The problem that you describe can be related to multiple factors. One of which you discuss which is feed incompatibility. Feedstocks become incompatible when there is a mixture of cracked and paraffinic materials. The mixing of these two materials allows for the saltation, precipitation or dropping out of the asphaltene material. This can also happen when the feedstocks contain a cracked stock and a virgin stock, such as you have referenced above. Once the asphaltenes have precipitated or dropped out, then they lead to accelerated coke formation. This occurs to the unreacted portion of the feedstocks whilst in the bottom of the main fractionator (aka main column).
Additional factors are:
Unvaporized feed , due to the amount of the high boiling fraction. This is especially problematic in units processing heavy/high-boiling point resid feedstocks. One normally examines the boiling-point of the feed and compares it to the regenerator temperature, feed pre-heat temperature and the pressure at the feed injection zone (this is the unit pressure plus the catalyst head of the riser, plus the cyclone P plus the P of the reaction termination device (aka the RTD)). This is analogous to performing a flash calculation for a crude unit. The higher the boiling point, the lower the regenerator temperature, the higher the pressure at the feed injection zone and the lower the feed pre-heat then there is a much higher likelihood of not vaporizing a portion of the feed. This issue tends to manifest itself in coking of the reactor cyclone gas outlet tubes, coking in the reactor overhead transfer line, coking in the reactor overhead line to the main fractionator inlet, and coking within the bottom of the main fractionator.
Also, there is unvaporized feed due to too low of a feed temperature. This prevents proper atomization of the feedstock, especially the heavier/higher boiling fraction. The results are the same as the unvaporized feed in the paragraph above.
Damaged feed injectors or too low of feed atomization/dispersion steam can also lead to poor feed vaporization. Gas oil feedstocks generally require between 1 and 2 wt % of feed atomization/dispersion steam. This amount of steam increases as the feedstock becomes heavier/higher boiling. A target for light resid feeds is 4 wt%, and for really heavy/high boiling feedstocks between 6 and 8 wt% atomization/dispersion steam can be required.
There is also an issue with higher conradson carbon residue and low catalyst circulation. However, this usually manifests itself as coking within the feed injector region.
Additionally, the coking within the main fractionator (aka main column) can be impacted by both the temperature within the bottom pool of the main fractionator and the residence time (level and volume) of the main fractionator. Many refiners have found that the temperature of the main fractionator bottom is unit specific. This temperature is impacted by: feed quality, process variables (Cat-to-oil, reactor temperature, riser residence time, atomization/dispersion steam rate, catalyst type), and the volume or residence time of the main fractionator liquid pool. Refiners with good quality gas oil, higher severity operations and small to medium main fractionator bottoms liquid pools can operate the main fractionator with a pool temperature of ~695F (~368C). This pool temperature drops significantly as the variables listed above change. It is not uncommon to find main fraction bottom pool temperatures as low as ~635F (~335C).
Finally, for those refiners who process resid feeds that contains high contaminant metals (for example – Ni, V, Fe), the need to passivate the contaminant metal Nickel (Ni) usually requires the passivation agent Antimony (Sb). If the refinery attempts to deposit the antimony too quickly, it is highly likely that a portion of the antimony will not deposit onto the cracking catalyst but instead be carried into the bottom of the main fractionator with hydrocarbon products and unconverted feed. Once in the main fraction bottom pool, the antimony catalyzes coke formation.