I first noticed that TNT can destroy blocks farther away than minecarts. To test this properly, I recorded the damage that I receive at point-blank range. I used Resistance IV so these explosions would deal damage but not kill me.
How to reproduce:
Surround yourself with obsidian, or another explosion-resistant block. Be sure to place a block above your head to prevent fall damage.
Use
/effect @s resistance 1000 3.Use
/summon tnt ~ ~ ~and notice that it deals 10 damage (this would be approximately 50 health without the status effect).Use
/summon tnt_minecart ~ ~1 ~and notice that it only deals 5 damage (roughly 25 without the status effect).
For comparison, creepers deal 8 damage at point-blank range with Resistance IV (roughly 40 damage without).
I expected TNT minecarts to deal 10 damage, like the TNT. This damage reduction may be intentional, but I figured I would report it.
Edit: According to the Minecraft Wiki, a TNT minecart has a more powerful explosion if it is moving faster. When I tested this bug, the minecart was always stationary. So this could be completely intentional.
The edit could be removed. This bug still exists in the current version; 1.21 and falling TNT minecarts seem to not get any increase in power at all for Bedrock/PE.
I have uploaded differences when using a TNT minecart contraption between Java and Bedrock/PE. While I didn't expect both versions to be exactly the same, the difference is quite sad for Bedrock/PE.
Setup:
7x7 Crafter(minecart)>Crafter(TNT Minecart)>dispensers hooked up to a redstone clock, timing both repeaters at max delay, repeaters along the way to extend the signal and slightly stagger minecarts.
The mountain in the image had a peak of y150 and was near solid of stone variants (no spaghetti caves, barely any air. Seed at bottom with co'ods of mountain demo'd) from there to around y-3 before hitting the deep dark cave that was air from -3 to -44.
Results:
Bedrock/PE -
To me it seems as if the game isn't taking velocity in to account. On Bedrock/PE, the TNT minecarts can barely ever do better than 4x4 and never over 5x5 clearance on stone blocks during the tests. In the screenshot, you can see massive spots missing when there is only 5 blocks between each cell. This should have made a clear hole. Even shrinking to 4 block distance doesn't clear all blocks on bedrock or help dig deeper using this contraption. With two cycles of the contraption emptying it's load (128 dispensed per cell) in any setup, we could never get deeper than about y60 using this mountain.
Java -
Almost hit deep dark after first cycle (64 TNT minecarts per cell). The second cycle wasn't even half way before hitting bedrock/water trap. I would estimate no more than 100 TNT minecarts per cell.
Conclusion:
I wouldn't expect features to be this different or that this feature has been so under powered in bedrock compared to java. It took less than 6k TNT minecarts to make that 4 chunk hole (roughly 32x32 blocks, if not wider) on Java. For Bedrock/PE, we barely went 80 of the 150 blocks needed (while missing massive chunks) to see the deep dark at the bottom with 6.2k, the total amount for two full cycles.
Attachments: mcstructure file for Bedrock/PE and nbt file for Java structure blocks to load the contraption fully stocked for one cycle of TNT minecarts in case anyone else wanted to test/confirm these results.
Mountain used
Seed: -1207704244 Coordinates: -2484, 150, -1714