DDTv3: The 3D Printed Nylon-Kevlar Combat Robot

DDTv3: The 3D Printed Nylon-Kevlar Combat Robot

 

I didnt have a chance to write up the Bot Blast 2015 event report earlier this year July, but if I did I wouldnt actually have much to say. Neither DDT or Silent Spring did very well and amounted for a total W-L of 2-4. Furthermore, DDT’s frame was done for and I saw this opportunity to revisit some of the design creeds that I’d been touting for for the past year such as the “big weapon entertainment” and “indirect drive”. These concepts are good for entertainment value and long-term reliability but are not necessarily the key to win fights. With a bit of TV appearance, I had a target on my back and many builders would love a chance to prove themselves against one of the “cream of the crop” robot builders as described by the Battlebots founders. I dont mind this one bit, but if this is the case I better live up to the hype. And so starting with DDT, I will evolve each robot in my fleet.

DDT v3 Isometric View

DDT v3 Isometric View

A New Design

Recently, Charles purchased a Mark One 3D printer from the company Mark Forged. This printer is unique because it has the ability to embed CONTINUOUS strands of fiber (fiberglass, kevlar, carbon fiber) within each layer of its print. Although it cannot place fiber in the vertical build axis, this is a monumental improvement in tensile strength.

In addition, it prints nylon as its base material, which is mechanically superior to ABS in our application. To further demonstrate the capabilities and applications of their technology, I have elected to print DDT as a nylon-kevlar unibody.

DDT v3 Unibody CAD

DDT v3 Unibody CAD

Designing for a unibody frame is a bit challenging because many details about the parts, their connections, and their assembly order must be known ahead of time. CAD became an invaluable tool in this process; each part was modeled with excruciating detail and a detailed assembly order was developed for the rear component bay.

I fully realize that this design is not optimal for 3D printing and perhaps I can explain in another post if people are interested. The reasoning for the design of the trusses were add-ons to stiffen long or thin features. A better design would have made the robot more blocky but would have been aesthetically boring in my opinion.

DDT v3 Internal View

DDT v3 Internal View

DDT kept its unique shape and trademark “huge-ass blade” but with a few modifications. First, the blade size was reduced to 6″ (instead of 6.5″). This decision was made to free up some weight to be allocated elsewhere, specifically the ribbing to the blades. I had two blades made for DDT v2 and both of them were horribly bent after the year or so they have been in service. The newer blade design beefed up the sections by about 1.5x and were still lighter than the old blades.

The new blade stacked atop the old blade demonstrates their difference in diameter

The new blade stacked atop the old blade demonstrates their difference in diameter

Second, more armor was added to the body. This decision was made after the number of hits to the pulleys and wheels. The frame parts would be thickened overall and side armor will be included.

Third, the indirect drive was ditched in favor of direct drive to 22:1 silver sparks with cushy foam wheels. The small brass micro gearmotors proved to be unreliable and the ratios I was using were far to mild to be maneuverable. The original DDT used roughly 20:1 gearboxes and had superior mobility.

Finally, the blade material was changed. This time I wanted to use prehard materials because my attempts to harden were subpar. The standard 6″ blades were waterjet from 60C hard 1075/74 spring steel. An alternate set of blades were made for specific types of opponents. These were also made from different materials and had different geometries.

Standard Blade (left) and Bash Blade (right)

Standard Blade (left) and Bash Blade (right)

4 DSC_0651

Fabrication

The frame took nearly 36 hours to print! The printer is currently admittedly slow, especially with the addition of fiber. While the frame could certainly have used more fiber, I decided to optimize for more “height” of fiber than depth and so all layers had fiber but only select layers had full layers.

