Get us a video.
I just dug into my WorkBench that I bought used. There are so many things that can go wrong that it is difficult to know where to start. The good news is that in a matter of a few evenings, I managed to fix them all without major modifications.
Item number one is the VREF especially for the Y axis. Don't worry about damaging a motor as that is unlikely. The reality is that Y motor is exactly the same motor used on SD 2,3 & 4's. It was under powered and overdriven on those models so a bigger machine with dual extruders isn't going to improve the situation. Open up the case (Another issue for later) and turn up the pot on the Y axis just a tiny bit (1/16-1/8) Clockwise. This will help but it won't compensate for the mechanical resistance of this machine.
Secondly is verify that the X gantry rods are actually perpendicular to the Y axis rods. I'll put money on it not being so.
The epoxy on one or more of the carriages has most likely let loose and will need to be repaired properly. This means taking the time to pull the gantry off the rods and buying some epoxy. This lead to the next issue that can be easily fixed.
Thirdly is fix the (previous) issue of the X rods not being parallel with each other and the bed. Also Perpendicular to the Z-axis. Once the gantry is out, find a good flat surface and lay the gantry (Minus the extruder and carriage) upside down and ensure things are flat before your 5 minute epoxy hardens. This is a god time to check the distance of the two Y-carriages. Then reassemble ensuring everything is clean of all grease and use a thin oil instead.
The fourth thing will be proper alignment of the belts and drive rod pulleys. Tension the belts evenly so they are snug, do this while they are clamped to the Y carriages. Then loosen the grub screws on both drive pulleys and slide the gantry forward against the stops. At this point if things are properly held in place both carriages should touch the stops (or real close to it) at the same time. This gantry binds very easily so alignment is critical. If things are moving smoothly then it is time to slide the gantry back and forth several times. You will need to apply pressure evenly with both hands on both sides. This is when and how you are going to feel for misalignment and binding. Once you get a feel for that happy neutral position you can tighten the set screws on the pulleys.
The Fifth problem is the rear drive rod itself. It needs to be lubed with a light oil on the bushings. The trick is to get it into the bushing and not all over everything else. Loosen the bolts that hold the interior support bushing bracket. This will allow you to oil it easier and align it with the rod as you tighten things back up. Check the motor drive belt and ensure that it is lined up with the pulley on the drive rod. This is true for all belts on the machine, if they ride to one side it can add serious resistance to the movement. The motor belt should be tight but not so much that it adds resistance to motion.
Any pulleys that idle on a bolt or shaft (front Y axis pulleys & left X axis Pulley) will benefit from a little oil as well.
The Sixth major issue is the location and routing of the (enormous) wire loom that goes up to the extruders. Anything you can do to lighten that up and allow nice, non interfering motion is going to be a must.
The Seventh major flaw in this design is the location of the control board. I appreciate SD's efforts to improve fit and finish but I am of the opinion that it should be outside the printer. There should be fans keeping those drivers cool and the little fan on the bottom of the frame would be enough if there was a source of air near the drivers. So cut some holes in that box and add another fan at the very least. Best would be to mount it on the rear with a good fan cover plate.
The Eighth major flaw is in regards to the Z axis threaded rod and nut. This setup has over 1mm of play in the nut and due to the huge cantilever platform, it will chatter. Clean the threaded rod of any debris and grease and apply a thin oil. Then find the proper well nut to match those threads. (I haven't located one yet) what you need is something that you can screw down from the top with some cushion in between the two nuts. This is a proven technique on other SD models and it only takes a couple of seconds to install. If your Z Screw is bent (mine is not) then you need to straighten it or replace it.
The ninth major flaw is using bushings on the Z axis guide rods with such a short space in between them. This is what causes most of the bed chatter as it moves down. SD was relying too much on gravity with this design. To fix this issue you want to clean and lube with light oil. If that still doesn't fix it then consider some form of bungee cord that can be attached to the platform near the guide rods. They will need to pull down to the bottom and most likely have to loop back up to the top of the frame. This will allow enough pressure through the entire range of motion.
The Tenth and (my opinion) largest flaw is in the dual extruder design and the autolevel garbage that hangs on it. Get rid of all that crap and switch it over to a good mechanical leveling and Z height adjustment system. The dual extruders have no easy way to tweak the nozzle height. I would guess this is the major cause of any shifting encountered given all the other issues above. Even with everything as near perfect as possible, the other nozzle still catches parts of the previous layer. There is nothing that can be done to remedy this issue other than switching over to a single extruder design. That is exactly what I did with a completely redesigned carriage. Over the next couple of days I will try and post up some pictures.
The only way this machine will be truly capable of dual extrusion would be with an E3D Cyclops run with dual Bowden setups.
I also was very disappointed with the butcher job SD did on the firmware for this machine and their Solidoodle MotherBoard. Thermal runaway protection is completely messed up along with a host of other things. That is a conversation for another day.
Hope this helps and as always
Happy Printing
wardjr
Printit Industries Model 8.10 fully enclosed CoreXY, Chamber heat
3-SD3's & a Workbench all fully enclosed, RH-Slic3r Win7pro, E3D V6, Volcano & Cyclops Hot End
SSR/500W AC Heated Glass Bed, Linear bearings on SS rods. Direct Drive Y-axis, BulldogXL
Thanks to all for your contributions