http://igor.chudov.com/projects/Brid...MC2-First-Run/

After wiring garage with Ethernet last night I moved on a bit. Worked
until 2am.

I have configured my mill PC to run EMC2 with "ppmc" configuration.

I tried to do a good job, so that all configs are in my CVS source
tree, there are shell aliases for everything, etc. The idea is that I
want everything to be based on computer settings, so that one computer
could possibly run several CNC devices and that I could run EMC2 on
several computers. I do not want to hard code everything so that the
only thing I run is my Bridgeport Interact from that PC.

Any of my Linux computers for which CNC_ROLE is set via host settings,
loads these shell aliases and requires a default machine to be
specified.

Just a little planning for the future, so to speak.

Anyway, I created a setup based on 'ppmc', and, lo and behold, it
talks to Jon's PPMC board and displays positions. I also wired the
speed command of the drive to PPMC's analog output for Y, so that PPMC
could drive the Y axis. I verified that it is able to do so, using
ppmcdiags program. Based on voltage that I specify, the servo motor
spins just fine.

Right now, the only axis able to move and equipped with encoder is Y,
and even on it, I have disconnected the timing belt. So it is just a
servo motor spinning back and forth. Nothing could possibly crash into
anything or pinch any fingers and other body parts. Very safe
environment to work on software and on estops and things like that.

Because this is all Linux and X11 based, I can do fun things like ssh
to the CNC computer, run EMC2 from ssh, etc. So, I can work on setting
this up, without even being in the garage.

The first task of the day, is to disable all estop functionality that
is in the default configs, so that I could get the servo motor to
move.

I will, of course, re-enable estop when I have that stuff wired
properly.

i

Ignoramus2398 wrote:

http://igor.chudov.com/projects/Brid...MC2-First-Run/

After wiring garage with Ethernet last night I moved on a bit. Worked
until 2am.

I have configured my mill PC to run EMC2 with "ppmc" configuration.

I tried to do a good job, so that all configs are in my CVS source
tree, there are shell aliases for everything, etc. The idea is that I
want everything to be based on computer settings, so that one computer
could possibly run several CNC devices and that I could run EMC2 on
several computers. I do not want to hard code everything so that the
only thing I run is my Bridgeport Interact from that PC.

Any of my Linux computers for which CNC_ROLE is set via host settings,
loads these shell aliases and requires a default machine to be
specified.

Just a little planning for the future, so to speak.

Anyway, I created a setup based on 'ppmc', and, lo and behold, it
talks to Jon's PPMC board and displays positions. I also wired the
speed command of the drive to PPMC's analog output for Y, so that PPMC
could drive the Y axis. I verified that it is able to do so, using
ppmcdiags program. Based on voltage that I specify, the servo motor
spins just fine.

Right now, the only axis able to move and equipped with encoder is Y,
and even on it, I have disconnected the timing belt. So it is just a
servo motor spinning back and forth. Nothing could possibly crash into
anything or pinch any fingers and other body parts. Very safe
environment to work on software and on estops and things like that.

Because this is all Linux and X11 based, I can do fun things like ssh
to the CNC computer, run EMC2 from ssh, etc. So, I can work on setting
this up, without even being in the garage.

The first task of the day, is to disable all estop functionality that
is in the default configs, so that I could get the servo motor to
move.

I will, of course, re-enable estop when I have that stuff wired
properly.

i

PCs are cheaper than dirt, and Linux and EMC2 are free. There is no
point in trying to have one PC control multiple machines *unless* you
are also multitasking the drives and I/O boards between the machines. I
do this with one PC and set of drives controlling either CNC plasma or
CNC mini mill. If you aren't sharing the drives and I/O boards, it makes
little sense to try to share a lousy $99 PC.

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

http://igor.chudov.com/projects/Brid...MC2-First-Run/

After wiring garage with Ethernet last night I moved on a bit. Worked
until 2am.

I have configured my mill PC to run EMC2 with "ppmc" configuration.

I tried to do a good job, so that all configs are in my CVS source
tree, there are shell aliases for everything, etc. The idea is that I
want everything to be based on computer settings, so that one computer
could possibly run several CNC devices and that I could run EMC2 on
several computers. I do not want to hard code everything so that the
only thing I run is my Bridgeport Interact from that PC.

Any of my Linux computers for which CNC_ROLE is set via host settings,
loads these shell aliases and requires a default machine to be
specified.

Just a little planning for the future, so to speak.

