I've never picked or bumped a lock, but from what I gather the basic principle is that you bump, pick or rake each pin to the shear line while applying a turning force, so that once in the correct place, the drive pin and key pin separate. The turning force keeps them in the correct position until all the pins are aligned, and the plug is free to rotate.

I wish to build a lock where instead of the pins being solid cylinders, they are a stack of metal washers. The top and bottom part of each pin would be connected through the centre holes by a small metal chain (possibly only with one or two links).

In normal operation the pins should slide nicely in their tubes, and as they're under compression between the springs and the key, vertically they should be identical to normal pins.

When a wrong key is inserted, the lock would twist about five degrees before stopping, either way. The pins would shear, then the linkage between top and bottom disks in the stack would be clamped at the normal shear line.

When someone attempts to pick or bump the lock, instead of a single point in the drive/key pin column being able to shear, it is able to shear at all the gaps between washers. You would only know if that particular pin is in the correct place once all the other pins are in the correct place. To adjust the vertical alignment of a pin, the lock must be twisted back to vertical and it would be hard to stop all of the pins returning to their default positions.

Essentially this turns the lock into a combination lock. To the user it would be a normal key, but to the lock picker there would be thousands of combinations of pin positions where the lock would turn the first few degrees. Even if all but one of the pins was in the correct position, you would have no idea* if the lock was any closer to being open.

The obvious downside is that the whole assembly would have to be slightly larger, so the linkage between disks is strong enough. That's a small price to pay though for exponentially increasing the time a lock takes to pick. Also, bumping would be almost impossible, since there are so many shear points the pins could stop at, not a single point like in a conventional lock.

Plus the clear advantage that regardless of all this, the pins *will not* move vertically unless no turning force is applied, which is the key part to any lock picking method.

(*If you had a special tool for it, you could probably measure the amount of play the pin has vertically, which would be much more if you'd managed to separate the drive pin from the key pin. Especially if the key pin was very short.)
-- mitxela, Jun 11 2009

This is a great idea, but extremely baked. Many good locks have split pins (and yes, they do make it much harder to pick them). Usually there's just one "split", but that's enough for moist porpoises.
-- MaxwellBuchanan, Jun 11 2009


I believe the typical system is now a "Mushroom shaped pin" Wherein one of the pins in the device is shaped like a mushroom, rather than a vertical bar. If pressure is applied on the cylinder, and such a pin is moved, it will extend to the mushroom stem portion, and that will get jammed at the shear line.

I've often felt that bar-bell shaped pins, which would tend to jam, and stay in the jammed position, would be more effective.

I suspect your chain would be a bit insecure, as it might break more easily at any link, but you might be able to make a bar-bell shaped pin, with a series of metal loops loosely set around the center portion. These would then jam in sequence, causing no end of trouble for the lock picker.

I also suspect that magnetically attractive pin sets would help to keep the two pins from separating at the shear line, but since breaking a magnet causes one portion to reverse polarity, such a device would likely require special maintenance checks, to be sure the magnetic pins had not cracked, changed polarity, and started forcing each other apart.
-- ye_river_xiv, Jun 12 2009


<pedant> Although exponentiation is the underlying principle, for any given number of pins your system would cause the number of combinations to be tried and the time taken to pick the lock to be mutiplied by a (very large) number, rather than //increased exponentially//. </pedant>.
-- spidermother, Jun 12 2009


maybe this would be good for rarely used locks but the increased friction would wear the key and the pins exponentially. Frankly a fine conventional lock, one built with as much complexity as you describe should be completely immune to reading. While I read the idea the whir of the drill filled my ears. The lock should always be the weakest part of the box.
-- WcW, Jun 12 2009


Maybe one day I'll come up with an idea that's both practical and hasn't been baked. Maybe.

[ye_river_xiv], the chain doesn't have to be small and weak, it could be a metal bar with a loop at each end, just something strong and moveable. Even if the whole thing had to be built three times bigger than a normal lock, it should still be harder to pick.

[spidermother], the very large number it's being multiplied by would be increasing in two dimensions, though, both in the number of pins and in the number of segments per pin, shirley?

