Saturday, March 6, 2010

Need to catch this blog up to reality


So many many successes since the last update. I've gotten my extruder reliability up pretty high, and I have had hours of successful printing (even joining two seperate ABS filament sections using a lighter, instead of having to stop to reload).

First things first, I went ahead and replaced the ball chain on my x and y axis. The difference is huge. My ball chain was slowly stretching and the chain was no longer meshing well with the laser cut gears that it originally fit. I wasn't able to print smaller items as the movement wasn't smooth and occasionally I could even have an axis jump an entire tooth. I ruined a long print this way and so I decided to go ahead and tear down that part of the machine and rebuild it around a "ladder" style cable. Don't know the official name for it, but it works great. No more stretch, it meshes well, and my prints are hugely improved. No more waviness between layers and the constancy of motion is great, especially visible in the evenly spaced rafts. The only huge trouble with it is that I didn't originally have gears for it (and couldn't print accurate small parts before I installed it. chicken vs egg scenario). I ended up actually "carving" my own out of larger laser-cut acrylic cylinders. I didn't have to drill the hole, and centering was fairly easy, but I did have to form the teeth & spacing. Took a while since I had to make three just using simple files, but they are now working very well.

I also went ahead and exchanged my thin acrylic base for one a quarter inch thick. My build surface stays consistently flat now; the older 1/8th sheet was getting wavy even though it was bolted down.

This kaizen approach is how I envision my work to continue for a while. At least until I've managed to print an entire mendel. I have a decent list going as to what aspects of my darwin I need to upgrade before I start printing in ernest and most of the improvements will actually carry over to the next gen machine pretty well. I've already printed a new pinch wheel extruder which should decrease my print times by quite a bit. I'd like to run that through the bowden cable configuration that some people are implementing so that I can at least lower the inertial issues that my top heavy machine already has. Hopefully I'll hit up against the upper limit of my head travel speed. also, before I can attempt some of the larger mendel pieces I need to deal with the warping that occurs, as my extruder body and large spur gear showed some level of it, and it would only be worse as the parts I attempt get bigger.

I'd like to design a heated bed that would cover up to 50% of my current print area, but easily removable so that I could throw it on to print a single large piece (such as the mendel x-carriage), but still be able to print an entire bed of smaller (more warp tolerant) parts.

I'm excited at the rate this is all picking up. I've printed sucessful parts and now a sucessful mechanism. next step is to integrate it into a sucessful system. Eventually I'll transfer to an entirely self made printer (and improve it further from there.)

Sunday, February 21, 2010

Big couple of weeks, eh?


Well, I have been busy lately. Obviously I got my printer working as you can see in earlier posts, but I also got to visit my younger brother in Boston and several day after we got back I proposed to the most beautiful and wonderful woman in the world (she accepted). No offense to nerdish people, but I was quite a bit more excited for the latter most point than I was when I saw my printer first spring to life after a year of technical challenges. I'll leave it at that and just say I've been happily busy with life, but I've still have had time to work on printing and optimizing the Hapsrap.

In fact, yesterday afternoon a fellow local enthusiast, Chris Keller, came by and helped me figure out some of the key skeinforge settings I needed to tweak. He's just embarking on his project (and was quite pissed when notification of makerbot's sellout of the build 12 units came out.) and so I hope the first hand observation of some of the systems and issues helped him. I definitely know it would have been great for me, never the less, I managed and now hopefully more of a community of reprappers can start coalescing here in Chicago.

As an aside, Chris and I discussed what would be the best way of reaching operational status. Obviously I went the route of lasercut repstrap, but a lot has changed since I chose my path. The biggest thing should be obvious to anyone who follows the community. Makerbot has sold 500+ units of their cupcake system and that install base and forum community makes troubleshooting much much easier. there are definitely limitations to their system (one of which is their constant lack of stock at any given time. good for them, bad for Chris). Plus once you get it working, you can go ahead and build your own Mendel, AND thats after the first learning curve, though more slight than if you built a unique system, you will learn quite a bit and carry that forward.

Now that I have better parameters (probably not the best I could have, but certainly decent) I've started printing more practical parts, with more complex geometries. I've printed a couple of the part shown above, a bearing block used to make a component of Wade's geared pinch wheel driver. Should certainly be functional (even with one being warped), but the best part was actually getting to click the "I made one" button in thingiverse.

Below is a video showing the part mid-build (another end-build vid showing the decent surface is available on my youtube page.)


Now I'm going to move ahead and try to print some of these other parts.


Wednesday, February 10, 2010

Some proper iterations

So the fact that my first real print (see previous post) turned out so clean was total luck. My axis were off by about 30%, my feedrate settings was also wrong due to my not understanding the cryptic setting names, and it seems that there are even sweet spots to print in on my machine.