Anyway, I created a setup based on 'ppmc', and, lo and behold, it
talks to Jon's PPMC board and displays positions. I also wired the
speed command of the drive to PPMC's analog output for Y, so that PPMC
could drive the Y axis. I verified that it is able to do so, using
ppmcdiags program. Based on voltage that I specify, the servo motor
spins just fine.

Right now, the only axis able to move and equipped with encoder is Y,
and even on it, I have disconnected the timing belt. So it is just a
servo motor spinning back and forth. Nothing could possibly crash into
anything or pinch any fingers and other body parts. Very safe
environment to work on software and on estops and things like that.

Because this is all Linux and X11 based, I can do fun things like ssh
to the CNC computer, run EMC2 from ssh, etc. So, I can work on setting
this up, without even being in the garage.

The first task of the day, is to disable all estop functionality that
is in the default configs, so that I could get the servo motor to
move.

I will, of course, re-enable estop when I have that stuff wired
properly.

i

PCs are cheaper than dirt, and Linux and EMC2 are free. There is no
point in trying to have one PC control multiple machines *unless* you
are also multitasking the drives and I/O boards between the machines. I
do this with one PC and set of drives controlling either CNC plasma or
CNC mini mill. If you aren't sharing the drives and I/O boards, it makes
little sense to try to share a lousy $99 PC.

Maybe, but doing things in a configurable way like this, makes for
clean, understandable computer settings. I could run two devices off
of one PC if I wanted to. I am not saying I should. What I did is I
avoided hard coding anything.

i

http://unitaig.jalling.dk/index.php?id=115
http://wiki.linuxcnc.org/cgi-bin/emc...ixingSMIIssues

You can improve reliability and response on your EMC2 system
by decreasing hardware latency caused by the System Management
Interrupt.

I ran Neils Jalling's hack on my EMC2 system and it decreased
jitter by ~ 5% and caused the dreaded "following
error" to disappear.

The thermal control hardware continued to work properly on my
system after the hack but there are rumored to be systems that
will overheat because of thermal routines handled by the SMI.


--Winston

On 2010-07-14, Winston wrote:
http://unitaig.jalling.dk/index.php?id=115
http://wiki.linuxcnc.org/cgi-bin/emc...ixingSMIIssues

You can improve reliability and response on your EMC2 system
by decreasing hardware latency caused by the System Management
Interrupt.

Thanks. I did not think about it. This is an older PC "designed
for Windows 2000" and thus is unlikely to be exposed.

In addition, I close the velocity loop with tachometers (changed my
mind again and I am using tachs as of now), so that alone could help
with jitter. Even if the CPU is taking a millisecond nap, the servo
drives maintain velocity.

I ran Neils Jalling's hack on my EMC2 system and it decreased
jitter by ~ 5% and caused the dreaded "following
error" to disappear.

The thermal control hardware continued to work properly on my
system after the hack but there are rumored to be systems that
will overheat because of thermal routines handled by the SMI.

If/when I get my mill to work, I want to buy a mini-ITX case with a
new mobo, SSD etc and will put it inside the control cabinet. When I
look for hardware, I will double check the SMI issue. Maybe I will get
AMD hardware.

i

Ignoramus2398 wrote:

On 2010-07-14, Winston wrote:
http://unitaig.jalling.dk/index.php?id=115
http://wiki.linuxcnc.org/cgi-bin/emc...ixingSMIIssues

You can improve reliability and response on your EMC2 system
by decreasing hardware latency caused by the System Management
Interrupt.

Thanks. I did not think about it. This is an older PC "designed
for Windows 2000" and thus is unlikely to be exposed.

In addition, I close the velocity loop with tachometers (changed my
mind again and I am using tachs as of now), so that alone could help
with jitter. Even if the CPU is taking a millisecond nap, the servo
drives maintain velocity.

I ran Neils Jalling's hack on my EMC2 system and it decreased
jitter by ~ 5% and caused the dreaded "following
error" to disappear.

The thermal control hardware continued to work properly on my
system after the hack but there are rumored to be systems that
will overheat because of thermal routines handled by the SMI.

If/when I get my mill to work, I want to buy a mini-ITX case with a
new mobo, SSD etc and will put it inside the control cabinet. When I
look for hardware, I will double check the SMI issue. Maybe I will get
AMD hardware.

i

If you want to go compact and in the control cab, look at PC104 form
factor systems and some of the other industrial PC variants many of
which are not expensive and are specifically designed to run Linux. Try
the advertiser index on circuitcellar.com if you need a starting point.

