Erik, since we're making this personal ...
Where I went personal, I directly quoted you, so who started that?

(a) you haven't corrected me.
of course not, since you have talked apples, which, as far as I recall, are different from oranges.

(b) you have yet to show where the negative supply you insist on has a role in receiving RS232 signals, let alone impvoing noise margins.
Where have I claimed that a "negative supply" improves anything> I have discussed input thresholds. OK, if you can prove - which you can't - that a noise margin of, whatever you can achieve with gnd and +3 (less than 3V due to saturation) is as good as a noise margin of 6V, I'll give in

(c) you've already stated in your own words that you can't produce any support for your notion that using a negative supply increases noise margins
Richard, come on I have said that a -3 input threshold "increases noise margins", I have NEVER stated ANYTHING about a supply except that the negative supply - if not used - does not do anything. If you want to refute my arguments, refer to what I said, not something else.

I've invited you to show a specification of any sort that indicates that even the MAX devices use their negative supply for that purpose.
Please find where I have stated the above. Again I discuss noise margins, you keep discussing chips ONCE MORE you reply to a statement about apples by discussing oranges.

This is getting far out of hand, you keep babbeling about "supply" I have not discussed supply, I have discussed threshold. You keep discussing chip, I discuss the merits of a negative threshold, as opposed to ~ +0.6V, I do not know where it comes from that you keep discussing chips and claim that argues agains an input voltage threshold statement.

If you can not live with me not agreeing with me, do continue this thread, I have no problem with you not agreeing with me if you present the disagreement as such, not by stating "the claim of twice the noise marging is better is false". You keep stating that I am wrong, where, in contrats, I have stated that your "solution" very well may work, just has a lesser noise margin.

Erik

OK a proposal
I agree that many chips use 0 and +3 as input thresholds.
YOU agree that +3 and -3 will give a better noise margin.

That should settle it, you get your chips (apples), I get my science (oranges).