Sunday, March 2, 2014

New printer, do as I say, not as I do.

New printer

I just finished building a new mini Kossel. This one I designed using my visual calculator Excel spreadsheet and decided to use 10" or 254mm horizontal pieces instead of 240mm. This was because I wanted a slightly larger build area of 125mm square. I also wanted to use printed sliders instead of rail guides. If you follow my Youtube channel you'll know that I was experimenting with these. Well here is the proof that they work for a printer too. Here is the Thingiverse thing for the sliders.

The endstops are not mounted on a plastic piece. I found some #4 screws and matching nuts in my bin. The 2mm nut has enough width to span the extrusion, so the endstops are mounted directly to the extrusion. The wires are routed down the inner "track" of the extrusion which is the only side of the slider that I left open. The other three sides have a "rail" that rides in the "tracks".
Otherwise, the printer uses all of the same parts as my other mini. Well, that's not true. I bought the screw kit from for my first mini. Since then, I've been buying extra screws via Amazon from Small Parts. I have purchased 6mm, 8mm, 10mm, 12mm, 16mm, and 20mm stainless M3 screws in boxes of 100 (anywhere from $3.50 to $6.75). Since I had so many 20mm, I decided to use those wherever I could in place of the 35mm ones. I still needed 3 35mm screws for the top bearings, and fortunately I had 3. I did have to redesign the effector to make the net traps deeper to use the 20mm there, but the carriages worked as is.  I'm also using the 40mm silent fan I like so much like my other printer.

I don't have an extruder for it just yet, so I'm using the one from my Prusa. I'm running a 3mm setup for this printer. So far, it oozes a bit more at idle, but the print quality is excellent, and no strings anywhere.

The short of it

I have spent the last two days finishing my new mini Kossel and running through the First Calibration and testing. I wanted to video the process so others could see what I was doing. I did video the process, but man did I screw it up royally on my configuration settings. Keep reading to find out how it unfolded.

The long of it

I started out using the calculated values from my mini Kossel calculator/visual calculator. My delta diagonal rods were built using a jig and set to be 220mm pivot to pivot. The calculator gave me 112.6mm for the Horizontal radius (Delta radius for Marlin), and I see all 3 tower offsets to 0. I went through 3 rounds and finally got the 3 positions in front of the towers to Z=0 and the center. However, when I tried the other 3 positions they were off (one was low, another high). I scratched my head, and proceeded. WHAT? Yep without having a valid explanation I plowed on.

I started with a simple calibration object, 20mm ID, 22mm OD, 1mm solid, and 1mm hollow for a total hight of 2mm. The measurements were too big, 22.75 ID, 24.8 OD, and 2.54 tall. The voice in my head was screaming TOO BIG, THE DIAGONAL IS 220, IT CAN'T BE ANYTHING ELSE ... but I ignored it, and proceeded to modify both the diagonal rod length, and the horizontal length. INSANE! Yes I know. After another 6 rounds of calibration (no I didn't film these), I got all 4 spots dialed in. The other 3 were still off, but I plowed ahead. This time the print was closer in X and Y, but the Z was still 2.54mm. The only thing that controls Z is the steps per mm. It's set to 200. 1/32nd stepping, 200 steps per, GT2 belts and 16....OH @#$%$#$ ... NOT 16 TOOTH PULLEYS, 20 TOOTH. So not 200, 160.

Oh yeah, I had copied my other mini Kossel Repetier firmware over and neglected to change the settings. Not only that, but I never bothered to check/confirm my print surface radius, or look to see if 10mm was moving 10mm with my steel rule. UNBELIEVABLE ... or not. I was in such a rush to get to the good stuff, that I neglected my own advice.

Back to Square 1

I reset everything, diagonal rod to 220, horizontal radius to 112.6, and all 3 tower offsets to 0. I then measures the Z max with my steel rule and could estimate 242mm build height. After several rounds I ended up with a horizontal radius of 113.5 and a Z_Max of 245mm. And all 7 spots on the build surface wer dead on. So I fired up the hotend and printed another test square. This one was accurate, so I decided to print a mini Kossel top vertex as a real test. It came out dimensionally accurate (15mm tall, the holes were 3mm, etc...) but the infill was too low. Yep I admit it, I had not calibrated my extruder :/

After finishing the extruder calibration, this is what the print came out like.

The one on the left is showing that the 1515 extrusion fits like a glove (15.01mm) as well as the top infill. The second is showing the bottom and how perfect the 3mm nut is captured. I would say this calibration is 100% and I can start printing new sets to distribute.