Question to the physics guy: why the rear suspension of E36 implemented as double wishbone, and doesn't represent real structure of E36 rear suspension? It is not that it cannot be done. It can make a notable difference in car handling/behavior.
That also makes me wonder if the rest was done right...
Semi-trailing arm + 2 links (per side) was considered as very good suspension in terms of handling and its simplicity.
It started its life in BMW Z1 (1989), used in E36,E46,E85/86, and lasted till Z4 E89 (2009-2016).
In motorsport - E36 STW/GTR/DTM, E46 DTC/ETCC/WTCC, E46 M3 GTR, E86 Z4 GTR, E89 Z4 GT3.
About a decade ago as a personal research, I have redone SimBins Race07 E46 WTCC physics and remade the suspension according to real E46 dimensions and geometry (with corrections to retain roll center dive/squat characteristics) and it made a huge difference in handling comparing to SimBins "double wishbone e46" suspension model. It does make a difference.
I don't know if you're wrong or right, not commenting on that, however:
Remember that we do not have any trailing arm geometry in AC. Multilink also has some issues, which is why it's preferable to use DWB.
DWB can create *almost* anything from the common suspension types. The E36 is essentially a multilink + trailing arm.
You *cannot* just put in the link connection points for a multilink or trailing arm. For a multilink due to the nature of the virtual pivot separation, for a trailing arm due to the nature of the wheel connection point.
The E36 is essentially a trailing arm (Wheel connection point not supported in AC so needs to be compensated for) AND it has a single upper camber arm also attached to it, effectively turning it into a hybrid or multilink type suspension.
I actually don't know how to even begin to calculate properties for this layout; it would need to be received from the manufacturer, or via measurement. Unless I am misunderstanding and the lower arm has a joint connection on the trailing arm and it is not fixed.
The correct way is to graph the curves IRL and figure out the IC/RC, then match that behavior. Simply putting in the link positions is extremely unlikely to work at all like it should; I know, I have tried in the past some half a decade ago.
Then you must also compensate the internal force error for vanilla AC, but it is doable while retaining behavior.
So unless you have all of the curves and compare it to this version in AC, it's difficult to tell how much error there is. You can't just look at it.
Also
@Patrik Marek like JPG said, you don't really need to make "a lot" of compromises, mainly just design the geo around the internal force. However if making an unconventional geo like this and not a normal DWB, it requires some special design knowledge.