RepRap: Blog

Labels: darwin, NextGen Science Fair, Průša Mendel, RapMan, reprap, Sampo

I've finished the hardware work on upgrading my home machine to Arduino control. I've done a few hacks:

1. I distributed the circuit boards round the machine next to the devices they control. This actually cuts down the wiring cross-section, if not number, in the sense that - for example - two power wires and three signal ones go to a stepper controller, as opposed to four power wires from it to the motor. Using ribbon cable for the signals keeps it neat.
2. I added 78L05 regs to the stepper boards so that everything only needs a single 12v supply. I just soldered them into the same holes as the power connectors, which I made 2-way screw connectors. There are already smoothing caps on the boards.
3. I reconnected the 5v input to the extruder motor controller board to its internal 78L05 to make a 5v output on it. This, and the ground connector next to it, are just at the right pitch to solder the little thermistor PCB on there facing backwards. It therefore gets a local 5v supply, and just needs a single signal wire to take the voltage from the potential divider to the Arduino analog input.
   
4. I don't have axis max optoswitches. If you leave them off, pin Analog 2 on the Arduino (Z max) needs to be grounded. X and Y seem happy open-circuit.

Everything checks out and seems to be working in the sense that I can drive the motors about. I think I still have to do a bit of debugging on the heater/temperature sensor...

Postscript: Number 3. proved to be a Bad Idea. The temperature sensing works a lot better if you drive the thermistor board off the Arduino's 5v line...

Labels: arduino, darwin

It took me 3 goes. One cable pulled, and a stuck filament roll; but I got it printed in the end. Thanks to Steve for helping with the STL file. The object took 5.8 hours to print in PLA using Darwin hardware and Adrian's latest fixes. It has really picked up steam now and the quality has improved as the nozzle moves faster - as predicted.

No hint of curling on this part either.

If you look inside the holes on the left, you can see some stringing, but it is now more like annoying cobwebs than a barrier that needs drilling out with a cordless drill. Just as well - I did stick an 8mm drill bit in my hand last time.



Here's the view from above, with a clearer view of the stringing. The thick stuff is caused by a plotting error we're looking into, and the fine, spider-web stuff is what the more persistent stringing issue now looks like.

I've tried to bend the part, and it doesn't break despite the plastic flow being a little low, so we should be good to give it a go in a real Darwin.

Vik :v)

Labels: bed, corner, darwin, pla, reprap

No, not a Catfab 4000, but a humble Darwin taking shape in New Zealand as it is at time of posting - and, of course the ginger cream Burmese 'Chad', who is kindly demonstrating what I have to put up with in my workshop. Thought some readers of the blog might like to know I'm now employed by Catalyst IT, stauch supporters of Open Source in New Zealand, who have kindly made it possible for me to work one day a week on RepRap.

Vik :v)

Labels: catfab, darwin, new zealand, reprap

For those of you wondering why I've been so quiet, a couple of things have happened. First, I lost my job because the New Zealand offices closed. Not happy about that. Nice bunch of people and we worked really well as a team. So, jobhunting is now in order.

Also, because there is a limit to how much jobhunting can be done in a day, I have been assembling a Darwin. I've had to stop due to a shortage of M8 threaded rod, and midnight looms...





Yes, the bits really are that colourful. Adds a sort of 60's psychedelic feel, I find.

Vik :v)

Labels: darwin, psychedelic, reprap

I've run off a couple of PCB clips and a couple of brackets from PLA. You can see the stringing on the X constraint bracket - it looks like it as been enveloped by a spider. The thread can easily be picked off with a sharp point, fortunately.





The blob on top of the constraint bracket is where the software crashed. Fortunately it waited for the last layer...

Vik :v)

Labels: darwin, pla

Here is the official Darwin-finished pic:



Ed's powering it up the wrong way caused less smoke than my typing an extra zero in the extruder temperature box by mistake...

Here's the very first test.



As you can see, the fibreglass insulation round the extruder makes it look slightly like Cousin It...

Labels: darwin, first-test

I've had a go at printing one of Ed's Darwin corner brackets. On about the 5th layer, the output detached itself from the stage and got pushed around a bit. While this wasn't exactly a huge success, it does look very much like we're on the right track so I've been persuaded to post the video. As per the last one, there are approximately 2.5 seconds between each frame.



