Which knot to use when two ropes are connected and "lower" contingency setting

I often go with beginners and set a releasable figure8 block for contingency options. Often, we'll have a long rappel that requires connecting two ropes together. In order to avoid "passing a knot" when activating this contingency plan, I set the knot connecting the two ropes below the anchor point (i.e. load is on the knot). In the link below (min. 2:28), Rich C suggests to use "retreaded-figure8" or "double fisherman" bend in such situation, but I was wondering if using a square-fisherman bend (image posted below) is also ok (I would think so, but want to hear from the experts).
I know it has more chance of getting stuck, but the reason I prefer this knot sometimes is because it's very easy to construct as well as easy to undo.


​​https://www.youtube.com/watch?v=GQn9m8ehffw (min 2:28)


http://canyoncollective.com/attachme...1e9409dea3d373

Cheers

I'm far from an expert, but I've always used a double edk, at 4:55.

If I'm just tying two ropes together so I can rap on one and use the other to retrieve I use a simple overhand knot.

If I'm tying two ropes together because one is not long enough I use a rethreaded figure 8, or a rethreaded figure 8 with a bright so I can tie off while passing the knot.

KISS is my golden rule when canyoneering.