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

On 2010-07-14, Winston wrote:
http://unitaig.jalling.dk/index.php?id=115
http://wiki.linuxcnc.org/cgi-bin/emc...ixingSMIIssues

You can improve reliability and response on your EMC2 system
by decreasing hardware latency caused by the System Management
Interrupt.

Thanks. I did not think about it. This is an older PC "designed
for Windows 2000" and thus is unlikely to be exposed.

In addition, I close the velocity loop with tachometers (changed my
mind again and I am using tachs as of now), so that alone could help
with jitter. Even if the CPU is taking a millisecond nap, the servo
drives maintain velocity.

I ran Neils Jalling's hack on my EMC2 system and it decreased
jitter by ~ 5% and caused the dreaded "following
error" to disappear.

The thermal control hardware continued to work properly on my
system after the hack but there are rumored to be systems that
will overheat because of thermal routines handled by the SMI.

If/when I get my mill to work, I want to buy a mini-ITX case with a
new mobo, SSD etc and will put it inside the control cabinet. When I
look for hardware, I will double check the SMI issue. Maybe I will get
AMD hardware.

i

If you want to go compact and in the control cab, look at PC104 form
factor systems and some of the other industrial PC variants many of
which are not expensive and are specifically designed to run Linux. Try
the advertiser index on circuitcellar.com if you need a starting point.

Well, I picked a few newegg items, if all goes well, I will buy them:

http://www.newegg.com/Product/Produc...82E16811154091
http://www.newegg.com/Product/Produc...82E16813121381
http://www.newegg.com/Product/Produc...82E16819115056
http://www.newegg.com/Product/Produc...82E16820145184
http://www.newegg.com/Product/Produc...82E16827135204
http://www.newegg.com/Product/Produc...82E16820167025

I asked my friend, who is a computer build expert, if they will fit
together, but I think that they will.

i

On 7/14/2010 8:27 AM, Ignoramus2398 wrote:
On 2010-07-14, wrote:
http://unitaig.jalling.dk/index.php?id=115
http://wiki.linuxcnc.org/cgi-bin/emc...ixingSMIIssues

You can improve reliability and response on your EMC2 system
by decreasing hardware latency caused by the System Management
Interrupt.

Thanks. I did not think about it. This is an older PC "designed
for Windows 2000" and thus is unlikely to be exposed.

SMI has been on Intel processors since Feb. of 1992.

It couldn't hurt to check. See Message 3 under:
http://wiki.linuxcnc.org/emcinfo.pl?TroubleShooting


--Winston

Another minor bit of progress: I was able to set up the Y axis
parameters enough, so that when I turn "machine on", it holds position
and when I press up and down arrows (which control Y axis in EMC), the
position changes and is then held.

I was doing this from away from home, so I am not totally sure that it
worked right, but it looks to me like it did.

The only problem is that the position that is held, is not very
stable, and changes from, say, 0.2345 to 0.2344 to 0.2346 and so on. I
think that it is because I have not yet done PID tuning, have not
adjusted gains and Test/Offset on the drives etc. I am not losing
sleep over it, because I am supposed to fix it once I hook up the
servos to ballscrews.

I think that the next thing I should do is wire limit switches and
estops and test all that with the servo motor disconnected from the
leadscrew.

Things are progressing much faster right now than I expected 2 days
ago.

i

Ignoramus2398 wrote:

Another minor bit of progress: I was able to set up the Y axis
parameters enough, so that when I turn "machine on", it holds position
and when I press up and down arrows (which control Y axis in EMC), the
position changes and is then held.

I was doing this from away from home, so I am not totally sure that it
worked right, but it looks to me like it did.

The only problem is that the position that is held, is not very
stable, and changes from, say, 0.2345 to 0.2344 to 0.2346 and so on. I
think that it is because I have not yet done PID tuning, have not
adjusted gains and Test/Offset on the drives etc. I am not losing
sleep over it, because I am supposed to fix it once I hook up the
servos to ballscrews.

I think that the next thing I should do is wire limit switches and
estops and test all that with the servo motor disconnected from the
leadscrew.

Things are progressing much faster right now than I expected 2 days
ago.

i

Try entering some G code commands in MDI mode, something like G01 Y10
F25 or G01 Y0 F10. It's more fun to watch it do a move on it's own than
to just jog it with the keyboard.

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

Another minor bit of progress: I was able to set up the Y axis
parameters enough, so that when I turn "machine on", it holds position
and when I press up and down arrows (which control Y axis in EMC), the
position changes and is then held.

I was doing this from away from home, so I am not totally sure that it
worked right, but it looks to me like it did.

