I don't think having the solenoid within the loop is a problem, in fact it would probably help the durability.. Its is highly unlikely the the solenoid will fail in a way that would vent gas out of the casing, it is much moe likely that the internals fail, either open or shut.. in the open case it doesn't matter if its in or out of the loop its goinfg to flood the loop with o2.. In either way external or internal you could add a whitey valve to manually shut off the o2 and still have access to the gas for oc deco bailout (assuming a 2nd stage attached) or just reachback and shut off the o2 valve the flow rate is not that high, granted it will rise quickly but it can be managed.. under normal operation you operate the inspiration with the o2 valve just cracked open , so there is little flow into the first stage(also the ip on the 1st stage is lower than usual), and the solenoid is a restricted opening, even if did come apart, unless you were real deep you have time to manage the gas.. in the case of an external unit, unless you have a valve to shut the gas to the solenoid, you cant even fly the rebreather manually adding o2.
Every rb has its pluses and minuses, The main thing I didn't like about the prism is not being able to set your own setpoints on the fly, your only way is to fly it manually, and it many cases you have to shut down the electronics.
My next RB is an infinito by powerbreathe, its a commercial/military unit designed for use to 250meters/820fsw (about $18.5k), I already have it on order and hope to have it for the beginning of dive season.. From all the units I have examined its the best.. Its primarily designed by stuart clough and is an evolution of the ut-240..
I put a few links up on a msg somewhere on this board, they are in italian and spanish but are easily translated using babelfish..
Joe