A useful question and one for which the general lack of knowledge causes the likes of OS, Mutt and me considerable angst ...
climb gradient should all intersect at the same point
As OS has noted, there may (and generally will) be a number of obstacles .. you may have one which is critical for all RDS data or, indeed, a different obstacle for each OLS. Solving the trig for the published data quickly tells the story. It is worth noting that the trig model required is pretty straightforward .. if you have any difficulties, either OS or I can give you a hand ..
you may by simple geometry calculate the exact height and distance
Generally, the original data is subject to a bit of round off and the runway slope can introduce some approximation errors. However, in general, the simple trig analysis (which must include an approximation for slope to be at all useful) will give results which are of more than sufficient accuracy to be useful for runway analyses.
where the intersections are actually "phoney" obstacles
The intersections may be conservative artefacts of the trig (ie "phoney") or they may be the actual obstacles, depending on the runway obstacle profile .. doesn't matter at the end of the day (unless you are ducking under the planned departure profile .. in which case, as passenger Bloggs in row 95, I would prefer the phoney obstacle).
EXCEPT for a possible more distant obstacle
One needs to keep in mind (at least for Oz) that the bulk of RDS data (perhaps all ?) is based not on inclino survey but, rather, discrete obstacle survey within the prescribed trapezoid. The surveying agency then works backwards to come up with the relevant inclino data. If this is the case, then distant obstacles will not be considered. (For an inclino survey, generally the distant obstacle is not identified as such and will be included by default .. although there would be nothing to preclude known distant obstacles being excluded).
Outside the trapezoid (both laterally and on the extension) .. the pilot is on his own if the operator doesn't do the sensible thing.
did I learn well
.. indeed, old son ...
if I pick the most limiting I will clear all other obstacles?
Yes, and no. Keep in mind that all we are doing here is defining bumps in the departure profile. It is up to you to make sure that the aircraft NFP keeps above them. Main problems -
(a) third segment acceleration which will need to be massaged into the picture.
(b) different segments have different gradients
My understanding is that the above still only applies within the surveyed area
Depends on the survey .. refer to other comments on the matter.
At this stage I am not concerned about out of the surveyed area
I suggest that you should be as that is where 90 percent of the work is involved.
I do remember back in the F27 days
As do we all .. (are you three parts deaf like the rest of us ?). Worst situation will be twin-jets with a large V2/Vcl split. For instance, the DC9, on a similar basis .. is looking at something in the order of 50nm, as I recall (a while, now, since I have played with DC9 RTOW numbers).
I knew he'd take the hook
We can't help ourselves, unfortunately ..
I wonder how he deals with those pesky slope elevation problems
Generally, one can obtain a surveyed centreline profile. Either the actual profile can be used or, more usually, it can be linearised.
Keep in mind that it is usually no great problem to find out who did the survey for a given aerodrome and liase directly with the surveyor to find out what was, or was not, done. While the big airports sometimes can be a little public service in style, most of the smaller airport owners (and the individual surveyors) are a delight to work with and I have no hesitation in singing their praises.
.. and the best bit is that the surveyor, often, will just give you the survey RL data for the obstacles of interest.... and Bob's your uncle ....