The only problem is that the position that is held, is not very
stable, and changes from, say, 0.2345 to 0.2344 to 0.2346 and so on. I
think that it is because I have not yet done PID tuning, have not
adjusted gains and Test/Offset on the drives etc. I am not losing
sleep over it, because I am supposed to fix it once I hook up the
servos to ballscrews.

I think that the next thing I should do is wire limit switches and
estops and test all that with the servo motor disconnected from the
leadscrew.

Things are progressing much faster right now than I expected 2 days
ago.

i

Try entering some G code commands in MDI mode, something like G01 Y10
F25 or G01 Y0 F10. It's more fun to watch it do a move on it's own than
to just jog it with the keyboard.

I think that it cannot do it until I home it. It says "Cannot issue
MDI commands until homed". And I cannot yet home this mill.

i

Ignoramus2398 wrote:

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

Another minor bit of progress: I was able to set up the Y axis
parameters enough, so that when I turn "machine on", it holds position
and when I press up and down arrows (which control Y axis in EMC), the
position changes and is then held.

I was doing this from away from home, so I am not totally sure that it
worked right, but it looks to me like it did.

The only problem is that the position that is held, is not very
stable, and changes from, say, 0.2345 to 0.2344 to 0.2346 and so on. I
think that it is because I have not yet done PID tuning, have not
adjusted gains and Test/Offset on the drives etc. I am not losing
sleep over it, because I am supposed to fix it once I hook up the
servos to ballscrews.

I think that the next thing I should do is wire limit switches and
estops and test all that with the servo motor disconnected from the
leadscrew.

Things are progressing much faster right now than I expected 2 days
ago.

i

Try entering some G code commands in MDI mode, something like G01 Y10
F25 or G01 Y0 F10. It's more fun to watch it do a move on it's own than
to just jog it with the keyboard.

I think that it cannot do it until I home it. It says "Cannot issue
MDI commands until homed". And I cannot yet home this mill.

i

You should be able to manually home it, i.e. tell it "this is home".

Ignoramus2398 wrote:

My takeaway: Run the latency test before buying your mobo, if possible.
Some configurations just aren't compatible without significant
blood sweat and tears.

http://issaris.org/rtai/list.php?off...der=machine_id

I do not know how to do that -- no one I know has a Mini-ITX motherboard.

Which one do you have? I have several
different Mini-ITX boards at work.

On 7/14/2010 12:24 PM, Pete C. wrote:

Ignoramus2398 wrote:

On 2010-07-14, Pete wrote:

Ignoramus2398 wrote:


(...)

I think that it cannot do it until I home it. It says "Cannot issue
MDI commands until homed". And I cannot yet home this mill.

i

You should be able to manually home it, i.e. tell it "this is home".

From a booted, idle screen,

Press F1 (Which turns off eStop) then,
F2 ("Toggle Machine Power")
then F3 (Manual Control).

Locate the Home key on your keyboard and press it once.
Select the next axis (Y) and press the Home key once
more. Select the final axis (Z) and press the home key again.

Continue selecting and Homing until you run out of axis'.


Welcome Home.



--Winston

On 7/14/2010 12:27 PM, Ignoramus2398 wrote:

(...)

I do not know how to do that -- no one I know has a Mini-ITX motherboard.


Please read section 3.1 beginning at "To specifically test the
operation of the RealTime kernel, use the kernel latency test
supplied with RTAI.":

http://wiki.linuxcnc.org/cgi-bin/emc...roubleShooting

It runs on the Linux command line.

As it is running, exercise the system *without running EMC2*.
Open windows, edit a file. Pop open a spreadsheet.

Let the program run for say 30 minutes, gathering
latency statistics for you.


--Winston

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

Another minor bit of progress: I was able to set up the Y axis
parameters enough, so that when I turn "machine on", it holds position
and when I press up and down arrows (which control Y axis in EMC), the
position changes and is then held.

I was doing this from away from home, so I am not totally sure that it
worked right, but it looks to me like it did.

The only problem is that the position that is held, is not very
stable, and changes from, say, 0.2345 to 0.2344 to 0.2346 and so on. I
think that it is because I have not yet done PID tuning, have not
adjusted gains and Test/Offset on the drives etc. I am not losing
sleep over it, because I am supposed to fix it once I hook up the
servos to ballscrews.

I think that the next thing I should do is wire limit switches and
estops and test all that with the servo motor disconnected from the
leadscrew.

