2012 New York City Regional

[Tristan Lall]

Originally Posted by Bill Tompkins

This will be my only post, and I am only going to comment on the rules and the need to comply with them.

[R73] Compressed air on the Robot must be provided by one and only one compressor. Compressor specifications may not exceed nominal 12V, 1.05 cfm flow rate, 120 psi maximum working pressure. Off-board compressors must be controlled and powered by the Robot.

• Whether or not anyone feels these are valid or needed rules they are in fact, the rules. Rules are put in place by FIRST for the safety of everyone and to keep the matches competitive. In NASCAR you could go faster and further with a bigger engine and fuel tank but it is against the rules. You are given a bunch of parts and set of restrains (rules). You are then asked to build a robot and compete in adherence of these rules. This is the game.

• Relief Valves (16-004-011) help prevent over-charging but too often these do not work probably and can be tampered with. As part of the robot’s inspection the safety Relief Valve is checked. However, it can be changed easily with a wrench. In addition, the Relief Valve does have limitations in its ability to release pressure. If you are adding pressure faster than the Relief Valve can release it then you have an explosion. A FIRST specification for on-board and off-board compressors is 1.05cfm. The picture of the compressor shown in this thread has 4 to 5 times that flow rate. The Relief Valve and pneumatics systems used in FRC are designed to safely operate at the 1.05 flow rate. Exceeding this could be dangerous.

• A feedback path to the robot is required in order to shut the compressor down when the pressure gets too high. The Pressure Switch (SM-2B-115R/443) is designed to electrically open at 115psi and close at 95psi. This is also checked at the robot inspection. However, without feedback to the off-board compressor the entire volume of the compressor could be dump into the robot at high pressure and rate.

• Pressure gages have been known to stick. They are mechanical in nature and sometimes malfunction. This is the reason for the secondary protection of the Pressure Switch and the Relief Valve and the need to have feedback to any off-board sources. Just visually watching an air gage and assuming the pressure reading is correct is inherently dangerous.

• Items not part of the KOP are required to be inspected for compliance to the rules and safety. If an off-board compressor not part of the KOP were to be used it should have been inspected, its operation demonstrated, noted on the inspection sheet and would become part of the BOM. It also becomes part of the maximum unit cost restriction.

• When charging the air tanks the battery is drained. The larger the on-board air storage capacity the larger the drain on the battery and the longer it takes to charge the tank. This is a design consideration trade-off. You choose to have the added air capacity knowing your battery and air charge time will be inhibited. This is the reason why off-board compressors need to run of the robot’s battery. Not doing so gives a team an unfair advantage.

• The time periods between finals matches are timed. You have the option of using timeouts if additional time is required. If your robot cannot be serviced in this allowed time period then you just have to do the best you can. This goes for everything from broken wheels and chains to battery changes and air charging. The design trade-off mentioned above gives you more air to work with on the field but lengthens your air charging time. Given the short time period between finals matches you may not have the time to fully charge you tanks. However, supplementing this with an additional air source is a violation of the [R73].

• "Compressed air on the Robot must be provided by one and only one compressor". This part of the rule is pretty clear.

In closing, this “seemingly needless rule” is designed to keep things safe and competitive. We can debate (which I am not) whether breaking this and other rules gave a team an unfair advantage or whether they could have won without it. The fact is, having a secondary compressor on the field does violate the rules for all the reason mentioned above.

While many of these things are substantially accurate, I think there is some question about the degree to which this is intended as a design constraint, the degree to which it merely functions as one, and the degree to which it is actually useful. Also, it isn't clear when FIRST is demanding something for safety reasons, and when they're doing it for competitive reasons (or the degree to which both considerations are represented).

As a design constraint, this is of limited effectiveness. After all, if you really wanted to, you could swap out spare tanks already pre-charged with the requisite quantity of pressurized air (from the KOP compressor) and overcome the delay. (If they're true spares, they don't violate the module rule. This assumes that stored air is not a robot part for the purposes of the rules.)

And to come to think of it, if you wanted to run a legal off-board compressor at a higher flow rate or pressure, the rules don't actually prohibit it. (Assume the robot on the field and at inspection is otherwise legal. If the air was provided by a device with the proper nominal specifications, it is legal—the restriction is not on the actual performance of the device at the time of filling.) So you could theoretically immerse the compressor in a bath of cold distilled water (properly protecting the intake, of course), operate it at 24 V (under robot control), and see what happens. Note also that during filling, the robot is neither competing nor being inspected, so it would be tough to argue that it must meet the robot rules at that moment.

In terms of battery capacity, an untold number of teams trivially overcome that by installing a fresh battery prior to every match, but after filling their tanks.

As for safety, that's a matter of pressure and flow. The flow is principally determined by the geometry of various components. While the compressor might be able to supply that much, what's the actual flow given the orifice sizes provided by a legal FRC on-board pneumatic system? Does that exceed what's safely releasable by the relief valve? (I realize that the inspectors are rarely in a position to determine these things exactly—and the rule effectively avoids dealing with that uncertainty. But that's different from a particular robot actually being unsafe.)

The valve we use (Norgren 16-004-011) can release up to 5 scfm when set in the range dictated by FRC. That's in the ballpark of what that compressor is likely capable of (indeed it's probably less for continuous duty at high pressure like that). And even if the relief valve is misconfigured, the highest it can be set is 150 lb/in²—it will pop at that point.

In terms of pressure, with a typical industrial compressor, there's an adjustable relieving regulator built in (or at least a relief valve set for a high pressure). If present, this must fail or be set incorrectly for a safety issue to arise. Lacking the regulator, the system indeed depends on the robot's relief valve.

Lacking the relief valve (which should have been noticed at inspection), we're now depending on the strength of the components. As far as I know, all of the mandatory components on the high pressure side of the FRC pneumatic system are rated to around 250 lb/in² to 300 lb/in² working pressure at room temperature, and are designed with additional margin. And when they do fail, a true explosion is unlikely—more often a seam or tube will burst, venting the pressure. (And what's the likelihood that that compressor can hit 250 lb/in², at any reasonable flow rate, and for a sustained period?) Team-supplied tanks, especially the PVC ones, may not have quite this margin of error—so in that case, maybe the issue of safety has more traction. (But let's not forget that this rule predates the introduction of PVC tanks into FRC, so probably wasn't intended to address them.)

As for the pressure gauges, they depend on having an operator to monitor them, and do (occasionally) fail in a way that is non-obvious to the operator. You wouldn't want to rely on a pressure gauge if it was the only thing keeping the system from going out of its safe limits.

Putting that all together, what's the most likely failure mode for truly unsafe operation using an illegal off-board compressor? A poorly equipped compressor (no regulator or built-in relief valve), an operator not paying attention or ignoring warning signs (or gauges missing entirely, and no clue about aural cues from compressor), a missing relief valve (or a very slow fill with the valve venting the whole way), and non-KOP components that fail unsafely at unusually low pressures. And then they have to do this without being noticed.

To me, that's too implausible to presume that the off-robot compressor rule exists as a meaningful safety measure.

Also, I would definitely call into question the idea that an industrial compressor (not present during matches) is a robot part subject to cost accounting restrictions. Is a battery charger subject to those same restrictions?

In summary, I think this rule is enforceable and valid, but doesn't do anything appreciable as a safety feature or as a limit on robot performance. That's why it's silly.