Unexpected XYZ Movements Destroyed Prints

I would really appreciate your expert opinion on the below issue.

What happened on those two pictures is that the Z axis movement started drilling into the model in the middle of the print. I stopped it only by shutting down the printer. In the case on the second picture, it reached the bed and carried on pushing so hard that the Y belt fall off.

I noticed the similar behaviour on X and Y axis, but in those cases the consequences were not so serious. For example, while printing "Robot", which is the model with very small footprint, the print head moved 4-5cm to the left, leaving the active printing area, then came back to the same point it departed from and carried on printing as nothing happened. The same happened on Y axis. And it does not happen that often, but often enough to be classified as the showstopper.

I managed to print the same Robot.gcode model more than once without the problem, so it is not the problem with the gcode file itself.

In addition to that I noticed:
- occasional pauses in printing (I read it could be due to the overheating...)
- occasional pauses in printing with the message on LCD display saying "Waiting User Response", where I have not got anywhere close to the printer prior to that.
- occasional stops in printing without carrying on at all.

This is the printer in question: [www.alibaba.com]

It has MKS Base V1.1, with MARLIN 1.0.0 on it.

I assembled it 4 weeks ago and before it started behaving like this it gave me around dozen very nice prints including 2 quite large and complex objects.

From my experience those unwanted moves are caused by a bad USB connection, often together with gorund problems on the controller or bad filters in the power supply.
Could not see anything about SD support so if USB is your only option:
Make sure the firmware, print software and computer use the same com port settings!
Most setups won't allow for flow control, so if that is turned on in your Windows settings for the com port try without.
Check the USB cable and if in doubt try a shorter one with ferrite cores to eliminate RF interference.
Do a simply check during a print with the help of someone:
One person watches the printer while the other switches lights and and other cumsumers like fridge, mixer or aircon on and off.
If the print fails during the switching you have interference.
Proper grounding of printer, controller, power supply and computer should help, otherwise try one of these fancy protection boards with filters and plug everything in there.

Another thing that happens is buffer problems due to wrong com port settings in Windows.
Usually this happens mostly at very slow speeds and with prints that have very long lines in the print - either long straights or big circles.
The software sends out the next move to the buffer but the current move takes longer than expected.
If this delay is not transfered and handled correctly the printer assumes the position was lost at the end of the move.
So depending on the direction the head will now move all the way to a soft- or hardware limit before returning to the print.
Often the actual print move planned is lost and it looks like the print is missing a line.
On most Windows systems it should enough to uninstall the com port through the system manager while the printer is attached.
Plugging the machine out and back in again will get com port back with the default Windows settings.

Thanks for the feedback. That must be it as I was rearranging the cables day before this started happening.

I forgot to mention that I am not using USB connection to the PC. Instead this machine has Card Reader as part of the LCD board. There are 2 40-50cm long flat cables going from the LCD board down to the motherboard. One of those two cables transfers GCODE instructions from the Card to the Board. I packed those cables up just next to the bunch of step motor cables, so it is possible that there were some false readings due to the interference.

I also contacted the supplier and got the similar answer. They offered me the solution which I am providing to this post. Basically, they suggest to shield the flat cables with separate wire and to ground it to the power supply casing.

I will try this over the weekend and let the community know if this solved the problem.

Shielded cables are always a good idea, but also often over rated.
Take older game PC's and you often see serial cables from the motherboard to the connector that are already longer - and then comes the cable on the device attached to it.
If you recently wiggled on cables there is a good chance you just lost ground somewhere.
The card reader requires 3.3V from an onboard regulator.
If the main controller is not supplying it properly the 5V come from the USB connection.
So check if display and card reader work properly without a connection to the computer.

In some cases a problem on the power connection itself can be the source of the trouble.
The motors require quite a bit of current to work and if the power connection is not 100% the controller can get a "hick up".
By the way: wrapping some wire around the flat cable and grounding this connection won't help you
Shielded cables that do as promised in this configuration are quite expensive as they are rarely used.
If you have a box with old wall warts there might be chance you find one with a ferrite core in the cable that is made of two halfs.
They are quite good for the job and you can push the flat cable into a round shape to make it fit.
But I doubt the cables are your problem, if you have too much interference than it usually comes from the power supply - checking it with an oscilloscope might be a good idea.