I have a setup problem that I think should be a fairly common issue
that I would like to bring up to the development team.
I have a milling machine that uses open loop steppers to operate the
screws, and scales to monitor table position. I want to use the
scales to monitor the table position wh
Brian wrote:
> I have a setup problem that I think should be a fairly common issue
> that I would like to bring up to the development team.
>
> I have a milling machine that uses open loop steppers to operate the
> screws, and scales to monitor table position. I want to use the
> scales to monit
Jon,
You idea would work, along with a number of others, so long as I
never disable the motors. The problem occurs when I disable the
motors and move the table manually.
The problem occurs because the only place to feed position feedback
that will work when the motors are disabled is through
Brian wrote:
> Jon,
>
> You idea would work, along with a number of others, so long as I
> never disable the motors. The problem occurs when I disable the
> motors and move the table manually.
>
>
I think there is a stepper-driven system that works like this, using my
Universal Stepper Contr
On 6 August 2010 18:13, Jon Elson wrote:
> OK, that is a different problem, and I guess the screw compensation has
> been put at the wrong place in the logic.
It is in the right place if the position feedback component is the
motor encoders.
A workaround would be to set up a pyvcp panel to disp
This is what I am currently doing, but it doesn't really do much for
me. I even have a bar graph of the current position error. I think
it would be a natural progression for EMC to have the option to use
feedback from a source that is subject to screw comp, or one that
isn't.
Brian
On Fri, Aug
On 6 August 2010 23:08, Brian wrote:
>
> This is what I am currently doing, but it doesn't really do much for
> me. I even have a bar graph of the current position error. I think
> it would be a natural progression for EMC to have the option to use
> feedback from a source that is subject to scr
Brian wrote:
> This is what I am currently doing, but it doesn't really do much for
> me. I even have a bar graph of the current position error. I think
> it would be a natural progression for EMC to have the option to use
> feedback from a source that is subject to screw comp, or one that
> isn'
Jon Elson wrote:
> Brian wrote:
>
>> This is what I am currently doing, but it doesn't really do much for
>> me. I even have a bar graph of the current position error. I think
>> it would be a natural progression for EMC to have the option to use
>> feedback from a source that is subject to s
Stephen Wille Padnos wrote:
>>
>>
> Nothing has been removed, you would just use the stepgen in velocity
> mode, with a PID driving it (making FF1=1 should give you the equivalent
> of open-loop mode, plus or minus a step here and there).
>
>
I have to retract my previous erroneous st
I may give that a shot as a band-aid to the problem. I suspect that
when reversing direction, the backlash is going to drive the PID nuts
and cause the table to jerk.
I appreciate everyone's interest to find a workable solution for me,
but what I would like to focus on more is making EMC2 a more
On Sat, 2010-08-07 at 12:07 -0500, Jon Elson wrote:
> Stephen Wille Padnos wrote:
> >>
> >>
> > Nothing has been removed, you would just use the stepgen in velocity
> > mode, with a PID driving it (making FF1=1 should give you the equivalent
> > of open-loop mode, plus or minus a step he
On 7 August 2010 22:47, dave wrote:
> Just to confuse the situation and possibly display my own ignorance;
>
> It is possible to take position off the stepper for P and D and feed I
> the linear scale to get final settling?
Not directly, but you could sum two PID blocks which took their
feedback
Yes, that is an alternate way to incorporate the scale into the
movement of the machine, but still leaves me with the problem when the
drives are disabled.
Brian
On Sat, Aug 7, 2010 at 5:47 PM, dave wrote:
> On Sat, 2010-08-07 at 12:07 -0500, Jon Elson wrote:
>> Stephen Wille Padnos wrote:
>> >>
Brian wrote:
> I may give that a shot as a band-aid to the problem. I suspect that
> when reversing direction, the backlash is going to drive the PID nuts
> and cause the table to jerk.
>
>
If you have sizable backlash in your machine, then it becomes very hard
for any CNC system to work right
Jon,
If I had encoders on the screws, the method described on the wiki
would work. This is basically what Dave suggested a few replies back,
and what my config already does to some degree. This method will also
work with steppers, until the motors are disabled, and you move the
machine by hand
Ok, I think I may have a perspective that will be clearer.
In the HAL component axis, joint position and motor position are
thought of as separate concepts. Joint position is related to motor
position through the compensation logic, but otherwise they are their
own entities. This is good desig
Hi
I don't know anything about the inner workings of EMC. I do know a bit
about control systems. I think I understand the question raised here.
If you want to combine the measured joint position and use this
information to figure out what the motor needs to do to maintain control
of the j
Steve's suggestion of using a velocity stepgen with a PID loop and then
using the scale feedback to close the loop - works. I have that running
on an actual machine and at high speeds - 1100 ipm. The machine is
running two shifts right now and is very reliable.
