I noticed that the buttons are very unstable with respect to the runtime in the bedrock:
They emit a charge above or below (19 to 21 for stone buttons, 29 to 31 for wood buttons) at random. They rarely execute the exact value of ticks (which is 20 for stone buttons and 30 for wooden buttons).

To analyze the existence of this error, I compared the execution time of the two types of buttons (stone and wood) in Java and Bedrock. To do this, I used a command_block in two situations on both platforms:

• FIRST SITUATION:
  1) I put command block in repeat mode, requiring redstone, and used the command "/ say a";
  2) I put a type of button (suppose stone) and pressed once;
  3) I counted each of the command lines manually.

1. In Java, totaled exactly 20 ticks in all tests.
2. In Bedrock, they totaled values ​​between 19 and 21, which caused misalignment in succession tests.

• SECOND SITUATION:
  1) I created a chat goal with "/ scoreboard add add Bug dummy";
  2) I put in the command block: "/ scoreboard add @p Bug 1";
  3) I put one of the button types (suppose wood this time) and pressed it once.

1. In Java, there was an exact score of 30 ticks.
2. In Bedrock, as well as in stone buttons, there was a range of 29 to 31 ticks, damaging, as is supposed, the following tests.

Still with the same error, this made some attempts to run very complex redstone circuits difficult, as noted in ticket MCPE-15793.

This error causes a number of breaks in Minecraft Bedrock's redstone, which is perhaps the most damaging to engineers.