3DTouch pin dropping midprint

Prokktor

Hi,
I installed a 3DTouch in my A5 and it mostly works as intended.
But sometimes (seems to be more often in the last days) the pin drops midprint, hits the print, bends and causes a layer shift.
I am basically unable to print anything that takes longer than 60min.

What I found out:
- it only happens when the printer moves fast, never during the first layer or so
- I am using octoprint, so there are 2 5V sources for the 3dtouch, the problem also occurs when printing from usb stick with only the internal 5V source
- sometimes the pin goes up again, but mostly it stays down hitting the print
- I could reset the alert and move the pin up via gcode, but its highly impractical to closely watch the printer for hours^^
- I re magnetized the grub screw and the pin itself with a neodyme magnet, it now holds up even without power  but still drops midprint
- I tried sepparating the servo signal line from the heater cable, no change
- I tried cutting the servo signal line right after starting the print job (there are 8cm left in parralel to the heater cable 1cm appart, cant get rid of that), no change
- I tried hitting and shaking the print head manually while the probe is powered, it does not come down this way

What else can I do?

Regards
Martin

Samuel Pinches

My guess would be electrical noise causing a command to the servo to drop the pin... but you already eliminated that. Can you try to make some gcode commands and see if you can get it to drop? e.g. 

G28
G0 x100 y10 f1200
G0 x100 y290 f5000
G0 x100 y10 f6000
G0 x100 y290 f7000
G0 x100 y10 f8000
G0 x100 y290 f9000
G0 x100 y10 f10000
G0 x100 y290 f11000
G0 x100 y10 f12000
G0 x290 y100 f1200
G0 x10 y100 f5000
G0 x290 y100 f6000
G0 x10 y100 f7000
G0 x290 y100 f8000
G0 x10 y100 f9000
G0 x290 y100 f10000
G0 x10 y100 f11000
G0 x290 y100 f12000

if that code doesnt cause it to drop... then it is unlikely to be a vibration thing....

Prokktor

Oh thanks i will try that.

Yesterday I installed a buck converter and gave the probe its own power source.

Strange behaviour: When i switch power on while the pin is up the probe goes into error mode and does not do anything, if the pin was down while receiving first power it makes a self test and runs normally... probably because the signal line is powered first and  the power rail later.

But it does not matter pin dropped 1.5h after starting the print :-(

Next try:
I wil use a longer grub screw and after homing i will attach a small neodyme magnet to the screw. That should prevent the pin from dropping,  whatever the magnetic coil does.

Der_Muck

The question is why your BLtouch doesnt pull the pin back when it falls down. I thought it should do that normaly.

Prokktor

Its all very strange...

I attached a neodyme magnet after homing so that it is impossible for the pin to fall down.

I then startet the print and after an hour I got multiple y layer shifts.

But the pin just cant be the problem, since it absolutely cant move down, the y motor runs at 935mA which is the maximum possible.

At 900mA the motor would run at about 35C so I do not believe it overheats and the belt is tight enough...

Only other symptom: The nozzle scrapes very audibly on the plastic when the printer is combing, is that normal? (it has alway done that)

Samuel Pinches

Layer shifts could be dodgy bearings?

Does the pin drop doing the test file I provided?

If you repeat the print, do you get the same results?

Prokktor

Layer shifts could be dodgy bearings?
I do not think that is the case but I will open the printer and check them, moving the bed while trying to stopp it with my hands does not reveal any problems.

Does the pin drop doing the test file I provided?
I did not run the test because I already know the problem must be within the 3dtouch device. I cut the signal wire and gave it a sepparate power source, the problem still persisted. The neodyme magnet makes sure the problem cant occur, so I do not really care what the exact problem is since it is not in the printer itself.
I may do that later out of curiosity but at the moment I want my printer to print again^^

If you repeat the print, do you get the same results?
I will try that today, with less mAh for the motors. I increased the mah because of layer shifts before I saw that the dropped pin was stopping the movement and therefore is responsible. 
But so far I have absolutely no clue where these layer shifts come from now...

Samuel Pinches

Ahh, I see. Good luck, hopefully the stepper tune solves it.

Der_Muck

@Prokktor
I only had once layer shifts because I was running the travel speed at 120mm/s. The JG cant do that as it looks.
As I see, you have a 3Dtouch, better go on with the original one. The new smart BLtouch has a plastic pin and I read about the old version that they also had the problem with droping pins. The new BLtouch doesnt have the ALU pins any more.
The current doesnt cause the problem, I ran my JG lot lower.

So how is your speed setting?

Prokktor

I print at 100 travel 80 rest outer walls 40

But sometimes it seems (hard to see) it moves faster than 100, but there are no more speed settings in cura...

Der_Muck

If you are over 100mm/s the hybrit threshold is overdrawn maybe it rattles a bit and cause the 3dtouch to fall down?
I set it to 120 because I had bad rattling cause of it.
I checked my BLtouch and find out that the magnet is so strong that it rather dont fall in any case down. So you maybe have a weak magnet, a to heavy needle or rattles. 
Try to limit the speed to 100 in the firmware.

Prokktor

This is what I get, the first 10 layers are fine then nearly every layer has a shift, x and y are affected.
Reduced travel speed to 90 and tightened the belts and lubed the y assembly nothing seems to be wrong there.

I then repeated the print and got a near perfect replica of what you see above...

Checked the y motor... its cool checked the Stepper Drivers with M122... no overheat

I have absolutely no clue what I can do next...

Samuel Pinches

If you print and then turn down the print to 50% speed manually, is the problem the same?

Prokktor

Ok
I did the following:

Speed 100/80/40  --> 80/60/30 
Combing off
ZHop when retracted 0.5mm

Then it printed fine although very slowly...

So I changed:
Speed 80/60/30 --> 100/80/40 
Acceleration 300 --> 350

And it still printed fine.
Its a little stringi now, but that happens when activating ZHop...

So I am quite sure the nozzle hits the print during travel and that causes the pin to fall down and then both cause layer shifts.

The question is why?

A  40x40mm calibration cube is less then 0.2mm off in all directions so that cant be the problem.

Samuel Pinches

All I can say is that I haven't heard anything but problems from the clone bltouches .

Prokktor

Thats probably true, Clone V6 and Titan are ok except for the heatbreak, BLTouch should be the original.

But in my case the problem is the nozzle hitting the print so hard it causes layer shifts, without the 3dtouch pin I now have small layer shifts instead of big ones, that not really an improvement^^

Samuel Pinches

Edit: Sounds like a TMC problem or possibly a fractured stepper motor wire causing intermittent problems? I'm not so familiar with what to troubleshoot for TMC though.

Prokktor

So I reduced ZHop to 0.3mm and allowed combing through infill, I guess hitting soft borders (1 layer) is not the same as hitting hard (3 layers) borders... (ok ok I watched to many brexit news^^)

Now my print speed is more or less back to normal, I probably can get rid of the magnet too but th bltouch mount is nearly broken away so a have to print a new titan mount before I try that.