Lithion-Ion Batteries

[Jon Stratis]

Originally Posted by ToddF

I'm a fairly new mentor, so some of my thoughts may not be as seasoned as the veteran folks out there. I'm seeing some assumptions in some of the responses that I don't understand.

There seems to be an assumption that for something to be legal for use on a FIRST robot, 1) its performance needs to be completely characterized, 2) it needs to be supplied to all teams in the kit of parts, and 3) it replaces whatever components were previously legal.

Where I can see how this MIGHT apply to control system components running software, I don't see why it should apply to things like actuators. I don't see a downside in allowing brushless motors and drivers to be included in the list of options we consider. Maybe they aren't right for certain applications, but for others, they might be. Why not allow teams who want to use them to save money to do so?

AdamHeard rightly pointed out that the Jaguars (and Victors) are far oversized for the shooter wheels. So why do the rules limit us to using those and only those motor drivers? Why not allow teams to use other options?

Reading between the lines, it seems that at some point in the past, cheap motor drivers were not available, so in order to ensure a supply of reasonably priced drives, FIRST worked with suppliers to come up with custom motor controllers. To ensure that the market was large enough to keep prices down, they made rules that limited teams from using anything else. I would argue that the commercial marketplace for inexpensive motors, drives, and controllers has caught up and bypassed the needs of FIRST. Inexpensive, rugged motors and drives, both brushed and brushless, are commonly available in the RC marketplace. For the $285 cost of the cRIO FRC II Controller, you can buy EIGHT Raspberry Pi computers (I have an order of three arriving today.)

During build season, I'm all about doing the best you can with the limited set of components allowed. But, with cheaper, possibly better, components available in the commercial marketplace, why not expand the rules to allow us to use them?

Again, I apologize for hijacking the OP's thread. If anyone knows how to split this topic off into a new thread, please do so...

In order for there to be more options, FIRST needs to understand and support the characteristics of those options. Otherwise, you could run into some dangerous situations. Some examples:
- What if a team figures their mechanism only needs a couple of amps, so they use a cheap 5A speed controller? Who puts out the fire when something goes wrong and it draws 20A during a match?
- How do you handle the situation where a team picks a motor that has a grounded chassis? having any sort of short to the frame can cause some major issues, both for your robot and for safety on the field interacting with other robots.

Further, how do you ensure sufficient quantity for all the teams that want a particular part? It's great to say that teams can use any motor, but what happens when the veteran teams buy up the limited stock of the most powerful motors, leaving the rookie teams to suffer? Companies like Banebots, IFI, Vex, and AndyMark are very invested in the FRC program, they know the typical demand every year, and can help to ensure that all teams have access (even if there are occasionally shipping delays for a week or two) to the parts we need. If we go with "common" parts from suppliers that aren't invested in the program, you run the risk we've encountered with game pieces (like Orbit Balls!) in the past - the company stops producing them, causing problems for teams during the build season!

Also, please remember that one of FIRST's primary characteristics is Coopertition. How can teams help each other if their systems are all different? How can you borrow a speed controller from someone if they use something completely different, with a different footprint and different characteristics? How can FIRST possibly manage a FMS (field management system) that ensures equality and safety for everyone if everyone uses a different robot controller?


Some advanced individuals or teams may certainly have the knowledge, experience, and ability to test, develop, and utilize alternative controls. Most rookies don't. Even some older teams don't. FIRST does its best to provide a consistent platform for every team, so everyone competes on an equal footing.

[ToddF]

Eric and Jon,
Thanks for the educational replys. You are speaking from a background in FIRST I just don't have yet. I appreciate the discussion.

Originally Posted by EricH

The second assumption is VITAL. If you're going to make a new technology legal, ALL TEAMS must be able to obtain it, preferably with minimal research. The easiest way to do that--and cheapest--is for FIRST to buy about 2000 sets of the item and distribute it via the KOP. Otherwise, you get the vouchers which may or may not be of any practical use. If it isn't in the KOP, it should be available from AndyMark/IFI/BaneBots/local retailers.

