Weak　faults　in　planetary　gear　trains　are　difficult　to　detect　due　to　the　interference　of　background　noise　and　the　amplitude　modulation　effect.　In　order　to　detect　weak　faults　in　planetary　gearboxes,　this　paper　proposes　a　comprehensive　diagnostic　methodology,　referred　to　as　improved　maximum　correlation　kurtosis　deconvolution　(IMCKD).　The　proposed　diagnostic　methodology　combines　MCKD　with　variational　mode　decomposition　(VMD)　and　the　grasshopper　optimization　algorithm　(GOA)　to　solve　the　difficulty　of　selecting　MCKD　parameters.　An　optimal　mode　obtained　from　VMD　is　selected　to　determine　the　deconvolution　period　of　T　in　MCKD.　The　filter　length　L　and　the　number　of　shifts　M　are　determined　by　GOA.　The　effectiveness　of　the　IMCKD　methodology　is　proven　by　the　successful　detection　of　an　incipient　root　crack　in　the　sun　gear　of　a　single　stage　planetary　gearbox.　The　advantages　of　the　proposed　IMCKD　method　are　further　highlighted　by　a　comparison　between　two　illustrious　weak　fault　diagnosis　methods,　MED-SK　and　PSO-SR.