Slic3r may be sending G92 commands to reset the position after each layer.
That has been reported to work in the past with Sailfish by funbart. But
it's not something which I test. However, if it's not working then I suppose
you could get slight skews if the position isn't reset the way it thinks it
will be. HOWEVER, I thought that slic3r and KISSlicer only reset the
E (or A & B) positions with G92.
But, and here's the interesting part, there's an issue with using G92 when
connected over USB to the bot with RepG: it doesn't work. When RepG sees
G92 Exxx
it is being told to set the E axis position to xxx. Unfortunately, RepG
is too clever by half and when connected to the bot wants to send the
coordinates for the X, Y, and Z axes as well. It doesn't need to -- it
just wants to. So, since it is connected over USB to the bot, it queries
the bot for the current position and then changes just the E axis part of
that position and sends that back. Unfortunately, the position it got
initially is likely garbage…. So rather than just set the E position,
it tries to set X, Y, Z, and E and ends up getting garbage for X, Y, and Z
which it then sends back and thus causes a problem.
Why is there garbage when it asks for the current position. Well, tracking
the current position from RepG's space of millimeters in absolute coordinates,
into the bots space steps in absolute coordinates, then into toolhead space
with offsets, then into and through a queue of planned segments is a big
headache in a bot with limited code space. So, MBI back around their 5.3
or 5.4 firmware stopped doing that. (When they first added acceleration to
their firmware, they dropped support.) We retained it in Sailfish just
enough to make G92 work. But we didn't also make querying the position work.
So, if you print connected to the bot over USB and use G92 then you can
see problems. (Comes up with Cupcakes which is how I first analyzed this
issue.) So, if you're printing over USB then this is just yet another
reason to not do so and instead print from an SD card.
And, on the other hand, this may not be the problem you're seeing.
Dan