[WcW], when the lock is in use normally there should be no extra friction to the pins or the key, unless you inserted the key while the lock was at a slight angle - which wouldn't happen if a simple spring was on it to default it to the vertical.
-- mitxela, Jun 12 2009


OK i guess i get that. One problem that you should know about with "unreadable" lock sets is the tendency to get hung pins that cause critical failure. Anything that makes the lock "unreadable" seems to dramatically reduce lock life.
-- WcW, Jun 12 2009


In fact, lock picking isn't really an issue. Criminals these days are much less artful than in the past, and usually don't bother picking a lock when they can kick in the door, smash a window, or whatever. Does anyone know anyone who knows anyone who's been burgled by lock-picking?
-- MaxwellBuchanan, Jun 12 2009


even a safe cracker is fairly likely to drill or grind rather than attempt to pick a lock in-situ. High voltage power tools have leveled the safe cracking field of play.
-- WcW, Jun 12 2009


The best way to design a pick-resistant lock, I would think, would be to have a mechanism that "samples" the key at one point in the operation of the mechanism, and then checks the sampled value for correctness at a different point. If the mechanism were constructed to limit the speed with which it could cycle, picking could become a very slow proposition.
-- supercat, Jun 12 2009


//best way to design a pick-resistant lock, I would think, would be to have a mechanism that "samples" the key at one point in the operation of the mechanism, and then checks the sampled value for correctness at a different point. If the mechanism were constructed to limit the speed with which it could cycle, picking could become a very slow proposition.//

Wouldn't locking and unlocking the door also become a slow proposition though?

I think the basic principle is a good one, I just don't think the chain part is essential. Indeed, I think that using a chain would give more play than necessary to the whole setup, and since a chain is only as strong as it's weakest link, I suspect that fewer moving parts there might be better.

I also think that a larger chain will require larger pins, and these in turn will require larger keys, and a larger keyhole, so that there would be some upper limit to the size of the lock mechanism. after all, we can't have people sticking their hand in the lock to pick it, and giving them enough room to see what they're doing would probably be unwise as well.
-- ye_river_xiv, Jun 13 2009


[MaxwellBuchanan], I've never known anyone who's had a lock picked, and the only people I know of who can pick locks do it for their own amusement. But houses have windows that can be smashed and so on -- in places where security is more important, lock picking might be more of a worry.

[supercat], that's pretty much what I was trying to design here. My first idea was to have hydraulic pistons which completely separate the 'input' pins for the key away from the 'output' pins which actually keep the cylinder in place. But I thought that was far too complicated.

//Wouldn't locking and unlocking the door also become a slow proposition though?//

Not at all, since you'd have the correct key. If it takes two seconds to open the door instead of half a second, I doubt many people would care - but if it takes two seconds per iteration for someone opening the lock by brute force, it would take them much much longer.
-- mitxela, Jun 13 2009


<i>but if it takes two seconds per iteration for someone opening the lock by brute force, it would take them much much longer.</i><p>

Actually it wouldn't be necessary for the mechanism to add any time to the first access attempt if the delay were built into the reset cycle. I would expect that some automatic lock-picking devices could try many combinations per second; allowing one reset per second would mean a five-pin eight-level lock would require nine hours to test all combinations. A six-pin eight-level lock would require 72 hours.

With a little more sophistication, the mechanism could be constructed so that each access attempt would add 15 seconds to a timer (which would count down when non-zero), but access attempts would be refused any time the timer was over, say, 60 seconds. So the first four access attempts would function immediately, but attempts beyond that would take 15 seconds each. Even if a joker tried a denial-of-service attack on the lock, though, the legitimate user would be able to use his key 15 seconds after the joker's last attempt.
-- supercat, Jun 15 2009


I'd like to see you build a completely mechanical version of that, [supercat]...
-- mitxela, Jun 15 2009


Not overly difficult, actually, if a little space is available. Probably wouldn't take much. Mount a spiral spring between two gears; attach one of the gears to a ratchet mechanism which turns a certain distance every time someone cycles the lock, and the other to a butterfly-style speed regulator (similar to that in a music box). Use an inline spring on the lock-operation control to limit the force that can be applied. If the timing mechanism is wound up too much (as it would be after repeated failed access attempts) it won't be possible to cycle the mechanism until it winds down.
-- supercat, Jun 15 2009


[FlyingToaster] had a good workaround, if I remember correctly.
-- 4whom, Jun 15 2009


I like this idea.

I think you could make the pin segments themselves into the chain, which might reduce complexity and therefore the necessary size.
-- Loris, Jun 15 2009



random, halfbakery