I love the notion of using lingusitics to deduce the root cause of garbled instruction manuals. That said, I can't comprehend why manuals are not submitted to a native reader for a final proofing step. Compared with the hardware and software development costs, the native reader review would cost nearly nothing. Furthermore, it would enhance the reputation and functionality of the device if the manual made sense. I'm sure that I'm only one of many scientists and engineers who would be pleased to provide their consulting services for the greater good.
This speaks [heh] to one of the enjoyable vestiges of my time studying Linguistics: I like to guess at the native language of the author(s) of a manual. Even if the English is perfectly correct, English has several WAYS of being correct, and the details of the structure, the options used, the STYLE, hints at original language. The position of modifiers, verbs, and using or eschewing complex tenses or indirect objects or subordinate clauses, etc. are telling. If there ARE errors, it becomes easy.
Given the presumed national origins of documents, I can tell that Japanese is very polite, but has some kind of weirdness with verbs (relative to English), but different from German, and Romance languages handle adjectives differently. Germans love their modifiers, Chinese hate tenses, and Scandinavians might just consider getting rid of prepositions altogether.
I have a collection of hilarious accidents (as well as deliberate ones like the "Damn Fast Buffer Amplifiers" and "Write-Only Memories" of legend). There's plentuy of hilarity on "Engrish.com" but adding the geek factor by collecting examples from my field amuses me.