With a desire to keep several of our robots past the season for demonstration…but the high cost of supplying each with a complete control system…I want to suggest that the team design and build a “modular” control unit this season with the Creo, sidecar, power distribution board, jags, and spikes. We could then modify the previous robots to accept this module and be able to switch it back and forth. They would need some kind of multi-conductor plug and jack system - and I haven’t had much luck with reliability. Is anyone doing something like this? What kind of connectors are you using?
For signals, you could use something like a parallel port DB-25 to make a harness for that robot’s sensors and pwm connections. You would solder all the leads to the appropriate pins on the plug(s). Last season, we used an aircraft-grade DB harness plug for our shooter deck.
For the cRIO and digital sidecar, you would keep them plugged together with a regular harness set. For power, we’ve put Anderson Power Poles between the motors and motor controllers for ease of rewiring.
We’ve used velcro on Baltic birch plywood to hold down all our motor controllers, cRIO, and digital sidecar.
What you are asking is actually a fairly good design practice for a competition robot, even if you had extra control system in your inventory. The ease of access and serviceability will save you a huge hassle later on.
Plus, we usually create our drive base first, then we find that we have to cover everything up with a big towel when we drill or cut on a new superstructure. It would be nice to pull all the sensitive electronics out for protection.
Thanks for those suggestions…I’ll look into that.
Please remember that the Anderson Power connectors can be assembled into blocks and are also available in different colors. We have assembled a connector with more than 15 circuits in the past for removable appendages.
For the past few years, we’ve used Anderson PowerPoles between the electrical board and the rest of the robot. It lets us build the board separate from the robot, so it only gets attached at the very end of the build season with a couple of bolts and plugging the connectors together. This is also making it easy for us to swap out the board this fall as an Alpha Test team, so we can hook up a completely new control system!
One of the TechnoKats’ best-designed robots in recent memory indeed used a DB25 for sensor connections. That even let us use an extension cable between the control box and the robot a few times when we wanted to test some code while the robot was surrounded by people doing mechanical work.
A couple of blocks of Anderson PowerPole connectors were the disconnect for the motor power wires. If we do that again, we’ll make sure the person responsible understands the importance of using the little locking pins between them.
Here is a link to some of the Anderson Power Connectors we have used. You can order new pins (contacts) and reuse the cases on wires no longer being used to help offset costs for future builds.
The crimper that we use for these connectors is a great one.
You can also connect the blocks in such a way that they only fit in one orientation instead of using different color cases. This eliminates the “need” for stocking different color cases.
Interesting problem here.
Designing a control board (sizes, mounting holes ect) before kickoff to be used on a competition robot doesn’t appear to satisfy R16 per 2013 rules.
R16
ROBOT elements, including software, that are designed or created before Kickoff are not permitted, unless they are publicly available prior to Kickoff.
You can get around this by say, posting a CAD model on Chief Delphi, but be aware of it.
No one says that the board will be the in terms of layout and components. Rather if you know the old robot’s control board mounting configuration, you can design a new board that incorporates both the old mounting configuration and the new mounting configuration. It may even be possible to use the same mounting holes for both robots. I don’t think this violates the rule, or the intent of the rule.
Now if you were taking last year’s (old) board and installing it in the current season’s (new) robot, this would be an issue. Although, I see nothing wrong with you using an old board on the robot for testing purposes until you can get a finalized design for the new board.
My plan is to do exactly this…with a twist, Use the design project as a CAD training activity in November and December. Post the design for all to use, modify, improve - then adapt it as necessary when the new game is revealed. Hopefully we can come up with something that will work with the new robot, but still operate Logomotion and Ultimate Ascent - the two bots we’re keeping for demonstration purposes.
To fully conform to the rules: Your design can be published & reused. Physical parts (non COTS) have to be made during build season. So your mounting board can be used for development, but you should make a new board for competition.
Exactly…wouldn’t want to build this before kickoff anyway - it might have to be waterproof! My hope is that it can be backward compatible…not that something we make now becomes forward compatible.
My main concern was finding some way to create the modular connections that would allow us to plug it in and remove it easily and reliably. The PowerPoles look like the answer!
I’ll second Anderson Power Poles for power, but we’ve also been using Tyco CPCs (Circular Plastic Connectors) for signal connections with great success. You can see the catalog page here:
In the past we’ve used DSub connectors, but they required good soldering skills and weren’t very forgiving. They can be found cheap though, while the Tycos will cost upwards from $10 for connectors and pins. You’ll also want to get their crimping tool ($225).
Our design for the last two years has split the electrical board in two - a signal box (cRio, digital sidecar, etc.), and a power box, each of which can be removed quickly from the robot. The system still needs some work, but it has been a big plus for modularity, and showing students how things should be done in the “real world”.
One option is to use D-subs that work with ribbon cable such as the example below. I have always used a vise to gently “crimp” connectors such as these for the prototypes I have worked on.
www.digikey.com/product-detail/en/1658612-3/AFR15B-ND/825359
If I recall correctly, many of the cRio interfaces will work with connectors that also take the same type of ribbon cable.
