Mark Feldman's Home Page Blog

Second complete set of RepRap parts

Last weekend I switched back to red plastic and printed another set of parts for a mini-Prusa. This set is considerably better quality so I've decided to build my second RepRap using these instead:

Earlier in the week I assembled the basic frame, the following photo of it sitting next to its parent shows how much smaller it is:

Also on the table in that photo are new pieces for a Mendel x-carriage which I printed last night, I've been experiencing a number of problems with the Mendel including bad back-lash that I'm having difficulties getting rid of. I suspect the culprit is the original x-carriage pieces which are heavy, inflexible and somewhat warped. Hopefully these replacements will go a way towards continuing to improve the quality of my prints.

I also spent some time this week packing mini-Prusa pieces onto 150x150mm plates i.e. the size of its own print area. Turns out you only need 2 plates to print all pieces for both a mini-Prusa and a NEMA 14 extruder:

What I'm left with now is a complete set of teal-coloured RP parts, plus numerous spares in red and green. It's not the best quality set in the world, so I wouldn't feel comfortable about trying to sell them, but functionally they should work fine. If anyone has any suggestions as to what I should do with them then I'm all ears.

First complete set of RepRap parts

They're not the prettiest pieces in the world but my Mendel has achieved a new milestone by printing a complete set of parts for a whole new machine:

The target machine is a much smaller mini-Prusa, but it's a complete set none-the-less. Unfortunately the teal plastic I used turned out to be a poor choice and the quality of the prints suffered greatly as a result; I suspect the coloring agent wasn't mixed with the original PLA properly because the extruder kept spitting out large globs of teal-colored liquid that dried to a crusty, flaky mess. In any case with a bit of clean up they should be good enough to use, and if any aren't then I can always print out replacements for them. The only parts I wasn't able to print were the two pulleys; the quality just wasn't good enough, so I picked out a few earlier prints that I did in green.

Studding is cut, nuts and bolts have all been built, and bearings are on their way from China. I'm ready to begin...

Heat bed woes

This past weekend was another RepRap Users Group meeting, there was a great turn-out this time round and I got to meet a lot of new faces. I had my 4-month old son with me (get him into RepRap early, I say) and no car, so I wasn't able to take my machine along, but it was a good chance to see so many other people's builds.

With regards to my own build, my prints this week began exhibiting warping along the bottom few layers and "wavy" edges, as shown in the following photo:

To cut a long story short it turns out both my heat bed and extruder were running about 20 degrees too hot, and I was also extruding plastic about 40% too fast. The nozzle was basically pushing plastic down into the lower layers and causing them to bulge out. The following shows prints of the same piece before and after calibration:

In other news I've started work on my second RepRap, this time I'm making a "Mini Prusa". Prusa is a simplified version of the Mendle I built, and local RepRapper Auzze has designed a 25% scaled-down version which I saw over the weekend and instantly decided I had to have. Rod, studdings, nuts and bolts were purchased this morning, printing of the RP parts begins tomorrow. My obsession with RepRap has entered a new phase...

New extruder and heat bed

This week was spent outfitting the printer with the new extruder. The Herringbone extruder was quite a finicky device to build in the end. It's considerably more difficult to mount than the other two extruders I built, but the results have been well worth it by way of better quality prints:

(And yes, it works!)

I also finished the heat bed and mounted that to the Y axis:

Now that I have a heat-bed I've had another attempt at printing ABS plastic which is supposed to be stronger than PLA. Unfortunately results haven't been promising so far. In the following photo the first piece is one I printed a few months ago in PLA, the next two are versions I printed this week in PLA (the shiny side was in contact with the heatbed glass) and the last piece was one I also printed this week in ABS:

I haven't calibrated the RepRap for ABS yet but even so it seems to extrude "blobby", has more trouble sticking to the bed and just doesn't produce good results. Of more concern is the fact that it jammed my extruder....again! ABS is normally quite soft but when it's heated and cools down it seems to go very hard. This seems to have happened in the PTFE liner in my heater nozzle, and since the drive bolt couldn't push the filement through it proceeded to tear it to shreds:

For the time being I think I'll stick to PLA and maybe give ABS another go with my next machine. My next task is to finish off calibration and then start work on printing the parts for a whole new RepRap. Exactly what I do with those parts, I haven't yet decided...

