Saturday, December 31, 2011

Duckon 2012 blinkie prototype finished

The 3D printer figured big in this year’s blinkie design.  Not only is the battery holder printed, but I used a gerber-to-gcode conversion script to enable me to use the RepRap to lay down the ink on the blank PCB to make the prototype, and by the time June rolls around I may have a printed sleeve to go on it that diffuses the LED light.

I used the Gerber to GCode converter, tuned up a bit (added some experimental backlash compensation.
That failed the first time, the extra fine sharpies don’t lay down much ink. The 2nd try I printed each side 3 times, then went over traces with a regular (fine point) sharpie as much as possible (my hand is too shaky to do the very close traces). That went OK.
Then a quick drill of 52 holes and about 90 minutes of soldering (soldering tightly spaced LEDs when you have to solder top and bottom due to the hand-etched board is time consuming) and it pretty much fired up right away. I did miss a couple of solder joints and had to touch those up.
All I have right now is the test pattern for the LEDs, so I need to do a bunch more programming to get a wide variety of patterns. This one will allow selecting of patterns as well and remembers settings when turned off (since it never actually goes off unless the batteries are removed, it actually starts right up where it left off).
The battery holder for the 3 LR44 cells is printed on the RepRap. I am going to play around with the idea of a printed diffuser for this blinkie, to be sold as an add-on for a few bucks. IMO a diffuser makes any blinkie look cooler.

Thursday, December 29, 2011

Gerber to G-Code conversion

I want to use my RepRap to draw ink on bare PCBs for prototype etching (toner transfer has never been good to me). It’s been done, multiple times. When I looked into the projects I could find, I found they all were broken with the current EagleCAD, or generated horrible, basically unusable traces, or involved a dozen steps involving importing and exporting multiple file formats across a half dozen applications, or they didn’t work at all.
I’d never seen the inside of a Gerber file before, and I’m barely competent to look at GCode, but I never let ignorance stop me before! Google has my back.
After a few hours of bashing away, I present a pretty much usable Gerber-to-GCode converter, written in PERL. I implemented the bare minimum of what I needed to print this board, and I haven’t actually tried it with a pen and a PCB yet but it looks usable to me.
I printed this one on paper with a graphite pencil, so the traces are a bit light, and it’s definitely not optimized; a little more work will have it not lifting the pen and putting it right back down in the same place at least. That’s probably as far as I’ll take it.
The board is actually not quite ready to print anyway but this bit is done now. I will post it online somewhere once I get a proven usable board out of it.

Thursday, December 22, 2011

Clears clogs fast!

I had seven nozzle clogs on Sunday, giving up at midnight.  Monday after work I started printing again and all went well until Tuesday when I had another clog.
I’ve been getting worried that I’m tearing the hotend down so much that I might eventually tear the threads up too much.  Using the nut lock technique to turn the nozzle on particularly wrecks threads.  So this time I just put the nozzle on as tight as I could get it by hand – I figure by now the barrel and nozzle have a close working relationship anyway.
When my nozzle clogged again Wednesday evening, I just pulled the insulation and thermistor off and removed the nozzle.  Then I turned the heat back on and extruded about 100mm of filament through the barrel, hopefully clearing out any burned crud in there.  I shut the heat off, cut the plastic at the end of the nozzle, and finger tightened another nozzle on.
It worked fine.  It’s not particularly faster than tearing the hotend down since I’ve gotten so good at that, but it wears things out less and with a little practice (which I do not particularly want to get) I might get faster at this method.  I need to recompile the firmware to allow extrusion while cold, since the thermistor is just hanging there it thinks it’s cold.  I don’t necessarily recommend this unless you’ve got some crappy plastic that’s clogging on you too, but it’s working for me.  Not heating the barrel up red hot and not having to re-thread the barrel into the thermal barrier should make things last longer.

Tuesday, December 13, 2011

3D printed molds

I’ve thought about using a 3D printer to print molds to make things out of materials that the printer can’t normally print, or shapes it couldn’t normally print in one go.  Here’s a great example of someone doing exactly that

repaired audio plug on earbuds (from

Monday, December 12, 2011

Nozzle clogs

I’m having a lot of nozzle clogs. I think it’s this ABS plastic. Over the weekend after clogging both my original MakerGear and the new JHead, I put on a brand new shiny MakerGear nozzle. It printed for about 6 hours, then this morning it clogged too. There’s got to be something in the plastic; perhaps some foreign substance or the plastic isn’t blended well.
When it clogs, it goes from having the filament drop straight down from the nozzle to having it spiral into weird, deformed curls, and I have to press like hell to get anything to come out.  I’ve tried pushing the temp to 270 and it still doesn’t clear.  The propane torch clears it but I have to tear the hotend down for that.
I’ll keep using it but it’s frustrating. I shouldn’t really have to keep spare hotends on the shelf in order to keep printing.
I’m getting good at cleaning nozzles.
I’m probably not buying plastic from this supplier anymore unless another cause for this problem turns up. I don’t think I had a single clog with UltiMachine plastic. This ABS was a “good deal” but not when I factor in that I’ll spend 10 hours cleaning nozzles before I use it all up.