I shrunk the axis, and my next several prints (3, 4a and 4b, and 5) were gross blobs with way too much plastic. The reason why there were so many blobby versions is that I was actually changing a parameter in the wrong direction.

my next print was decent with the amount of plastic, but it seems that there are snags in my XY, (though mostly in my X). my print ended up with ridges down one side. One of the four vertical walls were perfectly straight so I figured it out pretty quickly. My next print was the one you see above. it measures about 19.96x19.87x20.33mm so I dialed in pretty well. The bit that it is off is probably due to the flowrate change or measurement error.

I almost had seal on the top and bottom faces so I tried to go for something a little bigger and a bit more challenging. I adjusted some settings and went after making a small cup. I didn't change them enough so it won't hold water, and the larger print definitely showed more of the ridge behavior.

It certainly isn't the prettiest, but I've learned a lot.

Unfortunately for the momentum I've built up, I'm actually leaving town for the next several days to visit my brother in Boston. I'll relax some and come back with some ideas of what to design and print next.

I'll certainly take recommendations. I imagine several of the people reading this are people from the community here in Chicago. But I guess with Buzz, more and more people will be able to find this.


hi.

Monday, February 8, 2010

test block one

well, today I had my first (test) print! I ended up actually breaking my extruder yesterday (stripping the the threads on the PTFE insulator) when I tried to improve that barrel-insulator gap. It is really good that I actually have two sets of heater barrels/nozzle/nichrome and I was able to swap it pretty easily. In my recent rebuild I've made the extruder system more compartmentalized. The extruder motor assembly can be taken off quite easily from the heater/nozzle assembly without any rewiring. My cable harness is easily detachable, both from the hapsrap system as a whole and even between the main body of the extruder and the heater section. This will really help as I begin printing experimental parts, such as a stepper/pinch-wheel assembly. I'll be able to swap parts to test each iteration pretty quickly.

That actually brings me to one of the negatives of my set up. I've managed to work through many of the issues associated with this kind of DC motor screw thread extruder. But it is just that the flowrate is simply very slow. I'm able to extrude at about 10.6 mm/sec, which is pretty low compared to other systems. My cartesian system can certainly be pushed a bit more than it currently is. In any event, I'm too excited by printing at all to make too huge of an issue about it.

One other thing that is pretty obvious from this build is that my scales are way off, especially on my x and y axis. The box I built is actually supposed to be a 2cm cube. right now the measurements are 2.89x3.10x1.80. I was wondering why my print was coming out so sparse, the top and bottom surfaces are supposed to be water tight, but are definitely not. It actually took me a few layers to realize that it was much bigger than I expected it to be (mostly due to my excitement to be printing at all.)


There was some warping and my raft actually lifted from the bed on one corner. Also, probably due to the warping, there was one layer that simply over-hung on the right hand side, creating a lip about 5 layers up the print. I'll have to adjust the raft, but I should probably fix my scaling first, as the extruder/time shouldn't be affected, and I'll have much more (~30%) material per cm. Both on my raft and on my print.

Also, in the background you can see my little extrudate mountain from when I was dialing in this new extruder set-up.

Last but not least I'll leave you with a quick video of the machine running. Kind of obligatory and unnecessary, but as I mentioned before, I am kind of excited. A year is kind of along time to bang your head against problem after problem.





Sunday, February 7, 2010

Well, I went ahead and swapped my gen2 electronics with the gen3 electronics I recently built (I was using them to exercise my extruder as a stand alone system.) and lo and behold, I sidestepped my software issues.

I also discovered another problem with my previous set, so my x-axis used to skip steps, and over-heat very easily. I figured that the axis must have had some extra built in friction somewhere. I put a bigger heat sink on and even point a small case fan at it. I decided to change all electronics in this recent switch, not just the motherboard and extruder board, and figured I'd have to tweak the current on that one stepper board. Well, it seems that there must have been something wrong with that particular board, because I have my x and y at exactly the same settings with no overheating problems. Oh well, it is one hurdle that I no longer have to worry about.

Anyway. I had been spending all of my energy trying to get this one sanguino controller to be recognized by replicatorG, and was having a hell of time. The official reprap host software is set for a stepper extruder, and I didn't have that. ReplicatorG is much more of a flexible system, but I couldn't figure out why it wouldn't see my machine. It was giving me errors saying it couldn't see my serial connection, even though I could see and hear that my hapsrap was reseting. Really, quite frustrating. When I first got the new electronics set up and mounted, I tried running the most recent mendel version of the gen3 firmware. Still no dice. Then I decided to try to emulate the cupcake, as there is no reason why the system would know what my electronics were actually hooked up to. RepG has built in firmware and firmware push abilities and had no problems with updating my micros. And there you have it! Working system! It really wasn't a problem configuration wise either, RepG has a machines.xml file where I can put in my axis resolutions, and the firmware has onboard settings to flip directions where necessary (X and Y). I was even able to start some prints! (as you can see above.) So I still need to look into it more, but it seems it was a firmware issue. Seems strange still as there are plenty of people using replicatorG with mendel firmwares.