Printer upgrades

The past two weeks have been spent making some upgrades for my printer. First up, I've finished the Herringbone extruder (designed by rys-jones). I haven't had a chance to test it yet, that'll happen over the next couple of days, but so far it's looking good. It incorporates a number of things from other extruders including prime-numbered herringbone gears, support for an Adrian's hot-end (and also the J-Head which I'm using) and the Wades hobbed drive bolt.

I also made an "Open X" carriage (designed by Buback) which unfortunately turned out to be incompatible with the Herringbone extruder. It's a great design though so I'm planning to modify it and see if I can get the two to fit together:

Finally, I've done some initial tests of my heat-bed, which was designed thanks to some tips from Auzze. Using only a top sheet of aluminium I'm able to maintain 75 degrees C, certainly hot enough to print PLA and hot enough to print ABS according to some. I'm planning to attach a single-sided PCB to the bottom, the fiberglass should provide insulation and help bump the temperature up to over 100 degrees which I can then program my Repic electronics to control with a thermistor:

Printing again

After a month or so break my printer is up and running again with a new J-Head nozzle, which so far has been behaving extremely well; my first task has been to print off the parts for a new Herringbone extruder. The extruder I've been using for the past couple of months was an experimental one printed from my original Adrian's geared extruder and like many other experimental technologies its had its fair share of problems. It's done its job well so far though, and with some calibration has been printing its replacement parts well:

I'm also in the process of building a heat-bed, the design of which I've stolen from Auzze. If it works well then it should allow me to start printing ABS and should also reduce some minor warping problems I've seen with PLA:

One other thing I've been quietly experimenting with in my spare time has been trying to make a working extruder hot-end using parts sourced only from local retailers and requiring no specialized skills or equipment other than a hand-held cordless drill. The motivation for this has been driven partly by my recent hot-end failure and a corresponding determination to reduce my dependency on the relatively small number of people that sell these parts online, but it's also a first step into an area that I'm hoping to experiment more with in future.

Quality slowing improving

Contrary to popular belief, maintaining a RepRap hobby is in fact possible after the birth of a child provided you're good at balancing the whole "sleep-time vs hobby-time" equation. The number of hours I've been able to put into my build has been substantially reduced but I've plodded along and made particularly good progress on the software side of things.

The past six weeks culminated in last weekends RepRap User's Group get-together, hosted again this time by Cefiar. The night before my extruder started leaking plastic (more on that in a moment) but I was able to get it fixed in time and managed to pull off some small prints the next day.

Once again I came away with a lot of newly acquired knowledge and tips on how to improve my builds which I put to work immediately. Within just 24 hours my print quality had improved considerably, here's a before-and-after:

Part of the improvement was due to switching over to Skeinforge to generate my gcodes, until now I've been using the official RepRap software. I still have a lot of tweaks to make but I definitely feel that I'm back on track and gradually increasing the quality of my builds.

And then, just as everything was going great, disaster struck again. The leaking plastic that I experienced the night before the RUG get-together was due to a failed seal between where the PTFE insulator screws into the brass nozzle. Over the past few days I'd noticed an occasional dark blob in my extrudes which could have only been the tape used to maintain that seal, and that in turn suggested that the PTFE had somehow been compromised. Sure enough, plastic started pouring out the seal again tonight, and when I disasembled the hot end this time the whole thing fell apart, even though I did it at full extrusion temperature:

Still scratching my head as to exactly what's happened but I suspect that the PTFE part is simply more fragile than I realized and was damaged during the 2 or 3 times that I had to disassemble it due to over-heating and extruder jams.

Unfortunately this means my RepRap is out of action until I can order in a replacement, but again it's all part of the learning experience.

Wades extruder finished.

It's a great feeling when you use your 3D printer to print a replacement part for itself, it's an even better feeling when it prints itself an upgrade that works much better than the part it's replacing. My Wades extruder is now finished and working thanks to some help from local RepRapper Auzze. While moving over my hot end I accidentally broke a wire to the thermistor, not surprising really as I hadn't done a very good job of securing it (I have this time round with some Kapton tape). RS Components Online have spent the past 2 weeks dicking me around (despite their web site promising a 6-day delivery) so Auzze was able to provide me with a thermistor at very short notice. He also provided me with with a few different colors of PLA, and I managed to get some black and white ABS from a local supplier so I have plenty of plastic to play with now. This isn't a very good photo because the print in question failed due to a minor problem I won't go into here, but the Wades works perfectly and has been making much better quality prints even though I haven't had the chance to calibrate it properly yet:

In other news, my partner and I are happy to announce the birth of our first child yesterday. I mention this here only because it was largely due to this fact that I started working on a RepRap in the first place. When Heidi and I discovered we were expecting I immediately lost my favorite drinking buddy and I also realized that I needed a hobby I could work on at home in the evenings as my partner gradually assumed a semi-reclining position on the living room couch in front of the TV. My goal was to get my RepRap finished by the time he was born; I completed that task three weeks early and also managed to print and assemble a working Wades extruder in the nick of time.

