In reality the SUU-64/65/66 canister dispenser associated with the CBU- 87, 89, 97, 103, 104, 105 weapons dispenses according to its burst fuze device which is almost always a radio proximity device but older strictly-time based fuze devices also exist. The USAF appears to use the FZU-39 while the USN tends to use the FMU-140. Both are roughly equivalent with the -140 having the option for an optional timed release and is more complex. This means it’s “AGL” although if you dropped it next to a sheer cliff it would trigger based on terrain presence in any direction. The FZU-39 has a barometric sense to prevent triggering until going downward which prevents premature triggering during low altitude tossing such as over hills. Even the WCMD type weapons have no communication with the proximity distance setting. There is no way to change the sensor in flight and aircraft data entry is for informational purposes only. WCMD deliveries that involve spin (CBU-103) can be changed over 1760 interface. All WCMD canisters fly without spinning and those that use spin to disperse begin to spin shortly before. The remaining types (CBU-105) use inflatable gas bags to dispense and never spin in any phase of flight.
The submunitions themselves are pretty simple. The CEM class of weapons consist of several CEBs (BLU-97) which are combined effect bomblets. The effects are: HEAT, frag, and incendiary. The HEAT penetration of the BLU-97 is substantial equal or better than the Mk 118 of the Rockeye II. They were definitely a threat to MBTs of the era of their development. All of these HEAT (and the SFW) are hit-to-kill. The secondary (tertiary) effects of the CEB/RE/SFW are pretty weak. They aren’t area weapons in the sense that they do their primary damage to an area. Their primary effects are extremely concentrated and secondary effects can be treated as a uniformly distributed “damage cloud”.
The problem, as I understand it, with BMS handling SFW is that it doesn’t simulate the guided(ish) nature of the weapon. If BMS treated the SFW as a simple hit-to-kill munition then at the number of SFWs in a weapon and the expected coverage area then actual hits on vehicles would be impossibly rare. There are only 40 SFWs in a CBU-97/105 and are only effective because they trigger based on detecting that they are pointed at something worth attacking. Instead BMS seems to model the 40 SFWs as a sort of equivalent unguided damage potential against a wide area. This is an imperfect simulation because in order to be properly destructive against a reasonable target array it is much, much too powerful against a concentrated or numerous target area. In the same way a Mark 84 will destroy all the trucks within a certain radius no matter if that is two trucks or two hundred, the damage isn’t “kill limited”. It will happily destroy as many or as few targets that are present within the blast radius. SFWs absolutely do not work this way in reality. A destroyed target effectively removes an SFW’s worth of destructive potential. Forty SFWs can at most destroy 40 armored targets. Drop 40 SFWs over 40 targets and you might get 40 destroyed vehicles but deployed over 400 targets cannot destroy 400. The Mark 84 bomb type thinking doesn’t apply.
Ideally the whole mechanization of the SFW chain would be modeled in detail a la DCS but short of that there’s really a no-win situation in how to model the thing as a simple cluster bomb. Without the “canceling” of a target damaging a target somehow not damaging a different target how can you make destruction that won’t go insane when the targets are densely packed?