Z Nut Eject Solved

Problem was that the Z nuts would randomly eject on a Z home at the beginning of the print because the Z micro switch failed to trip. Failure was being caused by the Z/X assembly binding on the Z rails as the extruder approached the bed. Root cause was the spacing between the Z rails was 3mm smaller near the bed, than the spacing between the linear bearings in the Z/X assembly. As the Z/X carriage was lowered, the Z rails were being force apart, sometimes forcing the linear bearings to bind and momentarily jam.

In my case the fix was to cut the X rails to the correct size. I used a lathe, but a dremel would work. I grabbed a 12" digital caper from Harbor Freight, which can open enough to measure the inside distance between the rails. Move the Z/X assembly to the very top of the Z rails so just some rail is exposed out the top bearing. Measure the Z rail spacing just above the bed and measure the Z rail spacing just above the linear bearings with the carriage in this position. They should be the same. Any difference will start to cause binding.

If you find you need more space between the bearings, you can add some shim filler material inside the X axis rails before they insert into the Z/X housings. Interesting design. The tension on the X belt is the only thing holding the X carriage together.

In my case, the X rail diameters were 0.09mm undersize from the nominal 8mm and I opted to replace and cut them to the correct size with new linear motion rails from VxB. As a side benefit, the X carriage slop went from 0.38mm to nil. While I was at it, also replaced the Z rails themselves because the too were undersize and sloppy. Running many weeks now without a Z nut eject. Print quality is much improved.

This test procedure should be added to the standard Robo 3D quality check before machines leave the factory.

 

the .09mm undersize is because they are not using metric 8mm but imperial 5/16'' (with some variance in size, and caliper accuracy)

 

like imperial drill rod? linear motion rails usually spec'ed +0.00/-0.xx some of the Robo rods and under and over 0.3125 But could be as you say...

 

Yeah it's not proper linear motion rails, it's just cheap drill rod, I wanted to find 8mm rod locally for the y axis mod, but could only get 5/16'' drill rod and it's same diameter as the robo rods.

 

For Y Axis -8mm x 406 Linear Motion http://www.vxb.com/page/bearings/PROD/kit11867

Replacement "8mm" rods for X and Z from VxB were above 5/16 but below 8mm diameter. Worked great in existing Robo bearings on Z. Had to press out the X carriage and use an adjustable reamer to shave a bit off the plastic to get a happy fit when compressed.

Almost like you have to order extra and mix and match.

 

Nickster, this binding, was it causing your printer to make almost a grinding/stuttering type noise and drop the hot end too far after a lift and move? My hot end has piled right into the bed on my pyramid test prints about 20-25 layers in each time. The nuts haven't ejected, but the hot end ends up tearing through my blue tape and damaging the bed surface. Been a long night, will have to check this in the AM.

I did have to flip the z-axis switch around as the screw was touching directly over the micro switch rather than out on the lever a bit further for better leverage. Other than that all my settings are as per the starter video for PLA. Guess the calliper s will tell the tale in the morning!

 

Don't necessarily want the screw at the end of the microswitch lever arm. Farther out you move the contact point, the more leverage and over travel you have but you are trading that against loss in sensitivity. On my setup, the screw is about 13.5mm out from the Z rail and the microswitch trip pin is 18.5mm. So the screw is about 5mm out from the trigger pin. From the pivot to the trigger pin is about 4mm. My screw has to displace approx 9/4 *0.1 = 0.225mm to get a 0.1mm displacement of the pin. If the screw were contacting the end of the lever arm, it could be out around 14mm from the pivot so you would need 14/4 *0.1 = 0.35mm screw displacement to get a 0.1mm displacement in the trigger pin. Guessing weight of the carriage is more than enough to trip the pin directly without a lever arm.

Grinding/stuttering noise sounds like stalled steppers. Does it make the noise going up or down? Going up could mean too much load, bumps on the rails or lead screw, stepper current is too low (pot adjustment), or you are trying to slew too fast. Try duplicating the problem by running G1 gcode with different speed Z moves from the command box in Repetier. Since it is not ejecting the nuts, after this happens try using the manual Z motion buttons and see if Z=0 is where you expect it to be and hasn't changed. Also does the leveling of the X carriage change indicating one Z stepper stalled but not the other.

My problem was very subtle and difficult to diagnose because the carriage would still slide all the way down on the Z rails, bending them outwards in the process. Sometimes they would momentarily bind until something got jiggled, but this was not obvious under normal operation. Crappy print quality when I slic3d with Z retract on filament retract, because sometimes it would not return to the same Z for a bit.

 

So I have tried again this morning after checking things out again. It seems that my Z-axis switch bracket is broken from over tightening, however the unit zeros out and goes home accurately dozens of times manually from all heights. The screw is contacting the lever about where you mention, I never did say I paced at the end of the lever. Upon watching it closely what seems to be happening is the Z-axis steppers lift the extruder for a move, then when it goes to reverse direction to lower the extruder the right side motor jams, doesn't turn like the nut is seized, tries a few times rapidly then moves on. It then loses all track of where the extruder head is located on the Z-axis and tries to continue unsuccessfully. The left stepper lowers as normal. Perhaps there is a burr or rough spot on that particular section of the screw rod? I am checking that next.

I am new to 3D printing but not new to diagnosis and troubleshooting. Could you explain briefly what you mean by G1gcode? Is that sort of a dry run without filament being extruded? Thanks so much for you time.

 

Go to the Manual tab of Repetier. Upper right, just left of the send button. Assuming Z is homed, you can type in Gcode move commands. G1 Z80 F500 would to a speed controlled moved to 80mm at a speed of 500mm/m. G1 Z50 would retain the last speed that was set and move to a height of 50mm. If you enable the Commands radio button on the status bar, you will see your command you entered going out to the printer. Max speed is being limited apparently by marlin to 300mm/m or so.

 

Well, have had several successful calibration prints and a couple other items that printed very well. The main problem seemed to be the threaded rods, with the top off both were VERY bent and wobbling. Thus dragging on the top hole in the brackets. I straightened them out and while I had them out lubricated. Them with a light silicone spray and dried them off. Also ran wire brush on the Dremel over the inside of the nuts. All seems to be well and working as well as it can. The bed itself has a noticeable dip in the Center. I have to pick up a dial indicator and see how much the difference is from edge to edge. As long as I calibrate for the center my first layer is sticking ok. If I calibrate to the outer edges the middle print area will not stick as it isn't getting that slight squash as it extrudes.

Really enjoying the learning process, thanks for the tips and advice.