Little Ascher was born yesterday at 12:51pm. Mother and child are both doing well and looking forward to coming home tomorrow.

A Toast To Mendel

It's been a fun couple of weeks playing with my new toy. My attempt to print a Wades extruder revealed a few minor problems with my firmware, but they've been corrected now and I'm very close to declaring my Repic electronics out of alpha and officially at the "working" stage. My modified Wade's extruder is half-built and I can already tell it's going to work a lot better than the one I've been using so far:

One of the customs in the RepRap community is to toast the machine with a mini-mug that it has printed itself. Personally I regard this step as the defining moment at which you can honestly say you've successfully built a RepRap, so I caught it on film and also used the moment to show some of the pieces I've printed so far:

One more video, this one is of my Mendel printing the main block for the modified Wade's extruder in the photo at the start of this blog entry:

My Mendel prints its first replacement piece.

Been spending the past few nights printing various test pieces, most of which has been reasonably successful. An attempt at printing the large gear for a Wade's extruder started well but then the print board started slipping ever so slightly so I had to cancel the print:

Looks like I'll have to find a way to secure the board to the underlying bed a bit better. A glass-topped heat-bed is looking pretty good about now.

One notable success was the printing of my first replacement piece, which has now been installed on the extruder. It's a fairly important piece which holds the bearing that presses the plastic rod against the drive shaft. My original piece (white below) was warped and bent, and the holes for the nuts hadn't been cast properly so they just spun around in the depressions that I'd drilled out for them making it very difficult to adjust the tension in the extruder. The new piece accommodates them perfectly:

Accuracy improves.

After some tests and adjustments I've managed to print out a replacement drive gear for my extruder:

My extruder currently has a bit of a problem stopping the flow of plastic, so I did have to clear out the middle with a drill bit, but other than that it didn't need much clean up. For fun I tested it on the extruder...it worked perfectly! The piece is 2cm across so I estimate I'm currently achieving a resolution of roughly 1/3rd of a mm. Not bad given that it's barely 2 days old and running custom hardware and software.

There are many things I can try to improve the resolution further, for starters I'm still homing each axis at full speed into the micro-switches. The y axis needs tightening and adjusting, I haven't yet implemented 5D extrusion. And I still need to experiment with extrusion speeds and temperatures to find the sweet spots. I have all the parts I need to build a Wade's extruder, and my RepRap is probably accurate enough now to make the printed pieces for it, so I might put it to work on that next.

First successful print!

It's not perfect, and I still have lots of fine-tuning and problem-solving to do, but tonight my Mendel made its first successful print:

The piece is the gear used to drive Mendel's Z axis, with a little clean-up it would probably be good enough for use in a real RepRap.

The quality of the print is nowhere near what I'm ultimately aiming for but this is a huge leap forward regardless. With a working printer it's a lot easier to gauge the effectiveness of modifications I make in future.

Next goal: get it to the point where it can make a water-proof mini-mug for the customary toast.

Extruder cleared

Another lesson learned, this is what gets left in your extruder nozzle if you allow the temperature to get away from you (note the bits of wire that broke off when I was originally trying to clear it):

I've managed to clean out the extruder, oven cleaner was (unsurprisingly) very good at that. I also turned the heater on and had a play with driving plastic through by hand, so I've got a pretty good idea now of just how much force is required for successful extrusion.

Extruder jam!

This is frustrating beyond belief. After getting so tantalizingly close to a working printer the extruder has decided to jam. I completely disassembled it but the jam seems to be in the brass nozzle itself; everything up to that point looks clean as a whistle. The jam happened when I raised the nozzle temp over 250 degrees, I'm guessing that some plastic burned and hardened. Everything I shove in there to try to clear the blockage breaks off and makes the jam worse. Fortunately I can stick a small piece of wire through the opening easily, which means the blockage must be in the main cavity. I've jammed some plastic in there with the nozzle at 250 degrees in the hope that I can manually pull the jam out once it's cooled, if that doesn't work then I'll have to hit it with a drill bit.

