1023 Clippard Tank Explosion at IRI

Are you planning on donating metal tanks?

The cons of removing plastic tanks entirely have been repeated several times now.


This is all I found in the FRC Pneumatics manual. It states to mount the tanks so they are protected from other robot impacts. (Not that everyone looks at this document.)


However, the inspection checklist ignores that requirement.

Should that be fixed first? Or is that a different can of worms?


P.S. Edit. I found no robot construction rule that requires these tanks to be protected. Maybe that’s the place to start?


There are no requirements in the pneumatics manual. That is an assistive manual intended to help guide teams, it is not part of the game manual or robot rules.


I mean, we allow teams to have metal flywheels of various weights spinning anywhere between 3000-12000 rpm (maybe even more), and that’s just a shaft or bearing failure away from becoming a much more dangerous projectile than any plastic air cylinder. We’ve also had teams build mechanisms with >600lbs of spring tension in them before (popular in 2014), so long as we’re talking about forms of stored energy. No one’s been seriously hurt by either of these yet, but that doesn’t mean it couldn’t happen.

FIRST Robotics inherently comes with a certain degree of risk just due to the nature of the competition (whether it be in the workshop, pits, or playing field). It’s certainly important to mitigate risks wherever possible but you can never guarantee 100% safety. At the end of the day, FIRST and the FIRST community have to find a balance with what is, or is not, an acceptable risk.

Personally, when it comes to plastic Clippard air tanks, requiring some sort of fabric safety covering seems like the most reasonable approach to take, given the costs and weight penalties of all the mitigation options. No doubt there would need to be additional testing done in a more controlled environment (shooting a tank with a .22 rifle is illustrative, but perhaps not the most scientific) to find the best solution, but it seems clear this would be an easy way to improve the safety of pneumatic systems without imposing a significant cost (both in terms of money and weight) on teams. Not to mention you could potentially use such shrouds as decorations, on top of their practical function (I totally could see someone producing these in a variety of colors and with iron-on decals), especially since it’s not legal to decorate the pneumatics themselves. This would allow teams to use the tanks they already have and not pay an arm and a leg for new ones, as well as prevent a lot of perfectly good tanks from ending up in landfills should they be made illegal.

1 Like

I have refined my proposed rule in response to some feedback. Obviously we could quibble endlessly about the details of each bullet. My intention is to elevate protection of air tanks to a robot construction rule and then give suggestions for how teams could convince an RI that they’re in compliance.

Protect air tanks Pneumatic storage tanks must be appropriately protected from the risk of impact and fragmentation.

Blue box: Examples of such protection include:

  • Positioning sufficiently far from the frame perimeter that legal extensions from other robots cannot reach them.
  • Being protected by rigid structures (e.g. within the chassis) or shields (e.g. polycarbonate) that will dissipate the force of a horizontal intrusion from outside the frame perimeter.
  • Use of tanks (e.g. metal) with a low risk of fragmentation.
  • Use of a fragmentation-mitigation technique, such as full enclosure of the tank in a sleeve.

This should have the effect of making our use of air tanks considerably safer, without specifically requiring higher cost or weight, and without limiting creativity.


I broadly agree with your post, but just to nitpick: Shrapnel from this incident apparently entered non-PPE sections of the venue.

1 Like

2 posts were merged into an existing topic: 4907 ThunderStamps 2022 Season Recap Video

Related question - do these plastic tanks have expiration dates?
Polypropylene degrades in UV radiation, the more these tanks age (and are exposed to UV), the more likely they are to fail.

Understanding that most of FIRST is conducted indoors, but my hard hat is required to be replaced, even if kept indoors.


I don’t see anything about that in the datasheet. My understanding is that the tank in this incident was new this season and of recent (2019-2020) manufacture.

I’ve been following this discussion with interest. Beginning my career in the nuclear industry, moving into industrial robotics, and having spent over the last decade in the aerospace industry, I’ve had my share of exposure to various claims about what is or is not safe. Though there are whole disciplines devoted to safety, with various industry and government agency standards addressing the subject, there is vast disagreement about acceptable levels of risk. This is compounded by the widely accepted fact that humans are terrible at assessing risk.

This is especially true when asked to assess the risk of a singular event, like has been discussed in this thread. When faced with questions like this, I find it far more useful to evaluate the range of possible risks imposed by the activity as a whole. It’s far easier to rank risks against each other than it is to evaluate them in the abstract. This allows us to prioritize them and address them in order of priority from most dangerous to least dangerous. At some point in that continuum, there is a reasonable line which can be drawn between really risky situations, which should be mitigated, and not so risky situations, where we choose not to bother. My proposal is that the risks be addressed according to priority, and that once you reach one where you don’t feel the need to mitigate the risk, it is illogical to spend resources debating or mitigating lesser risks.

It is with that approach, and with an engineer’s eye, that I approach the entire spectrum of activities and situations that arise in FRC robotics. I also note that there are common practices in FRC robotics which are completely unacceptable in some professional working environments. Let me throw out just a few examples:

  • Allowing two adolescent humans to lift a 150 pound robot and carry it onto and off of the field. At my workplace, participating in or permitting this activity would be grounds for termination for violating safety standards. I know of multiple individuals in FRC who have suffered injury from this activity.
  • Building mechanisms containing metal chain and sprockets. I’ve witnessed varying levels of injuries caused by sprocket and chain mechanisms ranging from severe pinching all the way to a student at one event having a finger amputated.
  • Wearing safety glasses for protection of eyes from flying debris. I’ve lost count of how many times I’ve seen eyes (including my own) protected from stuff that bounces off safety glasses rather than causing eye injury.
  • Allowing students and mentors to use various power tools, hand tools, and cutting tools. All of which cause varying levels of injuries, from minor to severe, every year.
  • Attending events where participants are transported in motorized cars/trucks, in which thousands of people die every year.
  • Participating in robotics competitions in which the vast majority of attendees and participants are unmasked, even as thousands of people die from COVID every day in this country.
  • Stepping over the field barrier, which might result in tripping onto soft carpet, and possibly scuffing one’s knees and wounding one’s pride.
  • Putting a plastic tank on a robot, which might explode if damaged while pressurized.
  • Plugging an extension cord into any other non-permanently installed electrical outlet, such as another extension cord. A violation of electrical and fire codes.

Feel free to add any other risky or hazardous situations to the list. My point is that when evaluating risk in FRC, we need to look at ALL the risks, and address in them order of most risky to least risky. My contention is that situations which are known to have caused harm should be addressed first. And if we are not willing to correct those situations, we have established our collective acceptable level of risk as far above those situations where there is only hypothetical harm.

If we’re seriously considering changes to the rules to ban plastic tanks, which thus far have only theoretical harmful effects, we should darn well be considering banning sprockets and chain, which have amputated fingers. Both are safe when used properly. Both have easily usable alternatives (metal tanks and belt drives). Both could possibly cause harm if implemented improperly. One has a risk of injury which thus far is only theoretical, and the other has hurt people. I’m totally willing to consider banning plastic tanks when they are our most pressing concern. But on a prioritized list of harmful things in FRC which need mitigating, plastic tanks are way, way down on the list. Let’s start by fixing the situations that are actually hurting people before getting worked up over things that aren’t.


This is interesting, I think I’m about to learn something. I work in largely unmanned field robotics, so while I do deal with some human safety hazards my experience is biased toward protection of expensive hardware or preservation of mission objectives (e.g. lunar science instruments). In my work, we rank things by both the “riskiness” (likelihood and severity) and the ability to mitigate (cost-benefit tradeoffs). So something that is lower cost (not just money) to mitigate may get attention even if its risk rating is lower than others that are higher but harder to mitigate. Obviously there’s a threshold at which even a hard-to-mitigate risk must be mitigated because it’s so dangerous. But in the gray zone, we’ll also grab the low-hanging fruit to drop the risk profile incrementally (and it is properly incremental). Is this not the case in more human-centric safety handling?


I was thinking along those same lines when reading Todd’s post - if you identify a risk, it’s worth considering the cost of mitigating or removing that risk, not just the severity of the risk. Yes, there can and should be a severity limit that ensures a risk is addressed. But a risk that doesn’t rise to that limit can and should also be addressed if the mitigation or resolution is cheap. On the scope of the cost of running a team and building a robot, mitigating (locating tanks in a safer, protected location) or resolving (using metal tanks) this issue really is a fairly cheap one.


There have been other failures. There was a reservoir that exploded in 2019 at Champs that folks posted video of earlier in this thread. One exploded at the SCH district back in 2016-ish, but I do not recall the specific team or match number.

Moreover, when analyzing risk, you need to consider both severity and likelihood. While you’re making the argument as to why the frequency, and thus likelihood, is low - there can still be unacceptable severity levels regardless of what the frequency is. I think many of us would indicate this is one of those cases.

Even ignoring that this argument is a whataboutism, it makes sense when you consider the fact that one of these is integral to making this activity function while the other can be addressed without any major consequence. Using machine tools poses a higher risk than anything we do on the field, but using machine tools is an essential part of the fabrication/mechanical/design portion of the FRC experience, part of what is distinct to it. We implement what safety measures we can to mitigate the risks of using machine tools, such that it can be done as safely as possible. Without it, you’re removing part of what makes the program worthwhile for a large portion of its participants. Similarly, getting the robot to/from the field of competition is a pretty important part of competing, and many value these hanging/climbing-style end games as a core part of the design challenge and match play that inspires FRC participants. On the other hand, having to pick a different pneumatic reservoir or implement proper safety mitigations (such as sleeve or protection) would not have a negative impact on the FRC experience.

Do I think we can take steps to make carrying robots onto/off the field safer? Absolutely. But that’s not a reason to skip out on addressing this other safety risk.

Another whataboutism. More should be done about the increasingly dangerous flywheel implementations (especially as vendors are now openly marketing devices for use as flywheels and team performance/game design now openly incentivizes maximizing kinetic energy storage to maximize fire rate), but action or inaction on that front doesn’t mean we cannot address pneumatic issues.

Disagreed. The cost (including the metaphorical “cost” to the program experience) is also a factor in addressing risk. Just because you are willing to accept a risk that is essential to the operations of the program (such as the risk posed by power tools, machine tools, or operating 150lb machines running at 20 fps) doesn’t mean that everything with the same severity or frequency is now an acceptable risk. It means the cost of fully eliminating that risk is too steep for it to be eliminated, and instead we are taking steps to mitigate that risk.

The downside of banning plastic tanks - a completely optional design choice (and one that is typically not used by teams with pressing financial concerns) has more financial cost associated with it.

The downside of banning chains/sprockets - large swathes of prior FRC designs are now illegal, many teams are no longer able to build the types of designs that allow them to address the design challenges without access to precision machining that allows for C-to-C belt distances or proper gear meshes. Not to mention all the same complaints about organizational inertia and old supplies not being re-usable.

This is an example of why you need to consider the cost associated with eliminating a risk. The costs here are different by orders of magnitude.


Random question but why aren’t teams allowed to bring carts onto the field? Or are they usually and I’ve just been misled? Seems like that would be free and reduce accidents.


I can think of a few reasons.

  1. The gates themselves. Many carts are wider than the robots, which are just a bit wider than the gates with their bumpers on. I would also include the ramp but any cart that can’t handle that shouldn’t be in use anyways.
  2. Fields can be crowded. 12 robots with 3 people each PLUS a dozen resetters and 3 FTA/FTAA/CSA types is already pretty hectic, but when you add 12 carts? Yeah. You’d need to completely revamp load-on/load-off, and there goes the schedule when you do that.
    2b) Because of the crowding, it actually increases the risk of accident. if you’ve ever been whacked in the shin by a robot cart, you’ll understand where I’m going with this.
  3. Access. It’s not uncommon, particularly in smaller venues, for it to be rather tricky to load on and off without hitting the scoring table. Now you do that with a longer unit (the cart).
  4. I touched on this in 2, but time/delay of match. Navigate cart on, unload robot, clear cart–but maybe the team starts working from the cart on something they didn’t quite get time to do, which adds more time.
  5. Field damage risk. Not that this is particularly high up there, but there is the chance that the cart wheels picked up something that could damage the field… or that they get locked by accident and drag.
  6. Forgotten cart. Teams forget stuff on the field on a regular basis; sucker’s bet that someone forgets a cart on the field the first season they’re allowed. 50-50 on that sucker’s bet that the match starts before it’s spotted.

I can’t say I disagree that there should be some sort of allowance for carts on the field, but there’s just too many factors that render them a potential issue to allow it, IMO. Maybe if they were shoved through the gate to unload/load?

1 Like

Traffic in and out of the field, as well as the cart may not even fit through the gates

Please use this other thread for discussing carrying robots.

1 Like

Little bit of anecdotal data


Aftermath of a consumer Air compressor tank explosion https://youtu.be/sm_FJ6Pat4I

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.