Team 67's incredible shooter

[whophil]

I just stumbled across this very informative thread. Adam, have you guys experienced any problems with side loading of the 550 motors? I'm looking to set up some Banebots RS-775 motors in a similar belt drive configuration with the pulley cantilevered on the shaft, but had some doubts about how the motor's internal bearings would hold up over time. (The 775 motor only has a 7.6 mm long shaft, which is another problem in itself...)

Cheers,
Phil

[Adam Freeman]

Originally Posted by whophil

I just stumbled across this very informative thread. Adam, have you guys experienced any problems with side loading of the 550 motors? I'm looking to set up some Banebots RS-775 motors in a similar belt drive configuration with the pulley cantilevered on the shaft, but had some doubts about how the motor's internal bearings would hold up over time. (The 775 motor only has a 7.6 long shaft, which is another problem in itself...)

Cheers,
Phil

No problems to date with any of the motors we used on either the practice bot or the competition bot. We tried to take special care to not over load the motors with excessive tension on the belts.

I would guess that a 775 would be even more robust to a setup similar to ours.

[DampRobot]

Originally Posted by Adam Freeman

Our drawings and CAD models are so crude, it would be pretty embarassing to show them to anyone. But, I'll see what I can find.

-Adam

That's very interesting to me. I've always seen detailed CAD as the mainstay of every highly competitive team, and I'm interested in how a multiple time Einstien winner can perform extremely well and not do a detailed robot CAD.

For instance, how did you know that your climber could grab levels correctly, that nothing would break under stress, that screws would clear other parts, that there would be enough room for electronics or other mechanisms, or that you'd be under weight? Frankly, I don't think I could design a robot without detailed CAD. Things are so many times different sizes than we think them to be that simply fitting everything together would be a huge challenge, at least for me.

Can you shed a little light on your design process?

[Adam Freeman]

Originally Posted by DampRobot

That's very interesting to me. I've always seen detailed CAD as the mainstay of every highly competitive team, and I'm interested in how a multiple time Einstien winner can perform extremely well and not do a detailed robot CAD.

I would say that a detailed CAD process is the rule to follow, but there are always exceptions. When it comes to our design process, we are not the model that should be emulated. The designs that come from our process...maybe. But, the actual process...no!

We have two main designers, myself and Jim Meyer. Our robot is designed 99% in 2D AutoCAD. I sometimes use Solidwork to model already designed parts, but almost never for the design process itself.

Our 2D drawings are very much like very detailed sketches. Almost all the components are sketched out in an assembly view, but the very fine details are usually put in the print sent to the machine shop, and not in the assembly view. So our CAD files are almost impossible to decifer unless you are the designer currently working in the sketch.

Since we have full access to a water-jet machine, we try to design the majority of our parts to be manufactured that way. Since we only need a 2D file to export to the water-jet, it works out that we do all of our design in AutoCAD.

From a strength or weight analysis, our process is very experience oriented. We are confident in our abilities to estimate how small or large a part needs to be to survive in a FIRST environment. Along the same lines, we try to re-use similar designs for the frame, wheels, etc... so we usually know if a part or assembly is going to put us over for weight.

This style of design relys heavily on the designers to be completely involved in the manufacturing and assembly of the parts. Since they are really the only ones that know exactly how the parts are designed to work together.

As I said above, it's not a process that should be copied by anyone. We have made attempts to streamline the process...or add in aspects of other peoples design process (148, 1114, 254, etc...). But, every time when things start getting tight for time during the build season, the process reverts right back to where it has always been.

At this point, we have pretty much accepted that, like it or not, this is who we are...and how we operate.

The time we have to work on our designs seems to be getting less and less each year (between family and work responsibilities), so we don't do much more than is required to get the parts machined and assembled.

We are in the process of adding some more mechanical/design mentors that hopefully will help improve our process and provide more time to further enhance our designs and be able to teach more students/mentors how we create our parts.

-Adam

[Akash Rastogi]

Originally Posted by Adam Freeman

I would say that a detailed CAD process is the rule to follow, but there are always exceptions. When it comes to our design process, we are not the model that should be emulated. The designs that come from our process...maybe. But, the actual process...no!

We have two main designers, myself and Jim Meyer. Our robot is designed 99% in 2D AutoCAD. I sometimes use Solidwork to model already designed parts, but almost never for the design process itself.

Our 2D drawings are very much like very detailed sketches. Almost all the components are sketched out in an assembly view, but the very fine details are usually put in the print sent to the machine shop, and not in the assembly view. So our CAD files are almost impossible to decifer unless you are the designer currently working in the sketch.

Since we have full access to a water-jet machine, we try to design the majority of our parts to be manufactured that way. Since we only need a 2D file to export to the water-jet, it works out that we do all of our design in AutoCAD.