For my next RepRap I think I'll try out Wade's extruder instead....

First attempt at controlled printing

Had a very busy long weekend and managed to finish the firmware and host software needed for printing gcode files. One of the outcomes of this weekend was that I removed the DLL containing the official firmware. A bug in Microsoft's deployment code results in problems if you try to link a C++/STL DLL to a C# WPF application. As it turns out, the only thing I was really using the official firmware for was gcode parsing, and since there are only a dozen or so codes actually used by RepRap it only took an hour or so to replace it with a few dozen lines of C# Regex code.

In theory this should have meant that my RepRap was ready to print, but my first attempt wasn't as good as I was hoping:

I stopped the print after the first layer. The lower-right half of the piece printed first and it was pretty obvious that there was a problem with the extruder. For the upper-left half of the piece I was manually pushing the plastic rod through the extruder, so it would seem that the extruder isn't gripping the plastic rod anywhere near tight enough. It might be because I'm not using the proper metal insert, or it might be because the springs are too weak, or possibly the mechanism hasn't been tightened enough.

On the plus side, it looks like the extruder was physically moving around correctly, which is an encouraging indication that my custom Repic electronics is working correctly and therefore a cheap and viable alternative.

First "print"

Another productive day. Thanks to some help from others RepRapers (particularly NopHead...again...) I was able to troubleshoot my extruder problems; basically it was a combination of the nozzle temperature being too hot and trying to extrude too fast into air. Once fixed I did a quick test onto some painters tape:

Again, very happy with the results. The blobs at the corners are simply due to the fact that I was using the controlling software to manually drive the XYZ axis and the extruder stopped at the corners without putting the extruder into reverse to stop the flow of plastic.

Today I also developed the code to store and retrieve parameters in the motherboard's EEPROM. This means that any adjustments that need to be made (e.g. nozzle temperature, rate of extrusion etc) are stored on the RepRap device itself and can be changed by the user in the host software running on the PC. This is an important step needed for the next and final stage: getting the official firmware code to send its data over the USB link for printing. Most people wouldn't need to do this last stage, but I need to because of the custom electronics I'm using.

First plastic extrusion

It works! Or at least I think it does. This morning I allowed the nozzle temperature to go to 250 degrees and modified the firmware to allow for control of the extruder stepper. Here's a photo of the results, both ABS (white) and PLA (blue) plastic extruded fine:

I do seem to be having a few minor problems though, so I'll have to jump onto the RepRap forum this afternoon and get some advice. The PLA seems to extrude a bit better than the ABS but I suspect that's due to the fact that it's less slippery and more rigid which results in greater pressure being exerted by the drive mechanism. What I am a bit concerned about is the fact that the plastic has a tendency to bunch up around the nozzle:

It also keeps running for a few seconds after I halt the stepper. I suspect both problems might be due to incorrect extrusion temperature, but I'll have to see what the experts say.

This post wouldn't be complete, of course, without the obligatory video of plastic extrusion in action:

Extruder nozzle testing

Finished wiring up the power this week which means my Mendel is now physically complete. All work remaining now is purely software, apart from the usual on-going physical adjustments.

Tonight I performed my first test of the extruder nozzle which basically involved testing that the micro-controller was able to accurately control the nozzle temperature and relay temperature measurements back to the controlling software (nozzle temperature needs to be controlled by the board as a safety feature in case the USB link is broken). Everything worked exactly as intended and I filmed the test with my IPhone:

I forgot to zoom in on the extruder LED blinking in sync with the heating element so I'll show that in my next video when I'll hopefully be able to demonstrate the extruder actually extruding plastic filament.

Extruder wiring complete

My build got a bit of a boost last week thanks to the local RepRap Users Group BBQ. Many thanks to Cefiar who hosted the event and sorted me out with some PLA and ABS plastic, Auzze who has 2 working RepRaps and was a gold-mine of practical information, and Scorpio with his hardware knowledge and formidable-looking RepScrap. It was an extremely educational afternoon and also lots of fun! Unfortunately I didn't get to see any RepRaps actually printing, but the ones that were there gave me a lot of ideas for the future.