Wednesday, December 7, 2011

Hobbed bolt redesign needed

UPDATE: this turns out to be completely wrong, the problem was a broken idler on the extruder.
I got some ABS from a new supplier last night. It looks pretty good on the spool and when extruded, but it’s a bit softer than what I’ve worked with in the past. This is generally OK but it’s slipping on my hobbed bolt, and once it starts to do that it doesn’t stop slipping on its own (not unless I notice and push on the filament) and prints get ruined. Luckily it died on a small print, I was just making a batch of electronics brackets, not something like a plate full of X ends.
I made this batch of hobbed bolts with an M4 tap, and it’s been good for UltiMachine ABS and PLA, though it HAS stripped once in a while (rarely). Tonight I’m going to make an experimental one with an M8 or M10 tap so it’s got more aggressive teeth, and try that. I’ll have to recalibrate my esteps. I’ve seen people hobbing with both M8 and M10 taps before so it’s worth a try. It’ll be closer to the spur gear that MakerGear uses in their plastidrive, and MG’s is pretty bulletproof.  People who have used M8 and M10 say they’ve had good results.
I did buy a drill press vice last week so I could try this anyway so that’s sorted, I just need to print a pair of bearing holders so I can clamp the bearings in the vice.
If this works out I’ll have to re-do the hobbed bolts that I made for the batch of kits I’ve got nearly ready to go.
It’s all good though, if it results in an improvement in the design and an overall better printer in the end I’m OK with throwing out a half dozen bolts and re-making them.

Slic3r – check it out

I’ve mentioned it on G+ and elsewhere, but not here yet:
Slic3r is a new, fully rewritten, well structured gcode generator for RepRap printers.  The author has made tremendous progress in the last 2 months, and after checking out the latest version, I’m sending him some money and I’m replacing Skeinforge with Slic3r as my default gcode generator.
It’s clean, it doesn’t have leftover crap for dealing with equipment that nobody uses anymore (DC extruders, etc), it’s fast and it generates nice, fast GCode that follows more logical paths than what I get out of Skeinforge, and the resulting prints look better.
I gave it the acid test this evening; a Greg’s hinged accessible extruder.  It did wonderfully, in fact, I think better than Skeinforge.  The bridging detection is great; it localizes solid layers to only the areas necessary to support bridging and voids, rather than making a full solid layer as Skeinforge does.  This is up from really not even detecting bridging to speak of just a few weeks ago; the author is clearly putting a ton of time into this project.
Written in Perl, works under Windows/Mac/Linux.

I was completely wrong, multiple times

I stopped by the hardware store on the way home and bought a course thread M10 tap to try my hand at making a more aggressive hobbed bolt.
I came back from the garage with a passable (not great) effort and after careful cleaning, installed it.  Then while putting the filament back in, I realized that the extruder idler was broken at the hinge.  Dur.  I put in a new idler, put the old hobbed bolt in, and it’s running like a champ now.
Honestly, this is probably the reason with the trouble I had with the J-Head last weekend.  I will install the J-head in the new printer and see how it goes.
The upside to this is that I came across the idea (which I should have thought of myself) of using nylock nuts on the extruder gear rather than wedging two nuts together, which is difficult to do with one of them halfway into a nut trap.
Also all the parts I’ve printed for friends were done since I switched to 0.25mm layer height and they’re much stronger than the old one I had in my machine, printed at 0.33mm layer height (though it takes a lot longer to print things at that height)

Monday, December 5, 2011

Wearables / eTextiles

After reading specs and some bad reviews of the standard conductive thread that SparkFun offers, I found this link that looks like good info for those investigating wearable electronics.
This is going to require some investigation and experimentation.  Flaky/unreliable thread could be quite irritating.