Things are progressing much faster right now than I expected 2 days
ago.

i

Try entering some G code commands in MDI mode, something like G01 Y10
F25 or G01 Y0 F10. It's more fun to watch it do a move on it's own than
to just jog it with the keyboard.

I think that it cannot do it until I home it. It says "Cannot issue
MDI commands until homed". And I cannot yet home this mill.

i

You should be able to manually home it, i.e. tell it "this is home".

Yay, I figured it out, it works great!

I am still away from home, but I am sure the servo motor is spinning
etc. I executed my first G codes! I will verify it when I get home.

thanks!!

i

On 2010-07-14, Jim Stewart wrote:
Ignoramus2398 wrote:

My takeaway: Run the latency test before buying your mobo, if possible.
Some configurations just aren't compatible without significant
blood sweat and tears.

http://issaris.org/rtai/list.php?off...der=machine_id

I do not know how to do that -- no one I know has a Mini-ITX motherboard.

Which one do you have? I have several
different Mini-ITX boards at work.

Intel CPU: http://www.newegg.com/Product/Produc...82E16819115056
Motherboard: http://www.newegg.com/Product/Produc...82E16813121381

On 7/14/2010 1:00 PM, Ignoramus2398 wrote:

(...)

Yay, I figured it out, it works great!

I am still away from home, but I am sure the servo motor is spinning
etc. I executed my first G codes! I will verify it when I get home.

Kewl!

--Winston

Ignoramus2398 wrote:

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

Another minor bit of progress: I was able to set up the Y axis
parameters enough, so that when I turn "machine on", it holds position
and when I press up and down arrows (which control Y axis in EMC), the
position changes and is then held.

I was doing this from away from home, so I am not totally sure that it
worked right, but it looks to me like it did.

The only problem is that the position that is held, is not very
stable, and changes from, say, 0.2345 to 0.2344 to 0.2346 and so on. I
think that it is because I have not yet done PID tuning, have not
adjusted gains and Test/Offset on the drives etc. I am not losing
sleep over it, because I am supposed to fix it once I hook up the
servos to ballscrews.

I think that the next thing I should do is wire limit switches and
estops and test all that with the servo motor disconnected from the
leadscrew.

Things are progressing much faster right now than I expected 2 days
ago.

i

Try entering some G code commands in MDI mode, something like G01 Y10
F25 or G01 Y0 F10. It's more fun to watch it do a move on it's own than
to just jog it with the keyboard.

I think that it cannot do it until I home it. It says "Cannot issue
MDI commands until homed". And I cannot yet home this mill.

i

You should be able to manually home it, i.e. tell it "this is home".

Yay, I figured it out, it works great!

I am still away from home, but I am sure the servo motor is spinning
etc. I executed my first G codes! I will verify it when I get home.

thanks!!

i

If you had your netcam pointed at the motor, you could see it spinning
Of course if it wasn't spinning you wouldn't be getting encoder
counts, so you can be sure it's spinning.

Ignoramus2398 wrote:
On 2010-07-14, Jim wrote:
Ignoramus2398 wrote:

My takeaway: Run the latency test before buying your mobo, if possible.
Some configurations just aren't compatible without significant
blood sweat and tears.

http://issaris.org/rtai/list.php?off...der=machine_id

I do not know how to do that -- no one I know has a Mini-ITX motherboard.

Which one do you have? I have several
different Mini-ITX boards at work.

Intel CPU: http://www.newegg.com/Product/Produc...82E16819115056
Motherboard: http://www.newegg.com/Product/Produc...82E16813121381

Don't have one of those. Sorry.

On 2010-07-14, Winston wrote:
On 7/14/2010 12:24 PM, Pete C. wrote:

Ignoramus2398 wrote:

On 2010-07-14, Pete wrote:

Ignoramus2398 wrote:


(...)

I think that it cannot do it until I home it. It says "Cannot issue
MDI commands until homed". And I cannot yet home this mill.

i

You should be able to manually home it, i.e. tell it "this is home".

From a booted, idle screen,

Press F1 (Which turns off eStop) then,
F2 ("Toggle Machine Power")
then F3 (Manual Control).

Locate the Home key on your keyboard and press it once.
Select the next axis (Y) and press the Home key once
more. Select the final axis (Z) and press the home key again.

Continue selecting and Homing until you run out of axis'.


Awesome. Now, when I do G01 with a higher speed, speed up to F16 or so
works. But above that, say at f20, I get a following error. Same for
G00.

i

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

Another minor bit of progress: I was able to set up the Y axis
parameters enough, so that when I turn "machine on", it holds position
and when I press up and down arrows (which control Y axis in EMC), the
position changes and is then held.

