To continue the RepRap calibration process (yes you really need that much calibration) I printed an object to verify the Layer height.
The object printed perfectly but I noticed my wall thickness is 0.75mm instead of 0.40mm. Something that needs some investigation. I also verified that the X and Y axis calibration is perfect now; 20mm in any software gives an object of exactly 20mm!
Object: 4-mm-wall.stl by muelli from Thingiverse
Hotend: 245°C 1st layer, then 240°C
Heatbed: 110°C
Material: 3mm ABS (Silver from RepRapUniverse.com)
Layer height: 0.2mm
Wall thickness: 0.4mm
Infill: 100%
Static props: 0.4mm nozzle, E3D All metal hotend, MK2a heatbed, Gcode generation with Slic3r, Pronterface, Sanguinololu, Ice Blue Stepsticks, NEMA17 motors, Prusa i3, Achatz Edition, ATX powered 400W
I already experienced that a clean and well prepared print bed can make or break a print. Good preparation leads to good first layers, good adhesion(@110°C) while objects automatically pop off the printbed when it cools down (no banging on printed objects with hammers and breaking glass).
My procedure for printing ABS (for now):
Clean the glass (while cold) with Acetone and scrape off previous ABS remains (Nail Polish remover will work fine and smells less anoying)
Clean the glass with regular window cleaner
Spray 3 layers of extra-ultra-strong hairspray on the cold glass while letting each layer dry. (cut a 20x20cm hole in a shoebox to be able to spray while the glass is mounted on the Prusa frame)
Re-check Z-axis leveling (I really need to get a automatic leveling thingie)
During previous prints I noticed my hotend moved a couple of mm in the Y-axis direction (and this messes up the Z-leveling). The hotend was seated directly into the extruder and secured with a mounting plate. Now I've placed some fancy washers in the 16mm extruder hole (orange arrow below) and the E3D hotend. Now the hotend is firmly secured and not moving at all in any direction = better prints!
To make the heatbed heat up even faster I added 2 layers of PE-thermal insulation(95% heat reflection) under the heatbed.
I just learned how important the first layer is.
The Z-bed was not properly leveled (too close to the nozzle) and caused 2 prints to be worthless (guess which ones). The objects fall apart when peeling them off the heatbed because the first layers where too thin.
When the bed is leveled properly the printed object comes out fine and doesn't break (bottom-left).
The "hairspray-trick" gives me good adhesion for this size of objects. I'm curious how it will work on large objects. Hairspray trick = properly clean the glass plate with acetone (or nail polish remover) and spray 3 layers of hairspray while letting each layer dry. Repeat after each print.
The important first layer
Object STL: a letter B drawn in Sketchup for my friend Bjorn who asked to print a B
Material: 3mm ABS
Hotend: 245°C 1st layer, then 240°C
Heatbed: 110°C
Heatbed adhesion: 3 layers of Fructis Bamboo Extract Hairspray, extra hold nr 6
Layer height: 0,2mm & 0.4mm
Nozzle: 0,4mm
Gcode generated with Slic3r and printed with Pronterface
Object STL: 5mm Calibration Cube Steps by MCroucher on Thingiverse
Material: 3mm ABS (Velleman from reprapuniverse.com)
Hotend: 245°C 1st layer, then 240°C
Heatbed: 115°C 1st layer, then 110°C
Heatbed adhesion: 3 layers of Fructis Bamboo Extract Hairspray, extra hold nr 6
Layer height: 0,2mm
Nozzle: 0,4mm
Gcode generated with Slic3r
Object printed with Pronterface
Issues
the 10mm bridge is not that clean, the 5mm bridge is perfect though
object is warping after some layers, need to find out why (more adhesion needed?, heatbed temp too high?, etc?)
corners for short edges are more rounded off than corners for long edges (top 5 mm)
Good stuff
Extrusion calibration is near perfect
X, Y & Z-axis calibration is near perfect (calipers confirm!)
MK2a heatbed @ 110° + E3D hotend @ 240°C + powering the Sanguinololu = ATX power source wires melting!