From a strength or weight analysis, our process is very experience oriented. We are confident in our abilities to estimate how small or large a part needs to be to survive in a FIRST environment. Along the same lines, we try to re-use similar designs for the frame, wheels, etc... so we usually know if a part or assembly is going to put us over for weight.

This style of design relys heavily on the designers to be completely involved in the manufacturing and assembly of the parts. Since they are really the only ones that know exactly how the parts are designed to work together.

As I said above, it's not a process that should be copied by anyone. We have made attempts to streamline the process...or add in aspects of other peoples design process (148, 1114, 254, etc...). But, every time when things start getting tight for time during the build season, the process reverts right back to where it has always been.

At this point, we have pretty much accepted that, like it or not, this is who we are...and how we operate.

The time we have to work on our designs seems to be getting less and less each year (between family and work responsibilities), so we don't do much more than is required to get the parts machined and assembled.

We are in the process of adding some more mechanical/design mentors that hopefully will help improve our process and provide more time to further enhance our designs and be able to teach more students/mentors how we create our parts.

-Adam

In a way, this is one of the coolest posts I've read on CD.

I definitely agree that detailed 2D design work can be used to figure out the geometry of a design (I always think back to the 2D drawings of Simbot SS) but doesn't it get difficult to make iterations, even with a practice bot and the district model?

Thanks for the insight!

[DampRobot]

Thank you very much! A very interesting read.

[R.C.]

Originally Posted by Adam Freeman

Our nominal pyramid and autonomous shot was around 20-21 degrees. Any time we lost our position of the potentiometer, we would use a digital inclinometer to adjust the autonomous shot back to 20 degrees, then make fine adjustments from there.

Adam,

The digital inclinometer seems really nifty, do you have a part number for this guy?

Thanks!

[Gregor]

Originally Posted by R.C.

Adam,

The digital inclinometer seems really nifty, do you have a part number for this guy?

Thanks!

In a pinch most smartphone platforms have inclinometer apps.

[Adam Freeman]

Originally Posted by Akash Rastogi

...but doesn't it get difficult to make iterations, even with a practice bot and the district model?

Yes and no. I usually find it's a lot easier to copy over a bunch of lines in an AutoCAD drawing and cut/trim/modify a current design to a new iteration that I do trying to adjust a solid model or assembly. But, that could be b/c I use AutoCAD a lot more than I do Solidworks. It does take a lot more concentration to make sure all the little details are carried over to new drawing or sketch.

Originally Posted by R.C.

...do you have a part number for this guy?

I'll have to stop by the shop and look at the brand we have, but it was very similar in design to this one.

http://www.amazon.com/Digital-Inclin..._sim_sbs_hi_14

If you didn't have one around, I would say a cell phone with an inclinometer app would be just as useful for the level of accuracy we were trying to hold. We just opted for the digital inclinometer b/c the shop had one we could borrow.

-Adam

[Brandon Holley]

Originally Posted by Adam Freeman

If you didn't have one around, I would say a cell phone with an inclinometer app would be just as useful for the level of accuracy we were trying to hold. We just opted for the digital inclinometer b/c the shop had one we could borrow.

-Adam

Yep. The trusty "Level" app on my iPhone was THE measurement device used to set our shooter angle. We used the same app on the same phone through our many iterations of shooter to try and get fairly accurate results.

If the shooter was set to the proper angle (27 degrees I think?) we had no issues with scoring.

-Brando

[evanperryg]

Originally Posted by Brandon Holley

Yep. The trusty "Level" app on my iPhone was THE measurement device used to set our shooter angle. We used the same app on the same phone through our many iterations of shooter to try and get fairly accurate results.

If the shooter was set to the proper angle (27 degrees I think?) we had no issues with scoring.

-Brando

You must have quite the level app on your phone. Actually, considering how accurate your bot was, that is really surprising. We used a phone level app to get our shooter dialed in, but it just wasn't accurate enough to get us aligned to the 3 pointer.

[Brandon Holley]

Originally Posted by evanperryg

You must have quite the level app on your phone. Actually, considering how accurate your bot was, that is really surprising. We used a phone level app to get our shooter dialed in, but it just wasn't accurate enough to get us aligned to the 3 pointer.

So this is kind of one of those funny measurement things. Do you really care what the absolute angle of your shooter was? No, probably not. What you cared about was A. the frisbee going into the goal and B. the shooter being in that same spot everytime.

For this, the app on my phone gave a very repeatable result for us. The actual angle it was measuring was probably not extremely accurate, but we didnt care. We knew what number the app spit out that put the discs in the goal. The consistency came from a very heavily iterated shooter design (we built 8 complete shooter systems over the course of the season).

-Brando