABS (and we’d had no heated bed), and the extruder would just stop extruding right in the middle of a 2-hour-long print, ruining it.

But recently I had heard stories about why so many of the original MakerBot Cupcake CNC machines had been so bad (and had stopped extruding). The story I read somewhere was that the old DC gearmotor that they shipped was not that great a motor, and apparently over the life of the motor, it would actually change characteristics… later in its life the motor would draw more current than before. I also heard that the extruder controller on the 3rd gen electronics had a current limiting feature built in that would cut out after 2 amps, if I remember correctly.

From what I’d heard, the motor would still function if you gave it more current.

Hearing this made me want to see if I could fix up the machine we’d made, and use it to print out a new x-carriage for myself. I borrowed it back from work (some other guy had been using it after us, and had had very little luck because the extruder kept stopping). After so many hours of trying to figure out what the problem was, I wanted to see if indeed that was the problem, and whether I could fix it by making a relay board.

So I brought it home into my basement, once again.

Originally I planned to go into great detail here about attempts I made to fix the MakerBot that Chris and I had built so long ago.. but I’m a year behind in blogging, so I’ll try to cut to the chase, with a small amount of description and a few pictures at most.

First I tried mounting my Wade’s extruder on the MakerBot. I got pretty far into setting up the firmware to use a stepper motor for the extruder, but at some point realized I was spending way too much time on this MakerBot and not fixing my RepRap. But here was the mount I made for it, quick and dirty:

I have video of it barely extruding. In retrospect I think it was a poor-quality hobbed bolt which was cut deep enough to work well with PLA but not so much with ABS, but who knows.

So I moved on to trying to get the Makerbot plastruder driven straight from 12V instead. I found a simple design for a relay board circuit that would preserve polarity, then designed and built my own stripboard version of that (using a tool I found called diy-layout-creator), so the extruder controller would only be triggering a relay, and then far more current could flow to the motor.

That did seem to fix the problem, but then I ran into other mundane printing issues like a clogged insulator, which I then replaced. The final straw that made me give up was when the plastruder inexplicably raised about 4-5 layers at once, ruining a print.

As I type this a year later, I think I know what the problem was, after seeing a later video. The bowden cable used by the filament spool (which I hadn’t used before – it was added after my last use of the MakerBot) had fallen to the side somehow, and basically the feedstock stopped feeding smoothly, and the plastruder did such a good job not quitting that it pulled itself up off of the leadscrew nuts into the air. But I didn’t know this at the time, got frustrated, and realized that – again – I was spending time on the MakerBot and not my own RepRap.

So I stopped trying to fix the MakerBot, and ordered a printed carriage from Tony Buser. I originally asked for the same X-carriage I’d been using, but thankfully he suggested trying what at the time was a brand new carriage, the OpenX Carriage by Buback. I LOVE this carriage. Here is what it looked like as I was putting it together:

The three main features of this carriage are that it’s got a much wider hole in the center (my disaster is proof that that’s a good idea!), it CLICKS onto and off of the rails easily, and it’s supposed to grab onto the X-carriage belt without any belt clamps. Here’s video showing how cool the carriage is:

So then I mounted my Wade’s extruder:

April 2nd, I was printing again, and by the 10th I was printing well again.

Unfortunately I found that the third feature of the OpenX Carriage – the ability to grab onto the belt without any clamps, didn’t work for me as well (it slipped during a print or two), so I reenforce it with a new ziptie every time I take it off and put it on. That slightly robs from the whole idea of being able to put on and take off the carriage, but it’s still WAY easier than the old carriage, and I like it way more than the snappable PLA bushings that the Prusa Mendel uses (and also way more than LM8UU linear bearings, but that’s me saying that now in 2012.. let me get back into my 2011 mindset).

Then I tried some prints, with some mild level of success.

You’ll note that those last two prints (the hemi-demi-sphere and the octopus) don’t look that great. They’re perfectly lined up in X and Y and there’s no slipping, but plastic wasn’t extruding consistently at this point. Here was an example of just how bad this new extrusion problem was getting:

That got me to around April. In wanting to wrap this post up quickly I almost threw in here a picture or two of lathe work, a hobbed bolt, another octopus, a minecraft pickaxe, and a video showing what seems to be a bug in either RepRap Host or in the 5D firmware where pausing can sometimes render very slow speeds after a resume, but then I realized I have too much to cover. So this will just take more posts than I thought/want. Ending it here.

Trust me, it gets way better for me. Now at least I’m caught up to April of 2011 (even though it’s February 2012 now). Only 10 months behind! 🙂

But most importantly, I finally got that printing disaster story out of the way. Wasn’t looking forward to posting that. And, I finally got to praise Buback’s OpenX Carriage!

I do want to leave you with some sort of ending though.. So I’ll use this: while posting this blog that’s so far behind, I’m constantly reminded of this sketch. Watch it and enjoy: