Wednesday, January 31, 2007
All the systems work. The trick now is to get them to all work together. Wish me luck. I have a feeling that the hardest part of the job of making Tommelise a reality is ahead of me.
Tidying up the Mk 2.1
That's when the Mk 2.1 started pumping like crazy. I had the GM3 seated on #40 studding bolts not locked down. I noticed that the motor coupling was riding up pushing the GM3 up on its studding bolts. That didn't hurt anything and I've had the motor coupling, made of a cut coupling nut and nut with a lock washer between the two, come loose before, so I didn't worry too much about it. After I finished running the 400 mm of HDPE, I had a rush of good sense to the brain and decided to take the Mk 2.1 polymer pump apart before I tried running HPP.
That was rare good sense. When I disassembled it I discovered that the motor coupling hadn't broken loose at all. Instead, the thrust collar that I'd made of two nuts with a lock washer between had broken loose letting the whole threaded polymer pump try to work its way out of the pump.
One thing wants noting at this point. I use a thrust collar to absorb the tendency of the threaded drive rod to try to come out of the pump rather than the two bushings that the Mk 2 has. I have the a brass plate with a hole through it to admit the threaded rod at the top of the pump before you get to the motor coupling. The previously mentioned thrust collar butts against that. Basically, I used the same idea that Vik did with his drawing that is in the Mk 2 documentation on the reprap site.
Where the bottom bushing is in the Mk 2 I have a steel bushing that is the same diameter as the threaded thrust rod. This lets me avoid having to mill down the thrust rod since it seats right against the threads.
Back to the story. The threaded rod's thrust collar had come undone and let the threaded rod ride up in the pump till the bottom of the rod had slipped completely off of the bottom bushing.
That's when the pump started working properly.
What that meant was that with the configuration I was using all the bottom pressure springs in the polymer pump were doing was adding friction to the device for no good reason. The springs at the top where I was applying a touch of extra pressure with the c-clamp were effectively the only ones that mattered.
For the HPP I reseated the threaded rod in the polymer pump, really tightened the thrust collar and applied a bit of pressure to the top of the pump with the c-clamp in addition to what the springs could offer.
HPP pumped just fine with this arrangement.
This morning I sloped down to the hardware store and bought a more powerful pair of springs for the top of the pump in the hope that I could get rid of the c-clamp. That worked. You can see the Mk 2.1 sans c-clamp operating this morning in this pic.
The picture shows you something else interesting that is happening. The BBQ paint that I used to tack down the insulated nichrome 80 heating wire is perishing at the top of the heater coil. This lets it delaminate from the extruder barrel and then causes the insulation to overheat and begin to perish as well. The extruder barrel is still working, but I can see how this is going to go.
My first try at a fix for this problem is to use more BBQ paint coats to tack down the heater coil. I've got some other options if that doesn't work.
Some Observations on Wood's Metal
I was also able to draw some up into a plastic syringe. That gave me the ability to apply it wherever I wanted. Also, if I let it cool in the syringe, I ended up with a nice cylindrical slug that was easier to melt than the original lump I had. A bit difficult to get the cooled metal out of the syringe but not impossible.
As for actually applying it, I did notice that it tends to drip and bead when molten. No big surprise there.
Tuesday, January 30, 2007
Bloody HPP and HDPE extruded thread everywhere
I ran 500 mm of 2.9 mm HDPE filament and 750 mm of 2.9 mm HPP filament. Now I've got to tidy up the Mk 2.1 so that I can mount it on Tommelise's xz positioning table and then I can begin the long, painful process of getting it to make something useful.
Mk 2.1 extruder is operational
Monday, January 29, 2007
Mk 2.1: Defining the envelope
I first established that it would work at all.
It did with HDPE. The extrusion rate wasn't nice, but these are early days.
First I did a series of tests with HDPE. It's a fairly tough, but soft surfaced engineering plastic. With hand feed experiments that I did a while back it was slippery to a point where I had to grip it with pliers to feed it into the extruder barrel.
My first instinct was to really tighten the springs that press the filament into the threaded polymer pump. You can do that, but if you try to make the pump run too fast, that is, develop some pressure in the extruder barrel, it just excavates a groove in the side of the filament and you get nowhere. That was discouraging.
I then shifted over to polypropylene (HPP). It is a harder surfaced plastic and the threaded pump drive engaged it much better than it did the HDPE. After faffing around with HPP for a while I began to plan an experimentation campaign.
Right now I'm running the polymer pump with HDPE at 2 rpm and 2 amps power. I'm going to keep power constant and run the rpm up step by step and see what happens.
CuPro Cote - Preliminary Findings
I measured the resistance. In the megohms. Not good. I applied a second coat. Down to kilohms. The third coat got it down to about 10 ohms.
Odd thing: A thick coat didn't conduct any better than a thin coat but three thin coats did. Probably has something to do with how it dries.
I didn't bother trying any high-current tests. Unless the resistance can be made much lower, I doubt high current is viable.
Sunday, January 28, 2007
That allows you to extrude 1 mm/sec of 0.8 mm extrusion thread assuming a 2.9 mm filament.
I know I can go a LOT lower than that.
Firmware for the Mk 2.1
The Mk 2.1 has a number of changes in how it is operated. Probably the most important is the use of a channel on a 754410 dual Darlington chip rather than a transistor to run the gearmotor. What that means is that I can feed filament into the heated extruder barrel AND pull it back out. I've just tested that feature and it works properly.
I'm thinking that bidirectional control of the filament will make extracting tag ends of filament out of the Mk 2.1 easier and that it will let me avoid the dribbling that Vik has reported with CAPA by backing the filament up slightly when the extruder barrel is hot but we are not extruding.
The previously blogged low thermal inertia extruder barrel also makes heating response of the filament much more lively. I am also controlling the temperature of the melting filament indirectly rather than using a thermistor as per present practice. I'm controlling power input to the nichrome 80 on the extruder barrel as a function of filament feed speed rather than using a thermostat function.
Using the slow pluse "pseudo stepping" rather than PWM motor control also lets me maintain full torque down to zero rpm. We shouldn't be seeing stalling absent a complete extruder barrel jam with the Mk 2.1.
Running the Mk 2.1 from the Tommelise controller board
I don't know what the problem was with the little 16F628A test boards that I made up to run the Mk 2.1 extruder, but I could never get them to work properly.
This morning I shifted the gearmotor control over to the Tommelise controller board that I built up around the 18F4610 and it worked beautifully on the first try.
After breakfast I am going to run a few tests to make sure that the polymer pump on the Mk 2.1 is working properly. After that I will hook up the heated extruder barrel and see if I can fully control the Mk 2.1 as an ensemble from the PC.
Thursday, January 25, 2007
A new blog for machine vision...
I'm within a few months, I think, of having a working RepRap machine. I plan on improving Tommelise incrementally after that unless there are major operational issues with it. I plan on running it pretty much 24x7 afterwards to make things not only to improve Tommelise but also to advance work on the telepresence stream of development I'm starting.
I've begun a new blog that I've named Cyclops to track this work here and will be opening forums for others wanting to get involved. Drop by when you have time.
Wednesday, January 24, 2007
#$!!@#@% Radio Shack
I finally gave up trying to debug the dual darlington chip installation on the Mk 2.1 extruder test board. Instead, I disconnected the offending one, found another piece of landscape on the test board and put in another one.
This time the serial comms weren't affected by the installation. Just as I was checking out the circuit trace by trace my Radio Shack Digital Multimeter decided to give up the ghost. This happened once before. The head guy at the Radio Shack shop that I bought it from pretty much accused me of deliberately buggering it up so that I could get another one. I think that he knew, however, that I'm the sort of okie that will go to corporate headquarters if people don't stand by warrenties, however, so he got me another one and said that if I came in with another broken one he'd see me in hell before he replaced it again.
I just don't know what it can be. I'm working on a circuit board with no voltage higher than 12v and out of the clear blue sky it will decide, always at an inconvenient moment, to just die.
Anyway, this time it decided to wait until the warrenty was past, but I'll tell you, I'm finished buying multimeters from Radio Shack. This is the third one, one $24 one and two $48 dollar ones (one replaced on warrenty with considerable bad grace, mind), that have died on me. It's off to Potters' in the morning, I guess and I'll see if I can get a bullet proof one.
Tuesday, January 23, 2007
Glowy Things and Shiny Things
I've been considering the design of the motor modules. The serial ring idea is really neat, especially when each chip has it's own built-in UART, but I didn't like the connections... they're hard to key, require different kinds of wire(thin for signal, heavy for power), and besides, mixing power and signal feels like tempting fate. I've fried plenty of electronics already. So I split the power and serial signals.
For power, I'm going to use standard PC power supply MOLEX connectors. I've scrounged up an ancient AT power supply -- no hacking at all to get it running standalone since it's not ATX. Signal was a little more tricky, but it came to me while staring at a pile of dead modems -- I could just use their phone jacks and standard 2-conductor phone cable. Easy to get, come in pairs, cheap, keyed, with cables easily made or available premade anywhere. The picture depicts the signal board plugged into itself, echoing serial signals back to the computer.
Here's a tip for removing through-hole components from scrap circuit boards -- if you don't care if the board survives, don't bother with desoldering tools, use a blowtorch. Clamp the board to something metal, melt the solder, and rip the component right out with pliers. It's much more reliable than trying to suck or wick the solder out. But do it someplace with good ventilation, this can be smoky.
As for shiny things, I've picked up ten of these puppies from ebay -- 1.6A 5V 1.8deg/step stepper-driven linear actuators with 300lbs of force. These should have much less problem driving my syringes than my silly tiny steppers driving their homemade off-center linear threads.
Behrokh Khoshnevis' Building Machine
Monday, January 22, 2007
It looks like the laser is mounted on the X axis and the head just contains a mirror to redirect the beam.
(link to flickr album)
Sunday, January 21, 2007
A WinXP step
Although I like Linux and could function nicely there, I'm in the Windows world. Soooo.. I managed to get the RepRap program working on my XP box today. After I finally figured it out that I must be having a problem with the drivers. Couldn't get Java3d working without the drivers from Nvidia. Probably should convince people by having them try samples from j3d. Even that wasn't enough to make me realize what the problem was. Had to find a post stating this was my problem. Had to re-enable the acceleration too. The default XP drivers from Microsoft will not work. Next on to the fun stuff, Pic programming and the comm and controller boards :)
Lego Bricks in Action!
I had to glue down a guide piece; one of the axles still isn't quite straight (but the wobble has transfered to the now only partially attached motor... not sure how/if I'll reattach that, but it's secured on one of the two motor tabs, and it seems to be holding... It worked itself loose after I glued the guide piece down, and that tab slowly moves 'towards/away' a mm or two. This is the 'upper' axis, moving perpendicular to the camera.) My biggest problem with these is just getting a good connection between the motor and the axle... I'm using some Lego Bricks to do it, so they're true to the axle itself, but the motor itself had to be ground a little to fit, and it's not quite solid...)
The Z axis just plain didn't work... It's gone back to the drawing board, but maybe I'll have better luck with the actual pieces in front of me. The issue was that the axle wasn't 'solid' enough, and over the inch and a half or so that it spanned, it had enough give to move away from the rack pieces... I'm trying to come up with a better design; maybe with a guide block on either side of the rack piece to help hold the axle tight against it...
The only major hurtle I have left is figuring out how to get the two little controls on the operator's workstation to move in sync with the axis :)
Saturday, January 20, 2007
Debugging the Mk 2.1...
Mercifully, it still was. The Mk 2 is not exactly an easy device to modify for other uses. I had to move the motor around because I'd tilted the threaded drive rod and then there was, of course, the new extruder barrel to install. That part went fairly well. I had to open up the passage in the PTFE thermal barrier a touch more so that the filament would move easily.
With luck I will be able to get the firmware test program debugged and start extruding some plastic under PC control over the next few evenings.
Friday, January 19, 2007
The advice is good for newbies with marginal electrical experience who are trying to build Tommelise, Darwin and other Reprapper 3D replicator homebrews.
I also carried over the a section I'd written in the RepRap FAQ about sight recognition and schematics of basic electronic components.
Finally, I've made a hard first cut at a tutorial on making the braised copper tubing extruder barrels for the Mk II.
I've tried to keep it all short and simple to avoid overwhelming newcomers. For those of you who really know their electronics please let me know if I've made any mistakes or, paraphrasing Einstein who once put it so well, written it so that it is a bit simpler than possible.
Thursday, January 18, 2007
2 amp extruder barrel complete
I wondered why I was able to do the winding so well and finally remembered that when I was about 4 years old my father, an avid fisherman, taught me how to set guide loops on new fishing rods with thread and then set the thread windings with fingernail polish.
I've loaded most of the heating capacity up by the orifice and only a little further back. I also made the extruder barrel a bit longer. The theory is that the rear end of the extruder barrel will stay cooler and not cause the filament to melt at the entry to the barrel, something I'd seen happening when I tried to extrude polycapralactone at too high a temperature.
This evening or this weekend I plan to trim the threaded rod for the polymer pump and then assemble the Mk 2.1 for testing.
Wednesday, January 17, 2007
Making the 2 amp extruder barrel...
After I discovered that I couldn't braise brass without a Oxy/Mapp rig I went back to an old design notion that I had. That was to make the barrel, cap and a support flange of copper and thread it through the brass support plates. That worked just fine. It's not as elegant, but you can do it without a lot of drama.
The first thing I did was to use the tap set on the end of a thread die to dimple the copper cap for the extruder barrel. This was not something I did with the 1 amp test extruder barrel. I went ahead and drilled it for 0.5 mm as well. That went very smoothly.
It was also a total waste of time.
I set up the cap for braising just as I did with the test unit.
It braised beautifully. The copper/phosphorus braising material flowed into the extruder orifice, however. What's worse, the copper/phosphorus alloy is a lot harder than the straight copper so the plug was the devil's own job to drill out.
I pulled out my trusty #75 wire gauge drill bit and pecked at it for the better part of a half hour without making a lot of progress. I was getting tired and shaky what with the second day after surgery always being worse than the first, so I got the bright idea of using the broken #76 drill bit that had broken off with one turn of drill fluting left when I broke through the .5 mm cap on the test extruder. That turned out to have been a brilliant idea. The broken bit had hardened at the break and being shorter was a lot stiffer than a regular bit. It cut through the copper/phosphorus plug like it was butter in about 15 seconds.
The lesson is that if there is any fluting at all left on a wire gauge drill bit when it breaks don't throw it away. It can be very useful mostly because if you break it again you haven't lost anything.
After that it was a quick matter to slide the brass mounting plate over the extruder rod followed by the 0.5 mm copper support flange and braise that as well.
With that done it was a quick matter to braise the back copper support flange to a short length of 3 mm ID copper tube to act as a guide out of the polymer pump and into the PTFE thermal break. WARNING! If you use your vise and let the flange rest on the iron vise the flange will not heat properly and the braising rod won't melt or adhere properly to form a solid joint. Always leave some distance between what you are braising and a good heat sink like a cold vise.
It's not elegant but it works.
Now the extruder barrel and brass mounting flange are ready for trimming, painting with BBQ paint and wrapping with #32 nichrome wire. That's for another day, however.
One big word of warning. You see me braising things on a vise clamped to my worktable. Let me warn you that you NEVER, EVER point the butane torch at the worktable. I always point it parallel to the table where neither the flame nor the wash of superheated air hits the table. I also keep a fire extinguisher within easy reach in case I make a mistake. PAY ATTENTION TO THIS WARNING!
Tuesday, January 16, 2007
Failures and Setbacks
Tonight my goal was to attach the threaded rod to the stage. It only sortof worked. I realized there were some fairly major flaws in my design. However, none of them are showstoppers (i hope), and also the improvements will make the design much easier for others to do.
Securing the skate bearings / drive nut to the respective places requires epoxy and is a royal PITA. I dont feel entirely sure about the epoxy's performance (especially long term) and it is also quite messy and problematic to apply it (game over if it gets on the threads...)
instead of using epoxy to attach the nuts / bearings to the frame, i'll use a mounting bracket that goes over them and is tightened down until everything is snug and immobile. i attempted to do this with some wood, but it broke in the middle because of too much pressure. i'm not sure... but i think if i use a screw driver instead of a power drill that problem wont be too difficult. on the other hand, this would be a GREAT use for thermoplast, and they can even be hand molded very easily.
Mounting the threaded rod on to the skate wheel bearings is tough. The approach I took was to use bolts to attach a smaller screw to the bearing. That works really well and really awesome. However, the problem is that this screw then needs to be attached to the threaded rod. I tried using epoxy, however this has some problems. As said before, its messy. Secondly, its fairly tough to get the bolt to be exactly center. Thirdly, epoxy + threads = bad. Fourthly, it has to dry and is prone to drooping unless the rod is vertical, which leads to the rod falling over and throwing the bolt off (obviously that is bad).
I'm not really sure. I'm going to try multiple epoxy applications. First i'll just concentrate on attaching the bolt and making sure it is perfectly aligned / straight with the rod. Then once its dry I'll glob a bunch on to make it really strong. This will probably work. I would really like to find a better solution. The only thing I can think of is either drilling and tapping the rod so that i can bolt the bearings directly on the rod, or milling the end and threading it. Both of those options are fairly difficult and require yet more specialized tools (which i'd like to avoid.)
You guys have any suggestions?
Even though I ran into those problems I just mentioned, I've learned some valuable lessons and tomorrow I'm confident that I'll be able to overcome them and actually get a machine that works well mechanically. When the x axis was actually hooked up it turned pretty easily by hand, and the drill worked well on it (this is what killed it... =) The Y axis is shaping up nicely, and I'll probably lay the groundwork for the Z axis soon.
Check out my flickr photo set for more pictures of the work done today.
Upgrading the Mk 2.1 test board...
I've set the board up to feed 12v power into the gearmotor instead of 6v. That will make it run faster, but I will also get a lot more torque using the 12v Solarbotics motors and the use of shaft encoder monitored pseudostepping will let me run the motor at extremely low speeds will full torque.
I'm still not feeling tip-top so I'm not going to try doing any copper braising tonight.
Monday, January 15, 2007
will it cut?
i bought this acme threaded rod for the drive thread, but it came in 3' lengths. how do i cut it? aka what is the cheapest / least dangerous way. i'm pretty sure its steel.
i do have a hacksaw... would that be possible?
Well, to do brass to brass I need either nickle/silver or brass braising rods. Problem is that you need a flame that exceeds 1650 degrees F. to use those. In theory butane will do that. In practice, not.
That means that in order to do brass to brass braising I'd have to shell out $50 for a Oxy/MAPP braising rig. I'm not ready to do that just yet, so I will do what I was planning to do before I got the idea that I could do brass-to-brass braising, which is to use 3 mm ID copper tubing and an end flange and extruder tip of copper built in a brass flange. That's not elegant, but it should do the trick.
Sunday, January 14, 2007
I use Windows XP and noticed that Eclipse uses CVS instead of Subversion. Maybe that's switchable somehow. Anybody know?
I do have a problem with AOI. It displays the splash screen and then nothing else, although the javaw is still running. I found many of them. I'll look through the faq and google and see if I can figure it out but I'm hoping someone here already knows about this. I think I saw something about drivers and this problem but I don't remember where. There is so much to see and read out there.
I was hoping for a compartmentalized box, with the pieces randomly segregated into their own little packages; I didn't even get any Lego bags!
Still, it was just that much quicker before they pile had formed!
And the XY axis was quick to follow! I did most of this by memory/picture, and had it together in 10 minutes. I'd already prepped the motors while waiting for the shipment to arrive so they just popped right on.
I've already found a weakness in this design; the black 1x2 that the axle goes through, along with the two 1x2 red flats below it, can work loose and derail the pinion gear. I've tried to push it down as hard as I can, and it seems to hold, but I suspect it needs to be glued or something to ensure it doesn't pop off during use.
A little electrical tape securely holds it on for the time being. I'll replace it with hot glue once I'm ready to secure it all down. I used hot glue on my prototype and it worked perfectly; it also comes off cleanly with a fair amount of prying.
Aside from the guide brick coming loose, the XY axis moves back and forth nice and smooth. Once I decide how I'm attaching the limiter switches (and decide which of the two I've got to use), and glue stuff down, these axis should be done!
While the XY axis went together easily, the Z axis was a bit more of a problem. I decided it was too tall, with too much wobble potential, so I moved it below the XY axis, instead of on top.
The general setup and idea still seems sound, but there was a flaw in the design, well two. They all center around the rack gearing that runs down the inside of the inner piece. The motor mounts on the outside frame, through a guide, and meshes with the rack on the inner frame, pushing it up and down. I made some miscalculations on how the transition pieces (black pieces on the inside, with the nubs sticking towards the center) line up with the bottom of the white piece (the one with the rack attached to it). I did a little rearranging and was able to get the pieces to line up, but I only used two of the four transition pieces. In addition to that, the white piece is too thick and sticks the rack out beyond where the pinion gear sits. I'll need to use some flat pieces to line them up if I want more than two.
I've hooked up the motor to the frame (without a gear, and ran it on 'calibrate' for a few hours, to dry some epoxy I used to attach the axle to the motor), but I still need to replace the white piece to get the pinion gear to line up with the rack. The design still looks fine, but it needs some correction.
I'll be posting both the html directions (though they need work, and I'll hopefuly get a second or third version updated over the next couple weeks) and the Lego designer file, once I've corrected the issues with the Z axis, in the next couple of days. From what I've discerned from Lego, I can share this stuff freely, so long as I don't make any money off it, and I call everything Lego Bricks.
I reinstalled my computer around new years, and I don't have my video software up and running yet, so it'll be a couple days before I get video up. I should also have the Z axis sorted out by then, so I should have all three running!
I've still not gotten the IO working on the host software in the svn, so I'm do all this with the 'windows release' version from last year. I've been a little distracted making room for two new recliners that are arriving Saturday, but I hope to get back to that next week some time.
Parts fabricated for Mk II 2 amp extruder barrel...
As you see the build of the 2 amp extruder barrel was close to what I had visualised with AoI. I extended the brass flange on the barrel and replaced the ABS clamp that Adrian supplied with another brass mounting bracket. I cut the PTFE to a 1 inch cube and kept the mounting bolts outside the block. That simplified the fabrication of the PTFE block which was already a bit of a finicky operation.
I ran short of compression springs. I'll probably run out to Orchard in the morning to pick up some more.
I'm going in for dental surgery tomorrow. I'm told there is a bone graft in the offing to repair a jaw injury sustained in my teens before I discovered that you don't give a psychotic the chance to get in the first punch. The point is that I may be a few days offline, or not, depending on how bad the surgery is.
Saturday, January 13, 2007
Modified Mk II polymer pump working...
I sawed a slot in a 1/4 inch coupling nut to turn it into quite a nice little gearmotor coupling. You can see the result...
I was able to get a stainless steel 1/4-20 long bolt to serve as the pump. I haven't sawed it off yet and won't until I get the whole assembly working together properly.
