This thread will be about our off-season journey and our new robot which has been in the making. Creating a blog has been a last minute decision, so we have come a little way. Before we get technical, let’s start by discussing our goals and situations.
We have many new members to the crew!!! Our first priority is getting them engaged with the rest of the team.
The “???” in the title is about us not having an official team number after some unpleasant incidents. We are in some kind of an awkward phase for now.
Despite winning the first 2023 off-season event in the nation, we want to maintain our work capacity by building more!
The new bot may not get on the field ever due to the lack of a definite event to be held in Türkiye. But we do not care.
If we get events to compete in, our ultimate goal is to operate two bots and educate as much members as possible.
Although we have a new bot to build, our 2023 comp season bot “Centrifuge” needs some maintenance and has room for software upgrades.
CAD
CAD is almost complete other than some minor details. Let me list the highlights of this concept.
Side rollers and a long stick for shooting outside the community zone.
Wide range of motion for scoring mid and high.
1 meter ~ (39") long CF tube (The picture does not do justice on how long it is.)
Despite the claw and the stick are sharing the same axle, they move separately.
I don’t want to bore you with more for now. I’ll try to update this thread as close as possible to real time. Thank you for joining, and hope you have a nice day. Feel free to ask questions and we are all open for suggestions!
We were about 8 members total , so I am leaving this for the background. We had many things to do since we got many parts to work on before the remaining processes.
To be honest, the day was tougher than I expected. At least the new students seemed very well mannered and cooperative. Everyone was easy to work with and the faces were happy most of the time. Next stops are going to be welder, laser and anodizer?
Until next time, thank you for joining, and hope you have a nice day.
Next stop was laser. We were unlucky by arriving just after they started a long cycle. We ended up having lunch with our sponsors. Then everything went smoothly, and we got the parts.
We made a last minute change and had the electronics panel cut from 2mm (~0.08") steel instead of 3mm (~0.12") aluminum.
Reasons:
New thickness will work better with standard bolt lengths.
We probably need more weight at the base.
Material came for free, and we saved from our 7075 stock.
Since we can’t anodize steel, and are too lazy to go for powder coat just for one part, we will probably spray paint this.
The parts were supposed to be taken to the anodizing plant before the end of the shift. But we had to visit my home to clean the parts. Was not an ideal environment for the job, but desperate times call for desperate measures. Deburring parts in a shower cabin felt tragicomic.
Thankfully, we were on time and we called it a day after we dropped the pieces. We will start to build after parts get ready. Until next time, thank you for joining, and hope you have a nice day.
Do you think the structure (doubled up 6mm lexan plates, bearing blocks) will be strong enough to take impacts? We’re in a never-ending quest to find “the perfect reliable intake material”, so I’m interested in your experience.
I color coded the connected pulleys. Basically, we’llLimelight, an integrated vision coprocessor control the orange shaft with the yellow NEO 550. Blue pulleys will get connected with 860mm HTD5M belts. (Outer left one is hidden behind the structure.) (The reason behind the orange shaft being in the back is it would block the cubes’ path if it was in front.)
Last shafts will have the red gears which will drive the racks on the inner lexan plates. (No rack on CAD for now) (254’s comp intake was the inspiration)
We have no idea for now. Personally the lexan rack scares me a bit, but I think it will force us to utilize optimal current limits, so I see it as a win. I suggest you would try to reach 254 or some other teams which used similar intakes for further comments if you are really interested with this kind of a design. (we did not, i amAndyMark just being silly)
Called it a day since it was getting really late. Next tasks are the intake sliders and the claw’s pivot gearbox. Until next time, thank you for joining, and hope you have a nice day.
Get the intake sliders and claw’s pivot gearbox mounted.
Check what’s going on with the loose tooth on “Centrifuge”.
Sliders & Pivot
Sliders were easy to install. (They need to be one hole behind, going from memory is not the way. ) We were unsure if we could use loctite on delrin, so we went dry for now.
Pivot gearbox seems okay and runs smoothly. The only problem is that there is an enormous backlash between the hex shaft and the 74t gear. It still is pretty usable, but we wander around the idea of using a tape like thing to tighten the fitting. (Open for suggestions, I know that many teams have tricks for such problem.)
Back to the Mystery
After our last event (RoboCube Off-season), we realised that one of the Falcon shafts was acting like a loose tooth. It was easy to wiggle by hand. This motor was one of the drive motors on MK4 modules. Our initial thought was motor cap bearing failure since we were able to drive “normally”.
This is the last sprint at the last event’s last match. We scored 3 game pieces on auto, and teleop seemed normal. Bot is resting at the shop since that moment.
I have no idea on how, don’t wonder much either since I amAndyMark tired of BLDCbrushless direct current motor problems, but this is the story. Until next time, thank you for joining, and hope you have a nice day.
After assembling the intake plates and rollers, our first observations were:
Sliding action is not smooth since we had to hand file the rack on lexan. (Our CNC router sponsor couldn’t use an end mill as small as necessary.) We will go for water jet if movement doesn’t get better with some usage.
Assembly is very rigid against side loads. Much better than what we expected.
