Saturday 31 January 2009

Alternative Electronics mounting

After looking at the darwin, and winding it up and down a few times, I found there is quite a lot of dead space inside the frame at the bottom. The z-axis cannot travel all the way to the bottom, because the motor and belt assembly prevent it.

With a bit of measuring, I found that the PSU and the electronics could be fitted inside the dead space.

I cut a chunk of plywood to size that would take all the PCBs. I found some non-conductive solid foam in the garage, and cut a couple of strips. Sticking the strips along the plywood, the PCBs can be mounted with some small bolts, without the undersides of the PCBs touching. The plywood and PSU are mounted to the frame with cable ties. They seem to be holding things together pretty well so far.

The x and y steppers leads were too short, so I extended them with some 4-core speaker cable and re-attached the plugs. The extruder leads I also extended, giving it plenty of slack to move around and reach to the PCBs. The power supply leads are now situated right near the PCBs so they plug straight in.

The biggest risk is a stray cable being caught up in the belt and either being ripped out or jamming up the belt. Make sure all cables are securely attached with cable ties and can't touch the belt. I routed all mine along the outside, underneath the lower sqare of rods and then direct to the sockets. The arduino is tucked away pretty much in the center of the space.

This arrangement protects the electronics and means that from the outside, all you need is the USB cable for the arduino and the power for the PSU. I think it also looks slightly neater than the boards on the outside, but since the Darwin is unlikely to win any beauty awards (only engineering ones) it probably doesn't matter too much :-)

Blowing the IR home sensors

When my full kit arrived it contained limit sensors for home/end for all the axes - i.e. 6 sensors. Although I soldered them all, lazily I only attached the three home sensors to the build.

During testing with the stepper exerciser, I started getting some unusual symptoms. sometimes the stepper would only move in one direction (e.g forward and not back) and the 'home' buttons did nothing. Trying to find the range just sent the motors heading away from the sensor.

Finally, I worked it out and tested the sensors separately. I'd wired the 3-pin plugs backwards (+5v/grnd reversed) and this had blown the sensors. Even when there was nothing there, they reported closed. I guess I'd blown the IR source.

Luckily, since I'd been lazy earlier, I had three untouched boards. Once I'd rewired the 3-pin plugs properly, the arduino test program correctly reported open and closed when I blocked it with some card.

I replaced all three sensors and the homing function worked properly. The clear acrylic from the BfB kit doesn't block IR, so I glued some card squares to the moving pins to block the IR.

Tuning the stepper circuits

Once I'd plugged in the installed steppers into the driver circuits, initial testing showed that the heatsinks got *very* hot. After a couple of minutes, the x-axis started to 'jiggle' - oscillate about half-a step backwards and forwards making a juddering noise. I switched it off.

I moved each potentiometer to minimum, and the steppers stopped moving completely.

Using the stepper exerciser and the 'home' command, I set the axis moving. Turning the current up slowly, you reach a point where the axis just starts to move. I gave it 10% above that and then stopped. That gave it enough to move all the axes reliably.

The heatsinks still got *very* hot, so I took off the standard ones, and grabbed a few from my spares box. Due to a few years of rebuilding PCs, I had several PC heatsinks around (for those who are interested, they are a from a 486dx2, a Pentuim 75, and an Athlon 800). I drilled a 5mm hole near the edge and slapped them on. Now, even after a while, they only seem to get warm.

I also extended the stepper plugs using some 4-core ribbon speaker cable.

Friday 30 January 2009

Extruder (2)

After completing the heater, I warmed it up to 220 and put through some ABS.
The motor ground through a few cm of extrusion! it works!

Pretty soon, though the filament stalled and the screw ground down the filament. Feeding through by hand started it again, but it needed help.
I had to experiment a bit to get a consistent extrusion. Things that helped:
  • I bought a cheap tap/die set and ran a die down the bolt thread to sharpen it
  • tightening up the screws and springs pressing the filament onto the screw
  • upping the temperature to 240 *really* helped. I think at 220 it's only just hot enough. I'm still not sure that thermistor reading is accurate.
I still need to configure the temperature thermistor/heater power to get a more consistent temp - it tends to creep up slowly to over 255.

Extruder assembly (1)

I bolted together the complete extruder assembly.

I finished off the extruder and tested it. The barrel gets hot, the temp reader goes up and the motor starts on request. Result! or so I thought.
After testing, it seems that it wasn't getting hot enough. I measured the temp with a digital kitchen thermometer, and the difference was significant.
After a bit of investigation, I found I had the wrong include file for the thermistor in the arduino code. Switching it and re-loading the arduino gave me a much better reading - the kitchen thermometer (touching the outside of the barrel) now reads within 10 degrees of the thermistor (buried inside the fire cement).

I also insulated the barrel using a section of silicone plate mat. Tesco's had some reduced at 10p each so I picked up a couple. This equalized the temperature between the heater top and bottom.

Progress update

Time for me to put in a progress update on my RepRap: Here's what it looks like today:

It's mostly together and working - it just needs calibration and testing!

I'm going to write up some further posts detailing how I got to the current stage, including mistakes and changes, in case there's something that helps anyone else.

Sunday 25 January 2009

Normal service has been resumed

Sorry about the 'slight' (read 4 months) delay since I last posted. I had some serious problems with RSI/Carpal tunnel in both wrists, and ended up being off work for about three months. I've had surgery on both wrists and now things are much better. I've been back at work for a few weeks now, so I'm confident enough to restart working on the reprap and blogging again.