ausTIN CANs 2158 | 2023 Build Blog

[AndrewI]

Day 6 Prototyping and Testing

It was a day dedicated to testing and assembly of prototypes! Which is great because that means we are learning what works and what doesn’t work. Today we assembled the 2 jointed arm prototype to prove that the geometry works. It is just able to reach the high goal it will get a little longer reach to more easily reach the high goal.

Low Goal

Mid Goal

High Goal

Inside Robot

Next we spent more time testing the weed wackers. One thing we observed with the weed wackers was the speed the chassis is moving relative to the speed the weed wackers are spinning matters a lot. If the chassis is moving too quickly the cone doesn’t get enough hits to center or end up in the correct orientation.

Chassis Too Fast

Next we tested up right cones. This didn’t work great most of the work was done by the wood center beam. If we used this idea we’d have to have a passive bar ahead of the weed wackers to knock over the cones.

We also tested the cone against the field wall to see if we’d be able to intake it still. This worked better than expected the weed wackers were almost able to get the upright cones even. Once again chassis speed played a big role, if pushed too quickly the cone doesn’t end up in a repeatable location.

Probably Too Fast Here

To end the testing before we started trying to improve them we tested weed wackers with the cube. The results from this really surprised me. I’d expect the weed wacker to knock the cube away. But instead it pulled them into the center.

After the marathon of testing the weed wackers we set out to improve them. A basic cardboard funnel is being created and we are increasing the distance between the two axis’s.

Another group started creating dual overhead rollers to test intaking of cones. After quickly clamping the prototype to a dolly we started testing the cones in different orientations.

After a few tests we noticed a failure to grab the cones sometimes. We ended up lowering the prototype so it was closer to the ground, this fixed the problem. The prototype handles the worse case of the edge of the cone being perpendicular to the overhead rollers amazingly.

In the current configuration this prototype doesn’t work with cubes. There isn’t enough compression between the two rollers. We will be increasing the compression and testing again on Saturday.

This prototype worked amazingly well. It isn’t as sensitive to the chassis speed. It does however leave the cones facing front or back and any where along the length of the intake rollers.

Two Possible Cone Locations

In order to score the cone we need to put the bottom hole side of the cone onto the pole. With a single additional powered point of rotation on the dual over head rollers we can get the cone facing in a direction that allows us to score it.

Cone Rotated

The other problem to solve is how to handle the unknown position of the cone along the length of the rollers. The current plan is to add time of flight sensors used to detect which way the cone is facing and the location of the cone. With the location and direction of the cone we can construct an offset to add to the robot’s pose(gathered from april tags or reflective tape) that will allow us to place the cone on the pole. The other thing that makes this possible is doing swerve drive, the robot can easily move in any direction and move to the required pose to score. For the first time in a while we are creating problems for software to solve instead of solving them other ways.

The current plan requires the dual overhead rollers to be placed on the arm. This will be an engineering challenge to make them robust and light enough. There also might be a way to make it work with out having the intake on the arm, we’ll investigate this idea more at our next meeting. Our next meeting is on Saturday we hope to finish out the alliance grid, start cutting charge stations parts, more testing and prototyping.

[Ry3512]

What are the dimensions between the two spinners? We are trying to replicate this prototype and have it orient the cones the same way.

[AndrewI]

For that gif they were space 8.375" apart. That seemed to work alright. There might be a spacing that works better, we will be testing different spacing between the weed wackers tomorrow.

[Ry3512]

Is this centric or from the outside of the roller?

[AndrewI]

Ideally you want greater than the cone’s width, 8.375", spacing between each roller. So 8.375" plus some wiggle room on the inside. Don’t forget to take account any room that the hubs or what you use to mount the tread to take up.