I was doing this from away from home, so I am not totally sure that it
worked right, but it looks to me like it did.

The only problem is that the position that is held, is not very
stable, and changes from, say, 0.2345 to 0.2344 to 0.2346 and so on. I
think that it is because I have not yet done PID tuning, have not
adjusted gains and Test/Offset on the drives etc. I am not losing
sleep over it, because I am supposed to fix it once I hook up the
servos to ballscrews.

I think that the next thing I should do is wire limit switches and
estops and test all that with the servo motor disconnected from the
leadscrew.

Things are progressing much faster right now than I expected 2 days
ago.

i

Try entering some G code commands in MDI mode, something like G01 Y10
F25 or G01 Y0 F10. It's more fun to watch it do a move on it's own than
to just jog it with the keyboard.

I think that it cannot do it until I home it. It says "Cannot issue
MDI commands until homed". And I cannot yet home this mill.

i

You should be able to manually home it, i.e. tell it "this is home".

Yay, I figured it out, it works great!

I am still away from home, but I am sure the servo motor is spinning
etc. I executed my first G codes! I will verify it when I get home.

thanks!!

i

If you had your netcam pointed at the motor, you could see it spinning

The motor's shaft is smooth. I could not see it spin unless I put a
sticker on it.

Of course if it wasn't spinning you wouldn't be getting encoder
counts, so you can be sure it's spinning.

exactly!

Any idea what could cause a following error at higher speeds (f20
instead of f10)?

i

Any idea what could cause a following error at higher speeds (f20
instead of f10)?

Gizmo - AKA gremlins...

....Seriously though, good luck.
--


Regards,
Joe Agro, Jr.
(800) 871-5022
01.908.542.0244
Automatic / Pneumatic Drills: http://www.AutoDrill.com
Multiple Spindle Drills: http://www.Multi-Drill.com
Production Tapping: http://Production-Tapping-Equipment.com/
Flagship Site: http://www.Drill-N-Tap.com
VIDEOS: http://www.youtube.com/user/AutoDrill
TWITTER: http://twitter.com/AutoDrill
FACEBOOK: http://tinyurl.com/AutoDrill-Facebook

V8013-R

Ignoramus2398 wrote:

Right now, the only axis able to move and equipped with encoder is Y,
and even on it, I have disconnected the timing belt. So it is just a
servo motor spinning back and forth. Nothing could possibly crash into
anything or pinch any fingers and other body parts. Very safe
environment to work on software and on estops and things like that.

Smart thinking. I haven't used emc to to control anything yet but initial tests of drives
on anything I've built is watch the motor, not connected to anything do it's tricks.


Because this is all Linux and X11 based, I can do fun things like ssh
to the CNC computer, run EMC2 from ssh, etc. So, I can work on setting
this up, without even being in the garage.

You have a net cam pointing at the motor? You could be having a blast at work on breaks,
breaks, yah, just breaks. Maybe make up for coding stuff for work at home.


The first task of the day, is to disable all estop functionality that
is in the default configs, so that I could get the servo motor to
move.

I will, of course, re-enable estop when I have that stuff wired
properly.

Don't forget, easy to want to move to the next step skipping this one until later.

Wes
--
"Additionally as a security officer, I carry a gun to protect
government officials but my life isn't worth protecting at home
in their eyes." Dick Anthony Heller

On 7/14/2010 1:44 PM, Ignoramus2398 wrote:

(...)

Any idea what could cause a following error at higher speeds (f20
instead of f10)?

Winston ...and caused the dreaded "following
Winston error" to disappear.

A 'following error' you say?

Sounds serious.
Can anyone help Iggy with this?

--Winston

On 7/14/2010 1:58 PM, Wes wrote:
wrote:

(...)

I will, of course, re-enable estop when I have that stuff wired
properly.

Don't forget, easy to want to move to the next step skipping this one until later.


Um. (Cough.) [Stares at feet] Yes.

--Winston

On 2010-07-14, Winston wrote:
On 7/14/2010 1:44 PM, Ignoramus2398 wrote:

(...)

Any idea what could cause a following error at higher speeds (f20
instead of f10)?

Winston ...and caused the dreaded "following
Winston error" to disappear.

A 'following error' you say?

Sounds serious.
Can anyone help Iggy with this?

--Winston

I was thinking, and I think that I know what is happening, not yet
exacatly how to fix it.

I tell EMC what is my speed at max voltage, and EMC is measuring
actual position.

