:twisted: Here goes.
I dispute the automatic '1 in 6' die rule that you quote - go back a few pages to see why I believe an average person (END +0) would always survive - so it would be people with low END, illnesses etc that would have problems. But, let's keep on with your assumption.
If you have 5200 people in low berth, roughly 870 will die, just given your 1/6 ratio. So you'll need more than this number of bridge crew to ensure that some will be alive to pilot the ship at the end point. Not to mention that, if you want a 1 week watch by only one person, every single passenger will have to be woken up twice to do so. Given your 1/6 ratio, every time someone wakes up, they run that risk; you increase the death toll to around the 2600 mark - so, roughly half of your initial passengers, over a 200 year journey will be dead on arrival. Sort of. What is the breakdown rate of a piece of machinery in operation for 200 years continuously? Even if we assume 1% per 10 years (somewhat low as it would probably increase exponentially over time), that's another 1040 or 520 if we assume roughly half the failures were from the half of the passengers that will be dead anyway. We are left with roughly 2080, less accidents, suicide, other damage to the ship, etc.
Firstly, then, we need to work out a more usable mechanic than the one which arbitrarily gives us a 1/6 chance of death per occupant (see previous posts).
Secondly, to make the 'sleeper ship' idea work, you'd need a far higher level of automation; maintenance bots for all of the low berths would be a priority, autodocs if possible; and if you had that much advanced a TL, why are you bimbling around in STL 'sleeper ships'?