[AndrewI]

Day 8 Prototyping

To start the meeting we started working down the list of improvements to try with the weed wackers. The prototype moved off the dolly and onto a kit bot with bumpers(some old bumpers that didn’t quiet fit but got convinced with a hammer). The first change was increasing the spacing between each axis and longer tread. This didn’t seem to change the dynamic at all which was good. The longer tread is a different style of tread which is significantly “floppier” it still worked but I think ideally we’d want to move back to the blue nitrile tread.

The next item on the list to test was adding a ramp. We played around with different ramp configurations but we couldn’t get any consistent results with two weed wackers. More weed wackers might push it up further or a less steep ramp might work better. This is probably as far as we’ll take the ramp idea.

The next idea that was tested was adding a bottom roller to see if that would help at all. We didn’t have a strong idea of what this would do but it was simple to try. The under roller makes everything less controlled and it looks like the robot is trying to juggle the cone. Don’t recommend.

Dual Over Head Rollers

The dual over head rollers got the space between the compliant wheels reduced from .5″ to .25″. This allows us to grab the cube between the rollers. If the power on the drill is let off the cube pops back out. We were careful not to use full power because we didn’t want to harm another cube.

The next iteration of the dual overhead rollers had great progress made today. This iteration uses two PVC rollers repurposed from our 2020 robot. Not even winning robots are spared from prototyping(it’ll get put back together once we learn from the rollers). This version also has two motors and the pivot that allows it to rotate 90 degrees in both directions. We worked on this a good portion of the day and I forgot to grab any pictures of it. You’ll have to wait until tomorrow to see it I guess.

The next idea comes from a reader of our build blog who messaged me and I’d like to thank them for it!

Talking and showing our work in the open created this idea. I’m not sure we’d have come up with this idea on our own. The next thing we started testing was combining the weed wacker with the dual over head roller. The weed wacker will center the game object and the dual over head roller will grab and control it. Now instead of the cone being in 2 orientations and an unknown position it will be in one smaller location and one orientation.

This prototype worked amazingly well. I was surprised. It seems to be able to intake cones in any orientation. Chassis speed is still a slight concern because we are using drills and can’t increase the speed enough relative to the chassis speed. We’ll attempt to solve that on the next iteration of the prototype which will include motors, a change of spacing to fit in our frame size of 26″ x 26″, and it’ll be driveable!

There were a few odd failures that happened while testing. We hope the next iteration will be much more consistent with motors powering everything and increased rigidity. I’d also love to test the difference the stiffer blue nitrile tread makes.

Want Your Own Weed Wacker?

If you’d like to join team weed wacker I have some measurements taken from our prototype that you can use to start your journey. These numbers aren’t perfect and I can’t guarantee what you will build will match the results we’ve had. As always you need to do your own testing, debugging and improvements! Let me know if you need any other measurements and I can grab them at the next meeting.

Dual Over Header Rollers

• 4″ diameter 1″ wide Green Compliant Wheels 10 per axle(need two axles)
• Floor to center of axle 9″
• Axles spaced 4.5″ apart
• Distance between first roller and weed wacker 9″
• Back roller was 5″ from the kit bot frame
• Rollers are 10″ wide

Weed Wacker

• Arm 17″ from bumper (20″ from kitbot frame)
• 11.5″ from floor to top bearing plate of weed wacker
• Weed wackers spaced 15″ center to center
• Bottom of weed wacker’s hub is 6″ from the ground.

Hub assembly

• Hex shaft 8″ or so
• 2 hex hubs
• 10 – 32 bolts and nuts to hold the hubs together
• 13″ of 1″ tread(blue nitrile seemed to be less floppy and possibly better)

Field Build

Our laser cut actual charging station parts today! We had to kick everyone out of the shop for an hour so we lost a bit of productivity but that’s fine. We cut the first of three sheets of 1/8″ aluminum for the charging station build. The next two sheets will hopefully be cut on Monday. We also completed our first grid section today. The other two sections have all of the pieces cut and just need to be assembled now. CAD for the frame rails was released today and the stock was cut down to size. Our frame dimensions ended up being 26″ x 26″.