Lexan’s surface is not ideal for gripping the cubes, so we zip tied a mouse pad I had for further testing. This model has a shiny non slip latex like bottom surface. It is more than ideal if we glue it on our rollers, but I couldn’t find the same model or this kind of a mouse pad online. If you have a similar thing laying around, I would like to know the model name. If we can’t find something similar we probably will use rubber sheets with glue. (We also are open for other suggestions.)
Swerve
While all these were happening, crew disassembled the MK4 modules on “Centrifuge” for cleanup.
The following days are looking to be about electrical. We need “Centrifuge” to be ready as soon as possible since we have an exhibition where we will be demoing the bot. Until next time, thank you for joining, and hope you have a nice day.
Surface speed of roller at least 2 * robot max speed
If driving towards an object at max speed, you still need to be able to have net positive intaking speed
The 2* or more is needed because with a single top roller, the object will be rolling, so middle of object has speed V getting pushed, 0 at carpet (assuming in worst case no slip condition) and 2*V at top of object
Surface speed = motor free speed RPM * 80% (efficiency) * gear ratio * roller circumference (length units) * conversion factors for units
Maximize grippiness (contract friction) through all methods to maximize energy transfer, and thus intake speed and reliability
Friction = Coefficient of Friction * Normal Force between objects
Find the most grippy pair of materials, such as rubbers or fancy grippy tapes, to increase coefficient of friction
Have enough normal force to lift object against gravity and pull in quickly
Normal force needs to be generated between roller-object-ground, roller-object-bumper, or roller-object-something
If object is rigid (examples: 2019 hatches, 2018 cubes, 2017 whiffle balls) you need compliance/compression either in the mechanism (pivoting point allow floating wheel/roller and springs/pneumatic-pressure generates squeeze force) or in the wheels/rollers (flex wheels, thick rubber surfaces).
If object is compliant (examples: most balls, 2023 cubes/cones) then you can have a rigid roller and instead compress the object
Dont have too much normal force because it increases resistance/drag which can slow system down and make it drawing more power, this is why maximizing grippyiness through material selection is optimal
Maximize intake area to simplify driver/autonomous skill required
Ideally intake is as wide as entire robot, which usually means its better to pull object over bumper and then center/serialize inside robot
Also means intake has to deploy outside bumper perimeter which leads to point 4
Build robust intakes
Use 1/4" polycarbonate plates for good mix of stiffness and toughness (bend then spring back without cracking)
Big plates with plenty of material around holes. Ideally small holes, thats why the deadaxle roller design is so good, you only have a bolt hole in the plate
Have spares and design-for and practice quickly repairing
Use sensors
Hall effect sensors to zero motor driven deploy
Beam-break sensor (ex. retroreflective laser) to know that object is in robot and automatically trigger intake to stow, elevator to move, etc
Color sensor to detect multicolored game objects (ex. 2022 balls)
i’ve never run a beam break/hall effect for an intake before, is there any huge advantage vs current sensing on serializer/indexer to detect ball capture or home?
Not every robot design has an indexer that spins constantly, so there’s no current to measure to detect gamepices. Current readings are also ambiguous, many indexer interactions can produce higher currents, such as jamming (including many forms of mechanical antijam) or the indexer straight up being full.
One more thing to add, ensure there’s something in place to handle head-on collisions. On many intakes, a head-on impact will try to retract the intake which can put a lot of force on the pneumatics or motors used to deploy the intake. 1/4 polycarb won’t save you here, as the plates are loaded in compression rather than bending and a significant amount of force will be transferred straight through the plates before they start to buckle
This portion lives in my head rent free for 7 years after seeing “Missfire” in my rookie year.
Thank you for your time and suggestion, this really means a lot for us.
It is an immense pleasure for us to receive these words from you guys.
One example would be to make a 4bar to deploy the intake such as our 2022 intakes.. If doing a single pivot, look at 1323 and 4414 from both 2022 and 2023 for examples of where the pivot is versus where the roller is when deployed to see how the roller is not directly inline with the pivot so the whole intake will stow when hitting head-on.
The design of the actuator also matters. If using pneumatics, position the cylinder(s) such that they are not the hardstop to the max extended/retracted states. Instead, the polycarb plates should hit each other or the bumper or something like that to be the hardstop. You can see this plate design in our 2020 intake pretty easily.
If using a motor, I recommend chain for the final output stage, ideally #35 but 25 is ok if multiple runs because driving from both sides. The chain works great to absorb impact loads from a collision and prevent gear teeth from breaking. Belts can also work just need wide enough belts.
Also if using a motor, you can do fancy software to make the motor act like a spring and not be in an ultra-hold-position mode. I’m no software person, so perhaps others can share their implementation. You’ll see it when we release our 2023 code this Fall for our Laterator and Floor intake linear degrees of freedom.
Finally, as examples of what you need to do for robustness if you don’t do flexible polycarb: look at our 2016 and 2017 intakes. Big aluminum plate and tube. Lots of reinforcement. A metal tube to be a “beater bar” and protect the roller because the intake couldn’t collapse in a head-on collision, etc.
We are in some kind of an awkward phase for now.
, so I am leaving 