Link not found…
Copy/paste, he used the wrong HTML to create a link that made CD think it was an internal link… or, here ya go
http://www.digikey.com/product-detail/en/1658612-3/AFR15B-ND/825359
We did it for the 2013 season, in the form of a box. We won’t be doing it again. Here’s why…
The idea was that we could easily take out the box to service it, and everything was all connected together. (We didn’t expect to use it for a different year) So everything was together, in one place, and you could take all the electronics out to run them independently if you wanted to. As we went through the competition cycles and design changes, problems arose.
Our design had literally all the electronics that could be in the box, in the box. Including the router. We had a total of 7 Victors and 2 Spikes inside, along with the small CRIO and Sidecar, a couple lights to make it look cool, and 2 fans for airflow. The signal light and power switch were mounted on top.
Problem 1: Where it goes.
We chose an area of the robot which ended up right under the shooter. We ended up having to cut a hole in the top of the box, as we snapped a few connector tabs on our shooter motor. And being under the shooter created another problem: we had to undo the shooter to take out the electronics. Kinda defeated the whole ‘easy to remove’ part. It also prevented us from having our shooter at the correct height to smoothly receive frisbees from a slot, as we had to drop them slowly from the 2nd slot instead of smoothly drop them in from the lowest.
The previous shooter to the one shown here’s motor was [i]closer to the top of the box. The motor was dipping into the hole you can see under the current shooter motors.
Problem 2: Wires on the outside
PowerPole connectors for motor wires weren’t a problem. We were used to using them from previous years. Sensor PWMs were another story. The solution, ‘David’s Problem,’ was incredibly tedious. (I swear they were working on it for nearly a whole 9-5 meeting) Using a cable similar to the one in the above post, we created slots for each and every DIO port, but nothing for solenoids, so we had to unplug/reconnect solenoids every time we removed the box.
‘David’s Problem,’ connector mounted to robot end.](http://i.imgur.com/XdCOHMo.jpg)
Problem 3: Wires on the Inside
Picture says it all
Problems with our design specifically:
-Main power switch was on top, and the wires were oriented so that you had to undo one of them to get into the electronics.
-The router being inside didn’t seem to cause radio problems, but it wasn’t happy with the inspectors. They want to be able to see the lights on the router. We had to cut out a part of our L bracket holding it together to make a window to appease them.
–In the revision for CalGames I moved the router on top and the power switch to the side.
The good (It’s a small list):
-Built in cover for the electronics in the case of drilling.
-Occasionally, but very rarely, we could have one group use the electronics box and one use the powerless robot frame separately. Didn’t really happen much though.
So my advice is not a box. For next year, I myself was hoping for something that might bottom out of the robot on our team; the box was pretty more trouble than it was worth, really.
Thanks for this…it is the kind of information I need to make sure our idea does not do the same thing.
Thus far, my vision was for more of a “plate” of sheet metal with the power board, sidecar, cRio, motor controllers, spikes, radio power converter, and radio. We planned to have all the pwm wires remain on this board so there would be no need to use any kind of data connector - only the power. I’ll post an image of our first concept as soon as it’s done - but I think it avoids most of the issues here. Oh - and we always have our electronics all together at the base of the robot, so the “where to put it question” is not really a question in our case. I picture having it slide out from either the bottom or between the wheels through the front or back.
http://farm6.staticflickr.com/5490/10494053993_b82fe59f1d.jpg](http://www.flickr.com/photos/hotbotz2640/10494053993/)
control panel by Hotbotz 2640, on Flickr
The “donut shaped” pieces are rubber grommets so the wires can be fed through and run under the main board for neatness and ease of maintenance. The radio will, of course, be velcro attached, so it can be easily relocated in a more conspicuous spot if necessary. The battery and the main breaker will be elsewhere: we always position the battery somewhere to fine-tune the center of gravity, and the main breaker needs to be somewhere accessible. Our actual pneumatics will also be elsewhere (compressor and air tanks) so there will be one data cable running to the pressure regulator. All the solenoids will be positioned at their prospective cylinder to simplify the air lines. Thus, a two conductor APP for incoming power, a multi-conductor (28) PowerPole to power all the motors and other devices, and a pwm to the compressor high pressure shutoff.
Isn’t this being a bit over-engineered? For what you’re trying to do, you can make a simple teleop rig and cut out high-cost components like the cRIO altogether. This would allow you to affordably keep up a host of old FRC robots.
I’ve given some thought to trying to cut a cRIO out of the control system using the 2CAN. At that point, you’d be sending motor commands to each Jaguar wirelessly and emulating whatever drive code you have on an off-board laptop. There’s a slew of bandwidth / control timing issues I won’t get into here that I haven’t fully puzzled out, but it’s a relevant idea I’ve been playing with.
If you don’t care about using FRC components, there’s probably an easier / simpler off-the-shelf solution, but its an interesting problem nonetheless.
Things like pneumatics and Spikes are a little simpler, you can just use wireless relays.
Given that the FRC control system is changing next year anyway, I’d advise against putting too much effort into building a modular rig and modifying old robots when you could just get them all working for a similar or lesser effort using non-FRC parts.