So I decided to solder thicker wires that feed the Sanguinololu directly to the ATX printed circuit board(PCB). Results are good, the wires get a bit hot but don't melt anymore. I also added new spacers to the bottom of the ATX to be able to mount them to the Pusa's frame.
Next issue: the MK2a heatbed takes 30 minutes to heatup to > 100°C
So I took care of these 2 things: "create direct current paths" to spare the Sanguinololu PCB and "Balance the ATX power supply"...
1. Create Direct Current paths on Sanguinololu PCB
The pads and traces on the Sanguinololu board are not robust enough to handle the high current involved when controlling the heated bed with the dedicated mosfet. If left as is, the board will heat up in this area, could be damaged and the plastic connectors discolored.
To fix, simply add a direct current path by way of (preferably insulated) wire:
from the 12v input to the 12v pins on the Heated Bed (HB) connector;
from the center pin of the HB mosfet to the two ground pins on the HB connector;
from the ground pin on the mosfet to the ground input of the board (probably the ground side of the 12v connector)
I get these results (for now) that I need to configure in my Marlin firmware:
(In "Configuration.h" the variable: DEFAULT_AXIS_STEPS_PER_UNIT {80,80,2560,674.6521728} )
Edit: I have 5mm threaded rods for my Z-axis so there are 4000 E steps for 1mm (and not 2560!)
Thanks to some hints from the kind people on the RepRap.org forums I found out there was a bad connection on the bottom side of the PCB. Some solder bridged 2 connections due tho the short circuit I caused before(including smoke & flash). After cleaning up this bridge all electronics work fine and my X, Y and Z axis move smoothly!
This means I'm back on track and hopefully be printing something soon...!?
Next Step: Leveling the print bed
I'm pretty sure I fried the ATmega 1284P-PU Microprocessor on the Sanguinololu board and went ahead and ordered a new one from Richard at RepRapUniverse.com. A stupid mistake that will hopefully only cost me about 15€ and some extra days before actually printing something...
In the meanwhile I fiddled with uploading new Marlin firmware configurations. I'll need to do this a couple of times while calibrating the different axis stepper motors so I thought I'd give it a try (also my Z-axis is inverted and I don't know if I need to flip my wiring or change the INVERT_Z_DIR parameter in the Configuration.h file).
At first I could not connect the Arduino IDE to my Sanguinololu because the maximum standard baud rate in Arduino is 115200 baud. After some hex value editing in the windows registry I was able to connect the Sanguinololu to the Arduino IDE at 250000 baud, the current rate configured in my Sanguinololu. (Thanks to this post from Conseils on his NorthernHope blog!)
Now I can send G-code via the Arduino IDE and flash new Marlin firmware via the Arduino IDE.
That makes me happy.
Flashed my first new Marlin Firmware configuration
Wiring done for all three axis and the extruder stepper motors. Also the connections to both thermistors and the extruder heater and heatbed are ok. Still need to connect the fans to the ATX power source. (I connected the extruder fan to a free header on the Sanguinololu but I read this is a bad idea)
Setting the Vref to match my stepper-motor-driver configuration. My calculations don't match my measurements (calculated 2.24Volts and measured max 0.74Volts so my sense resistor value must be way off). But hey, the motors are working fine after some manual tweaking while listening to the motors' sound. I'll check what went wrong with the calculations later to fine tune the driver current...
This video shows the X & Z axis move during a test with Pronterface (the Z-axis moves really slow, this is normal).
Another milestone in the building process: To test the extruder I heated up the hotend gradually to 290°C and meanwhile extruded some blue PLA using Pronterface.
Open issues that need to be looked at:
Z-axis & Extruder inversion
The Z axis goes UP while going to the home position in Pronterface. Also the exdruder retracts while it's supposed to extrude and vice versa. So I need to reverse my stepper wiring from ABCD to DCBA or change the inversion setting in Configuration.h in the Marlin firmware and upload it to the Sanguinololu again.
Whoooops, I think I might have fried the Sanguinololu board :/
While fine tuning the driver current for the Y-axis i made a short circuit with my multimeter. Now my Y-axis won't move properly. Just when everything was going so good. Stupid me!