There is no need to add anythin
On 8 August 2010 07:57, Brian wrote:
>
> In the HAL component axis, joint position and motor position are
> thought of as separate concepts. Joint position is related to motor
> position through the compensation logic, but otherwise they are their
> own entities. This is good design, because in
Dave writes:
>The only way around this is to run your machine very slowly, which some
>would say, negates a lot of the benefits of CNC'ing your machine.
I would disagree, although it depends on what you're doing.
>At some point fixing the machine is more productive than trying to work
>around a
Andy,
You and I are now on the same page. That is EXACTLY what I would
like to see. It would be simple, and would solve all the other BS
that has been errantly discussed in this thread.
In addition to splitting the joint-pos-fb PIN as you mention, it
would be cool that add a PIN that has th
Ok... I guess I don't understand what you are trying to do.. or what
the issues are.
I thought this was a control issuethat you want the control to
compensate for backlash and possibly a screw that that has some pitch
errors, and also close the position loop with
some direct reading scale
Dave,
I simply want axis to be able to utilize my scale feedback as actual
joint position. Thats all.
Brian
On Sun, Aug 8, 2010 at 12:28 PM, Dave wrote:
> Ok... I guess I don't understand what you are trying to do.. or what
> the issues are.
>
> I thought this was a control issuethat y
Brian wrote:
> Jon,
>
> If I had encoders on the screws, the method described on the wiki
> would work. This is basically what Dave suggested a few replies back,
> and what my config already does to some degree. This method will also
> work with steppers, until the motors are disabled, and you
Dave wrote:
> Backlash is a difficult thing to handle and it just gets worse when you
> try and put a PID loop around it.
>
> Backlash is like having a loose shaft coupling with slop in it.
>
> Pretty much a control nightmare.
>
Yes, a simple control algorithm requires a linear function between
Andy Pugh wrote:
> Am I right in thinking that your request is basically that
> axis.N.joint-pos-fb should be read/write (or split into
> joint-pos-fb-in and joint-pos-fb-out?). Then what is displayed on the
> screen is the actual position from the scales?
> My impression is that you are happy with
Jon,
> Well, I simply don't understand why this doesn't work.
It quits working when the motors are disabled. Once you do that, the
wiki method wont track the table position with a stepper machine.
> Now, the screw mapping with backlash may not work at all for manual
> moves. I believe the co
> He wants the displayed position to have the screw error (and backlash)
> correction applied. Even if EMC2 doesn't do this now, making the
> displayed position show the corrected value sounds like a couple hours
> of work and a new HAL component. The most difficult part is to arrange
> the code
Brian wrote:
> Jon,
>
>
>> Well, I simply don't understand why this doesn't work.
>>
> It quits working when the motors are disabled. Once you do that, the
> wiki method wont track the table position with a stepper machine.
>
Wait a second. You said in your original email that you c
On 8 August 2010 19:04, Stephen Wille Padnos wrote:
> Wait a second. You said in your original email that you couldn't
> connect the scale feedback to axis.n.joint-pos-fb - have you actually
> tried doing that?
axis.N.joint-pos-fb is an output:
9 float OUT -9.566617 axis.0.joint-pos-fb
Andy Pugh wrote:
> On 8 August 2010 19:04, Stephen Wille Padnos wrote:
>
>
>> Wait a second. You said in your original email that you couldn't
>> connect the scale feedback to axis.n.joint-pos-fb - have you actually
>> tried doing that?
>>
> axis.N.joint-pos-fb is an output:
>
> 9 fl
On Sun, Aug 8, 2010 at 12:27 PM, Stephen Wille Padnos wrote:
> Andy Pugh wrote:
> > On 8 August 2010 19:04, Stephen Wille Padnos wrote:
> >
> >
> >> Wait a second. You said in your original email that you couldn't
> >> connect the scale feedback to axis.n.joint-pos-fb - have you actually
> >> tr
> The motion controller copies feedback to command while in the machine
> off state. The stepgen also updates its internal command position while
> it is disabled (which it should be in machine off). The motion
> controller shouldn't do anything when it's enabled (it might, and that
> would be a
>
> How about axis.N.motor-pos-fb?
>
> from the integrators manual:
>
> axis.N.motor-pos-fb (float, in) The actual position for this joint.
>
> as opposed to:
>
> axis.N.joint-pos-fb (float, out) The joint (as opposed to motor) feedback
> position.
>
Because the screw comp gets added to that value
Ahh, I see your point.
I know it doesn't solve your problem, but perhaps something like a PYVCP DRO
like the one shown here would be a good short term or alternative solution
http://wiki.linuxcnc.org/emcinfo.pl?PyVCP ( A big DRO from Goslowjimbo) I
haven't played with it much because my machine ca
Brian wrote:
> Dave,
>
> I simply want axis to be able to utilize my scale feedback as actual
> joint position. Thats all.
>
And that's what I SIMPLY don't understand. I have two servo-controlled
machines.
One is a converted manual Bridgeport mill, the other a converted 150Lb
minimill.
The
Brian wrote:
> I am quite happy with the way EMC handles screw comp and lash. I am
> not interested in changing my machine or EMC in these regards. The
> ONLY reasons backlash and screw comp have been mentioned at all in
> this thread is because it is the reason I can't feed my scale feedback
> i
On 8 August 2010 22:25, Jon Elson wrote:
> And that's what I SIMPLY don't understand. I have two servo-controlled
> machines.
And I think he has none, if I have been reading right.
I am fairly sure this is a stepper machine that is fitted with linear
scales. My understanding is that leadscrews
Brian wrote:
>> He wants the displayed position to have the screw error (and backlash)
>> correction applied. Even if EMC2 doesn't do this now, making the
>> displayed position show the corrected value sounds like a couple hours
>> of work and a new HAL component. The most difficult part is to ar
Do you really need screw correction AND backlash comp?
I'm not so sure you do.
I would turn off any backlash comp and screw comp and setup the feedback
loop with the linear scales and try and tweak the position loop to get
what you want. Use a velocity mode stepgen and feed that with a PID loo
> And that's what I SIMPLY don't understand. I have two servo-controlled
> machines.
>
> Jon
Jon,
EMC adds the screw compensation to the measured value from the
encoder before it is interpreted as actual joint position. Therefore,
if I want to use screw comp and scale feedback at the same tim
Brian,
If you have a closed loop system and are using scales then using screw
comp really doesn't make any sense since the outer loop is closed to the
scales.
Screw comp only makes sense if you are relying on the screws accuracy.
You will be relying on the scales accuracy, not the screws.
Yo
Andy Pugh wrote:
> On 8 August 2010 22:25, Jon Elson wrote:
>
>
>> And that's what I SIMPLY don't understand. I have two servo-controlled
>> machines.
>>
>
> And I think he has none, if I have been reading right.
>
> I am fairly sure this is a stepper machine that is fitted with linear
>
Dave wrote:
> Who really cares where your motors are at if you can accurately control
> the actual position
> via the scales?Backlash will not be compensated but the velocity
> mode stepgen won't care. The PID loop will try and compensate.
>
Yes, but with larger amounts of backlash, the m
Brian wrote:
> EMC adds the screw compensation to the measured value from the
> encoder before it is interpreted as actual joint position. Therefore,
> if I want to use screw comp and scale feedback at the same time, it
> wont work.
>
The whole POINT of the screw comp table is to correct for
> --
>
> Message: 6
> Date: Sun, 08 Aug 2010 23:10:36 -0500
> From: Jon Elson
> Subject: Re: [Emc-developers] Better logic for scale feedback,
>[Feature Idea]
> To: EMC developers
> Message-ID: <4c5f7fbc.1090...@pico-systems.c
On Sun, 08 Aug 2010 00:51 -0400, "Brian"
wrote:
> Why is there so much resistance to this? I must not be effectively
> explaining the situation. I suspect in person, I could better
> describe the benefits to making it work the way I propose. Over email
> however, it just doesn't seem to be
John Kasunich,
You are spot on with your description of with I have and what I want
to do. I am surprised that my configuration would be uncommon. It is
just a regular knee mill with a DRO. I put stepper motors on the X-Y
handles, and found out that I can wire the DRO's scales to my LPT port
I meant to say,
...The motor-pos-fb at home will become the current motor offset...
Brian
On Mon, Aug 9, 2010 at 12:50 PM, Brian wrote:
> John Kasunich,
>
> You are spot on with your description of with I have and what I want
> to do. I am surprised that my configuration would be uncommon. I
John Kasunich wrote:
>
>
> It sounds like Brian already has a fairly good idea of the
> flow of positions through the motion controller, but to help
> clarify things, take a look at
> http://linuxcnc.org/docs/2.2/html/code_Code_Notes.html#r3_2
>
>
Thank you John, I was looking for this!
OK,
John Kasunich,
I did a little thinking about the homing problem ,and I think I
thought of something that could be implemented that wouldn't break the
exiting homing logic.
According to the flow diagram, the current logic looks like this
motor-fb + (comp + offset) -> joint-fb
If this was imple
52 matches
Mail list logo