I agree that it should be commercially available, but don't see how it's necessary that it be in the KOP or available from those particular vendors. There are plenty of things that are legal that don't fall into those categories. Colson wheels, for example, or any gearbox you can dig up or cannibalize from a commercial product. And why should our collective entry fees go for parts we don't want or need? FIRST has recognized this issue and said that next year, teams may elect not to even get a standard KOP, electing to accept a voucher instead.

Originally Posted by EricH

As far as the first assumption, YES! You do have to have characterizations! (We'll assume fairly complete, but not completely complete.) Why, you ask? How else are you going to size motors properly for the applications? Not all teams have the equipment to test the motors, or the budget to replace ones that fry during testing. Not having even an incomplete characterization can result in teams frying motors left and right--and even with an incomplete one... See Tetrix motors on minibots in 2011.

It seems to me that FIRST has already departed from this premise. See last year's [R48], part I: "up to 2 window lift, seat, windshield wiper or door motors obtained through either the FIRST-Automotive Recyclers Association partnership or from a prior years’ KOP." Since we were already allowed to use junkyard motors, with no info on their performance characteristics, why not brushless, too?

Originally Posted by EricH

The motor system is the same way, except that it's limiting the maximum available power from motors instead of making the technical group's life easier. Same available motor pool = same available maximum power = no advantage unless you can put more motors from the pool on the robot and control them. It's been opening up a bit lately, but I think the main reason for brushless motors not being allowed is that the legal speed controllers can't handle them. But when you open up the pool of legal speed controllers, you have to make adjustments to the control system, somehow, or force the controllers to work with the system.

I think Palardy was right when he said pointed out that as long as PWD is used for signaling, the control system doesn't care if the motor driver is brushed or brushless. He also pointed out that using more powerful motors gives no advantage when all teams are electrically limited by the fuses on the power distribution board to a maximum of 480 Watts per motor (40 amps x 12 volts).

Originally Posted by Jon Stratis

- What if a team figures their mechanism only needs a couple of amps, so they use a cheap 5A speed controller? Who puts out the fire when something goes wrong and it draws 20A during a match?

This is a very good point. I wasn't able to think of a likely safety problem, and this is one. But, it could be simply solved with a rule specifying that any motor controller be required to handle a minimum of 40 continuous amps. Any more is unnecessary, because the fuse protection mentioned above.

You both brought up the issue of suppliers being able to carry adequate stocks of parts to supply all teams who might want an item. I'm glad you did. This is an issue that bothers me a lot, and has caused our team to redesign mechanisms when vendors run out of stock on popular items. The rules are very specific about how a vendor must perform in order to be "a legitimate business source for COTS items".

Specifically, last year's [23]:
C. "The Vendor must be able to ship any general (i.e., non-FIRST unique) product within five business days of receiving a valid purchase request. It is recognized that certain unusual circumstances (such as 1,000 FIRST teams all ordering the same part at once from the same Vendor) may cause atypical delays in shipping due to backorders for even the largest Vendors. Such delays due to higher-than-normal order rates are excused."
D. "The Vendor should maintain sufficient stock or production capability to fill teams’ orders within a reasonable period during the build season (less than 1 week)."

Notice that a small shipping delay is acceptable, but running out of stock is not. From my reading of Chief Delphi and the experience of our team, neither BaneBots nor AndyMark would remain legitimate COTS sources if these criteria were strictly applied. Does FIRST maintain a list of qualified vendors? Has anyone ever been removed from this list for non-compliance? Since FIRST seems to allow vendors to slide by with a "well, they tried their best" attitude, why should this standard be strictly applied to any other vendor? Our team has learned from experience to buy critical parts during the off season, so that when the build season starts, and AM and BB run out of their stock, we don't end up on the short end of the stick.

The low cost of brushless controllers and motors, and the wide range of available options (these days they are practically a commodity) makes the at-risk investment in a couple of sizes of motors pretty low. Pre-buying a half dozen $20 ESCs beats the heck out of stocking up on $120 Jaguars, like we do now.