As for my current build, Cefiar also gave me the much-coveted high-temperature Kapton tape needed to finish my extruder. I finished mechanical construction during the week and wired it up this weekend:

Probably a bit hard to see exactly how this thing works but basically the motor (that big square thing on the left) drives the plastic (that little white thing poking out the top) down and into the nozzle at the bottom. The black wires go to the heater element which heats the nozzle to around 250 degrees C. The white wires go to an embedded thermistor that the micro-controller uses to monitor the temperature. Here's a photo that shows the extruder mounted on the Mendel and ready for operation:

The extruder is now complete, and I've also installed the motherboard's power transistors needed to drive the extruder heating element and the heat bed. Technically I now have a mechanically operational Mendel:

There's still a bit more cosmetic work to do, the power needs to be hooked up and I'd rather that look neat and tidy with a proper power connector rather than wires running all over the place. Once that's done I'll be ready to start testing the operation of the extruder itself. Mechanically the device seems to be able to easily drive plastic, and multi-meter measurements indicate everything should be fine and ready to go.

Extruder contruction

Spent the past couple of weeks getting all the last little bits and pieces I need and putting together the extruder. I haven't wired it up yet, but mechanically it's pretty-much ready-to-go:

The only thing needed now is the high-temperature Kapton tape which I'll be ordering in this week. The one piece I didn't get for the extruder was the metal gear that is the heart of the device and drives the plastic rod through it. I've take a few ideas from elsewhere and modified them to create a replacement i.e. the curved black piece in the photo below. Don't know if it'll be quite up to the job, but finding out is all part of the fun.

Melbourne RepRappers are having a BBQ get-together next weekend, anyone interested in the scene should head on over to the Melbourne RUG page for details.

Extruder nozzle pieces arrive in the mail.

After a 1-month break in my build I'm back on track with the arrival of the pieces needed to build the extruder nozzle i.e. the hot bit on the end that melts the plastic. Many thanks to Brian Reifsnyder on the RepRap forums for getting these parts to me so quickly.

This photo shows the white PTFE insulator, the brass nozzle that the plastic feeds into, the grey PEEK block for bolting everything to the main extruder and the aluminium heater block. Also shown on the right is the resistor that slides inside the heater block and does the actual heating, and a thermistor used to monitor temperature. When assembled it all looks like this (I've left the thermistor out for this photo, there's a small cavity in the block that it gets glued into):

I'm now very close to having everything I need to get the RepRap printing. The only pieces I need now are a small metal gear for the extruder drive axis and a heavy-duty power transistor to drive the heating resistor.

New toys!

Santa has been exceptionally generous this Christmas by giving me a suite of brand new toys to play with including a Dremel rotary tool (with flexible shaft and mini saw), an oscillating tool and a 620W laser guided jig saw with various attachments. Needless to say I'm one very happy kid! :)

I really wish I'd had all this while building the first stages of my RepRap. At least I'll be able to use them on my next model :)

Initial axes testing

The steppers and end-stops are now properly wired up and I added code to the firmware and host software last night to do a test of all three axes. Very happy with the results, everything seems to be operating well with very smooth and quiet movement. Here's the video as promised:

Frame assembly and wiring nearly complete.

The last couple of weeks have been much better. The new belts arrived as promised (thanks Stephen George in Queensland) and I've installed them along with my Repic motherboard:

I'm currently wiring up the motors and micro-switch end-stops, it actually takes a surprising amount of planning and work if you want the end product to be neat and tidy. I've managed to test movement of all 3 axes with belts in place...Y and Z work perfectly but X occasionally slips a bit. I'm not too worried about it though, I'm guessing the belt's probably just a bit loose. Also my power supply isn't ready yet so I was testing with a cheap-n-nasty switch-mode transformer that could barely muster 500mA.

The wiring work will continue this week, hopefully I'll get a video up before the holidays showing the machine in motion. It's also time to order in the final pieces for the extruder and finish integrating the host software with the now almost-fully-assembled hardware.

Ripped Off!

I guess it had to happen sooner or later....I've been ripped off in an online transaction! A month ago I ordered my belts from Dwayne Moody aka "berry-bros" on EBay. A soon as payment was made all communication stopped. No confirmation of receipt (PayPal says it was accepted), no confirmation of shipping, no response to the several emails I've sent from 3 different accounts. A check on EBay reveals this seller has an approval rating of just 80%, with plenty of negative comments on his feedback page.

To be honest I'm really surprised, I've always found Americans in particular to take business transactions quite seriously, no matter how small. And the RepRap community is far too small to be pissing off any individuals with dirty tricks like this.

I've started a PayPal dispute resolution in the hope of getting my money back, in the mean time I've found a local supplier who should be able to get a set to me by the end of this week. Oh, and the power transistors arrived in the mail today, they're the last component missing in my electronics.

Waiting, waiting....

