Hmm, is it possible that this could be partly caused because the water gets sucked not just directly down the pipe, but around the lip of the pipe opening... and one edge of the pipe opening is further away from the center axis of rotation (of the whole sprinkler) than the other end of the pipe opening. I'd think that whether or not this would have an effect could be easily tested by varying sizes of the pipe opening or how far away the opening was from the center to see if it made a difference -- even if the bends in the pipe were the same (the authors seem to attribute the whole effect to the bends in the arms).

I'll add that this could become particularly important in the 4d case once the sprinkler starts moving. Cause when it's still, the pressure gradient will create a partial-torus like shape around opening, but once it starts rotating, the outer side (rotating around the sprinkler's axis) of the 4D version of this shape has a larger diameter than the inside. So the inner side will be more affected by the environmental water's momentum state than the outside which has more water mass in it's scope rotating around the sprinkler axis.

Why can't you just submerge the sprinkler in water and reverse the flow?

Yes that is the experiment. I think when you do that it doesn’t move at all because the effect is tiny so will normally be nullified by friction? And I think the research used very low friction so they could observe the small effect.

That’s what they did

I've googled for about 5 minutes trying to understand ....is the "reverse sprinkler" just a device that has been impossible to build so it was a strictly theoretical problem? "We found that the reverse sprinkler spins in the ‘reverse’ or opposite direction when taking in water" - really, this was that impossible to build that nobody knew what it would do? weird

It's a sprinkler with the pumps pulling water in instead of pushing it out.

Plenty of people built one since Feynman stated the problem, it simply doesn't rotate or rotates very weekly on the same direction a normal sprinkler would. A lot of people think this is anti-intuitive.

From stories I’ve read, I think people generally, and including Feynman, concluded that the sprinkler would not rotate. Because conservation of angular momentum. Which is almost true. So what’s fun is that this experiment shows it’s not entirely true, if you have a sensitive enough experiment that takes into account more nuance of the fluid dynamics.

The "inside out" rocket engine is a great explanation.

Doesn't this boil down to time-reversal? I doubt CPT symmetry is relevant here, so I'd be very surprised if it didn't simply reverse the rotation.

The rotation is not reversed though (at least not with the same speed), that's why it's always been interesting.

A normal sprinkler works by a pump pushing water out the nozzle. Wouldn't a time reversed sprinkler be a situation where water is flying into the sprinkler and depressing the pump? in that case I would expect the rotation to be reversed.

Well, yes. Isn’t that the situation the article is describing? Suction isn’t a real thing, so-

Wait. Is this meant to imply it’s running underwater?

Yes, in the thought experiment it runs underwater. It's fine though because it can run underwater in both cases, so that doesn't ruin the symmetry.

The article implies a situation where the pump is running in the opposite direction as normal, sucking the water in. That's not time-reversed though (eg in the time reversed version, the pump generates electricity instead of consuming it, because the water is pushing it instead of it pushing the water)

No. A nozzle behaves differently if you plug it the other way around, and that does not break CPT symmetry.

Related:

Centrifugal flows drive reverse rotation of Feynman's sprinkler - https://news.ycombinator.com/item?id=39128851 - Jan 2024 (19 comments)

Feynman Sprinkler [video] - https://news.ycombinator.com/item?id=38985308 - Jan 2024 (21 comments; top comment just added to https://news.ycombinator.com/highlights!)

Feynman sprinkler - https://news.ycombinator.com/item?id=14531297 - June 2017 (6 comments)

The Feynman Sprinkler - https://news.ycombinator.com/item?id=11720126 - May 2016 (1 comment)

An elementary treatment of the Feynman sprinkler - https://news.ycombinator.com/item?id=11626758 - May 2016 (1 comment)

Feynman sprinkler - https://news.ycombinator.com/item?id=9710119 - June 2015 (3 comments)

Why is this problem more relevant than say the “reverse explosion” problem? Or “reverse for X”

F=MA

Appling force to suck in water accelerates it and every action has an equal and opposite reaction

I do not get what the mystery is. What am I missing?