FRC Control System Future Direction

I know I saw a post on this somewhere this past few months, but I can’t find it.

I remember reading a post that a RFI was put out from FIRST for an updated control system and that NI was selected to continue with the roboRIO for the next like 5 years.

IF someone has a reference to this information, could you please post it here?

In making plans for purchasing new components, we don’t want to get stuck with a brand new controller only to have FIRST change it next year :slight_smile:

Thank You!

https://www.chiefdelphi.com/forums/showthread.php?t=164767&highlight=FRC+Blog+roborio

“RoboRio technology” extended through 2020 and 2021.

Hopefully HQ is starting to look now into ways to make the control system cost a fraction of what it does now. It’s a shame that every team doesn’t get a full control system every year in the “KOP”.

So the VEX V5 Robot Brain is $250 (vs. $435 for the roboRIO at the FRC team price). The full V5 control system is $450.

The questions:

  1. Is $250/450 really a the right price for that? We don’t know VEX’s margins on either, though I’m sure it’s profitable. But if they were in a different world where they weren’t selling all those other related parts that require a lot less development, I’m not sure what the price would be for a comparable controller.

  2. If someone’s suitable-for-FRC robot controller could retail for $250 (or system, at $450), is there a configuration of price and terms where FIRST could get one of these in the KoP every year? I imagine few companies would be interested in donating such a device annually for free, and if they’re not doing it for free then the funding has to come from somewhere.

  3. Is that even a relevant thing to want to pursue anymore? As much as I hear people chirping that FIRST should reduce or eliminate the Kickoff Kit and emphasize PDVs and FIRST Choice and other concepts, maybe there are teams that are okay with 1-2 robot control systems and don’t need six or eight?

The expensive part isn’t the hardware (although that can be plenty expensive if you use certain components), it’s the software support. You could probably run an FRC robot on a $50 computer board. The PDP/PCM/VRM are unlikely to become cheaper.
I would say having 2 control systems is good, and having 3 is great. That’s only around 2-3k over 2-3 robots. Not a bad deal, overall.

We have always used our First Choice points to get the non Rio parts of the control system. (PDB, PCM, & VCM) That still leaves the radio and Rio to buy if you want a full system.

I agree that cheaper would be better. What should be deleted from the current controller to reduce the price?

I’d be interested in pointers to $50 boards that have the reliability (conformal coating, resistant to shorting, etc.) of the RoboRIO, and the level of hardware support that NI provides. Can you share a link to some examples with pricing?

Cool, so it only costs $570 and the hopes and dreams that you get what you need from FIRST Choice to get a full control system every year?

FYI, we tend to do the same thing you do.

The logo.

I’ll add that the product needs to have acceleration and impact protection, and the connection points need to be resiliant.

Proponents of less expensive hardware also need to keep in mind that whatever is used must have a non-user changeable safety system built into it that is FIRST specific. That means some level of manufacturing / design support from the providing company, along with trouble shooting and other services.

I keep coming back to the Hero Board. With some modifications that would be a very interesting candidate.

M12 connectors would be amazing, but they are quite expensive unless you get Chinese ones.

A thought - would it be theoretically possible for FIRST to provide a standard or specification for how the safety system would work, leave it up to teams to implement as they saw fit, and then have a part of inspection be to validate the behavior?

My thought here would be to open up the controller selection to be fully team-driven. HQ could provide a reference design, and maybe some crafty suppliers could sell it COTS, but ultimately advanced teams could do whatever they wanted hardware-wise as long as they comply to some specific requriement of safety operation?

As I type this, a certan car company with diesel emission problems comes to mind…

I’m quite fond of mini xlr connectors (2 to 6 pin) for my industrial applications because they’re built to withstand the abuse of concert roadies. They’re probably a little big and expensive for FRC though: Search results for: mini xlr XLR Connectors – Mouser Canada

How about the up to $13k a year my team spends in registration fees?

At those prices and the quantity you’d need, the resulting control system would end up being more expensive that then current one! :slight_smile:

Maybe they don’t need to remove anything. FIRST could instead open up options for different suppliers to create their own robot controllers (with restrictions*), which would have two effects:

  • Teams can now pick their own controller based what features they prioritize (more sensor ports, more processing power, lighter weight)
  • More suppliers means that each controller will have to be competitively priced, now that teams aren’t forced into one specific one.

Of course, the KOP system will need to be tweaked a bit so teams would be able to choose what they want before the season.

*Restrictions like requiring certain safety measures, compatibility with WPILib, and making sure that suppliers don’t lock teams out of competing hardware

Conformal coat is quite literally less than a dollar to add to the production cost of a board. If you wanted to do it post-production, it would cost you 5 minutes and a bucket of lacquer with a brush. The Roborio also regularly has permanent hardware failures that result from shorting, so it’s not all that safe to begin with. Customer support is important, but I personally haven’t had any particularly good or bad experiences with them (except for them stating that they don’t RMA some hardware failures when we faulted a RIO power rail some months ago, which is understandable). Can you elaborate on what kind of “hardware support” you mean?
In any case, my point was that the computing power you need can easily be matched by cheap computer boards. We’re not doing anything particularly computationally intensive on these robots other than vision processing, which is usually run on a coprocessor anyway. In fact, 971 ran most of their robot off a BeagleBone Black back in 2014, before the RoboRIO came out. Coincidentally, it sits around $55 right now, and I don’t doubt that there are suitable alternatives on the market right now. It’s also easy to configure a HERO Board to run an FRC robot, what with its built-in CAN protocol and fairly good processor. This past year we did almost no processing on the RoboRIO, and had Talons slaved to run all the control loops instead.

The thing I don’t like about the HERO is its reliance on Visual Studio and C#. It makes being a Kotlin fanboy harder :wink:

Also, I remember back in early 2018 there was some speculation about a RoboRIO with an integrated AP/Radio, I personally highly doubt FIRST is going to stop using NI anytime soon.

I would imagine that if CTRE were to submit the HERO as a future option they’d expand the language list as part of that proposal.

I’ve never seen that speculation. Can you link to it?