Right at the end you can see the deposited part start to detach and go springy before it gets pushed around. You can also see the second head appearing in the right of the frame in the latter half of the movie. The camera moves around a little as the USB lead was dragging on things, but I fixed that with some tape about half way through; not a problem as it actually gave a better camera angle!

Adhesion to the stage (or lack thereof) is my big bugbear, and I'm going to try a light coat of varnish as previous tests have shown molten CAPA sticks to it like glue. If that doesn't work, I'm onto foam bases like the rest of 'em.

Vik :v)

Labels: bracket, capa, darwin, detached, fails, reprap, stage, zaphod

Appologies, this has been a bit slow coming - I've had the lurg. The final repeatability test for Darwin: here we're looking at the performance of the x-axis after mouding a gear diectly on to the motor (see the previous "Darwin's X-axis toothed pulley" post). Results are, again, looking promising with repeatability to ±0.025. Again, a further test is needed at a later date to understand how asymptotic the trends are, but as with the other two axes, we're in the right ball park here.



Next on Darwin's todo list.. home sensors, wiring & documentation.

Labels: darwin, gear, performance, repeatability, reprap, x-axis

The last gear problem for Darwin, casting a 12 mm gear to fit on a 6.35 mm motor shaft. A flat was needed to ensure no slippage on such a thin shaft. Final solution was to use the same kit used to make the Y gears (see previous post). The collar was prepacked again, but this time, becasue there was a bit more wall thickness, the hole was simply started by working a screwdriver into the capa. The collar was then put into the pot with a new blade part which fitted the motor shaft profile. The motor shaft was then poked in from the back. Worked first time.

This was then complemented with Vik's superb idea of a washer + bit of capa for the second lip.



Next: a repeatabiltiy test for the X. Though I need to solder up a universal PCB first...

Labels: casting, darwin, gear, toothed pulley, x-axis

Firstly, I've been issued with IE7 which means Blogger actually works properly on my computer... I cannot express the joy!

Secondly and more importantly, the toothed pulleys from the previous post were put on the Y-axis. Big thanks to Adrian for sorting out the electronics and the repeatability programming!



Super glue was used to lock the little gears into position on the 8 mm drive bar. [Incorporating flats are on the development list]. The same repeatability test used on the Z-axis (see previous post) was re-used, results below:



After the first ten runs the repeatability appears to be in the ball park of ±0.04 mm. I think that the first few runs balance the tension: this is important because the transmission isn't symetrical:



It's worth stressing that this is not bad, we're still within our targets, but we need more data to pin down excatly what's going on. The next test will need 100+ runs.

Labels: darwin, performance, repeatability, y-axis

I tested a bit of Darwin today. I asked the Z to go up and down 20 times, putting the motor through 1000 steps for each stroke. This, in theory, would only move the bed up and down by 3 mm, but there was some timeout error going on in my code, so that's as far as I could push it today. Testing over a big range is on the list, but it's repeatability I'm interested in, so 3 mm will do for now.



The motor torque was at 100%, speed 200 (see exerciser scale - approx 80%), and the results on the caliper bolted to the bed were... well... the graph can do the talking:



Happy days! I guess that puts the resolution at ±0.01 mm. There's still the X&Y to get moving yet, but it's good to know they're on solid foundations ;-)

Labels: darwin, performance, z-axis

I've hit a temporary snag with the PC software, so I'm going for the firmware while fixes happen by magic. Well, sweat & swears - er - tears from Simon and Adrian, I suspect, but let me carry on in ignorant bliss...

I've noticed that my half-washer bearings slip a bit (crappy job of smoothing the lands on the screw, I suspect), so I've packed a little bit of CAPA on top of them. This prevents exposed sharp edges engaging with the filament from dragging things around. In short, it stops your half-washers from popping out. I hope. Here's an old ABS one compared to the freshly-minted CAPA part. The CAPA part is the one on the left. The slightly fuzzy appearance of the bearing is due to Vaseline.



I've taken advantage of the squishy properties of CAPA a bit, I know, but it seems to work. Now we just hope the bearings don't heat up enough to melt the CAPA when you run them fast.

Keep on reading for our shameless fund-raising plug...

Vik :v)

Labels: bearing, capa, darwin, extruder, polymorph, reprap, washer