Bolt over base with no ejected case, jammed in, can also be caused by high cyclic speeds.
I built a 6.8 with 16" barrel, and was using an A1 stock and rifle buffer. Worked great, ejected 2 - 4 pm with various 6.8 loads. I later switched to a carbine stock, tube and buffer, and it constantly jammed up. I realized that bolt speed was greatly increased with the light buffer weight causing the round to be cycled so fast it didn't have time to escape the port. The bolt would rebound, with casing still attached and attempt to chamber the empty with the new round.
Added a tungsten weight kit to bring the buffer to an H3. It immediately returned to functioning correctly. No issues since.
In tuning buffer weight, it's better slightly heavy than too light. All it will do is slow the cyclic rate from the high speed 900 rpm to the better 600, giving enough time to eject and then feed correctly. In the OP's case, it aint broke, don't fix it. There are some who do use a heavier than rifle buffer - the 9mm blow back, as it's all there is to delay the action. No gas.
Crane is the contractor who sets up and modifies short barreled AR's for shipboarding etc and their recommendation with 10.5" barrels is up to H3 to slow the cyclic. It's almost unecessary with a 20" with rifle gas as the pressure is about 10k psi lower at that port than carbine. I've heard of other methods to control where the brass lands, one is clipping the ejector spring by one coil. For a range only rifle with close stations, prone shooting, it seems to have a following. But - they are attempting to eject closer to 1 than 4 to avoid hitting other competitors, and hoping the guy on their left is equally considerate.
Me, I bought the net bag to collect brass. It works. Tuning, as I have discovered, has it's pitfalls.