I print slow.  I print almost exclusively at 35mm/sec.  I can go 50% higher for short periods of time.

Flexies at 15mm/sec.

I got a E3D bowden setup now, should be even more lightweight than your E3D Titan direct drive! Would be interesting to know your current print / travel speed!?

When mine went bad, it wouldn't zero properly but didn't ever affect the print.  In fact, while printing, you can trigger that sensor all you want... it doesnt do anything from what I can tell.

My random shifts diassapeared when I switched to the Titan/E3D extruder.  With that I dropped 50-100g in weight if not more.

I agree with you that weight/speed/momentum are the likely culprits.

This works fine for smaller prints but when printing larger objects random X-shifts are back again!!!

To be 100% sure that this is not any trouble with the end-stop sensor I've connected a cable dirctly at the main board where the sensor is plugged in:

I run this cable into the inside so I could bypass the sensor detection by a switch. The switch will short the opto-output to GND - digital low here means light is passing to the gate!

I also had to remove all G28 G-code commands from my custom code at the end - but leave one G28 at the start so the printer can calibrate correct. After the first layer I switched on the end-sensor bypass so there is definatly no end-sensor command entering the main board.

To my surprise it is still x-shifting!!!! So if it is not the end sensor and i woud exclude the software than the only possible reason is the mechanics/stepper motor. My assumption is that the stepper motor heat up and can not follow the speed and skipping some steps ... to prove this I've run the same print but with 45mm/sec for max. print speed/travel speed without problems! I think 45mm/sec is the maximum that this machine can handle without the risk of a shift.

It was good to try the switch to be 100% confident that it's not the end-stop. I will remove the switch at the next removal of the main board then!

The point is that the recommend speed settings that can be found in some forums are way too high! The natural behavior of a stepper motor that it has an upper speed limit where the torque is decreasing massivly and finally stopping the motor to increase in speed. If you run a stepper motor above this point it will simply skip a few steps as it can not follow the speed. This skipping of steps leads to an error in position between the machine and the software! This can happen to all axis and is a different failure than the one with endstop sensor (BTW - thanks for refering to my post)!

When I had this trouble I did some reverse engineering on the stock firmware/software to see what XYZ is choosing for the speeds. I assume that they know their machine well and did intensive testing for ideal speed settings.

The speed settings from the XYZ stock firmware/software with quality setting "good" are:

    Perimeters:     15 mm/sec
    Infill:               30 mm/sec
    Support:          45 mm/sec
    Travel:            45 mm/sec
    Retract:           30 mm/sec

I know that these settings seam to be quite lame compared to other machines - but that's what the stock firmware/software is using. You may can go 10% above but if your speed settings are 50% above these settings the trouble is for sure! If you want a faster print I would suggest to do a hotend mod http://www.soliforum.com/topic/14386/mo … avinci-10/ and change to a bigger nozzle diameter - I would not recommended increasing the stepper motor speed with this machine!

No doubt I will have to reroute my wiring at some point

@Telemachus: I think that your trouble has other reasons - and maybe you try to print from SD-card to see if the server will cause this! The Y-axis sensor is fixed and therefore there is no mechanical stress at the contacts - the chance for a loose contact is very minimal. But nevertheless: I would highly recommend to do the X-axis fix like I did as this problem with the x-axis sensor will occure sooner or later and it's always a pain when ruin a longer print!

1) I removed the JST-plug on the sensor and directly soldered the cables to the sensor PCB. Then I secured the wires with a drip of hot-glue before mounting it - this is now a rock-solid connection!

2) I run the cables for the sensor and the stepper-motor in two separate cable tubes and secured them with cable ties very flat so they will not touch the side wall.

3) I re-route the wire harness - to me it makes no sense routing the cables up and then down again. I run them down and secured them at the bottom so they can only move as they are supposed.

4) I applied a ferrite bead on the X-axis sensor cables close to the main board. Due to the new routing I get some extra length and can have two turns in the ferrite bead. (no picture).

I've just started a 42 hours print - if this will pass I think I found it Keeping fingers crossed!

if you have a break in the wire harness and may only show up when printing in specific spots.

If your stepper motor gets to ho.t

If the stepper driver gets to hot.

so it may take a bit of investigation to track down what is the cause

There are many threads that discuss this issue, but here is a link to one of the more helpful threads:
http://www.soliforum.com/topic/10905/ma … -printing/

I have a Davinci 1.0A with Repetier 0.92! On longer prints I have random x-axis shifts, the print is shifted about 2mm to the right!

I run the travel moves with 45mm/sec that is the speed of the stock firmware that I found when reverse engineering the code - so this should not be the problem.

After some search on the web I found that this could be caused by an end stop giving a wrong signal. I looked at the X-axis endstop and found that the cables run in parallel with the stepper motor cables - I think this is not good as there could be some EMI introduced on the signal lines.

I will do a mod and run the cables separete from the motor cables plus will twist them! I also thought about putting a ferrite bread on the end before the main board would also give a benefit in signal quality!

A software fix would be to disable the endstop (M121) after the initial homing and enable before the final homing (M120) - but I don't know if this is supported by the firmware - I would need to check when my current print is finished.

NOW my question: Has someone of you already tried some of this stuff and could give some advice?

Take care!