Saturday Setbacks

I ended Saturday significantly farther behind than I started it.  That’s how these things go sometimes.  You just hope you learn something in the process, so that next time you can make new and more glorious mistakes.
I started with trying to use a few unbroken scraps of that clear PLA that I was about to throw away.  I figured I could make a few linear bushings or something with it.
Unfortunately I got into some areas where the diameter exceeded 3.2mm (it varied from 2.65 to 3.25mm, wow) and that was too big to fit through the hotend – and while it was jammed, what was left in the bottom apparently scorched and clogged the nozzle.
So I disassembled the extruder and spent quite a lot of time with a torch burning crud out of the nozzle, and with drill bits in hand scraping junk out of the shaft.  After 3 tries I was sure I had things sorted, and…it wouldn’t heat. I found I had broken the wire into the heatcore, just where it enters the ceramic.  Argh.
I had a J-Head sitting around that I’d intended to use on my new printer, so I installed that.  I’d never used resistor block heaters before.  I had little luck with it – it seemed like it was extruding OK but it also seemed to have little thermal mass, so when I started to go fast with it (over about 80mm/s) the extrusion started to peter out until it slowed down again.  It’s possible that I was doing something wrong, people on IRC assure me that they use J-Heads without trouble, but for now I don’t have the time to mess with it.
EDIT: I now believe that the J-Head is innocent – I later discovered that my extruder idler was broken.
I got out a spool of 28 gauge bare nichrome, broke the ceramic up with pliers to get down to the bare M6 threaded insert, wrapped it tightly with kapton, and wound a new core.  The 28 gauge needed two wraps to get to 5.6 ohms so I made a kapton/wire/kapton/wire sandwich, and covered the whole thing with leftover ceramic paste.  This eventually got me going.
I also had the thermistor fall of the bottom of the HPB during a print, which caused the HPB to get very hot and scorch and partially melt the extruder I was printing.  As a result I came up with the idea of drilling a hole in the center of the top platform and feeding the thermistor up through that, and laying it down in a groove carved out from the hole and putting a piece of kapton over that.  This allows the HPB to lay directly on the top platform which should increase performance (no loss to the air) – it did make the reading on the thermistor more accurate than I’ve ever seen it before.
I did manage to make a little progress by the end of the day Sunday – the frame build instructions are about done, I have a batch of Sanguinololu CPUs with bootloader and Marlin burned to them, I did finally get that extruder printed, my printer is slightly improved (and therefore all future printers will be improved)

Friday, December 2, 2011

PLA problems

A few weeks back, I bought two 5 pound spools of PLA from a supplier I hadn’t used before.  It looked OK on arrival though the translucent green stuff looked a little funny; like it had been kinked in places.  I transferred it to my standard spools and put it in storage.  A few days ago I took it out to use it and found that it was so brittle I couldn’t use it.  In fact it had snapped on the spool in multiple places.
The supplier provided a good resolution though.
This is a photo of the clear stuff now.  I did nothing to it other than wind it on a spool and let it sit for a few weeks, it just snapped into many little pieces in the center.  Clearly it does not like to be bent at any kind of an angle.
The green stuff is brittle in a different way.  It’s so bad that you can’t print with it; it snaps inside the extruder.
When I contacted the supplier, he came back quickly with an apology, and offered to either refund, replace with PLA, or replace with ABS instead.
I didn’t trust his replacement PLA to be any better, but I do have a use for ABS so I went for that option.  Kudos to him for being quick to do the right thing.  I will report back on whether his ABS is good
Also he is including a sample of a new PLA formulation which he says has an additive to make it less brittle.  I hope it works, because we need more good plastic suppliers in the US.  This guy has the honesty, if he gets his product problems sorted out I’ll be happy to work with him more.
FYI the supplier here is – and though I must recommend against buying this specific plastic, he’s trying to do the right thing and I think it’s possible his ABS and newer PLA will be OK.

Thursday, December 1, 2011

Dog steps

My dog is getting old and has trouble getting into bed (the top of the bed is 30 inches up).
A few months back I studied what was available, found that ramps were the most common, so I built one.
It turns out my dog hates ramps.  She wouldn’t use it at all.  She’s not getting any younger so I decided to try steps.
Tonight I broke down the old ramp, did some cutting and came up with what you see here.  It’s a little uglier since the ramp carpet wasn’t big enough to cover this configuration but it’s good enough to see if the dog will use steps instead of a ramp.
If nothing else I’m sure the cat will take to camping out under it.

New printer: JGAurora A3

This week I decided I'd had it with all the other printers in my stable.  The CTC is stable and decent but it just bugs me (can't st...