A redstone torch burnout clock seems to not work quite well: the torches themselves pulse, but the connecting redstone wire does not
To reproduce: Place two solid blocks next to each other on the ground. Place four torches, on opposite sides of the blocks. Put one solid block on top of each torch, and run a redstone line atop the two original blocks.
What I expected: Random and sporadic pulsing of the torches and a one tick pulsing of the redstone line above the torches
What happened: The redstone line stayed solidly powered through the cycles of the clock, causing this to not qualify as a one tick clock.
Description of the screenshots: In RedSideOn, the torches of the red side are both on, and in Bluesideon, the blue side torches are on, and in both the redstone lamp is powered. These two screenshots represent the only two states that the clock goes through in a cycle, where in neither state is the upper redstone (represented by the lamp) un-powered.
Add a comment with any issues regarding this. The closest I could find to this was MCPE-11547, but I feel that the other report is not quite about this.
Attachments
Comments 13
migrated
Confirmed on 0.13.0 / iOS. At first I thought this was just another example of MCPE-11871 but soon realized I was wrong. In this setup, torches behave as if they had 1-tick delay for turning off but no delay for turning on. That is,
Torches on the blue side are turned on. On the red side they are turned off. Due to the solid blocks above them powering the wire, the blocks on the ground are weakly powered at this point.
1 tick later, two torches on the blue side turn off. As soon as the wire gets unpowered, torches on the red side turn on and thus the wire gets powered again in the same tick.
1 tick later, torches on the red side turn off, and then blue side torches turn on instantaneously.
However, this is contrary to [MCPE-11869], which shows that torches have no redstone delay for turning off (despite a visual delay being observable) (Correction: In my comment to [MCPE-11869] I proved that torches have 1 tick delay for turning off). The reason for torches on sides being alternately turned on and off is unexplainable as well, because if my hypothesis above were correct, all of the 4 torches might stay turned on. This may have something to do with the 2nd bug reported in [MCPE-11855] but I'm not really sure.
Confirmed on 0.13.0 / iOS. At first I thought this was just another example of MCPE-11871 but soon realized I was wrong. In this setup, torches behave as if they had 1-tick delay for turning off but no delay for turning on. That is,
Torches on the blue side are turned on. On the red side they are turned off. Due to the solid blocks above them powering the wire, the blocks on the ground are weakly powered at this point.
1 tick later, two torches on the blue side turn off. As soon as the wire gets unpowered, torches on the red side turn on and thus the wire gets powered again in the same tick.
1 tick later, torches on the red side turn off, and then blue side torches turn on instantaneously.
However, this is contrary to [MCPE-11869], which shows that torches have no redstone delay for turning off (despite a visual delay being observable) (Correction: In my comment to [MCPE-11869] I proved that torches have 1 tick delay for turning off). The reason for torches on sides being alternately turned on and off is unexplainable as well, because if my hypothesis above were correct, all of the 4 torches might stay turned on. This may have something to do with the 2nd bug reported in [MCPE-11855] but I'm not really sure.
In my last comment I suspected that torches had no delay to turn on. This was proven to be wrong in the general case by the circuit shown in
[media]. It has a NOR gate where one of its input is directly connected to an activated lever and another is negated by a torch. When the lever is deactivated, there will be 1 tick of moment where both inputs being low if and only if the torch takes 1 tick to turn on, producing a 1 tick pulse from the NOR gate and thus initiates the clock connected to its output. The result was that the clock did get initiated so the turning-on delay was proven to be 1 tick, at least in this setup.
In my last comment I suspected that torches had no delay to turn on. This was proven to be wrong in the general case by the circuit shown in
[media]. It has a NOR gate where one of its input is directly connected to an activated lever and another is negated by a torch. When the lever is deactivated, there will be 1 tick of moment where both inputs being low if and only if the torch takes 1 tick to turn on, producing a 1 tick pulse from the NOR gate and thus initiates the clock connected to its output. The result was that the clock did get initiated so the turning-on delay was proven to be 1 tick, at least in this setup.
I think it is possible that [MCPE-11953] is the source of the glitch.