DDT v3 unibody frame as seen in "eiger"

DDT v3 unibody frame as seen in “eiger”

DDT v3 unibody frame as printed on the Markforged Mark One Printer

DDT v3 unibody frame as printed on the Markforged Mark One Printer

The frame needed some cleaning to allow the proper fit of parts. Any undercut features were finished with a rotary tool sanding bit to clean up areas were support structures were placed. This was critical for the drive motors, screw holes, and other components to be placed.

Fingertech Spark gearmotor embedded into the frame

Fingertech Spark gearmotor embedded into the frame

The critical electronics were packaged into a small cube that I designed using my expert level tetris skills. This small cube contains an Orange RX receiver, a Turnigy Plush 10 ESC for the weapon, and two Vex 29 motor controllers for each spark motor.

DDT v3 electronics cube contains three ESCs and a receiver and is about 1" x 1.125" x .75"

DDT v3 electronics cube contains three ESCs and a receiver and is about 1″ x 1.125″ x .75″

The electronics cube is intended to fit in the frame like so.

Placement of the electronics cube within the frame

Placement of the electronics cube within the frame

The motor was recycled from v2 and the robot was wired. No dedicated power switch was used since I have had bad luck with the Fingertech switches lately and was feeling skeptical; in lieu of this, I elected to use a direct connection to the battery.

The blade and pulley interface were modified as well. Instead of placing the bearing directly within the bore of the blade as done previously, both bearings would be placed in the hub and the blade would be bolted to the pulley. A round feature on the pulley would mate the blade concentrically with the pulley. This method is better because it allows the blades to be swapped without disassembling the entire shaft assembly, it reduces misalignment because the skew is no longer dependent on the alignment of the blade and pulley, and it reduces the shock of the bearings by placing a softer medium between the bearing OD and the blade.

Illustration of the blade-pulley interface

Illustration of the blade-pulley interface

Lastly, side armor was made to fit wrap around the sides. These were cut from 0.0625 polycarbonate which is tough yet flexible. However at this short length and thickness the sides were difficult to put on and heavily preloaded. I do not have especially high hopes for these; in anticipation, I made several copies.

DDT v3 with an older 4130 Blade

DDT v3 with an older 4130 Blade

DDT v3 featuring the newer 1074/75 blade

DDT v3 featuring the newer 1074/75 blade

Final weight was 15.5 oz! A very comfortable weight and some wiggle room to improve.

Preliminary Performance

DDT had the opportunity to smash a number of items. However none of them actually weighed a pound. Still, these tests demonstrated remarkable damage to UHMW, aluminum, and even other 3D printed Nylon parts. I couldnt think of a better way to send off the v2 frame permanently!

Blade Tested Against Metal

Blade Tested Against Metal

Blade Tested Against a Nylon 3D Print

Blade Tested Against a Nylon 3D Print

Dragon Con 2015 Microbattles

DDT v3’s debut was at Dragon Con 2015 where it had four matches and went 3 wins, 1 loss. I am extremely happy with its performance and the resiliance of the printed frame. I half expected the weapon shaft bore to open up after so many hits (including arena wall hits) but it remains as true as when it was first made, probably due in part to the heavy kevlar fill around the critical dimensions. DDT’s fights are embedded below. I was very lucky to not fly out of the pit on several occasions but unfortunately my luck ran out when a solid hit on Algos sent DDT flying into the pit.


 

DDT took some damage from Green Reaper when he climbed over the blade and began chewing on the frame. I was surprised to see that not much damage was done but only because my blade was keeping him from pushing into the frame otherwise more damage would surely have been done; some of the fiber is exposed which means he got about 0.6 mm deep. Both of the wheel guards exploded against Algos but I was able to leave a nice mark in his 0.0625 thick Ti before flying out.

Post-Algos picture

Post-Algos picture

Green Reaper damage

Green Reaper damage

Damage to Algos

Damage to Algos

Bent Ti side on Algos

Bent Ti side on Algos

I want to find more opportunities for DDT to challenge other antweights. There are two competitions nearing in October; I will try my hardest to bring DDT there. Stay tuned!