[grbl Issue#1611] Modify grbl for CNC model

#1 – whitetd 于 2020-03-02

create a post process for what ever your using to create your gcode to do
it, rather then in grbl

<https://www.avast.com/sig-email?utmmedium=email&utmsource=link&utmcampaign=sig-email&utmcontent=webmail&utm_term=icon>
Virus-free.
http://www.avast.com
<https://www.avast.com/sig-email?utmmedium=email&utmsource=link&utmcampaign=sig-email&utmcontent=webmail&utm_term=link>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

On Mon, Mar 2, 2020 at 7:44 AM SanZamoyski wrote:

> Hi and hello,
>
> I know it is NOT an issue, but I need your help as I’m trying to resolve
> this for two days…
> My goal is to change spindle behavior to:
>
> – when it is on – PWM 50%,
> – when off – LOW
> – when on AND Z-Work Position < 0, PWM 100%. >
> I was trying alot of variables and in many places of the code, but with no
> luck. I’ve ended with report.c file on line ~522 (under float wco[]) – it
> is place where WPos/MPos are calculated, before it is printed.
>
> I have this (dirty) code here:
>
> if(currentposition[ZAXIS] – gcstate.coordsystem[ZAXIS]+gcstate.coordoffset[ZAXIS] <= 0){ > pinMode(13, 1);
> digitalWrite(13, 1);
> }else{
> pinMode(13, 1);
> digitalWrite(13, 0);
> }
>
> I’ve tried
>
> sysposition[ZAXIS] – (gcstate.coordsystem[ZAXIS] + gcstate.coordoffset[ZAXIS] + gcstate.toollength_offset )< 0 > gcstate.position[ZAXIS] – gcstate.coordoffset[Z_AXIS]
> gcstate.coordsystem[idx]+gcstate.coordoffset[idx]
>
> and everything what came to my mind analysing the code, bot none worked.
> The effect is that LED (connected to 13’th pin) is on when Machine
> Position is under zero, not Working Position.
>
> Please, please, please help me. I just need to know is there any variable
> I can use here, and if not – how to calculate it.
>
> I’m quite sure that sysposition is Machine Position, so [ZAXIS] should
> give me what UGS shows under Machine Position. WCO is difference between
> MPos and WPos, so adding/substracting should give me right value. But it is
> not somehow…
>
> Best regards…
>
> —
> You are receiving this because you are subscribed to this thread.
> Reply to this email directly, view it on GitHub
> <https://github.com/grbl/grbl/issues/1611?emailsource=notifications&emailtoken=AATT5E33XQNTMMA2R5E63ZLRFPH4BA5CNFSM4K7W5LB2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4IRXP2ZA>,
> or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/AATT5EYQHW344HFPJWRCTTTRFPH4BANCNFSM4K7W5LBQ>
> .
>

#2 – SanZamoyski 于 2020-03-02

Unfortunately it is not an option. I forget to mention that I’m trying to modify grbl for the CNC model for students. X and Y axes are made with regular steppers. Z and spindle are replaced with laser – that’s why I need to use different PWM for over-zero and under-zero. Students will have to program moves by hand (lessons are focused on G-Code, not machine/program). That’s why postprocessing option is not an option.

I’ve already modified GRBL to use h-bridge with microstepping for tho steppers (the third stepper is possible sacrificing some pins but without microstepping) so I would never suspect that this goal was that hard for me to achieve but it is…

I know that report.c is not the right place to add this functionality. At the beginning I was trying to add this at the end of the stepper movement function but there was already mentioned problem so I moved it here to be sure that this is related to wrongly chosen variables. And it is.

#3 – langwadt 于 2020-03-02

could you just add a few lines to the top of the stepper interrupt that sets the pwm to 50/100% depending on whether the last direction was up or down?

#4 – SanZamoyski 于 2020-03-02

It would be almost a good solution, but unfortunately, it would set falsely PWM at 100% if “spindle” is over the material. Please keep in mind that it will be used by people who do not know almost anything about G-code, so this kind of scenario is quite possible.
Is this really very complex problem and we need to look for a workaround?

Also, there is WPos output from grbl (setting verbose in UGS and changing some settings). This would be a workaround for me. But it shouldn’t be that hard to achieve „normally”…

#5 – langwadt 于 2020-03-02

I guess something along these lines,

float zpos;
zpos = sys.position[ZAXIS]/settings.stepspermm[ZAXIS];
zpos -= gcstate.coordsystem[ZAXIS]+gcstate.coordoffset[ZAXIS];
zpos -= gcstate.toollength_offset;

#6 – SanZamoyski 于 2020-03-02

I believe I tried something like this by:
if(sysposition[ZAXIS] - (gcstate.coordsystem[ZAXIS] + gcstate.coordoffset[ZAXIS] + gcstate.toollength_offset ) < 0){ ... }
based on:
``
float wco[N_AXIS];
if (bitisfalse(settings.statusreportmask,BITFLAGRTSTATUSPOSITION_TYPE) ||
(sys.reportwcocounter == 0) ) {
for (idx=0; idx< N_AXIS; idx++) { // Apply work coordinate offsets and tool length offset to current position. wco[idx] = gcstate.coordsystem[idx]+gcstate.coordoffset[idx];
if (idx == TOOLLENGTHOFFSETAXIS) { wco[idx] += gcstate.toollengthoffset; }
if (bitisfalse(settings.statusreportmask,BITFLAGRTSTATUSPOSITION_TYPE)) {
print_position[idx] -= wco[idx];
}
}
}
`
Where printposition is memcopied and coverted by systemconvertarraystepstompos() sysposition. Which is, diggin deeper, steps[idx]/settings.stepsper_mm[idx].
print_position is then used here:
`
static void reportutilaxisvalues(float *axisvalue) {
uint8_t idx;
for (idx=0; idxCoordValue(axisvalue[idx]);
if (idx < (NAXIS-1)) { serialwrite(','); }
}
}
``
But I missed calculating to mm. And You are correct. Thank You for pointing that out. Also I thought that I don't need recalculating to mm/inch or anything. I'll check Your solution tommorow. Thank You.

#7 - SanZamoyski 于 2020-03-03

Thanks to @langwadt I found the solution to my problem. In the file report.c I put before //Report machine position this code:

``
float zpos;
zpos = sysposition[ZAXIS]/settings.stepspermm[ZAXIS] - wco[ZAXIS];
//zpos -= gcstate.coordsystem[ZAXIS]+gcstate.coordoffset[ZAXIS];
//zpos -= gcstate.toollength_offset;

`
Also commented //SPINDLEOCRREGISTER = pwmvalue; // Set PWM output level`. in spindlecontrol.c.

It seems to work very well now without any workaround
Probably there are a better places to put this code but I am not sure if it is needed.