I still need to work through the normal extruder work-in phase as I recently rebuilt it, but I've done that before and it shouldn't be a problem. I do, however, think I need to adjust my thermistor settings, as I have to run the extruder way too hot (around 235C vs 220C) to get smooth extrusion without stripping my filament. I also came across a problem I had the first time I built the extruder, there is a small gap between the brass heated barrel and the PTFE insulator. After running the extruder for a while (especially with stops) the melted portion of the ABS can reach all the way up to the cool insulator and plastic fills that gap, making a pretty good plug that pushes my force needed way too high. This is why I haven't been able to complete an entire built (only 10 or so layers).

Yeah, so real annoying, but easily workable. I can definitely say I am excited again.

Friday, February 5, 2010

refocus

So my machine has been sitting pretty idly by in the room where all of my project gear lives. Plenty of things have kept me from working on it this last three fourths of a year, from summer bike rides, an apartment move, work travel, and various holidays and smaller trips.

but it always bothered me that i was letting myself get stopped by these software issues. so in December i decided to start focusing again. i fired it up. set to rebuilding the few parts that gave me mechanical trouble (z-axis and extruder) and recommit myself to learning more about programming. of course i quickly hit the same point that stopped me earlier, my older configuration wasn't explicitly supported with the newer software and i would have to hack at it to get it to work. i decided it would be prudent to get a few generation 3 boards so at least i could get component systems going and have back ups in case issues arose with my original micro.

it feels like spring cleaning, where i used to want to power through and keep everything stock and deal with what i was dealt, i no longer feel that i am betraying anything by tearing down and rebuilding mechanical or electrical systems. I've also admitted to myself that i need help from others who stand so much more firmly on these software/firmware structures i can't make heads from tails with. I've extended a call to help to the local hacker space, and to others whom with which i have some tentative connection. (friends of friends, friend of brother, etc.)

i am still learning quite a bit, but i'd really like something to show for it. even if it is some lumpy piece of plastic.

Sunday, May 17, 2009

software stuffs.

I actually got into this project through the arduino microprocessor that the newer versions of the reprap electronics use.

I was struck by all of the art/hobby uses of the arduino and its clones across all sorts of different areas. from plants that can tweet when they need water, to a simple robotics platform. It's inexpensive, a breeze to program, and there are so many different physical packages, that you can certainly find one that fit's well with your project.

The newer versions of reprap electronics use the arduino platform, and the associated programming environment. It has done wonders to make the project more accessible to people who may have difficulty with such thigs.

Anyway, I wanted to mention a bit about the software side of the reprap project. Personally, I'm a very mechanically minded person, it's what I studied in college, and I can even hold my own electrically too. Basically, they are physical systems and I understand those easily, I can visualize voltages and torques and balance and currents, I can see when something I put together probably won't work. I'm also pretty good at seeing how things might/will fail, and that's what I do in my day job, actually.

When it comes to programming, however, I have a hell of a time keeping variables and functions straight. I've spent plenty of time dealing with more simple languages, BASIC back on my commodore64 as a kid, HTML/general web stuff in high school, C and C++ in college. I've been wowed as to what some reprappers have been able to do, specifically some of the JAVA and PYTHON work.

At some point, once you've decided that you actually want to build and run one of these things, you need to sit back and take stock of what you need to download/purchase/learn. Like i mentioned in a previous post, there are many options mechanically, and there are actually a few on the software side as well. Personally, I use a mac, which happens to be one of the less common platforms used. Linux and windows seem to be the defacto standards, and there is some tough going for mac users. There are options, but one in particular, the JAVA host software I linked to above, doesn't seem to work yet on OS X.

The tool path is this: 3d model in STL (and it seems like it is now possible with STEP files too!) is sent to the Skeinforge "slice and dice" program to be converted to gcode instructions for fabrication. These instructions are what the printer will follow to rebuild the solid layer by layer.

This gcode file then has to be fed into the printer's microprocessor. These are instructions like, "print a perimeter from (2.3, 4.4),(4.3,4.4),(4.3,7.8), (2.3, 7.8), and back to (2.3,4.4). then fill it with zig zaggy lines" This is slightly varying for every machine, line thicknesses for example, so there are pretty deep parameter fields in skeinforge to be adjusted. Normally, someone may use the standard host software. I don't have that option, but luckily there is another package available. This one is called ReplicatorG. It also seems to have some issues (i've heard accounts that it may hoard resources on the computer if you are doing very long prints) but it does what i need it to do. i can jog each axis, and run the heater and extruder motor manually. I can then load a gcode file after conversion in skeinforge, and it then feeds those instructions to my system. i'm not quite there yet, but right now it looks like everything is coming together.

My apologies for all this, if anyone has read this and my previous posts i hope you got some thing out of it. i do plan on doing some shorter, more frequent posts with my own specific build details, but it feels like i should do some kind of service to people just getting involved in the project. paint some big broad strokes in one place, to at least get you started. everyone has probably tried finding some specific information in a message board site, and the search output is usually terrible.

best of luck if you're inclined to try some of this stuff.