Two member onset consonant clusters with /s/ as the first member (#sC onsets) behave differently from other double onset consonant clusters in English. Phonological explanations of children's consonant cluster production have been posited to predict children's speech acquisition. The aim of this study was to consider the role of the Sonority Sequencing Principle (SSP), homorganicity, headedness, and factorial typology in the explanation of productions of consonant clusters by children with phonological disorders. Thirty monolingual English-speaking children with phonological disorders produced words commencing with #sC: /s/+stop (/sT/), /s/+nasal (/sN/), /sl/, and /sw/. There was wide individual variation between participants. Overall, there was an increase in accuracy of consonant cluster production as the sonority indices of the second element of the cluster increased: there was significantly greater accuracy of /sl/ than /sN/ and significantly greater accuracy of /sN/ than /sT/. However, /sw/ targets only followed the SSP for phonologically accurate (but not phonetically accurate) productions. Correct renditions justified the binary grouping of the SSP-following vs SSP-violating separation. Homorganicity of the cluster was not a factor in determining relative accuracy of the targets. The percentage of correct renditions justified the /s/+[-continuant] vs /s/+[+continuant] binary separation. There were several counterexamples of predictions of the headedness approach. Thus, sonority and factorial typology were found to be useful in predicting /s/ consonant cluster acquisition of children with phonological disorders. (Contains 2 notes and 6 figures.)