So what the consensus seems to be is:
DC 15 is a standard average lock, DC 20 is a good lock, but even an average person WITH PROFICIENCY will defeat a DC 22 lock given enough time. DC 23+ is a custom puzzle lock that no average person will succeed at picking, ever. And all of these DCs - excepting "this is beyond your skill" - are irrelevant if there is (a) no time-pressure; and/or (b) no consequence for failure. They only matter if popping the lock needs to happen fast.
Example: my rogue needed to "unlock" four elemental locks, each in a differently-charged elemental death chamber. Every round I took 2d6 damage of the appropriate element. The Fire, Cold, and Acid rooms, it mattered extremely how many rounds it took to pop the lock! (In the Lightning room, my particular lightning-immune rogue could take his time.) However, the lock on the chest in the next room, despite being DC 25, was just a 2-minute interlude (my bonus was +8; eventually I will succeed).
Even if you house-rule a failure penalty - like "you damage your picks or the lock on a Nat 1" - this mostly just leads to an exercise of dice rolling that is unfun. D&D is not a DyingLight [zombie video game] mini-game, where your twitch skills with a lock sometimes break picks and sometimes jam the lock. So this house rule needlessly complicates the task in 90% of the cases.
Having said all that, though, sometimes it can make "story-sense" that a lock is designed to counteract attempts, or break locks, or jam itself. In which case... you attack the hinges or the material of the door/container/wall, if you can! My PCs have cut the
backs off many a chest, leaving the trapped lock intact and undisturbed! (Didn't stop the one chest with a contained poison gas, but it worked on most of the rest!)
In my campaign, I decided to make lockpicking less generally successful, and reward the truly skilled (i.e. rogues, and people who invested Expertise or somesuch). So instead of making "plenty of time" = "eventually get a 20", I lowered it to "2 minutes = Take 10"; insane amounts of time (like hours instead of minutes) gives Take 20 and a simple roll to see how much damage you caused your tools. A standard door or footlocker is just DC 12; a "poor" quality lock that is generally available, and easily defeated - "keeps the honest people honest." The PHB lock, DC 15, therefore means Joe Average the locksmith (DEX 10, proficiency with lockpicks) has a chance of not being able to defeat the lock without some tool damage and lots of time, but Rosie the Rogue (DEX 16) can handle it in anywhere from 6 seconds to 2 minutes.
I also increased the number of successes. Common (DC 12) locks usually take only one success, sometimes two. The PHB lock at DC 15 takes three. A Dwarven Masterlock, DC 25, might take 5 successes. Still irrelevant if your Take 10 is sufficient; when rushing or stressed (being shot at by stormtroopers, for example), failures matter. Failure 1 has no consequence, but tells you the DC and the # of needed successes. Failure 2 damages your picks (they can take 3 "hits" before needing replacement). Failure 3 jams the lock. [Nat 1 = 2 failures; Nat 20 = 2 successes]. Even then, the lock can be cleared with the application of more time... unless ... but then... oh, but, instead, you can...
See, I tried to make a mini-game of it. Instead, I created two pages of rules that no one remembers, and even when pulled out, just means the DM and the Rogue get to play "dice rolling" for 20 minutes while the other players watch. [The Thief subclass did shine here, though, because they could potentially achieve
two successes a round, once with Bonus Action and once with main Action.] We used it... twice, I think? I still think it's a great idea, that just doesn't play well at the table.
