not entirely sure that this is the correct place to put this, but it is an issue. radio waves move in straight lines, and while it is possible to bounce signals off the ionosphere, you do not have much choice of where the signals end up. While it would be possible, with the late 1920s tech base that would arise after a collapse (based partly on your book and partly on the limits to growth study) to build very light and small heavier than air craft, they, and ships in the pacific and southern indian oceans, would need a dependable radio signal.
And so I come to the main point of the post; is 1920s technology sufficient, and is an economy of that sort of size large enough, to support a constellation of small satellites for communications with ships and planes far from land?
Certainly there were would be[url] transistors[http://en.wikipedia.org/wiki/Julius_Edgar_Lilienfeld/url], and possibly solar panels http://en.wikipedia.org/wiki/Cadmium_te ... tovoltaics. Rockets should also be around; Goddard was experimenting during the period, and given sufficient time the research would eventually be done again. Incidentally, Goddard also experimented with ion thrusters, so station keeping should be possible. But how to fuel the rockets? Goddard used KEROLOX, but there would be little to no available petroleum; but while KEROLOX is great as a fuel for its density, the specific impulse is not as great as HYDROLOX. But there is a fuel/oxidiser mix that falls neatly between the two; METHALOX. Methane would be abundantly available from anaerobic digesters.
So, the satellites would be feasible, and the rockets would have reasonably cheap fuel, but what of the rockets themselves? Alumium-copper alloy would be the easiest tankage material to make, given the comparative scarcity of lithium. Aluminium alloys can also be used in the engines, apparently http://www.nasa.gov/vision/earth/techno ... dient.html
But can anyone actually pay for it?
the total budget for NASA in 1966, the highest funding they ever had, was 39.3 billion 2009 dollars. to compare, the GDP of the US in 1923 was 867 billion 2009 dollars. this makes an equivalent program approximately 4.5% of the GDP of the US in 1923. It would probably cost less, but these are intended to be conservative estimates.
In conclusion, not only would it be possible for a post collapse society to set up a satellite network (assuming a sufficiently large population for the purposes of taxation and skill specialisation), it may also be possible to have a manned space-flight program.