Problem machining with Linuxcnc 2.8.0 and Linux Mint 19.3.

 Computer:
Power Supply 250W
Motherboard PCware Model IPM31
Memory: 4GB (2GB DDR2 800MHZ+2GB DDR2 800MHZ)
HD Sata Toshiba 80GB
Processor: Core2duo E7500 2.93Ghz 
Operating System: Linux Mint 19.3
Linuxcnc version 2.8.0

 Machine:
CNC Control Board: CNC3AX
Stepper motors: 9kg
Ballscrew:16mm 5 ppr
Linear guides 16mm linear with pillow block

 I have tested the mechanics using MDI, sending the machine to a
certain point and then returning to zero for each axis at various
speeds, it works correctly.

 I created a G code for her to machine a 2mm plastic square ('plastic'
photo file), to test the machine and set it up to go removing 0.5mm of
material at each step and in the third step, she lost step machining
in stitch incorrect and maintained this positioning until the end, is
highlighted in red in the photo.

 I generated the code to machine a pcb through Eagle 7.4 and
PCB2Gcode, when I tried to machine the board it also lost positioning,
overlapping the tracks as highlighted in yellow and blue in the photo,
which was where the problem was more visible.

 Some time ago I machined a wooden frame (naval plywood) to put a
speaker, the machine machined without problem and it can be seen in
the "wood" photo. I used Autocad to draw the part and then Artcam to
create the Gcode.
  I also machined some pcbs and had no problem, it's
in the picture "pcb ok", which incidentally, I'm trying to machine
another identical board using the same gcode file and now it's giving
the problem I'm reporting.

 What could be causing this error?

tommylight
  The rails, spindles, motors are well fixed, I can move the frame over the guides using only my fingers without problem.
  I can also move it by turning the spindle using only my fingers.
  I had already checked the current setting, it is correct for the motor I am using, which is 2A.
  Also already checked wire loose or poorly connected, everything is ok.

andypugh
  The coupling is also tight, not slipping.
  From what it seems to me, the error only occurs when a single axis is moving, in the pcb on diagonal tracks, there was no problem.
  The source is adequate yes, I made it following the recommendations of the control board.

  I ran the latency test script and it returned the following:
Core 0:       +31.0°C  (high = +76.0°C, crit = +100.0°C)
Core 1:       +27.0°C  (high = +76.0°C, crit = +100.0°C)
Note: Using POSIX realtime

    8 secs   base min: -11.800 uS max:  13.700 uS sdev:   1.300 uS
    8 secs  servo min: -18.400 uS max:  18.600 uS sdev:   1.900 uS
    9 secs   base min: -12.900 uS max:  13.700 uS sdev:   1.200 uS
    9 secs  servo min: -18.400 uS max:  18.600 uS sdev:   2.000 uS
Core 0:       +32.0°C  (high = +76.0°C, crit = +100.0°C)
Core 1:       +27.0°C  (high = +76.0°C, crit = +100.0°C)
   11 secs   base min: -14.100 uS max:  14.100 uS sdev:   1.200 uS
   11 secs  servo min: -18.400 uS max:  18.600 uS sdev:   2.100 uS
   12 secs   base min: -16.300 uS max:  14.100 uS sdev:   1.300 uS
   12 secs  servo min: -18.400 uS max:  18.600 uS sdev:   2.000 uS
   14 secs   base min: -16.300 uS max:  15.600 uS sdev:   1.300 uS
   14 secs  servo min: -18.400 uS max:  18.600 uS sdev:   2.000 uS
   15 secs   base min: -16.300 uS max:  15.600 uS sdev:   1.300 uS
Core 0:       +32.0°C  (high = +76.0°C, crit = +100.0°C)
Core 1:       +27.0°C  (high = +76.0°C, crit = +100.0°C)
   15 secs  servo min: -18.400 uS max:  18.600 uS sdev:   2.000 uS
Unexpected realtime delay on task 0 with period 25000
This Message will only display once per session.
Run the Latency Test and resolve before continuing.
   16 secs   base min: -24.800 uS max:  41.100 uS sdev:   1.300 uS
   16 secs  servo min: -19.200 uS max:  19.800 uS sdev:   1.900 uS
   18 secs   base min: -24.800 uS max:  41.100 uS sdev:   1.300 uS
   18 secs  servo min: -19.200 uS max:  19.800 uS sdev:   1.900 uS
   19 secs   base min: -24.800 uS max:  41.100 uS sdev:   1.300 uS
   19 secs  servo min: -20.300 uS max:  19.800 uS sdev:   1.900 uS
   21 secs   base min: -24.800 uS max:  41.100 uS sdev:   1.200 uS
Core 0:       +35.0°C  (high = +76.0°C, crit = +100.0°C)
Core 1:       +28.0°C  (high = +76.0°C, crit = +100.0°C)
   21 secs  servo min: -20.300 uS max:  19.800 uS sdev:   1.900 uS
   22 secs   base min: -24.800 uS max:  41.100 uS sdev:   1.200 uS
   22 secs  servo min: -20.300 uS max:  19.800 uS sdev:   1.900 uS
   24 secs   base min: -24.800 uS max:  41.100 uS sdev:   1.200 uS
   24 secs  servo min: -20.300 uS max:  19.800 uS sdev:   1.900 uS
   25 secs   base min: -24.800 uS max:  41.100 uS sdev:   1.200 uS
   25 secs  servo min: -20.300 uS max:  19.800 uS sdev:   1.900 uS

I didn't quite understand the reason for the red message. I ran the manual test, leaving several programs open with about 7 glxgears windows and latency did not exceed 33us.
  This control board I use is a chopper with L297 and 74LS09 connected to the PC's parallel port.

News:
I made some changes, changed the L297 default values in the time setting:
Step time L297 500ns > Step time 2000ns
Step space L297 4000ns > Step time 4500ns
Direction hold L297 4000ns > Direction hold 4500ns
Direction setup L297 1000ns > Direction setup 2000ns

  And I ran this code 3 times in a row, making the machine go to zero point just before starting to run the first time, at the end of the third time it ran, it returned to zero point correctly without losing position at any time.
 As it is now late here to do more tests, tomorrow I will try to machine another pcb and see what happens.

 

Looks like it got better, I made the board just now, the tracks were ok, I just think I set the forward speed too high and ended up breaking a little piece of the tip of the V bit.

Andypugh thanks for answering, can you explain to me in more detail, i can't quite understand your answer.

In the case of this L297 default step time setting, you mean that this value is very close to the component's limit and with that there may be failures, and that increasing this value will not interfere so much with the performance of the pc and the control board, right?
  About the latency time, I ran the latency test through the script here on the forum and how I got the impression that it ended quickly and also returned this message:
  Unexpected realtime delay on task 0 with period 25000
This Message will only display once per session.
Run the Latency Test and resolve before continuing.

  I ran the next Linuxcnc latency test and it returned a value of 33uS, so I put in 35uS.