DeltaArm2022

Ive wanted a general purpose delta arm for a few years, but I wanted to design and built it myself. Heres my journey from sketch to prototype.

It always starts with a sketch. I had to go through a few iterations before I found a frame design I was happy with. While I was designing the frame I kept in mind that I would need to fine tune the center for each servo post, dampen vibrations, and keep print times and costs super low. I was originally going to print the entire frame, but decided to use extruded aluminum posts for the vertical beams to provide extra rigidity and vibration dampening.

Good enough! Working with 60deg angles was a pain in the butt.

I did a terrible job at cutting the extruded aluminum. Next time I will use a mitre box.

This first layout gave me an idea of the working size of the arm: A 144mm sphere of travel isnt too bad!

Side struts and servo mounts, oh my!

2 hours for each strut, about an hour for each corner bracket; Getting to this point took about 3 days.

A super simple Raspi mount. I covered it in craft foam before mounting it.

After going through a few iterations on the servo mount I finally settled on the design to the right. It allows me to slide the servo mount horizontally, but also provides excellent rigidity.

Good enough! Each servo mount took about 2 hours. I had a few failed prints too…

The next thing to design was the 1st phalanges, or as some folks call it the Bicep’s of the delta assembly. The goal was to use magnetic ball joints, as opposed to the printed ball joints I was originally picturing.

The original 1st phalanges design. Total crap, bad idea, but sometimes I need to the see bad idea in meatspace before I realize its bad lol

Lol, nope. Terrible design, but a good start. Wasted 45min printing this awful idea, but Im glad I did. It highlighted a few interesting points I didnt consider.

Back to the sketch… I needed the magnetic ball joints to be simple but strong.

Luckily the redesign on the 1st phalanges was pretty simple. I created a keyed 90deg joint that I could put a m2 bolt through.

Much better! The servo mounting hole is also stronger because I can print it flat.

But then… A tragedy occurred!

Under Extrusion!!!! Nooooooo! I did a complete teardown on my hot end, only to then realize that the issue was a cracked tensioner arm on my extruder. Big giant pain in the butt, but a pretty cheap fix. Oh well… At least I have my SLA printer!

3 Hours later and I had all the 1st and 2nd phalanges built. Looking much better!

I threw together this very simple end effector. I will be redesigning this part first, as soon as I get my FDM printer rebuilt.

All done for the prototype phase. Now its time to redesign the end effector, print the 1st phalanges all in PLA, and replace the cap on the PWM hat (its got a ground fault and its not showing up when I run a “i2c-detect -y 0"). The action of the arm is very smooth and a heck of a lot better than my first prototype in 2020.

Lessons Learned:

• Check your extruder assembly before tearing down your hot end… I could have saved a few hours this way.
• Magnetic ball joints are a really cool way to build kinematic structures. Think about all the other places a magnet ball joint could replace a regular joint or hinge.
• SLA printed structural parts could come in very handy in the future.
• If your i2c device is not showing up, check for ground faults on any caps.
• Iterate faster by printing smaller. Think of ways to split up your structural pieces into smaller parts.