Hooking up a 6 v Solarbotics GM8 gearmotor to the assembly I was able to pump both HPP and HDPE without problems. I expect that I will have to sand off a half millimeter or so from the filament side of the pump to insure a snug fit with the welding rod.
I will be finishing the pump in the morning and building up the 2 amp heated extruder barrel. I will also be installing a dual darlington chip (754410) on the extruder test board so that I can run the pump as an ensemble.
Friday, January 12, 2007
X axis shapes up
unfortunately, my power drill broke, so i decided to wait until i can get a new one to proceed. in the meantime, i cut some boards for mounting when the time comes. it did however, give me a chance to mull over the process so far. here are a few things that have struck me:
this would be alot easier if there was a standard plan. measurements, lists, etc. i understand this is repstrap which basically means forging your own way, but once this is over, i would love to draft out improved plans for creating this device. that would make things so much easier.
1. look at the plans, cut all your boards to length in one go
2. drill holes at correct places
3. assemble pieces according to order.
i would love it if someday this thing were as easy to put together as ikea flatpack furniture. that would assuredly rock.
When I get the stuff, I'll try it out on various surfaces. Curious to see what materials is adheres to. I'll also try it with solder, on the off chance that it actually takes a solder bead. I've also got some Cerrobend sitting around. I'll see how the two interact. I'm also interested in what the practical limits are on trace width.
Any other suggestions, thoughts?
Thursday, January 11, 2007
Vik's polymer pump configuration works...
I again used pliers to rotate the threaded rod. This configuration seems to require less torque that the previous experiment.
I am observing the tendency of this polymer pump to slowly rotate the 3 mm filament. It appears to complete a rotation every 1-1/2 inches. This is the same slippage that Vik Olivier has documented with his Mk II.
Most of the energy used in this pump seems to be employed cutting the threads into the filament. I tried to attach my electric screwdriver to the threaded rod with little success. My electric drill, on the other hand, proved far more of a success. Here you can see polypropylene filament being pumped. Please pardon the blurriness. I had to operate the drill with my right hand and the digital camera with my left. Sadly, my son has gone off to university again, so I'm down to two hands again.
Cutting some corners on the Mk II
My first notion was just to rip Jim's idea, viz, replicating the Mk II parts in wood, directly. After thinking about it for a while I wondered if all the design ideas that Adrian put into the Mk II were really essential or if there would be some ways to simplify the design. While the Mk II has many fine qualities it's not particularly simple for a retard like me with two hands full of thumbs to replicate.
When my lovely 3 mm cobalt drill bits arrived today I took off a little while and did some serious experimentation. Going back to Vik Olivier's original idea, viz,
I wondered about several of its features. The most noticeable is that it has only a very short run of threads doing the polymer pumping. Dr. Bowyer's design, on the other hand, has about 40 mm of threads that the filament has to get past. Was that really necessary?
As well, Vik and Adrian both seem to have the idea that you need a pressure fit between the two halves of the pump. Why? I could understand it if you were expecting wide variations in the diameter of the filament. We solved that problem some months ago, however.
Here you can see my dumb little experiment. I took two scraps of 3/4 inch poplar. I drilled slightly overlapping 3 mm and 1/4 inch holes in one of them. The 3 mm hole is for the filament and the 1/4 hole is for the threaded drive. The second piece of poplar scrap covers part of the 1/4 inch hole so that the threaded rod actually pumps the 3 mm filament instead of just crawling out of the other side of the board.
Seating the 1/4 inch threaded rod and threading in the 3 mm polypropylene I rotated the threaded rod with a pair of pliers.
As you can see in this closeup the threaded rod bit into the filament and pumped it right through the poplar board.