Still waiting for my axis belts to arrive, I ordered them from a seller in the US. No word from him in weeks, and he didn't respond to my last mail either. Hmmm.....

In the mean time I've made another motherboard with a few fixes and done some initial testing of the driver boards and one of the steppers:

Not a whole lot going on at the moment, but at least I've confirmed that I can programatically control the steppers and that the motherboard seems to be up to the job. I've also confirmed that I'm properly micro-stepping the motors; if I adjust the drivers so that current limiting doesn't kick in then they effectively revert to full-step mode, as expected, and make a loud noise when they run (they're practically silent in the video). The only thing I haven't installed yet is the power transistors, I'm waiting on them to arrive in the mail too.

The machine that I'm building is a Mendel and I've been designing these electronics as a Mendel controller, but to be honest I think they're probably better off described as a Huxley controller. Unlike Mendel, Huxley is only meant to support one extruder, so it's probably a better target for all-in-1 electronics such as these.

Working electronics.

Finished constructing the first pass of the electronics last night. I haven't installed the power MOSFETs or the stepper drivers yet but everything else is working fine and I now have a functional USB device that I can communicate with via my host software:

Part of the plan for making this board was to identify any problems and/or changes I need to make, and there are a few minor issues. The super-bright blue LED is far too bright, but I also want to make some wiring changes that will require me to make a new PCB this weekend. Despite this minor setback I'm very happy with how the electronics are coming along so far and I have enough to start developing the final firmware.

Bearing problem fixed, PCB made.

There's been a few developments since my last update, for starters the bearing problem is fixed; turns out it was due to the washers rubbing against the inner seal. The high quality 608 bearings I've been using don't have a recessed inner seal, and since they're metal and don't deform much the washers cause enough friction to jam the z axis:

I fixed it by re-ordering the sequence of nuts and washers on the z axis so that only the M8 nuts are in contact with the bearing. The z axis is now rotating smoothly.

The Pololu driver chips have arrived in the mail, they're needed to convert the clocking signals from the PIC microcontroller into high-current driver signals for the stepper motors:

I now have everything I need for the electronics so I've begun work on that. I made the final design changes to the motherboard over the weekend and fabricated the PCB last night using the iron-on transfer method i.e. laser print the circuit onto an overhead transparency, iron it onto a PCB then etch with ferric chloride. I suspect I didn't get the temperature quite right on the iron so it needed some touch-up work before etching, but the end result was pretty good:

Bed finished, axis motors and x-axis attached.

Looking more and more like a Mendel every day...

There are a few minor problems that need sorting out at the moment (e.g. the extruder plate needs to be done again because I drilled the holes in the wrong place) but the biggest problem is with the z-axis. If I tighten the z-axis bolts in any way the entire axis jams. If I loosen them them they rotate with the axis and eventually tighten and again, jam.

After some experimentation I discovered the problem: if I place a 608 bearing on a spare piece of studding and tighten it between two M8 bolts with washers then the outer ring jams tight. Ergo, there's something wrong with these bearings. This is a bit of a surprise because unlike the 624's I paid extra money for these to get high-quality ones. Looks like I'll have to order some cheap ones in. In the mean time I'll have to start work on the extruder.

Y-Axis Assembly

Attached the y-axis pieces to the frog and mounted it on the frame.

It turns out that nopheads frog design was missing quite a few neccessary holes, maybe it was for an older version? In any case that was an easy problem to fix.

[Edit: nophead tells me this was deliberate, it's explained clearly on his thingiverse page. The holes are left out so you can drill them in the precise locations with the brackets in place and adjusted, which is what I wound up doing anyway the next day. Sure enough, it worked well.]

I'm having a few other problems at the moment though, namely the bolt heads that stick out the bottom of the y-axis RP pieces are hitting the threaded rods across the bottom of the frame. I can fix that by tightening them, but then it's too tight and the y-axis jams. I haven't done a proper tightening and adjustment of all bolts yet so hopefully when I do that problem will go away. If not, there are a few other things I can try to create a bit more clearence in that area.

I am noticing a trend in the problems I've faced so far: x-axis doesn't hold the carriage properly, y-axis bolts hitting the frame, z-axis bars jamming. The axis rails seem to be where problems are most likely to occur!

Lower bed complete

One of the things I'm learning about RepRap is that building it can be quite a challenge...and to remember that the whole process is meant to be fun, even when things go wrong.

