Gantry machine
I read many question about gantry machine with two motors.
I have big 5 axis gantry machine in construction. And now gather information of the best way to set it up.
As gantry is driven by 2 motors there can be sync problem.
the gantry kins can do that but I'm afraid that is not good way how it works. Actually I didn't find the best solution.
As stepgen is hardware type I'm very limited with hal options.
In gantry kins I see problem in homing. When one motor hit home switch they instantly reverse direction, while other motor still isn't reached that position. And there is very easy way to jog just one motor (joint) and screw up machine.
Does I miss something?
Please Log in or Create an account to join the conversation.
Does I miss something?
There is an experimental branch of LinuxCNC called joints-axes3. This has a much better internal representation of the differences between actuators and cartesian axes.
This branch also contains a "gentrivkins" kinematics module, which behaves much like the default "trivkins" module but supports multiple motors per axis.
There is some work ongoing to improve homing on gantry machines, though this is mainly aimed at sorting out how index-homing should be done.
Please Log in or Create an account to join the conversation.
...and I'm little scared to use development branch for production machine. I talking in big 5 axis machine 3m x 7m x 1.5m.
As gantry is 3 meters width, and morors are strong, with wrong command I see that ganry wrecked with bad command....
Please Log in or Create an account to join the conversation.
Fell stupid as I can't find any docs of joint-axes3 and gentrivkins.
I don't think that there is any documentation.
The code for gentrivkins is here:
git.linuxcnc.org/gitweb?p=linuxcnc.git;a...aa195be76616642e13b1
It appears to be pretty much identical to gantrykins, except that it declares KINEMATICS_IDENTITY. (which in practice means that there is no "joint mode"
Please Log in or Create an account to join the conversation.
I have no clue how then machine perform HOME operation.
Is it this described somewhere?
Please Log in or Create an account to join the conversation.
There are still individual motor position commands, and the system can home the joints individually (they _have_ to be in the same homing group, clearly).I have no clue how then machine perform HOME operation.
I don't know.Is it this described somewhere?
Please Log in or Create an account to join the conversation.
-Start motors toward home switch with speed 1
-when switch is hit stop the motor
-go in other direction to release home switch with speed 2
This is ok for one motor. With two motors does same thing on each motor. And as one motor will hit switch earlier then that motor will reverse motion before other motor will reach home. Depending of home switch gap (distance betwen hit/unhit) this can make a lot of stress on gantry. It can be solved with speed 2 (to release switch) to be realy small compared to speed 1. So the other motor will hit the switch when other will make just small motion.
I think the dual gantry homming should be like this:
-Start motors toward home switch with speed 1
-when switch is hit stop the motor
-WAIT to both motors hit the home switch
-go in other direction to release home switch with speed 2
So far I didn't find that in WAIT in any release.
It can be done in hal is software stepgen is used, but can't find the way to use with hardware stepgen.
I can't find hal file how I did it but it was done like this.
The motor's are wired with common dir and separate step.
The step's are tied together with hal in normal work. But in homming it was rewired.
If both switches aren't trigered the motor receive step pulse if home switch for that motor isn't triggered.
when both switches are triggered (the AND signal is pass to hal as home switch) the both motors start in other direction.
and similar they move until the corresponding home isn't reached and pass the home signal to hal as OR of both signal.
I had newer problem running this. The emc doesn't really know that there are two motors but works without problem.
So with hardware stepgen I don't have step signal to be rewired in hal. So I must do this in hardware if I can't find other (better) solution.
Please Log in or Create an account to join the conversation.
During my everyday forum reading many times the "gantry homing" was discussed but nerver solved as I understood .
As slavko wrote :
shold be correct from my point of view but usually the final homing position is done with the zero marker of the encoder (for every single axis) .I think the dual gantry homming should be like this:
-Start motors toward home switch with speed 1
-when switch is hit stop the motor
-WAIT to both motors hit the home switch
-go in other direction to release home switch with speed 2
On big machines as eslavko machine is usually are installed glass scale on the traverse axis (as I remenber) due to the precision needed.
Maybe this is not the case.
So for a "perfect" alignment the home cams position and alignment is critical , the hysteresis of the homing limit switches or proximity
is critical and the motor encoder phase is critical and if glass scales are installed the homing marker phase is critical too.
Because a gantry machine can be moved (main traverse axis) by stepper motor (open loop , small machine),
by servo drives encoder positioned (closed loop), by servo drives encoder+glass scale positioned (closed loop) maybe is not easy to have
a unique solution for the homing sequence for all the cases.
I'm no more close to machine tools (mainly industrial automation motion control) but is there anybody that can access to a
gantry machine builder user manual that describe a correct homing procedure for a servo controlled gantry machine ?
bigalex
Please Log in or Create an account to join the conversation.
Maybe this can be an interesting point of view.
bigalex
Please Log in or Create an account to join the conversation.
bigalex
Please Log in or Create an account to join the conversation.