Philosophical Question(s)

Lately I have been working on the separation of the Soyuz spacecraft into the 3 different components that compose it. The jettison of the Orbital module (BO) is fairly simple to work-out, altought small details are unknown (to us at least), like does it use pyrotechnic devices or springs and what's the value of the force applied to push it away from the vessel ... Now, the separation of the Service module (PAO) from the Reentry module (SA) is less straight forward to implement mainly because of the question of the whereabout of the SA's thermal blankets that shroud it during most of the flight. What is know "for sure" is that the blankets are attached at the forward part (around the hatch ring) of the SA and also at the forward part of the PAO.

Right now, I'm contemplating the following three different theories (as usual, it is difficult to find one source that doesn't somehow contredict what another source said 9_9 ...)

1. Prior separation of the BO, the blankets are pushed away from the SA after been cut loose. (as seen in a animation from Anatoly Zak)
2. After separation of the BO, the blankets are getting loose as now they are only attached at the aft of the SA, so they open like a flower (a bit randomly) before been cut loose when the PAO/SA separation occurs. They then drift away. (Urwumpe's theory)
   
3. After separation of the BO, the blankets are getting loose as now they are only attached at the aft of the SA, so they open like a flower (a bit randomly) but stay attached to the PAO when the PAO/SA separation occurs. (as seen on a drawing on a french site)

Now, they all sound plausible, but my preference goes to the third for the following reasons:

1. I can't tell by looking at all the pictures of the thermal blankets applied to the SA (and also at pictures of the SA without the blankets on) if there is any devices placed behind them to push them away as suggested by the first theory.
2. It is known that the PAO performs an avoidance maneuver after its separation from the SA, in order to get out of the way and avoid possible encounter with the SA later on. Now if the blankets are detached from the PAO during the separation, they are going to drift around the SA and possibly strike it (for example, when the SA change its orientation). Of course as these blankets are relatively light (guess-estimated at about 1.8kg each) it may not matter much, as the difference of speed between the SA and the blankets will be relatively low ... so the question really is, how cautious are the Russian designers?
   

Any suggestions or solid source will be very welcome to solve this mystery :-) In the meantime I'll go ahead with that 3rd therory which, the more I think of it, sound the more likely ...