I did not correctly specify the speed and thus EMC expects a wrong
position.

i

On 2010-07-14, Wes wrote:
I will, of course, re-enable estop when I have that stuff wired
properly.

Don't forget, easy to want to move to the next step skipping this one until later.

And might end up with a broken mill.

i

Ignoramus2398 wrote:

On 2010-07-14, Winston wrote:
On 7/14/2010 12:24 PM, Pete C. wrote:

Ignoramus2398 wrote:

On 2010-07-14, Pete wrote:

Ignoramus2398 wrote:


(...)

I think that it cannot do it until I home it. It says "Cannot issue
MDI commands until homed". And I cannot yet home this mill.

i

You should be able to manually home it, i.e. tell it "this is home".

From a booted, idle screen,

Press F1 (Which turns off eStop) then,
F2 ("Toggle Machine Power")
then F3 (Manual Control).

Locate the Home key on your keyboard and press it once.
Select the next axis (Y) and press the Home key once
more. Select the final axis (Z) and press the home key again.

Continue selecting and Homing until you run out of axis'.


Awesome. Now, when I do G01 with a higher speed, speed up to F16 or so
works. But above that, say at f20, I get a following error. Same for
G00.

Tuning...

Acceleration, deceleration, top speed, etc. Will of course be different
when you have machine mass moving. It's probably using too high an
acceleration rate currently.

Ignoramus2398 wrote:

On 2010-07-14, Winston wrote:
On 7/14/2010 1:44 PM, Ignoramus2398 wrote:

(...)

Any idea what could cause a following error at higher speeds (f20
instead of f10)?

Winston ...and caused the dreaded "following
Winston error" to disappear.

A 'following error' you say?

Sounds serious.
Can anyone help Iggy with this?

--Winston

I was thinking, and I think that I know what is happening, not yet
exacatly how to fix it.

I tell EMC what is my speed at max voltage, and EMC is measuring
actual position.

I did not correctly specify the speed and thus EMC expects a wrong
position.

i

Acceleration rate. You probably have it trying to go from stopped to the
specified speed with no acceleration ramp. You get away with it when
specifying a low speed since you only have the motor mass to spin up
currently, but even then when you try to do 0-20 IPM it can't accelerate
fast enough and errors. The same will apply to your deceleration rate.
Both of those rates will also be a lot lower when you have actual
machine mass to get moving or stopped.

Ignoramus2398 wrote:

On 2010-07-14, Wes wrote:
I will, of course, re-enable estop when I have that stuff wired
properly.

Don't forget, easy to want to move to the next step skipping this one until later.

And might end up with a broken mill.


I see you understand the big picture. Have fun and keep reporting. I sure am enjoying
your progress in a vicarious fashion.

Wes

--
"Additionally as a security officer, I carry a gun to protect
government officials but my life isn't worth protecting at home
in their eyes." Dick Anthony Heller

On 2010-07-14, Pete C. wrote:

Ignoramus2398 wrote:

On 2010-07-14, Winston wrote:
On 7/14/2010 1:44 PM, Ignoramus2398 wrote:

(...)

Any idea what could cause a following error at higher speeds (f20
instead of f10)?

Winston ...and caused the dreaded "following
Winston error" to disappear.

A 'following error' you say?

Sounds serious.
Can anyone help Iggy with this?

--Winston

I was thinking, and I think that I know what is happening, not yet
exacatly how to fix it.

I tell EMC what is my speed at max voltage, and EMC is measuring
actual position.

I did not correctly specify the speed and thus EMC expects a wrong
position.

i

Acceleration rate. You probably have it trying to go from stopped to the
specified speed with no acceleration ramp. You get away with it when
specifying a low speed since you only have the motor mass to spin up
currently, but even then when you try to do 0-20 IPM it can't accelerate
fast enough and errors. The same will apply to your deceleration rate.
Both of those rates will also be a lot lower when you have actual
machine mass to get moving or stopped.

but my motor is not connected to the milling table, it spins by itself!

Ignoramus2398 wrote:
Another minor bit of progress: I was able to set up the Y axis
parameters enough, so that when I turn "machine on", it holds position
and when I press up and down arrows (which control Y axis in EMC), the
position changes and is then held.

I was doing this from away from home, so I am not totally sure that it
worked right, but it looks to me like it did.

What, you were powering up the servos and spinning the motor while not
even in the same building????
The only problem is that the position that is held, is not very
stable, and changes from, say, 0.2345 to 0.2344 to 0.2346 and so on. I
think that it is because I have not yet done PID tuning, have not
adjusted gains and Test/Offset on the drives etc. I am not losing
sleep over it, because I am supposed to fix it once I hook up the
servos to ballscrews.

