Getting soft limits to work without homing

I use abs encoders so no homing will be done. I would still like to be able to use soft limits. I found this from command.c :

void refresh_jog_limits(emcmot_joint_t *joint)
{
double range;

if (GET_JOINT_HOMED_FLAG(joint)) {
/* if homed, set jog limits using soft limits */
joint->max_jog_limit = joint->max_pos_limit;
joint->min_jog_limit = joint->min_pos_limit;
} else {
/* not homed, set limits based on current position */
range = joint->max_pos_limit - joint->min_pos_limit;
joint->max_jog_limit = joint->pos_fb + range;
joint->min_jog_limit = joint->pos_fb - range;
}
}

If I change this to:

void refresh_jog_limits(emcmot_joint_t *joint)
{
/* Set jog limits using soft limits */
joint->max_jog_limit = joint->max_pos_limit;
joint->min_jog_limit = joint->min_pos_limit;
}

Would that "do the trick"?

Tero

terkaa wrote: Would that "do the trick"?

It might, until the next LinuxCNC update...

Ok, I would like to try. What is make line only for compiling and installing parts related to command.c ?

Tero

I don't know, I have never tried compiling anything bigger than a HAL module or smaller than the entire source tree.

Ok, I did compile the whole thing but trying to run sudo make install tells me "configured with run-in-place" configure without run-in-place. How do I disable run in place?

Tero

It's probably best to try it out with run-in-place first.

(Open a terminal, then
I still think you should run master. As far as I recall the absolute homing feature was added for you.

Hi,

I did try it with RIP and it does work. Originally my idea was to tweak source so that when NO_FORCE_HOMING=1 then soft limits would be tied to pos-fb and if NO_FORCE_HOMING = 0 the offset+pos-fb. Yes I understand your point of view and I will try 2.8 after I receive another similar PC. But you also have to see my point of view: After I get everything installed and working this machine will be ran 3 shifts per day in a machining shop. It will be ran by our CNC-machinists.(not by me) + Given the fact that I made the decision to install LinuxCNC instead of Fanuc. So I am not very eager to run untested software in this environment where machine users are in risk to get injured if something goes wrong(Also it will be on my responsibility). + If users have to deal with bugs they will request me to replace control with Fanuc. Also it is very nice to hear that this feature was added for our application but I did not request that. I requested SSI+sserial firmware for 5i25 and I am happy to have that. But yes I will try out 2.8 and report bugs if I see em. If it does seem to work we may even stick with it. But it is good to have a backup plan, in this case 2.7.6 with minor source mod.

Tero

Sorry, I didn't mean to sound like I was trying you guilt-trip you into trying 2.8.
And it wasn't you anyway, it was this chap.

forum.linuxcnc.org/forum/38-general-linu...imits-for-axes#74605

Oh I see I have been thinking about building a mini mill (X=250mm Y= 150mm Z=100mm) This would run LinuxCNC with Mesa cards and Fagor SSI absolute encoders. It would be nice to have in my garage, and would act as a very good test platform.


Tero

Something like this ?