Here is a short example of a test drive. In the meantime a lot of code has been written, (Also this is without the arm mounted) The problem initially was that we really don’t have a driver who could properly navigate the “maze” The FTC field is 12 by 12 feet and every 2 feet (24 in) there is a Pole (to put cones on) So you got to go straight or a true 90 sideways and if you are off you run into poles. The frame makes it difficult to get “hung up” which is good initially we had wheels getting snagged.
So our approach is what we call a “semi teleop” because our autonomous worked real nice. So we still have Joy sticks for “emergency” but we have defined buttons (and doing more) to do certain things. So for example we made sure the robot under Joystick control can only go fwd/back or strave right/left. We use an encoder on each wheel and have a button to go fwd/back/left/right for exactly 2 feet. And keep the “nose” pointed always in the same direction. So when we want to go to a certain square we might hit Fwd 2 feet 2x and right 2feet 3x to get there. As a side effect our drive code allone has grown to over 500 blocks/lines (using blockly and or Java). And that was with encoders. We now sometimes experience a little drift still so we are in the process of using the IMU to help with that. On the video posted. That drive has some problems on an old uneven wood floor as the tires sometimes spin out as you can see. It happens to a much lesser extend on the competition surface. So one of the tasks is to define and max acceleration/deceleration also to save time to avoid the overshoot by 2-4 inches and then the robot backing up to reach the pos defined by the encoders.
In the next 2 weeks we will also focus on automating the placement to avoid drivers to misplace the cone (next to the pole). I probably will update how that goes and post a video if we are successful.
This is a pic of it put together with most of the arm I dont have any pics of the frame without the motors and the electronics on it FrXdrive.zip (5.2 MB)
I got some .3mf files of the frame - not sure if they are all from the same version but good enough to look at in a slicer. We will release it all if its successful
Think is printed from petg about 1.4kg Its thinnest wall is 5mm solid printed with a .5 nozzle .55 outside perimeters .8 to .9mm inner perimeters up to 9 perimeters and .28 layer height. Reason why its printed on and ender 3 as we (the team) wants to print on the printer they have available. I can print it faster on a bigger printer.
Here are some pics of stuff that goes on it likethe LZ Zusan and some pulleys and stuff
I included the wheel hub to show how we interface with the D Shaft we print square sockets to slide in some square m3 nuts and then use set screws and of course a press fit tolerance D cutout for the shaft Seemsto work well we dropped the robot a couple off times and in testing ran it sometimes (unintentional) full speed in the wall or a table/chair etc leg. Its maybe overkill but I come from FRC and this is my first FTC year and I know 5mm wall thickness usually if used properly survives FRC (ab)usage
Very cool, thanks very much for the ton of pictures. In the VRC world, teams can’t 3D print. (Coaches and mentors are the ones that are against it, citing cost and time to print larger higher infill objects) You’ve presented a good example on how 3D printing could drastically change how robots are built.
Looking forward to hearing about your season, good luck!
That one uses 1/2 in 16ga al square tube to reinforce the 3dp plastic we later replaced some of them with steel as the robot was too light. But to print the whole skateboard frame (bottom part with the wheels. We started on kickoff and by tuesday afternoon we had “wheels down” meaning robot standing on the floor on 4 wheels driveable. with an electronics board and battery mounted.
With my 3 printers I can print the shooter in 24 hours.
As for cost…
For the first robot we used 100 feet of the al. tubing - that at $1 per foot is $100. Then we use 14kg of HIPS that you can get at $10/kg so 140 bucks if you use petg I can pick that up at $13/kg easily or less (I usually buy 50kg at a clip) so that would make it about 180 bucks. So for less than $300 you can do the whole superstructure of a robot.
The difficult part with 3dp is
a.) Knowing how to do it right
b.) doing all the cad. Even with lets say 5-6 cad ppl on the team I probably would not start that project during build season as they could not keep my 3 printers busy. How you do it is you design a System in the off season. Like the above frame system that prints in components Like wheel holders and transmissions and straight and angle pieces. Then when season comes you just crank them out. The shooter above wound up costing 24 hours in printing and about 70 hours in cad (with revisions) you can let a printer that is well set up print overnight and is successful 99% of the time or close to it. CA D to 3dp time often is 3:1 to 5:1 The FTC frame I can print on my 3 printers in about 12 hours but the team took over a week to cad V1 and probably 3 weeks of cad time to wind up with the current version and that is plastic only
As for strength PETG will at a 10x5mm cross section 3dp piece if printed to my best knowledge fail at picking up betwee 180 to 250 lb depending on printing direction hence on the frame everything is at least 5x10mm in each direction sometimes in one piece sometimes in multiple strands. Now the cad for the FRC robot pieces is available I think there are links on my website. So you could print your pieces and put a 3dp kit bot together by yourself and yes there are also gearboxes that will lift the thing
Here is an experimental differential planetary I did just to see if I could
and here is a 8:1 planetary in a wheel for about $25 -30 Nylon gears, TPU tire, ABS/HIPS frame
