This is simple. You don’t exactly reproduce on paper what any of your characters sounds like, because you would write incomprehensible gibberish all the time.
So when you go to represent the speech of a character with a regionally or ethnically distinct speaking pattern, you keep it as simple as you can. You might have your character use a few distinctive vocabulary words, and you might have him pronounce a few things differently, as captured in what you write. Whatever it is, two rules:
1. Keep it simple.
2. Keep it consistent.
Think of it as a little algorithm or a translator. You put normal speech in one end, apply the couple of rules for the character, and his distinctive speech comes out the other end.
Here’s an example from what I’m writing now. Bob speaks with some sort of unidentified Estuary English (Hackney / Essex / somefink), and I write his speech with a few simples rules, which you can see in action here.