Better start this off with a sweeping generalization: No true canyoneer has never gotten their rope stuck. It's a rite of passage into canyoneering so don't be embarrassed that you've stuck your rope. If you haven't then you need to go more often.
My goal with this article is to go over the most common reasons people get their ropes stuck for all the different ways of rigging. I will also go over some of my solutions to these problems.
There are two possible solutions to the problem of a bend getting stuck on a edge.
The first and most obvious is to put the knot in a place where it isn't a problem. If it is a short rappel, you can completely eliminate the need for the bend by going toss-n-go.
Setting up courtesy rigging over the first lip will eliminate this problem as well.
One can also set the length of the rappel side so that the bend is past all edges then it can't be a problem, either. These solutions will not work in every circumstance, though.
In non-ideal circumstances, we can use a flat bend instead of a traditional bend. A flat bend is advantageous because it doesn't have anything that can get caught on the lip as it goes over. Examples of a flat bend are the EDK, stacked overhand bend, flat double overhand bend, and flat fisherman's bend.
Flat bends aren't without disadvantages, though. All flat bends have can capsize, which is a very scary experience for the unfortunate rappeller. If the bend were to capsize enough times and the bend not have sufficient tails, the bend could roll right off the end of the rope!
The types of constrictions I can think of are between a chockstone and the canyon wall, or where the anchor is far from the lip and the rope travels through a skinny section before reaching the lip.
This is the best solution and the same as preventing the bend from going over an edge. Either go toss'n go or put the knot past the lip.
I discovered this solution the hard way. If you rappel over a chockstone and it looks like there's no way to not get the bend caught between the chockstone and canyon wall, you can likely just rig another rappel at that spot.
A stuck block is very similar to a stuck bend with all of the problems and solutions. Of course, the solution to removing the block would mean to go toss 'n go.
Yeah, I wish I didn't have to mention this one. Bend and block must be on the same side of the rigging otherwise the rope isn't going anywhere.
Along similar lines of thoughtlessness, remember to remove the safety before the last rappeller goes down. Safety is the "S" in my STEP acronym.
This is where things get complicated. The majority of rope pull issues k occur between the anchor and that first edge, where you can't see the ropes and often can't change how they're oriented.
This happens when the pull side somehow ends up on top of the rappel side. Ways to prevent this are as follows.
If the rapide doesn't twist when someone is on rappel (like it's attached to a chain), rig the rappel side by feeding it from the wall up through the rapide. The block, then, would go between the rapide and the canyon wall. This sets the block in the correct position so that when it's pulled it won't pinch the rappel side.
The simplest solution is to keep the lines separate for as long as possible. When you are going over that first lip, place the pull side in a different crack or rope groove than the rappel side. From there just keep them separated all the way down. Having someone at the bottom holding the pull side out of the way can help with this.
It is possible to do this with DRT to a limited extent, even with both ropes weighted.
In situations where it is impossible to separate the rappel side and pull side, the next best thing is to make a conscious effort to put the rappel side on the pull side. This is the natural thing to do since the pull side normally sits unweighted on the ground and the rappel side is folded onto the edge as the rappeller descends. Just make sure they don't accidentally get flipped around.
This refers to when the rapide is not resting on something while weighted and has the tendency to spin all on it's own and for whatever reason. I hate these. The best thing I have found is to see how they end up twisted when weighted and twist the rope down below accordingly.
This solution often fails to completely remove all twists but the good news is that the rapide can freely spin when pulling the rope so the twists at the top are rarely a real problem.
There are additional complications when rappelling DRT (Double Rope Technique) that mainly have to do with rope twist.
This happens when the last rappeller uses a figure 8 style rappel device with DRT where the strands are not isolated.
By far the best solution is to use a non-figure 8 rappel device such a totem in throttle mode. It completely eliminates this class of problems and forces the ropes to be untwisted when the rappeller reaches the bottom.
Other solutions which work to varying degrees of effectiveness are to clip a sling/carabiner combo into one of the strands above the rappel device. As the rappel descends and the rope twists in the device, the carabiner will twist around the ropes and it will be easy to determine how much the rope has twisted when the rappeller gets to the bottom. The sling can create a lot of additional friction if it twists a lot, though, and can even act as a friction hitch and stop the rappeller. For this reason I don't recommend this solution.
The rappeller can place their finger in between the ropes and force out any twists as they are coming down. This pretty much eliminates any fun that is involved with the rappel since you spend your time fighting the rope the whole way down. The rappeller could also add a carabiner to their leg loop and clip it into only one of the ropes to kind of accomplish the same thing. I haven't tested this yet method, though.
My usual solution, which is still terrible, is to count the direction and number of "bumps" as I go down, each bump corresponding to one half-twist. If you're below me, you might hear me mumble, "Three counter-clockwise", meaning once I get down, the ropes have to twist back three half-turns counter-clockwise for them to be untwisted again.
When free hanging it is common for the rappeller to spin a few times on the way down. If the rappeller is going DRT or they have the pull side bagged and clipped to themselves this spinning can cause the two ropes to twist on the way down.
My preferred solution to account for this is to count the number of times I spin going down. As soon as I touch down I then spin in the opposite direction that number of times. Then I de-rig and the two ropes should be correctly untwisted.
Once everyone is down and all the twists are removed, there are still several things that should be done to simplify the pull.
If the rappel side length wasn't set, it will likely have rope twists in it. As the rope is being pulled, someone should be holding the rappel side and remove any twists as the line goes up.
Often you can eliminate drag on the pull (and prevent rope grooves) by stepping back from the drop and pulling at an angle besides straight down.
Retrievable rigging always has extra stuff that must come down besides just a bend or block. Toggles have the toggle and the twisted rappel side that needs to go around the anchor without getting caught. Sand traps have the trap and also must have a place for the sand to empty to either side before it can be pulled down. Water anchors have the pull cord and must have room to empty the water bladder. Hooks are especially finicky on the pull and are rarely used for this reason. Even with rope-only systems such as the CEM there will be a looped rope and potentially lots of twists that must come down.
Each of these have special considerations that are beyond the scope of this article.
Don't ascend if either strand is out of reach. Let me know what went wrong and what you think you should have done.