synchronisation problem when using M3 command

Hello,

I am an newbee. I have some synchronisation problems;
setup is a dual core dell PC with parallelport. The mill has a frequency convertor with 0-10V input. All axismotor are steppers and they are controlled by puls direction.

The parallel port has an pwm output for the spindle speed ( there is an external pwm to analogue convertor)

But when I use the M3 S1000 command in the G code it seems that there is an synchronisation problem. Without the M3 command or (M3 S1) the x and Y axis are correct. M3 command has influence on the synchronisation between X and Y. I have measured the latency and filled it in the configuration but the problem still exist.

Does anyone have some tips ? That is the only problem that is holding me back. (I have solved all the others EMC problems on the cable to the spindle etc...)

thanks for any advice.

Hi

You will have to attach your .hal and .ini files, what you describe sounds very strange, but we have nothing to go on.

regards

these are the hal and ini files, Thanks for taking a look !

There is nothing obvious in the files, except that your scale figures are very low even for a metric machine.

It is usual for Z to be higher than X and Y, because of the weight of the spindle head, but is X gearing really half that of Y (or X using half the micro-stepping of Y) ?
X is moving 2.5mm per stepper rotation, that is a lot, something like 0.5mm - 0.25mm (scale of 400 - 800) would be more usual.

You will need to explain exactly what you mean by synchronisation and how this manifests itself.

regards

EDIT:

Not connected, but you also do not have any home switches specified and have both negative and positive limits set to the same input
According to your ini file soft limits, 'home' is somewhere in the middle of your axes travel

I would suspect this is wrong and the current setup probably gives you neither repeatable homing nor crash protection.
It is certainly capable of improvement.

Actually this sounds a bit like an electrical problem,
either interference (EMI) from the spindle or something wrong
in the breakout.

To check if it is interference, a trial run with the spindle drive
powered off is probably worth a try.

@PCW EMI is not the problem. When the frequency convertor setpoint is generated by a potentiometer and the M3 command is left out there is no problem. Yes we have had some emi problems but with 360° bounding shielding that is already solved. With synchronisation problem i mean the following: We are drawing a box with two diagonals. Without the M3 command diagonals are in the corners. With M3 the diagonals do not end in the corners but outside the box.

Indeed there are three possibilities now either Linuxcnc or the parallel port or the breakout (chinese brandless model).

@ArcEye : indead there is al lot of gearing and microsteps.

@all of you : is there a maximum PWM frequency it is now 500Hz.

and thanks for the quick response, always nice to have some feedback !

A potentiometer is quite a bit different EMI wise because in that case the spindle is not connected to the PC

I would still try with the spindle drive disabled

I have added some pictures to show the synchronisation problem. One with a low M3 spindle speed (nice one) and one with a high M3 speed (the one with the faulty diagonals). frequency convertor is not powered. In one of our test we have selected a free output pin for the pwm (one how is independent of the breakout board)
Second test was with al lower pwm frequency, no improvement.

I am out of inspiration. Thanks for any advice

i have had some problems uploading the attachements
here is the dropbox link for the files
www.dropbox.com/sh/r8zz78j5fhqrwxp/AABrw...iFowdV_E_LMnNAa?dl=0