I think the firmware will start a retract a soon as M103 is found, and will continue to retract until either the extruder is turned back on, or the retract distance is met. So if there is a very short hop to the next point, it will not retract fully. This also means it is retracting *while moving*!! This is partly why the wipe path helps.
Retract while moving is good in some cases, like over long traverses where ooze will develop a blob on the nozzle before it gets to the next layer start.
Then I am not sure of the logic used for the prime of the next section, but somehow it is affected by how much was actually retracted on the previous suck. When the Suck & Prime values differ, I have no idea how the firmware performs that scaling. Also, I have found no way to specify the retract / prime extruder speed.
In Axon/Skeinforge there are adjustable values for Prime and Reverse Amount in mm and Prime and Reverse Speed in RPM, so there must be some way to access these.
I tried triggering destring manually, since there are retract and prime commands, but the BfB firmware moves the head to the warm-up location before performing the retract or prime, so it's not overly helpful.
I'm betting that the BfB terminology is whacked. (Not a surprise) From looking at the values, it would seem that Reverse is the Destring but Prime is actually a mid-print Purge done at the wipe box. Default settings are:
Prime Amount: 18 revolutions Prime Speed: 60 RPM
Reverse Amount: 3 revolutions Reverse Speed: 90 RPM
18 Revs Prime (@1 mm/rev) is a lot of Prime, while 3 Revs of Reverse seems about right for a Destring. From watching the lead screws, this seems to coorelate. Prime (Purge at the wipe box after heating to temp) is a long, slow, steady run. Reverse between deposits is short and quick.
So, perhaps this is a window into that to trigger and what not to trigger in order to get what we call Destring.
