I see two "deep" issues here.

One of them is that it's damn hard to decide, in online communities, who should [not] be allowed to perform some action in a fair, transparent, and simple way. There's always some way to circumvent it, and always someone who should perform it but gets locked out.

For example: what would prevent me from subscribing to a comm, downvoting everything there, and then unsubscribing from it? Or just subscribing to comms to vote-brigade them, while newbies legitimately interested on the comm are unable to vote in it?

I have no good solution for this issue.

The second one is that this sort of Reddit-like voting system doesn't really work well. It's at most bidimensional (score vs. controversy, or up vs. downvotes); and yet there are a thousand reasons why people vote, and a thousand pieces of info that they can retrieve (or falsely believe to retrieve) from them. And depending on those reasons, the vote might be completely fine or not.

There are also more practical concerns; I believe that @davel@lemmy.ml's Hexbear example illustrates this well. If you anyhow hamper the ability to voice negative feedback through downvotes, people do it by noisier ways.

For this issue, perhaps a "reverse Slashdot" system would work better? Basically splitting the downvote (but not the upvote) into multiple categories (e.g. "disagree", "this doesn't contribute", "this is factually wrong" etc.). It wouldn't prevent this sort of voting brigade, but it would discourage it a tiiiny bit (you'd need more clicks per downvote), and make it more obvious.