Dragon print.

Just some pictures of the first successful print with the new fan configuration.
I printed this dragon from loubie on Thingiverse. Since the figure was printed at 50% scale (total height is about 8 cm), the front legs were a bit too fragile, so I added a little base to make the model a bit stronger.

Printed with 0.15 mm layer height, printing time about 3 hours.

Object cooling

The previous fan was used only to make sure the top of the hot end did not get too hot. I designed a guide to get the cooling air to flow both through the vents in the hot end and at the same time on the printed object around the hot end. The air guide  looks like this:

Installed, side view:

Installed, front view:

New hot end for Delta printer.

I wrote about the problems with the hot end before. Though I did get the hot end to work, it only worked at slow speeds (25 mm/s) and the extruder motor was visibly loaded more that what should be necessary. The extruder's hobbed bolt also slipped regularly leaving holes or weak parts in the printed objects.
The friction of getting the filament through the hot end (0.5 mm nozzle) was just too high. I decided to try another hot end, the J-Head, and this has been working fine for 2 weeks now. I bought the hot end here: http://reprap.me/ 
As the top part of the J-Head is bolted directly to a part made of PLA, I wanted to make sure the cooling was taken care of properly: a separate fan is blowing air through the holes of the J-Head to make sure the top part will not melt the PLA.

To cool down the printed PLA quickly I currently use another, stationary, fan. This caused a problem with reliable flow of plastic because the heater block was now being cooled too much. Too much cooling was prevented by insulating the heater block (not visible in the picture about yet) with some ceramic tape. I will make another solution for the PLA cooling later, but for now the quality (and speed) of the prints is greatly improved.

The Stanford Bunny from thingiverse:

0.15 mm layers, object about 6 cm high.

Moiré pattern, part 2

The problem with the moiré pattern which was visible in the prints last time has been fixed. I got several suggestions about possible causes. The one I focused on first was the possibly that the pulleys of the wires which are moving the "axes" up and down where not round. I got better quality pulleys, but this did not solve the problem. While trying to debug the issue I noticed that when the printer was running really slow, the sound of the printer was changing. I vaguely remembered a blog post about problems with the stepper motor drivers somewhere and I found the problem. It turns out the current which is running through the motor is too small  to actually get the motor to move in certain phases of a rotation. The solution was quite simple: increase the current through the motor. The current through the motor is adjustable with a tiny potentiometer on the stepper driver board. The previous setting was so that the stepper motor would have just enough current to not loose any steps (full steps).  I now increased the current so that it is enough to drive the motor in all phases (all 16 microsteps) of a full step. The blog post I found the information in is here:  hydraraptor Stepstruck.

The print quality increased a lot:

Moiré pattern

There are some artifacts showing up in the prints of the delta printer. Depending on the shape of the object they are more or less visible. I printed a flat vertical object, 80 x 80 mm, to be able to see the effect properly:

Most of the time this is no problem, but there may be some ways to tune the printer to minimize the deviations.

Interesting discussion going on on the google forums here.

Hot-end clogged

In the new delta printer I use this hot-end from RepRapWorld. I have used exactly the same hot-end in the previous printer and it has worked very well there. The hot-end should be made according to these specifications.
When pushing the filament through by hand before mounting the bowden cable, I noticed there was a lot of friction. When printing the first objects the extruder motor started to skip steps, or the hobbed bolt would slip the filament back before the filament got stuck completely and no more plastic was coming out.
When removing the plastic from the extruder, the plug looked like this:

There are 2 things wrong here: there is a bit of a flat end which is larger than the 0.5 mm hole where the molten plastic leaves the hot-end, and the angle of the top of the plug seems to be too big.  The specifications state a 90 degree angle drill bit should be used for the inside and there should not be a flat area at the end. This is part of the original specifications:

I don't have a proper 90 degree 3.5 mm drill bit and my local shop did not have one either. So I modified a normal 120 degree drill bit and drilled the hole a fraction of a mm deeper. After this, the plastic is flowing better. I would still like to have it flow faster, but at least now I manage to keep the printer running for more than an hour. The maximum speed is about 40 mm/second, I think on the other printer I manage to get more, but I don't have a reference figure at the moment.

Delta printer progress 2

Since last time, I added carbon fiber reinforced rods between the ball joints on the carriages and the ball joints on the extruder platform. 

I also added some switches to each X, Y and Z "corner" which will be used to home the extruder. The standard homing procedure which is used, is to quickly move each carriage over the axis up until the switch is closed. It then moves down a few millimeters and moves up slowly until the switch is closed again.

The switch at the top:

You can see the homing procedure, the switch and the carbon fiber rods in this video:

In this printer the extruder is split into two parts: the extruder motor which is mounted on the top of the machine and the extruder hot-end which is mounted on the extruder carriage. Between the motor and the hot-end, the filament travels through a teflon tube, just like a bowden cable.

The motor part is here:

The parts are borrowed from RichRap's 3DR printer which in turn is based on the Greg's Wade extruder.

The first half of the hot-end is mounted on the carriage here:

The first print of the printer failed, more on that later. But the printer is working as can be seen in this video: