1023 Clippard Tank Explosion at IRI

At room temperature the impact energy capacity of metals is much better than most plastics. That translates to better and less catastrophic failure.

BTW, on the vacuum subject, brittle materials are bad here too. When glass vacuum vessels implode the shrapnel keeps going; it doesn’t just collapse. Brittle plastics are likely to exhibit similar failures, but I don’t have data on those. Acrylic is easy to get, but PLEASE use maybe 1/4" or thicker walls. It really don’t have a lot of toughness… Also, thin walls are vulnerable to going to oval shapes and collapsing. Mesh and fabric to catch fragments will work a LOT better on vacuum than pressure.

For all the mentioning of the Bimba voucher, did I miss something? It seems like it went away in the last couple years.

So, just as another point, but was the transition from the metal tanks (steel/aluminum/chinesium) due to weight and potential corrosion? That’s something the plastic tanks have over metal, is the corrosion resistance. Now, I realize one FRC season isn’t an issue, but what about those tanks being reused over a span of say, 10 years, where they’ve possibly sat in a damp storage space. What type of drying medium (filters etc.) need to be, or should be considered for use, which would also add weight… Just something else to consider.

This is already allowed- there is nothing stopping you from using whatever COTS air tank you want provided it is rated for at least 1205psi.

At least 125 psi*
image
This is due to the relief valve physically allowing up to 125psi to be present in the high pressure side of the system before venting image

4 Likes

This is a good point and worth highlighting. Let me provide some more background on our thought processes because I think it’s a good example of being analytical about safety, like I was describing above.

Acrylic is reasonably tough, and generally considered to be “10 times stronger” or “10 times more shatter resistant than glass” but that’s not really saying much.

As far as containing vacuum (1 atmosphere), that’s not very hard. There’s evidence of people building submergible sensor pods out of 4" diameter 3/16" acrylic tubes and they failed at 10 atmospheres of pressure. Also, it was the end caps that failed, not the tube itself. The tube we used is 1/4" thickness.

But we know that acrylic stressed beyond the breaking point will shatter into small pieces. This was a hazard we identified early in our design. By “hazard” I mean a danger we could reasonably see occurring, but we hadn’t attached a probability to it yet.

The best solution is to get rid of the hazard, but we couldn’t find a suitable replacement (I remember we looked for metal or polycarbonate) that were available and within the tolerances we needed.

Since it didn’t seem possible to remove it, we turned to mitigating it. Here’s what we came up with:

  1. The tube shouldn’t be used as a structural item because that would put additional strain on it.
  2. The tube needs to be protected from mechanical impacts.
  3. In the unlikely event of failure, we want to contain as much debris as possible within the robot.
  4. We should limit the duration of time that it’s “charged” with vacuum.

So it’s no accident that the “big syringe” was mounted self-supported in the middle of the robot with structure all around it.

6 Likes

Drying should not be a concern in FRC for a few reasons.

  1. Metal tanks are almost exclusively SS and aluminum.
    1A) I have never seen a mild steel tank, but at one point some team has probably used one.
  2. Mount the steel tank with the outlet facing down, any water will be removed shortly after it condenses.

I don’t really want to do the math on this one, but I suspect the compressors we use are just such a low flow rate and low pressure that there isn’t much water to remove. Robots also don’t generally sit around at full pressure, they’re vented regularly, which prevents water buildup.

FWIW I have a 30 y/o compressor with a mild steel tank that gets drained once or twice a year. No appreciable rust comes out of it.

2 Likes

I’m aware of this, but even with the rules being very unrestrictive a large portion of the teams have chosen the exact same solution. Enacting an overall ban on a very common solution with the only requirement being that the replacement must be made out of “shiny/conductive material” might create a period of time where there is a large variety of air tanks types on robots until another solution that is good enough for most of us is identified and the community converges again. I think that avoiding this period of divergence is important.

2 Likes

The moisture comes from compressing humid air. I do not need to do the math :). From experience we get quite a bit of water in our robotics pneumatic system. Enough to be a nuisance sometimes. I once argued it was the reason for a 1lb weight gain on a robot re-inspection. Dryer climates maybe not so much. BTW my home compressor is also 30 + years old. I drain it at least once every decade. It has been a while since I have done any stress calculation on steel tanks. But at 125 psi and any reasonable wall thickness, the margin of safety is so high it is almost a why bother calculation. (Doesn’t mean the calculation doesn’t need to be done.)

That turns out not to be the case. After having been damaged by improper assembly, the white plastic ones with screw fittings were seen, at least once, to go boom just sitting there under pressure. I suppose you can weasel out of your statement by claiming that the overtightened fitting makes it not “randomly”, if you really want to.

This statement crosses the line from discussing a technical topic to attacking me personally.

I’m happy to have a heated technical discussion where we dramatically whip out new evidence in front of the jury like an episode of Matlock, but I insist that we be respectful of each other while doing it.

17 Likes

:white_check_mark:

To be fair. The white clippard’s had a known failure mode from a design deficiency and improper assembly. It was not random. The current tanks don’t have the failure mode.

The only failures I have heard of from the current tanks is blowing a fitting out. which is not that big of hazard. and from impact. You can greatly minimize impact failure by rules on tank placement and quantity of tanks allowed. I am undecided if it is worth the extra complication of the rules or would be better to use tanks without that failure mode.

5 Likes

FWIW - White Clippards also had issues with casting voids causing “random” failures.

1 Like

Yup, I’m pretty sure the one that I got a chunk of that fired into my back (not pointy end first, thankfully) and flew halfway across the pits was not overtightened. We got lucky on that one, another mentor’s head was about 6 inches from the tank ~15 sec before it popped.

4 Likes

I am a student on 3357, we intentionally didn’t use pneumatics in 2022 to try and mitigate what ways things can go wrong, as well as to try and help that pesky weight problem. I am in no way saying that we are the best of the best, however I am quite pleased with our performance this season. I personally believe that our robot is a decent example of a competitive robot not needing pneumatics to perform at some of the most competitive events. I agree with you in saying the competitive hit of not using things like limelights (or any vision), brushless motors, or even the resources for a full/partial field for testing is a much more tangible harm than that of not including pneumatics. While none of those will definitively mean you cant build a competitive robot, I personally believe they are more important than pneumatics.

4 Likes

I’d say a shiny new IRI banner soon to be hanging at the school is evidence of a pretty good season, with or without pneumatics! I personally didn’t realize that your robot wasn’t running any and was quite impressed with it throughout IRI. Congratulations to the entire Comets 3357 team.

1 Like

Fun fact, the 3357, 2075 alliance that won 2 events this year had at least 2 robots with no pneumatics this year. I am unsure if 6122 or 6128 had pneumatics.

For those wins, I believe 3357 also had no swerve and no turret. A team can win without having to join the arms race. (COMETS did add a turret later.)

2 Likes

To all the people saying that poorly made metal tanks will fail more: what failure mode have you witnessed of a low pressure tank like these? If it fails non-dangerously, I don’t see why it’s a problem.

I also have to ask, how many Clippard tanks have failed in the last ten years or so, apart from the 2014 ones? If we’re seeing a single failure in ten years, after tends of thousands of uses, then maybe pneumatic tanks aren’t quite that big of an issue. We’re playing with large power tools and 120lb robots - there’s injuries sometimes. A non-harmful on-field failure once every decade is relatively low-risk compared to those. There may be more, I just don’t remember hearing about them so I’d like to know more if they have happened.

Basically, I’m asking where the bar for “safe” is.

5 Likes