I think you are both right about the dangers of fragmenting the control system. That would just be a terrible idea all around. But, when the time comes to reevaluate the standard hardware, I hope a more open, less expensive system like a Raspberry Pi/Arduino controller gets a fair shake. The drawback to something like that is it might come with very little vendor support. Having the NI support staff available to work out problems is a big plus for a big name vendor.

[Jon Stratis]

Originally Posted by ToddF

I think you are both right about the dangers of fragmenting the control system. That would just be a terrible idea all around. But, when the time comes to reevaluate the standard hardware, I hope a more open, less expensive system like a Raspberry Pi/Arduino controller gets a fair shake. The drawback to something like that is it might come with very little vendor support. Having the NI support staff available to work out problems is a big plus for a big name vendor.

The time has already come - we'll have a new control system in 2 years. This should help give you an idea of how detailed oriented FIRST is being in determining the right system for us to use when they go about changing it!

http://www.chiefdelphi.com/forums/sh...d.php?t=108083

[MrForbes]

Originally Posted by Ian Curtis

Can you expound on this Adam?

I'd be interested in hearing this too, Adam.

My wild guess is that RC cars have very little load, and the RC car parts just won't be able to handle the mechanical loads we subject our robot motors to.

How did we get here from Lithium batteries? huh

[SamM]

I don't think it's been mentioned yet, but the derating for Lead-Acid batteries is a big factor. They're rated for 17Ah at 1 Amp but in FIRST use, drawing triple digit amps, their capacity is going to be closer to 6 or 7 Ah.

[Al Skierkiewicz]

OK,
I think it is time to enter a few real world numbers to the discussion both on motors and batteries since a lot of data has been quoted from various sources.

As to the motor controllers now in use, while rated at 40 amps, these devices regularly survive 130 amp peaks when driving CIM motors at or near stall. The Victors are a little better at this than the Jags due to the intelligent monitoring in the Jag for faults. The Jag monitor tries to limit over current by faulting.
While everyone likes to look at the limits of the electrical system via breaker size, all FRC breakers withstand 600% overload for a short periods of time (seconds) without trip and easily handle 200% for several seconds.
First has always attempted to keep some simulated real world limits in the rules to force creative thinking using a defined set of parts. It is for this reason that many "legal parts" choices are made. If a team were to gain an advantage over other teams by simply finding an esoteric part, that while stocked was not generally known or available, that team would have gained a real advantage over other competitors. While this can be viewed as "holding back" some teams, it allows everyone to compete using the same components for producing mechanical power on the robot.
As to the battery choice, the AGM SLA batteries now in use are 1/4 the price of other battery technologies while providing more than twice the peak output current at a fraction of the series resistance of these other technologies. One item that many teams ignore or forget is that our batteries have a defined life (charge/discharge) of 400 cycles. This life is shortened by severe use resulting from high current demands. Robots designed with inefficient mechanical systems will regularly exhaust a battery in one match. While I have no real data, it would seem that these teams shorten that 400 cycle life by 1/3-1/2. While the lithium technologies have some interesting specifications, please look at the series resistance or impedance. Then calculate the internal voltage drop for say a 400 amp spike and the 12 volt battery Li suddenly is making 0 volts at full charge instead of the 8 volts that will be available from the SLA battery now in use.
As to the use of brushless motors, it would be nice to make some types legal in the future, I agree. However they have serious limitations in general use as several people have already discussed. The teams need to be informed on their best use and implementation.

[EricH]

I don't remember exactly on the PWM front, but I remember hearing something about the Jaguars/Victors using a slightly different PWM than say a servo/ESC. I think it's PWM vs PMM, but can't for the life of me remember exactly which is which.

Palardy, there's a minor difference between the max power being regulated electrically and in the motors. The electrical max power is driven by the current limits (and subject to the breakers actually tripping when they're supposed to, and as we all know they can handle short overcurrents)--the motor max available power is driven by the motor characteristics. If I can use 20 BaneBots without tripping a breaker, and you can only use 4 CIMs and 2 FPs, but you need 5 BaneBots on top of that that you can't use without the breaker tripping, I have a higher max available power and somehow worked in a way to use more of the electrical max power. Unless, of course, I used very weak BB motors...

This is a very good point. I wasn't able to think of a likely safety problem, and this is one. But, it could be simply solved with a rule specifying that any motor controller be required to handle a minimum of 40 continuous amps. Any more is unnecessary, because the fuse protection mentioned above.

A couple of notes on speed controllers: I would like to see some USB-capable ones available, and I recall that FIRST is looking to add those. However, I would actually raise the minimum from 40 to 60 (and if you meant maximum, so did I). Some of those controllers have ridiculously low safety shutoff settings, which don't account for the startup current that a motor can draw. (That current is why a breaker is used on the compressor spike instead of a fuse.) Trust me, I've seen a controller that was rated for at least the current my team was putting through it shut itself off 5 seconds into a 10-minute run--and the fuse protecting the circuit, rated for the same or less than the controller, didn't blow or show signs of blowing. A lower-rated controller, with the safeties shut off, went about 7 minutes in the same position before releasing smoke--it could have gone longer but wound up with an unusual load when the terrain shifted. And that controller was rated well below the expected current...

Oh, right: How it's necessary that a part be in the KOP. Teams were supposed to go buy a Kinect last year? That was also a new and readily-available technology. In all seriousness, it's not, but I'm willing to bet that the economics of scale are better for 2000 than for 20. Motor technology is one of those things where it's better to introduce it in the KOP--then teams will have a better collective knowledge on how to handle it, and where to buy it.

I wouldn't quite say that the ARA allowance last year was quite without performance characteristics. For one thing, van door and seat motors have been in the KOP before, and the van door motors were available from AM (which does provide the spec sheets); for another, if you need to look up the characteristics and you have a motor part number, that's pretty trivial with an internet connection.


I'm thinking that squirrel is right on not putting R/C components on FIRST robots due to the loadings. If you're stalling a R/C motor, you're doing something very wrong. At least, that's the aviation side of the theory--I don't know about the ground vehicle side. How often do FRC motors stall or come close, especially in a pushing match or a robot autonomously running into the wall?

[ToddF]

Originally Posted by squirrel

How did we get here from Lithium batteries? huh

Well, depending on your point of view, I very rudely pirated Aaron's thread, or I nudged it to include discussion about a wider variety of technologies not currently allowed by today's set of rules.

[Ether]

Originally Posted by Al Skierkiewicz

all FRC breakers withstand 600% overload for a short periods of time (seconds)

1/10 of a second max for the 40 amp breaker @ 600% overload (700% of rated load) according to the snapaction spec sheet for the MX5:

http://www.snapaction.net/pdf/MX5%20Spec%20Sheet.pdf

... am I looking at the wrong doc?

[Jon Stratis]

Originally Posted by EricH

I wouldn't quite say that the ARA allowance last year was quite without performance characteristics. For one thing, van door and seat motors have been in the KOP before, and the van door motors were available from AM (which does provide the spec sheets); for another, if you need to look up the characteristics and you have a motor part number, that's pretty trivial with an internet connection.


I'm thinking that squirrel is right on not putting R/C components on FIRST robots due to the loadings. If you're stalling a R/C motor, you're doing something very wrong. At least, that's the aviation side of the theory--I don't know about the ground vehicle side. How often do FRC motors stall or come close, especially in a pushing match or a robot autonomously running into the wall?

I wanted to expand on these two paragraphs, real quick.

First, on ARA motors. I inspected at two regionals, and didn't see a single ARA motor. Even though it gives teams a wide variety of choice, is readily accessible, and was completely legal, a vast majority of teams decided to stick with what they knew. Using a motor in the KoP has two advantages: you know it's characteristics, and you can test it since you have one sitting right in front of you. With the ARA, you go to the junkyard, get a motor, then have to take it back and see if it'll do what you want. Those extra steps and unknowns make it not worthwhile for most teams during a time-constrained build season. It'll be interesting to see how many teams used their voucher to get motors in the off season, tested them and understood them, then decide to use them this upcoming year.