I sure wouldn't worry about it. I think we figured out you have 4000 x
5 = 20,000 counts/inch. (Maybe I am neglecting a belt ratio there.)
So, that is .00005" per encoder count. If it is holding position to
+/- .0001", that is only two encoder counts! VERY good.
I think that the next thing I should do is wire limit switches and
estops and test all that with the servo motor disconnected from the
leadscrew.
Well, once the servo is holding position without runaways, you can just
hook it all up, as long as you have at least a manual E-stop that is
easy to get to.

Jon

Ignoramus2398 wrote:


Awesome. Now, when I do G01 with a higher speed, speed up to F16 or so
works. But above that, say at f20, I get a following error. Same for
G00.
Right, that means you need to do the servo tuning.
See http://wiki.linuxcnc.org/cgi-bin/emc...rvo_Amplifiers
for some info. This was written for my PWM servo amps, but almost
everything there will be perfect for velocity servos, too.

Jon

Pete C. wrote:
Acceleration rate. You probably have it trying to go from stopped to the
specified speed with no acceleration ramp.
Impossible. It is possible in EMC2 to set an acceleration higher than
your servo can deliver, but it will accelerate at whatever rate you have
specified in the .ini file. It is impossible to command a move without
acceleration. Even if you give the MDI command G01 X25 F200000, it will
ignore the ridiculous feedrate and move at the specified max velocity,
and accelerate and decelerate from max vel at the max_accel rate.

If Iggy's servo tuning has too low a gain, then there will be a greater
servo lag as speed increases. The longer you move, the farther the axis
lags behind, eventually tripping the following error limit. Increasing
the proportional gain in the PID function's P term will reduce the
error. Also, the axis' MAX_VEL must be set right so that a commanded
velocity will generate nearly the right DAC output without waiting for
position feedback.

Jon

On 7/14/2010 3:51 PM, Jon Elson wrote:
Pete C. wrote:
Acceleration rate. You probably have it trying to go from stopped to the
specified speed with no acceleration ramp.
Impossible. It is possible in EMC2 to set an acceleration higher than
your servo can deliver, but it will accelerate at whatever rate you have
specified in the .ini file. It is impossible to command a move without
acceleration. Even if you give the MDI command G01 X25 F200000, it will
ignore the ridiculous feedrate and move at the specified max velocity,
and accelerate and decelerate from max vel at the max_accel rate.

If Iggy's servo tuning has too low a gain, then there will be a greater
servo lag as speed increases. The longer you move, the farther the axis
lags behind, eventually tripping the following error limit. Increasing
the proportional gain in the PID function's P term will reduce the
error. Also, the axis' MAX_VEL must be set right so that a commanded
velocity will generate nearly the right DAC output without waiting for
position feedback.

One thing is for absolutely certain and that is
that there is no issue with operating system latency.

Even without dismissing unnecessary BIOS functions
and running the realtime kernel latency test, I
can tell from here that the computer returns from
the System Management Interrupt in *plenty* of time
to keep track of encoder values.

(Yes, I will try anything.)

--Winston

Ignoramus2398 fired this volley in
news
but my motor is not connected to the milling table, it spins by itself!


That's what he just said, Iggy. Even the motor mass is excessive if you
have no acceleration time.

LLoyd

On 2010-07-15, Lloyd E. Sponenburgh lloydspinsidemindspring.com wrote:
Ignoramus2398 fired this volley in
news
but my motor is not connected to the milling table, it spins by itself!


That's what he just said, Iggy. Even the motor mass is excessive if you
have no acceleration time.

Well, I do have max_acceleration set.

i

Ignoramus2398 fired this volley in
:

Well, I do have max_acceleration set.

The operative question would be, "To what?"

It may be excessive, even for the motors own inertia.

LLoyd

On Wed, 14 Jul 2010 15:43:40 -0500, Ignoramus2398
wrote:


Continue selecting and Homing until you run out of axis'.


Awesome. Now, when I do G01 with a higher speed, speed up to F16 or so
works. But above that, say at f20, I get a following error. Same for
G00.

Tuned your amps yet?

Now is the time.

Gunner

One could not be a successful Leftwinger without realizing that,
in contrast to the popular conception supported by newspapers
and mothers of Leftwingers, a goodly number of Leftwingers are
not only narrow-minded and dull, but also just stupid.
Gunner Asch