I haven’t done much work trying to get the dual extruder i3 working. I was frustrated by the curling I was seeing with ABS prints and the mess created using ABS juice (ABS dissolved in Acetone) so though I would try a sand blasted Ikea mirror
Going back to basics I wanted to get decent quality prints in ABS with one extruder. I have replaced the cheap 0.35mm jhead clone with a decent quality one. Initial results were promising but I was still seeing some curling.
I tried printing with brim and a light coating of ABS juice. That improved things, with only minimal lifting.
Getting the print off the bed was another matter. The brim came away easily enough but the print was well and truly stuck.
Prising the print up with a thin knife freed the print, but also some of the glass bed! I’ve gone back to glass with Kapton tape (wiped with Acetone before use) and brim and that seems to be working, for now.
You can get your hands on the files for the dual hotend mount here http://www.thingiverse.com/thing:108273 and the adaptor to make it a quad hotend mount here http://www.thingiverse.com/thing:108274
There hasn’t been much to add about the bowden extruder recently as I have just been using it to print i3 plastic parts with varying degrees of success. I’ve had to switch to using a brim due to excessive ABS print curling and I’ve broken the teeth off a printed gear.
Broken gear against replacement part.
I think the print quality I am getting from my Huxley is not so good any more which was part of the cause of the gear failing. The other is that I am running the extruder retraction on the i3 bowden fast ( @ 120mm/s) and long ( @ 9mm).
The bowden extruder clocked up about 36 hours print time before the failure. I was able to continue the print that was in progress at the time, with the remaining 4mm of teeth at the bottom of the gear by moving it up the stepper motor shaft. That wasn’t enough to save the print as it had printed a layer of air and the resulting top part of the print pulled away from the body easily.
A bag of i3 plastic parts.
While the extruder was running, I was able to produce an (almost) full set of i3 plastics at reasonable quality – functional if not pretty. I am still having problems with blobs and strings and need to work on the retraction settings.
I’ve added the mount to the extruder so I can mount it on my i3.
I did include M3 captive nut cutouts on the inside of the plate mount so it could be secured with M3 screws but I’m not sure they are needed which is good as they were completely filled with support material. I switched to honeycomb support material which, when printed on my Huxley is too strong to remove easily.
So now I need to run some long prints with this setup and see how reliable it is. I think this printer is about ready to print a set of it’s own plastic parts. If they are useable and as it has become a bit of a tradition, the first set of parts will be given away to someone who plans to build a printer.
Things have slowed to a crawl on the bowden extruder. So going back in time a bit, here are the gears I designed for the extruder drive. Once I had the gears at the right size, I was able to add the fittings for the hex head and holes to reduce the amount of material needed to print it.
The slot for the captive nut and screw was added after the plain gear was designed.
After chasing problems with the 0.35mm hotend all morning I decided to take it out of the equation and concentrate on proving the bowden design works with the well proven 0.5mm hotend.
Printing with one hotend and the test extruder setup on the bench I was able to get some reasonable prints for a first try. I need to tweak the settings for the bowden setup and I will do that as I go.
The aluminium 0.35mm JHead clone proved to be continuiously problematic and after spending far too long trying to clear blockages I decided to bin it. I plan to replace them both with jheads from hotends.com or reprapsource.com in the not too distant future.
With two Bowden extruders hooked up and the firmware updated to reflect my calculated esteps value, I tried to run a test print and the result, while messy, is promising.
The first red layer is very neat, the second is a little too widely spaced. Again, I need to properly calibrate the esteps before I can expect good prints, so this is just to see if the extruders will work initially.
I discovered another problem that I recognized as an issue I was having on my previous extruder. That is the jheads I have are different lengths and so I can’t get them level in the z-axis. I didn’t build in adjustment for the jheads as I assumed they would be machined to the same lengths.
You can just make out the blue has been printed nearly 1mm higher than the red. It is also messy because it was no where near the bed for the first layer and so didn’t stick to the bed.
I measured the difference in height for the jheads from the mounting to the end of the heater body and shimmed the shorter one accordingly but it was still short. Turns out the nozzle shapes are also different lengths so I need to take into account the length from mount to tip and try again. In doing this I also discovered that the bearing holder on the x motor end was broken and the bearings were slipping out causing the x axis to move around, so I need to replace that before I go any further.
This is just an update really. I’ve been designing the gears (when I say designing the gears, I really mean tweaking the script settings so I have gears that I think should fit). Now I have gears that are (hopefully) the correct size, I just need to add the fittings for the nuts and bolts, print them and then check they work.
If my maths is correct then the e-steps vaule (number of stepper motor steps per mm of filament extruded) then I should be looking at about 705 steps per mm.
I know the diameter of the hobbing on the bolt (measured at 6.9mm diameter, circumference = pi x diameter, 22/7 X 6.9 = 22.69mm), the gear ratios (55:11 or 5:1), the number of steps per revolution of the stepper motor (1.8 degrees per step, 360 degrees per revolution, 360/1.8 = 200 steps per revolution but with 1/16th micro stepping that is 16 x 200 = 3200 steps per revolution), so we can say that 3200 x 5 (gear ratio of 5:1 means 5 revolutions of the stepper motor for one revolution of the hobbing bolt gear) = 16000 steps = 22.69mm of hobbing bolt travel so that means 16000/22.69 = 705.16 steps per mm (approx). I’m not taking into account the size of the filament or how much it sinks into the teeth. I just wanted an approximate value to configure the firmware for calibration.
I have also made the motor mount thicker as my test print proved a bit too flexible. The picture below is of the original NEMA 17 motor mount test. The next version is twice as thick.