Next, on stalling. I don't think there's a single motor my team has ever used that hasn't been stalled. CIMs stall all the time in pushing matches. We've stalled (and burned out) FP and BaneBots motors. We destroyed quite a few Tetrix motors. We stalled the window motors during prototyping on our BreakAway robot, before we figured out our final design. We briefly stalled the AndyMark gear motor this last year almost every time we used it. With RC cars, you're more likely to suffer wheel slip than stalling - the cars are so light and the motors spin so fast, even driving it straight into a wall will just result in the wheels spinning while you go nowhere.

Finally, one more thing to consider: The more options you add, the harder inspection gets for everyone. If you add a speed controller that can only be used with 20A breakers, then inspectors need to check that. If you open it up completely and allow any speed controller that meets certain requirements for current, voltage, thermal shutdown, etc, then the team would need to bring documentation for that speed controller, and the inspector would need to be able to understand that documentation. This is something teams are notoriously bad at doing (I've encountered a lot of situations where a team uses a pneumatic part that we need to verify the operating characteristics of, and they don't have a spec sheet with them. With no internet at the venues, getting one can be difficult!). The last thing we need to do is introduce additional inspection headaches!

[pitzoid]

Getting back to Batteries, there are really good LiION technologies available that would be ideal for the FRC bots. I've been working with Enerdel the last couple years on various projects, we put some of their 17aHr LiION cells in our Combat Heavyweight (220lbs) which easily spikes 400amps at 24v and had no problems, it would run for ~hr when not in a fighting match during testing. LiION don't have the same safety concerns as LiPO, like any stored technologies, they can still be dangerous, but won't spontaneous combust like LiPO, we’ve accidently poked holes and shorted a couple of the cells and never had a fire like LiPO are prone to do and hence have caused airplane restrictions. As mentioned, Li batteries are typically rated by a “C” factor which is discharge amperage proportional to capacity, and the discharge capacity of Li technology is also related to the actual surface area of the Li anodes. So unlike the “hobby” batteries made for short high discharge burst, Li technology made for the automotive, military and commercial applications are large flat packs to have maximum anode area for sustained higher amperage discharge without cell damage, say a equivalent to the SLAs used currently in FIRST FRC in an automotive type LiON would be a ~6X7X1” flat pack and would weigh ~2lbs in packaging.

But as many of the senior mentors have noted in this thread, the issue is cost and logistics. The technologies that are key to making the competition "fair" for all teams are the batteries, motors and controls that everyone is allowed to use. Simply supplying some of these newer technologies to more than 2500 FRC teams is HUGE for a non-profit trying to do things “inexpensively”. If FIRST was a big company doing billions in business, the challenge wouldn’t be as great, but regardless of some gripping about cost, FIRST does a lot with the budgets that they have to work with and switching to a different battery tech right now is cost prohibitive, IMHO.

FIRST FRC isn't necessarily a "use the most cutting edge" technology challenge like some technology competitions, but more like a “here’s what you have to work with, make it happen” situation. Believe it or not, this is more real world realistic. Some deal better than others, but all have the same opportunity.

[Al Skierkiewicz]

Ether,
You are looking at the correct document. The graphed trip times are the guaranteed "must hold" times. Typical is a little longer. As a reminder, as the temperature goes up, the trip times will go down.

[apalrd]

Originally Posted by EricH

I don't remember exactly on the PWM front, but I remember hearing something about the Jaguars/Victors using a slightly different PWM than say a servo/ESC. I think it's PWM vs PMM, but can't for the life of me remember exactly which is which.

You can just call it a Servo in code and be done.
The only thing that is different is the width of the PWM pulse. Servos use a pulse between 0.5 and 2.5ms, Victors use 1ms to 2ms, both have neutral or center at 1.5ms.
In LabVIEW (I assume C/Java are similar), the Motor cluster type includes a definition of each of the pulse positions for scaling, adding additional motor controllers that fit between a Victor and Servo are quite easy.

Originally Posted by EricH

Palardy, there's a minor difference between the max power being regulated electrically and in the motors. The electrical max power is driven by the current limits (and subject to the breakers actually tripping when they're supposed to, and as we all know they can handle short overcurrents)--the motor max available power is driven by the motor characteristics. If I can use 20 BaneBots without tripping a breaker, and you can only use 4 CIMs and 2 FPs, but you need 5 BaneBots on top of that that you can't use without the breaker tripping, I have a higher max available power and somehow worked in a way to use more of the electrical max power. Unless, of course, I used very weak BB motors...

In any case, the mechanical power available is a factor of electrical power and motor efficiency. Any brushed motor significantly larger than a CIM used in a FRC drivetrain will likely trip the breaker - There are many teams who underpower their drivetrains and already trip the 40a breakers.

In addition, the 120a main breaker and battery limit the total available power of the machine. Drawing too much instantaneous current will lower the battery voltage due to battery resistance, and too much sustained current will trip the 120a main breaker. It's also possible to melt the 50a battery connector. I have personally driven a robot to do both in a competition match (although the battery connector was in the off-season).

Also, every motor, speed controller, gearing, and wire is weight. Weight (and often time) is the most precious thing on a FIRST robot.

Originally Posted by EricH

Oh, right: How it's necessary that a part be in the KOP. Teams were supposed to go buy a Kinect last year? That was also a new and readily-available technology. In all seriousness, it's not, but I'm willing to bet that the economics of scale are better for 2000 than for 20. Motor technology is one of those things where it's better to introduce it in the KOP--then teams will have a better collective knowledge on how to handle it, and where to buy it.

I don't think it's a requirement to put the part in the KOP if it's commonly available. If FIRST were to allow certain types of motors (e.g. all 5xx-sized brushed motors), there would be many places to get them from (e.g. many cordless tools use 550/540 sized motors, you can get them at home depot if you want). The only exception is the control system - Since the FRC control system must be used as is, the base control system should be included (ideally every year).

[EricH]

Originally Posted by apalrd

I don't think it's a requirement to put the part in the KOP if it's commonly available. If FIRST were to allow certain types of motors (e.g. all 5xx-sized brushed motors), there would be many places to get them from (e.g. many cordless tools use 550/540 sized motors, you can get them at home depot if you want). The only exception is the control system - Since the FRC control system must be used as is, the base control system should be included (ideally every year).

Which is why I asked about the Kinect. You can go into any store that sells video games and buy one if you think it'll be useful. So why bother putting it into the KOP, like FIRST did last year? You could just give every team a voucher good at the local Best Buy or similar store that says "Good for one free Kinect".

And, just to beg the question: If base control system components should be included every year, why isn't the cRIO in every kit the last couple of years?

The way I look at this sort of thing is: If there's a change to the control system, especially a major one, it needs to be in the KOP. If there's a change to the motors, start with one or two in the KOP--it helps to have something in your hands that you can monkey around with. If there's a change to the hardware, make it available in the KOP/FIRST Choice--but teams will change that anyway.

tl;dr: If an electrical or controls change needs to be made, as it would with a new battery chemistry or brushless motors, put it in the KOP for one year. After that, it goes in rookie KOPs only; everyone else gets to use vouchers or local sources.

[Jon Stratis]

Originally Posted by EricH

tl;dr: If an electrical or controls change needs to be made, as it would with a new battery chemistry or brushless motors, put it in the KOP for one year. After that, it goes in rookie KOPs only; everyone else gets to use vouchers or local sources.

I would only change that to read "put it in the KOP for at least one year"

There are some cases where it makes sense to keep things in the KOP year after year. For example, the robot chassis, with the included CIM motors, gear boxes, chains, wheels, and sprockets. I love how FIRST addressed removing it from the kit for those who opt-in, though... it makes sure that every team can get something driving.

I have mixed feelings about the control system, though. While I want everything (including the cRio) included in the kit every year so we don't have to dismantle old robots, that gets to be extremely expensive and wasteful. How many 20 year old teams do we have that have shelves of robots at their build space doing nothing but gathering dust? I would rather dismantle some old robots to get control system components back than have to sacrifice anything else in the KOP just to get a new cRio each year.