The idea was that one ring would provide about 2/3 Gs of acceleration to a 100kg (this should have been 10kg, I forgot space otters were teeny) object for 100 years.
Here's the problem -- acceleration doesn't matter. What matters is 'power' or 'work' or 'kinetic energy' which also depends on the velocity of the object in question. Intuitively, this makes no sense to me -- if I'm floating in space and want to push myself with 1G of force, why does it take 10 times as much power if I'm going 10 times as fast relative to (I guess) whereever I started?
But, blah, screw it. I'll do the math. Let's assume the max speed they'll be used at is what they'd get up to after three days of acceleration, and go with that.
Okay, so each ring needs about 48 kg of antimatter.
If we assume the 100 year lifespan was for 'normal use' which was one three-day constant-acceleration each-way trip or so a month, 4.8kg of antimatter.
Obviously, repulsors have to cheat. They already cheat by being reactionless, so screw it, they'll cheat by giving proportional acceleration to the power input regardless of current speed. Using Schlock Mercenaries' math about the safe storage of antimatter, and aiming for a 2kg ring or less (that mostly hovers itself around, so it's not like the space otter feels that weight), I get a break-even speed of 1000m/s.
So if parts of your machinery can move faster than that without flying apart, you can generate power from nothing using repulsor feedback. This is a LOT faster than conventional jet turbines, but it just might be plausible for them to generate power that way...
Oh, and it means a typical modern-day nuclear power plant could generate a repulsor ring in about 10 minutes. This is good. It means that it makes sense that the Reef really couldn't manufacture them -- not because they don't have the energy, since energy is free, but because moving that much energy around safely is not going to be easy.
Now, assuming you start by building giant power plants based on repulsor feedback technology, how long would it take you to make the tiny black holes that power the spaceships? Aiming for a 2500 ton singularity, I get an output of about 3 singularities per year from a ring of giant power plants circling the equator of your average earth-sized planet (specifically, 5 gigawatts per 100 meters worth of power plants). Even with ludicrous amounts of automation, we're talking like 100,000 man-years of work per singularity. That's good scarcity right there. >:)