Take for example the Greek letter chi, which could also be spelled khi (usually pronounced in English to rhyme with "high", but in Greek it sounded more like "key"). The sound it represented is "K" with an aspiration or puff of air -- actually the same sound as our initial "K" sound, again as in "key". It differed from another hard "K" sound without aspiration, which was represented by the letter kappa "K", and sounded like the "k" in "skate". (If you can't hear the difference, try holding your hand in front of your mouth as you say the two words. When you say "key" you should feel a puff of air, but when you say "skate" you should not.)
Standard English spelling of Greek names is filtered through Latin, which is where some problems develop. The Greek letter chi has the same basic form as our "X". Latin, however, used "X" to represent the sound "ks"; and because the Latin alphabet did not have a letter that corresponded exactly to chi, it used the combination "ch" to represent it. It did not use "kh" because it cropped "k" in favor of "c" to represent that particular sound. (In Greek "C" or "sigma" represented our "S" sound; and modern English "C" varies between [K] and [S] accoring to context.) So, where the Greeks wrote XAOS, the Romans wrote "Chaos", which has come down into modern English as "chaos".
Many Hellenists (=students of Greek language and culture), however, do not see why we should continue to spell Greek names as if they were Latin. Some of them therefore transliterate Greek words in a way that they feel represents the original spelling more exactly. This is why you will see "Khaos" as well as "Chaos".
So, the situation is not quite as chaotic as it might seem....