This was screwed together I just forgot to take another picture….

First Charging Station Part

Next team meeting is Tuesday, can’t wait!

[roboticsqueen]

What were the measurements for the post placement? Directly centered?

[AndrewI]

This picture could clarify any measurements. I am thinking by “post placement” you are referring to weed wacker axles? If so yes we tried to center the gap created/between the weed wackers.

[Osby]

I might be not seeing things correct in the video. Does the tread on the weed wacker violate the frame perimeter rules posted in update #2. See G109h.

[AndrewI]

As it exists currently on that robot it would be in violation of that rule yes. It’s currently on a spare kit bot chassis for testing purposes. Tomorrow we’ll be shrinking it to fit with in our planned frame perimeter of 26" and testing everything again.

[AndrewI]

Day 11

To start the day we tested the dual over head rollers which are using PVC rollers from our 2020 robot. The PVC rollers didn’t work well, the first roller doesn’t seem to have enough grip to pull the cones in.

We decided to take a small break from our weed wacker prototype to explore using vectored intake wheels. Our team saw open alliance teams having great success with them. If we are able to center the cone reliably with out weed wackers our robot gets simpler and lighter.

Our first attempt used the dual over head roller prototype, the first roller was changed out for vectored intake wheels. Unfortunately for us the spacing on the first iteration was too tight, with a .25″ gap between the vectored intake wheels and the second roller. Too much compression means the cone didn’t center well. In the below gif the wheels were also on the wrong side which vectored the cone to the edge of the robot.

The second attempt used a different test bed with .5″ spacing between the vectored intake wheels and the back roller. This had slightly better results, it took in the cone easier but there was too much spacing between the vectored intake wheels on the shaft. This allowed the cone to get stuck between wheels. Tomorrow we will increase the number of wheels and play with the distance between rollers to see what works.

The buddy carry prototype got assembled on the swerve drive chassis today. We ran out of heavy and easy to secure things to add on the swerve drive chassis. Tomorrow we’ll find more heavy stuff to strap down and attempt to move around with a full weight robot on the forks. Our plan is to test driving up the charging station with a partner.

Charging Station

Tonight we attempted to bend some of the aluminum parts that we cut out on our laser over the weekend. This didn’t go well, the pieces kept breaking during the bend. There is some speculation that we might have gotten/used the wrong alloy for the first sheet of parts. We will be cutting a smaller piece out of our recently purchased(known alloy) to test tomorrow and see if we can bend it. Oh the joys of doing something for the first time during build season.

[eterna1c0d5]

How did you get the tread to stay on the shaft? Are there just screws through the tread and the shaft?

[AndrewI]

For the prototypes the tread is wedged behind the bolts that hold both hubs together. We haven’t invested time into finding a more permanent solution yet.

[AndrewI]

Day 12

Field Build

The alliance grid is almost complete. The last three grid pieces are assembled, they need to get bolted together and moved into place. There was good news on the charging station front today, we bent a small aluminum test piece and nothing broke. The charging station build is now full steam ahead again, the poly carbonate parts are being cut on the router. We are still waiting for our metal supplier to source us 1.75″ x 1.75″ aluminum square tube(which is taking surprisingly long).

Continued Prototyping

A group continued to test vectored intake wheels and their possible use to center the cone from the flange. The spacing is the biggest challenge right now. Too much space between the rollers and it won’t pick up the cone well enough, not enough space and you have too much friction to center cone. They tested .75″ of compression first, this had too much friction. They removed the grip tape on the PVC roller and retested. Still too much friction. Next was 1″ this was barely able to grab the cone, not enough compression.

Our next step is testing different spacings between .75″ and 1″. They will also be adding more vectored intake wheels to prevent the cone from getting stuck in the gaps between wheels. Another thing to try is adding a slippery material on the back PVC roller to reduce friction.

The buddy carry prototype today concluded their testing by carrying our 2022 practice bot up the 11 degree ramp(poor substitution for charging station). The practice bot didn’t have bumpers or a battery so it only weighs around 70 lbs. From a slow start the swerve drive chassis struggled on the lip. A little bit of speed conquered the lip and the two robots drove onto the ramp.

We are happy enough with the results currently that we will be starting to make the first real iteration of the buddy lift. We don’t expected to get the first iteration perfect. There will be a few problems that will pop up and we will need to solve but we’ve learned a ton from this prototype and its time to start refinement.

[jedgar6]

@AndrewI

To think I scoffed at a shifting swerve.

[chrisfl]

What were the dimensions for your smaller weed wacker? We have been having great success with the design and need to begin prototyping a compact version that fits the frame perimeter.

[AndrewI]

Glad to hear about y’alls success! The only weed wacker that we built that is smaller is the first version we made and we didn’t take a lot of measurements of it. The only real change was the spacing between the weed wackers. That was smaller at ~8" across(cone’s base didn’t fit through easily). I’d recommend changing that number to be larger than the cone’s base, maybe try 8.5" and see how it works? The other change would be the length of tread on each weed wacker. With 8.5" between the weed wackers I think 9.5" of tread per weed wacker would be a good starting number.

[jbarry987]

How do you address the pick up of cones parallel to your top rollers?

[AndrewI]

Day 13

Assembly and Field Build

The competition robot assembly began today. The swerve modules were mounted to the new drive rails. Next the frame was squared up and secured. The top brain pan and bottom belly pan were cut on the router and are ready to be installed. Electronics will start going in on Saturday.

The alliance grid was finished tonight. Manufacturing will be working on the cable protector and barrier next. Charging Station parts are getting laser cut on Friday.

Scheduling and Prototyping

Today we had a discussion in the CAD team about the schedule and where we stand. Currently CAD zero subsystems complete. If we were to unrealistically say we’d get CAD complete and ship the first version by next week it would take manufacturing about two weeks to complete the parts and assemble. That would put us around week five for a functioning robot. We have a week one competition so we need time before then to program and drive the robot.

It is impossible to keep all of our same goals for the first version of the robot and complete CAD in a week. So we talked about which items we could remove from the first version of the robot. The current plan is to finish the buddy carry, and the arm for the first version. The team agreed to hold off on refining a floor intake for the first version.

We plan on making use of a released intake(not floor intake) for the first version. We will continue to prototype a floor intake but it won’t hold up progress on the entire robot. This will allow us to get a robot to programming faster, and allow us to start practicing sooner.

With our schedule cleared up the prototyping group continued to test vectored intake wheels. They increased the number of vectored intake wheels on the front shaft. This improved the prototype significantly. It was able to some times center the cone now. The position of the back roller still causes too much friction and the cone will get stuck sometimes.

Next we took an old prototype and put it on the table to simulate replacing the back roller with a smooth surface. This worked amazing well to center the cones. On top of the table is a piece of poly carbonate to reduce friction. This worked well enough so we’ll probably end up exploring removing or changing the location of the back roller more on Saturday.

That brings us to the next problem we’ve discovered with the vectored intake wheels. Currently our prototype requires the game piece to be close to parallel to the robot in order to intake.

This is a significant down grade in performance over the combined dual weed wacker or even the dual overhead rollers. The combined dual weed wacker was able to intake cones in all orientations besides tip pointing towards robot.

We’d like to see this performance with the vectored intake wheels. So we are exploring ways to increase the angle the vectored intake wheels can intake. On Saturday we will be trying 4″ vectored intake wheels to see what that changes. Saturday is a long meeting and we have a big list of things to try. See you afterwards!

[Potttaaatttooo]

What type of vectored intake wheels did you guys use for these vids? TTB ones/3DP ones (where theyre all plastic) or do they have rubber rollers. Also what material was the back roller as well? Is it just PVC?