The past few days have been the most fun of the build so far...by which I mean I've encountered the most challenges. First of all there appears to be a design flaw with respect to not enough clearance between the RP parts on the y axis and the support studding that runs across the base of Mendel. This clearance is already very small, and if you use 6mm MDF for the bed then the bolt heads actually extend about a millimeter further down. With perfect RP pieces there might just be enough space, but with the pieces I have there isn't. I've switched the frog to 3mm MDF and it's all good now.

As for the frog itself, I wound up making that 3 times. The first version was 6mm MDF and my pieces were a bit too warped to fit it properly. The second was also 6mm and the bolts were hitting the base studding. The final version was made from 3mm MDF (only the lower piece, the rest is still 6mm) and I drilled most of the holes with the RP pieces in place. I also had the axis bars in place and tightened all screws so that everything was perfectly adjusted before drilling the remaining holes in the MDF and bolting everything down. The result is a perfect fitting board that slides smoothly with no unwanted rotational or lateral movement.

Initial Procurement

[Many thanks to Cefiar on the RepRap forums for referring me to the suppliers I used to aquire the bolts, bearings and motors for this project.]

I spent most of today running around town picking up various bits and pieces. Bolts and nuts were bought from CostLess Bolts for http://www.ppl-pilot.com:80/$100.

(Note this photo has the bearings as well, I took it a few days later once they arrived in the mail).

Threaded rod was purchased from Bunnings for about $30. They don't look very nice, and they were quite tight so I've had to clean them up with some CLR. They should be adequate for the job though [Edit: CostLess Bolts sell BZP threaded rod, for my next machine I'll buy it from them instead].

Steel bars were purchased from Home Hardware for about $10. I originally bought some bars from a local steel specialist but when I got them home I noticed they were bent and covered in a rough mottled pattern. Buyer beware!

Nichrome wire was purchased from Dick Smith for about $5, I also picked up some micro-switches for a few bucks each which I plan to use for the end-stops in place of optos.

I've managed to salvage a small 200W power supply from an old shuttle PC. Unfortunately it can only deliver 120W @ 12V...that's enough for the 60W Mendel but not a 100W heatbed as well. That said, I read on the wiki somewhere that you can get more accurate control by using two heater circuits i.e. one that's always on and another than you control with PWM. What I might be able to do is run 50W on 12V with a constant supply and the other 50W on 5V with PWM.

I'm stunned at the disparity in pricing between different bearing manufacturers! The store I went to today, a bearing specialist no less, tried selling me FAG bearings for $8 each, that's over $400 for the full set! In fairness though I should point out that they seem to target the automotive industry and their bearings are very high quality, far above what's needed on this project. I decided to instead ordered in my bearings from AusXMods. Total cost: http://www.ppl-pilot.com:80/$70.

Thick Sheet Cut

Picked up some 6mm MDF from Bunnings over the weekend and cut it by hand tonight. It's pretty easy to work with, although drilling holes will probably be a bit of a PITA. I'm using nopheads simplified frog design, and since I plan to replace all the electronics with my own I'll probably wait until they're done before drilling holes for them. Pololu drivers are in short supply at the moment so my electronics are on hold until the new batch arrive in December...or until I figure out a way to replace them.

Z-Axis and Initial Frame Assembly

At last! Something that's finally starting to bear a vague resemblance to Mendel! I tried mounting the x-axis but the z-axis rods are pretty tight and will almost certainly jam :( I suspect it's because they're galvanized steel rather than BZP, they don't seem to rotate within an M8 nut very easily when the force of the x-carriage weight is applied. It's easy enough to replace those two bars with BZP but it'll require another trip to the shops this weekend. I've left the x-axis off for now.

One thing I have done that's different to the instructions is to mount the 624 bearings into position on the the zaxis pieces before attaching them to the z axis rods. Accurately positioning them and tightening the six M3 bolts that hold them in place was just a bit too tricky otherwise.

Y-Axis Piece Assembly

Nothing too exciting here, these are just the bits that connect the moving bed to the y-axis bars. I still need to purchase the board materials and cut them to size.

Electronics Circuit and PCB Design

Spent the past few nights designing electronics and a PCB, I figure I can get everything onto a 4.5" x 3.5" board using all through-hole components with no between-pin traces or jumpers.