Keep in mind that I am applying torque to the rod from the bottom of the metaphorical polymer pump instead of the top, so the second scrap of poplar absorbs the thrust at the top of the pump. I will need to apply torque to the 1/4 threaded rod at the top so I will have to adopt a scheme somewhat like Vik did in his drawing where he used a lock nut seated against a thrust collar at the top of his pump. Doing that there should be no need for any machining of the bottom of the threaded rod. It can simply seat against the 1/8 inch brass or aluminum plate that I will have between the poplar pump and the PTFE thermal barrier.
There you have it. No springs holding two halves of the pump together. No machining of the threaded rod. I might use a file to make some flats on the rod where it meets the gearmotor. I'll probably use some fish tank tubing to connect the gearmotor and the threaded rod, though I might simply cook up one of those coupling nut connectors that I documented some time ago.
It might not work, but it won't hurt to give it a try.
Tuesday, January 09, 2007
Over the last few days I've read very intently pretty much every single page in the reprap wiki, most of them more than once... and I decided to start building a reprap/strap. I'd like to blog the build process for the benefit of others! I started 3 days ago, and I now have: * the "stripboard" comms controller fully working* my first (X) axis unipolar stepper controller built and working (buggy, but the stepper turns some of the time!) - also made as a "stripboard version".* all my stepper motors, and about a dozen spares of assorted sizes(been collecting them for a while)* a lot of old printer hardware (rails, belts, etc) that has been stripped down in preparation for "the build".* all the software installed and running in an existing linux server I also use for my http://www.mythtv.org/ tv server.
Here's a few pics:
(small square board is my max232 based comms board)
(retangular board is the X -axis stepper board)
(a pic of the two of them wired to a small stepper)
(some of my gutted printer rails with beefy steppers - I have 3 of these!, two shown)
Monday, January 08, 2007
Oh yeah, also got my roller skate bearings and my stepper motors =)
Anyway, heres what she looks like:
More here: http://www.flickr.com/photos/29642607@N00/tags/reprap/
Sunday, January 07, 2007
The extrusion is taking place using my low thermal inertia extruder design with a 0.5 mm orifice and producing a 0.8 mm polymer thread. It is being run at 12 v and 0.95 amps. I was able to measure the temperature of the copper collar braised to the back of the extruder at 90 degrees Celsius during this exercise. I can make the HPP extrude while the collar is as cool as 70 degrees. I am hand feeding 3 mm HPP filament using pliers to grip the filament. The filament is a bit slick to hold onto with fingertips alone.
You can see a video of the extrusion here. In the video clip you can see the extrusion oscillating laterally at times. This occurs when I put too much pressure on the HPP filament. When I stop putting the little pressure that I am on the filament (estimated at about half a kilogram) extrusion stops immediately. I have yet to detect any dribbling whatsoever with HPP.
HPP is a robust engineering plastic. It handles bending and fatigue extremely well. A short video of it's behaviour demonstrates this characteristic.
Saturday, January 06, 2007
The last two pieces of the puzzle...
I haven't attempted to make them simply because Adrian was kind enough to forward me a set of Mk II parts early this year.
Jim Wilkins, a RepRap builder faced this problem back in September of last year and solved it very straightforwardly, however.
Jim simply took the dimensions of the Mk II and made them out of a mixture of polycapralactone, medium density fiberboard and thin plywood. The result is brilliant.
Friday, January 05, 2007
Mk 2.1 extruder
It works rather well, I think.
I've basically thrown away Adrian's ABS bottom plate, which he calls the "Clamp" because it clamps onto the PTFE rod that provides the thermal barrier between the heated extrusion barrel and the ABS body of the polymer pump of the Mk II.
I've replaced the "Clamp" with the new extruder barrel which is fabricated out of 5/32 inch hard brass tubing and braised onto a 25x40 mm support plate (all in yellow). A 25x25x40 block of PTFE (pink) fits over that and is backed with another 25x40 mm support plate (also in yellow) that bolts directly onto the bottom of the Mk II polymer pump using the existing mounting holes.
This reduces the width of the overall Mk II assembly from 76 mm to 40 mm. That should, if I want, mount two of them side-by-side on the xz vertical positioning plane on Tommelise.