On Fri, 10 Apr 2020, Leonardo Marsaglia wrote:
Date: Fri, 10 Apr 2020 22:08:05 -0300
From: Leonardo Marsaglia <ldmarsag...@gmail.com>
Reply-To: "Enhanced Machine Controller (EMC)"
<emc-users@lists.sourceforge.net>
To: "Enhanced Machine Controller (EMC)" <emc-users@lists.sourceforge.net>
Subject: Re: [Emc-users] Home use index problem (In step + dir mode)
Sorry for not catch it earlier Peter. I thought no matter the feedback,
when index was detected it will behave as a second home switch.
I'm thinking now about making a custom component to use as limit switch and
drive it with the actual limit switch for home search and with the index
pulse triggering for the final detection (latch velocity) of the home
position. This way, I can home with the accuracy of the index pulse, but
with the normal behaviour of only having the limit swtich.
My idea is to use positive search vel and latch vel so when I hit home the
following should happen:
LinuxCNC will see that the switch is pushed, it will back off a little
until the switch is released. After that it will return to move towards the
switch in latch velocity, that's when I intend to bypass the switch and
instead use the index enable edge as the switch signal for the final home
position. I suppose I can use the joint.#.homing-state to know when to
bypass the switch easily.
It seems very doable. I'll always have the limit switches in case of a
problem.
Yes I think this is possible (basically using the index hardware as a latch to
catch the quick index pulse) but note that the actual home position will vary
with the velocity when the index is detected (and with a servo thread period
uncertainty)
El vie., 10 abr. 2020 a las 21:16, Peter C. Wallace (<p...@mesanet.com>)
escribi??:
On Fri, 10 Apr 2020, Leonardo Marsaglia wrote:
> Date: Fri, 10 Apr 2020 20:32:23 -0300
> From: Leonardo Marsaglia <ldmarsag...@gmail.com>
> Reply-To: "Enhanced Machine Controller (EMC)"
> <emc-users@lists.sourceforge.net>
> To: "Enhanced Machine Controller (EMC)" <emc-users@lists.sourceforge.net
>
> Subject: Re: [Emc-users] Home use index problem (In step + dir mode)
>
> I'm playing with the homing at slow speeds so I can hit the home switch
>always half a turn away from the index pulse more or less.
>
>One thing is happening is the following:
>
>If I let the mark (keyway in this case) of the motor at 3 o'clock and then
>I turn off LinuxCNC, whenever I home the axis again (with the switch and
>index and home position and offset both in 0) the axis goes to this same
>position after the home sequence stops. Does that make sense?
>
>I was hoping that with both the home offset and home position in 0 the
>motor would stop near the index pulse with some overshoot may be. But
>always there and not in the last position when the machine turned off. By
>the way I'm not using the POSITION_FILE option and I also tried to use
>VOLATILE_HOME with the same results.
>
>Sorry for all the questions but this is the first machine I'm using with a
>proper homing sequence.
As I tried to explain, you cannot really do homing to index without using
the
encoder as feedback, because only the encoder hardware has the suported
ability
to zero the position at the edge of index.
> You can use homing with index if you use the drives encoder output
>> for feedback, but this is a bit more work and more difficult
>> to tune
>
>
> Do you think is a good practice to use them as feedback? Because I'll
> be coupling the servos to the ballscrews with a timing belt, so this
might
> not be that accurate as having the encoders mounted directly on the
> ballscrews. In that case anyway I would be using the servos with analog
> voltage for velocity command because I can't use them in speed mode with
> the speed reference being pulsed.
This can also be done with step/dir, though tuning may be difficult
if there is substantial lag between step --> motion
>
> Also there's the difference in diameter of the pulleys, but that can be
> scaled so it would be no problem.
>
> By the way, I finally succeeded in using the home switch + index. I think
> it was me trying to imitate the sequence in a wrong way pushing the
switch.
> I'll try a little more to be sure.
>
> El vie., 10 abr. 2020 a las 19:23, Peter C. Wallace (<p...@mesanet.com>)
> escribi????:
>
>> On Fri, 10 Apr 2020, Leonardo Marsaglia wrote:
>>
>> > Date: Fri, 10 Apr 2020 19:01:51 -0300
>> > From: Leonardo Marsaglia <ldmarsag...@gmail.com>
>> > Reply-To: "Enhanced Machine Controller (EMC)"
>> > <emc-users@lists.sourceforge.net>
>> > To: "Enhanced Machine Controller (EMC)" <
>> emc-users@lists.sourceforge.net>
>> > Subject: Re: [Emc-users] Home use index problem (In step + dir mode)
>> >
>> >>> Do you have a home switch and is it working?
>> > If you want to home to index only you must set HOME_SEARCH_VELOCITY
to 0
>> >
>> >
>> >The home switch is working perfectly. I'm pushing it by hand because
I'm
>> >only testing in the desk but it works. If I only use home switch for
>> homing
>> >it works perfectly. The problem starts when I try to use the index
pulse.
>> >
>>
>> If you have homing setup for home switch + index, motions index-enable
>> output
>> should be set high when the home switch is tripped
>>
>>
>> >If I set HOME_SEARCH_VELOCITY to 0 and try to home using index only,
the
>> >joint.0.index-enable pin turns to TRUE and when the motor trips the
index
>> >pulse it stops but I get the following error message and the homing
>> aborts.
>> >Always in the same position of the motor, so the index is detected.
>>
>> The following error is expected unfortunately since motion expects the
>> position
>> to have been zeroed on index detection, but the stepgen position is not
>> affected
>> by the encoder index
>>
>> >>Do the servos have a emulated encopder output (you mentioned a Z
>> signal, is
>> >> that from the drive?)
>> >
>> >
>> >Yes the drive has the encoder output pins for the A, /A, B, /B, Z and
/Z
>> >signals. I'm only using the Z and /Z signals because I thought I could
>> home
>> >with index independently of the feedback I was using.
>>
>> No, index handling is intimately entangled with the position feedback.
>> This is because index detection effectively zeros the position feedback
>> at the instant of index detection so its is precise regardless of the
>> detection
>> speed. As I mentioned, there is firmware to support index detection with
>> the
>> stepgen hardware, but it is not currently supported by the hostmot2
driver
>> Also as you might expect, the stepgen index input pin is not an encoder
>> input
>> rather, its part of the stepgen.
>>
>> You can use homing with index if you use the drives encoder output
>> for feedback, but this is a bit more work and more difficult
>> to tune
>>
>>
>> Peter Wallace
>> Mesa Electronics
>>
>> (\__/)
>> (='.'=) This is Bunny. Copy and paste bunny into your
>> (")_(") signature to help him gain world domination.
>>
>>
>>
>> _______________________________________________
>> Emc-users mailing list
>> Emc-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/emc-users
>>
>
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
Peter Wallace
Mesa Electronics
(\__/)
(='.'=) This is Bunny. Copy and paste bunny into your
(")_(") signature to help him gain world domination.
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
Peter Wallace
Mesa Electronics
(\__/)
(='.'=) This is Bunny. Copy and paste bunny into your
(")_(") signature to help him gain world domination.
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
