Z axis with a mind of it's own

Hi

open OBControl
connect to the controller
click the 'grbl settings' tab
set '$1 Step idle delay, millisecond' to 255
click the save button at top left

 

If the controller setting is not 255, which GUI you use will never make any difference since it is inside the controller.

If the problem is in the Gcode, we can help with that if you post a sample file and tell use how you created it.

 

Dial back your Z max Rate and Z Acceleration to half of current then retest - too high values causes stalls which looks like a drop.

If that does not help, and everything is mechanically perfect, replace the controller with a proper controller like the BlackBox from openbuildspartstore.com (see docs.openbuilds.com/blackbox for docs)

 

That is because the setting is saved inside the controller (inside the Arduino), not in the GUI software, so each GUI just reads it from the controller...

I need the entire, original, gcode file, and a screenshot of the drawing in Aspire so I know what the gcode is supposed to create.

This 'extra 5mm drop', did it happen in one move, or spread over all of the cut?
if in one move, when? start or end of a cut? which cut?
once it is cutting too deep, does it pull out of the cut correctly or is the retract now also too low by 5mm?

I agree with Peter, change the Z rates and test, seams like Z is getting stuck somewhere and the current (high) speeds will make that worse. If reducing the Z max rate and acceleration provides a cure, then you need to check Z carefully for sticky spots, and also check that the driver is running cool enough and is operating correctly at all times.

You said you had 'some cheap driver', and replaced it with a MakerShield - is that this thing? If it is then I strongly suggest that the problem is the Z driver overheating and skipping steps when it shuts down to control the heat. The 8825 chip is simply not enough chip for the Nema23 size stepper motors, it is well suited to the Nema17's on 3D printers which move light loads, but routers are much heavier and need about twice the current from the drivers. I do wish that sellers would include a simple line like "this is for Nema17 motors and is not suitable for driving your CNC router" when they advertise this type of product.

 

Thankyou

set this
$11 Junction deviation, millimeters 0.010mm
to 0.02mm, which is the default, changing these settings ($11 and $12) from the the defaults has no usefulness.

These other values tell us some things.....

Code:

$100 X-axis steps per millimeter 40.00steps/mm
$101 Y-axis steps per millimeter 40.00steps/mm
$102 Z-axis steps per millimeter 40.00steps/mm

These indicate you have not calibrated, for distances and speeds to be accurate you have to calibrate first.
(round numbers never happen in reality (-: )

Code:

$110 X-axis maximum rate, mm/min 2000.00mm/min
$111 Y-axis maximum rate, mm/min 2000.00mm/min
$112 Z-axis maximum rate, mm/min 500.00mm/min

probably ok, but after calibrating you can probably tune these higher, esp after upgrading stepper drivers to ones that can actually supply the required current of the motors (assuming Nema23s, you havn't answered that question, among others).

Is this 500mm/min for Z the original value or after halving it for testing as we suggested earlier?

Code:

$120 X-axis acceleration, mm/sec2 100.00mm/sec²
$121 Y-axis acceleration, mm/sec2 1000.00mm/sec²
$122 Z-axis acceleration, mm/sec2 50.00mm/sec²

3 widely different values will confuse the internal math a bit, or at least make it harder to calculate proper step rates.
X and Y should be the same unless the drive mechanisms are wildly different, like belt on X and screw on Y.

Z is of course under scrutiny and conservative values are what we want until we identify the issue.
Z acceleration will normally be lower than X and Y because it is lifting weight. combine that with dirt under a wheel or in the leadnut and we have a recipe for skipping steps on a lift movement, especially with your current limited drivers.
This will appear like a 'drop' but is in fact a 'failure to lift'.