Detecting large extra dimensions via "mini black holes"?

[Coin]

Occasionally I come across something explaining that one of the
possible discoveries at the Large Hadron Collider might be the
production of "mini black holes", which if found would demonstrate the
existence of large extra dimensions. The idea, at least as I
understand it, is supposed to be that gravity in an more-than-four-
dimensional universe could be actually much stronger than we measure
it to be; but in a way that only has any effect over very small
distances, because the gravitons leak out into the extra dimensions if
given the chance. The "very small distances", if I'm understanding all
this right, could possibly be large enough that colliding particles at
a modern particle accelerator would be able to trigger the creation of
microscopic, short-lived black holes. The reason I usually see given
for why people are excited about this possibility is that it would
provide strong experimental support for string theory, by showing one
of string theory's more controversial elements-- extra dimensions-- to
be not just a theoretical convenience but actually a physical fact.

An example of a writing on this subject, picked at random from google,
might be:

http://cerncourier.com/cws/article/cern/29199
or:
http://quasar9.blogspot.com/2006/08/new-dimensions-at-lhc.html

The thing that confuses me about this, though, is the frequent
implication that this effect-- the effect that allows for mini black
hole production in particle accelerators, that is-- wouldn't exist in
all theories of more-than-four-dimensional spacetime, or even all
variations of string theory. It's usually left very vague, however,
*which* theories would or wouldn't produce this behavior. This seems
to make it very difficult to evaluate what the detection or non-
detection of these black holes at the LHC would or wouldn't be telling
us. So, I would like to ask:

** What attributes would a string theory have to have, in order for
the extra dimensions to be in principle detectable through black hole
production in a particle accelerator? The one thing everything I find
on this subject seems to agree on is that the spacetime which the
strings inhabit must have more than four non-compact dimensions, but
are there any *other* conditions which either the spacetime, or the
string theory acting within that spacetime, must satisfy for the "mini
black hole" production to be possible? **

I have a few slightly more specific versions of this question I'd also
like to ask, but I unfortunately do not know very much about String
Theory so I am not certain I know how to ask them correctly. So please
excuse me if the following turns out to be gibberish:

1. As far as I'm aware, the preeminent variation of string theory with
"large extra dimensions" would be the "brane cosmology" models, where
ALL 10 or 11 dimensions are taken to be arbitrarily large, but the
perceived universe is lower-dimensional because the movement of
strings is restricted to the surfaces of lower-dimensional "branes".
(I somehow got the impression that the string theories with compact
extra dimensions can be approximated within this model by folding the
branes properly, but I am not sure about this.) Within a model of
braneworld cosmology, is it possible to see mini-black-hole production
at a particle accelerator within some individual brane-world embedded
in the bulk? And if so, would the mini-black-hole effect be *required*
by the use of braneworld cosmology, or is it possible to design a
brane-world where the effect which results in the mini-black-holes
does not exist?

2. If the answer to the last question is "yes, but only in a model
with the correct circumstances", then is the mini-black-hole effect a
property of the theory which describes the branes, or is it only a
property of some individual brane-world described by that theory? In
other words, would it be possible for there to be a higher-dimensional
bulk, contained within which were BOTH brane-worlds which allow the
"mini black hole" production, and also brane-worlds which do not?

3. The specific large-extra-dimensions model I seem to hear about the
most is this Kachru-Kallosh-Linde-Trivedi, or KKLT, construction,
defined by the paper the arXiv has at hep-th/0301240. My *incredibly*
limited understanding of this construction is that it assumes a
braneworld cosmology model (i.e. all ten-ish dimensions are noncompact
but strings live on a lower-dimensional brane) and then describes a
way of configuring branes within the bulk so that you create a nice de
sitter space for the strings to live on. (I somehow got the impression
that the "string landscape" you sometimes hear about specifically
refers to all the different ways of configuring the KKLT construction,
but I am not sure about this.) Does use of the KKLT construction allow
the mini black holes effect to be present within a brane-world? Does
KKLT *require* the effect be present?

Thanks!