I do not know exactly what (s)he wants.
However, the question states that it is about server responses in a log file, and the fact that, when non-printable character are in the response, TS has trouble identifying what exactly was written.
If you convert everything that is written into the logfile into hexadecimal representation of the string, then evaluating server responses by inspecting the log file (just looking at it) is going to be a rather teadious task. Not many of us can translate '48 65 6C 6C 6F 20 57 6F 72 6C 64' back into 'Hello World' just by looking at it (and that is with deliberate spacing added).
OTOH having 'Hello Word[DC1]' in your logfile could easily be understood as 'Hello World' + Device Control 1.
[DC1] is the offical 3-letter ASCII representation of that character (I looked it up when I implemented Utf8EscapeControlChars).
The "problem" with Utf8EscapeControlChars() is that it absolutely not designed with speed in mind (it is meant for use in the Lazarus IDE, which does not do a lot of logging).
If you have a server that uses logging a lot, then having to use Utf8EscapeControlChars() on every log-string may cause slow down.
AlexTP has made suggestions (but he did not provide a patch yet) how to speed up this function (see
Issue #39573).
Bart