Virtual Mendel

I've spent the past couple of days porting the official RepRap firmware to a managed C++ class library which is linked to my front end. I've also written some graphics code for rendering plastic extrusion and hooked it up to the firmware. The result is a 3D "Virtual Mendel" which I can now use to load gcode files and visualize the animated print before committing to plastic. This is extremely handy for examining the internal features of print parts, which is something I need to do with my RP parts because many of the holes are filled in and need drilling. It also means I can be reasonably confident that my plan of running the official firmware host-side and transfering the binary drive data across my USB link will actually work. Finally, it will come in handy if I ever plan to work on modifying/improving the official gcode generation code.

While working on all this I discovered a null-pointer dereference bug in the official firmware which I reported to the dev group. I provided a patch on request and it's been applied to the official firmware.

X-Axis Assembly

The new pieces are indeed much easier to work with, the x axis is now mostly assembled and I haven't had any breaks or any other problems of that nature. There's no belt or extruder yet, and the BOM didn't mention anything about needing M3 studs for attaching the gears to the motor shafts so the motors aren't ready either, but at least it's a start. The only thing that bothers me is that the bearings on one side of the x carriage aren't sitting squarely on the bar, they're a few millimeters off to one side so only the edge is making contact. Not sure what's going on there, maybe they're supposed to be propped out with a few washers or something? In any case it doesn't seem to be affecting movement of the carriage...yet...

New RP Pieces Arrive

Fontonlabs was true to his word and a new set of RP pieces have arrived in the mail, I can tell right away that these are going to be MUCH easier to work with. For starters they aren't warped and they have a reasonable amount of flex but not too much. They don't smell as much as the polyester did either!

Motors Arrive

Four NEMA 17 motors arrived in the mail today, I purchased them from AusXMods for $95 inc shipping:

The shafts are round, so they'll need to be filed back on one side before attaching them to the pully gears.

Host Software Development

I've started working on the host software while I wait for the new RP pieces to arrive. I've taken Capo's Mendel SketchUp model, exported it to Collada, run it through some old DAE-parsing code I wrote years ago and written some new code to convert it to XAML. The result is a user-interactive 3D model of Mendel that I can control programatically:

It's not perfect, and I'll eventually want to whip up a better model in Max, but it'll do for now.

Multiple failures!

The first night of assembly hasn't gone well, with 5 pieces snapping in half despite my every attempt to be as careful as possible. Mendel relies on the RP pieces having at least some flex so that proper tension can be applied to the bars but these pieces are rigid and crack when subjected to even the slightest pressure. Even drilling holes (which is needed for this set because it's a cast) causes large chunks to crack off. To make matters worse many of the pieces are warped to the point of being unusable, and of course they can't be bent into position without breaking.

I mailed Fotonlabs to let him know and he said he must have accidentally sent me an older set made from polyester which is too brittle for Mendel and warps badly. I wasn't asking for my money back but he offered to send me a replacement set made from polyuerethane with a flexablizer, which was good of him. It's a minor set-back as I've spent an entire evening assembling pieces and then disassembling them again to get the bolts back, and I've also wasted quite a few nylocs in the process. I've learned a lot about Mendel along the way though.

Bearings arrive

The bearings arrived today, I pretty much have everything I need now to commence construction of my Mendel.

RP pieces arrive

The RP pieces arrived today! Yay! All I need now are for the bearings to arrive.

Ordered RP parts

Ordered parts off Ebay from Fotonlabs. Total cost: $140, including shipping. That's a couple hundred dollars less than a typical set of plastic parts sell for.

Initial Design Decisions

After much research and deliberation I've decided to buy a set of Mendel parts off EBay but replace the electronics and driver software with my own, at least for now. This decision is driven primarily by cost: a full set of electronics on EBay costs about $350 inc shipping to Australia but I could make them myself for a fraction of that. Plus there are some features that I would like to add such as full USB support etc. I'm going to make the electronics PIC-based (specifically the 18F4550) which will make USB 2.0 support fairly trivial. For the front-end I'll use WPF (Visual C# Express), mainly because it's free and quick and easy to develop on. The firmware will be run host-side in a managed C++ class library (Visual C++ Express) which I'll statically link to the front-end. For the end-stops I'll try nopheads trick of using microswitches, I can always replace them with optos later if they don't work out. For the motor drivers I'll probably stick to Pololus, but I'll keep an open mind about that...there's a shortage of them at the moment and it'd be nice if there was a cheap and easy way to replace them with discrete components and standard ICs. Collectively, this should allow me to develop a full set of USB-driven electronics for around $30, plus the cost of the motor drivers. As an added bonus the firmware will run host-side, which will make it much easier to develop and debug future extensions with Visual Studio.