Messages

2931

Mach4 General Discussion / Re: GoToWorkZero

« on: March 10, 2019, 04:24:16 PM »

Hi,
then use the screen load function GoToWorkZero().

I have made three suggestions and you have ignored all of them......

Craig

2932

Modbus / Re: My experiments with Mach3 ModBus and Arduino, Part I

« on: March 10, 2019, 04:21:17 PM »

Hi,
Modbus is a serial communication protocol, there is no way it can send accurately timed Step/Direction pulses
they are way to fast.

Craig

2933

Mach4 General Discussion / Re: GoToWorkZero

« on: March 10, 2019, 04:02:48 PM »

Hi,
apologies, have just noticed a typo in the code, it should read:

Code: [Select]

local rc=mc.mcCntlGotoZero(inst)
if (rc== -18) then
        wx.wxMessageBox('Machine not in idle state')
        do return end
end
I have had another thought, the API you are using calls a standard Mach Action which is programmed in the core and you can't
view or edit it. If it calls for an A axis move you will encounter the same fault, namely that the machine attempts to move
the A axis to zero despite you not having an A axis defined.

If that is the case you will HAVE to use the modified GoToWorkZero() function in the sceern load script as I originally
suggested.

Craig

2934

Mach4 General Discussion / Re: GoToWorkZero

« on: March 10, 2019, 03:51:28 PM »

Hi,
you should attempt to use the signal library instead.

If you wish to use the PLC script it will still work.

Have a close look at the API you are using:

LUA Syntax:
rc = mc.mcCntlGotoZero(
      number mInst)

Description:
Move the X,Y,A,B,C and then the Z axis to zero of the current fixture offset.

Parameters: Parameter Description
mInst The controller instance.


Returns: Return Code Description
MERROR_NOERROR No Error.
MERROR_INVALID_INSTANCE The mInst parameter was out of range.
MERROR_NOT_NOW The operation could not be completed at this time.

Note in particular the return codes. Alter you code to test the return code. I will almost guarantee that the machine
is not in 'idle' state and cannot therefore GoToWorkZero because its busy.

Code: [Select]

local rc=mc.mcCntlGotoZero(inst)
if (rc== -180 then
        wx.wxMessageBox('Machine not in idle state')
        do return end
end
Note that if you use the SigLib{} approach you avoid this difficulty because the GoToWorkZero() function is
in the Screen script and does not have to switch chunks as your approach does.

Try wx4.set, it is way more complete/developed.

Craig

2935

Mach4 General Discussion / Re: GoToWorkZero

« on: March 10, 2019, 03:27:53 PM »

Hi,
I notice that you have made your own screenset? You have very few Tabs to call on.

May I suggest loading wx4.set and see all those additional screens.....including diagnostics, probing  etc.
Then make a copy of it and THEN customize it. Its all very well making a new set from scratch but is there any need
to do without all those features that NFS has laboriously programmed into Mach?

Craig

2936

Mach4 General Discussion / Re: GoToWorkZero

« on: March 10, 2019, 03:23:12 PM »

Hi,
looking at the code I would guess you have that in the PLC script?
That works, but is not efficient. The PLC script runs many times a second, and your code will therefore
run many times per second. This type of input is called 'polling' and is common in Mach3.

With Mach4 however there is another way. Whenever a signal (input or output or even an internal Mach signal)
changes state the Signal Script runs, and the signal script looks into the Signal Library (SigLib{}) in the ScreenLoad
script for an entry corresponding to the signal which has changed. If it finds an entry it executes the inline function
defined there. If it doesn't find an entry, ie we are not interested in that signal then it returns without doing anything.

It sounds complicated and it certainly takes some effort to understand at first but thereafter is  very elegant
and programmatically efficient.

Craig

2937

Mach4 General Discussion / Re: GoToWorkZero

« on: March 10, 2019, 03:05:07 PM »

Hi,
if you only want  the screen button to work then edit the GoToWorkZero() function as I have
suggested and ignore the signal library addition.

Craig

2938

Mach4 General Discussion / Re: Call function in macros folder from button

« on: March 10, 2019, 03:00:23 PM »

Hi,

I'm still learning the nuances of the Mach 4 control structure. But it does make sense.

When we say ' we are trying to learn Lua', its not really correct. The real challenge is learning how the different
parts of Mach interact and the vagaries of Machs API not Lua at all.

Craig

2939

Mach4 General Discussion / Re: GoToWorkZero

« on: March 10, 2019, 02:39:41 PM »

Hi,
open the screen editor and open the screen load script. Scan down through the file till about line 225
until you find the GoToWorkZero function:

Code: [Select]

---------------------------------------------------------------
-- Go To Work Zero() function.
---------------------------------------------------------------
function GoToWorkZero()
    --mc.mcCntlMdiExecute(inst, "G00 X0 Y0 A0")--Without Z moves
    mc.mcCntlMdiExecute(inst, "G00 G53 Z0\nG00 X0 Y0 A0\nG00 Z0")--With Z moves
end

Note that the active line (mc.mcCntlMdiExecute(inst, "G00 G53 Z0\nG00 X0 Y0 A0\nG00 Z)) has an A axis move.
I'm guessing that you don't have an A axis defined in your machine. So either edit the line to delete the A axis
reference or comment it out and add this line instead:

Code: [Select]

mc.mcCntlMdiExecute(inst, "G00 G53 Z0\nG00 X0 Y0 \nG00 Z)

To add a physical button....define a pin on your controller as an input, say ISIG_INPUT10 and then put this in the
SigLib{} in the screen load script:

Code: [Select]

[mc.ISIG_INPUT10]=function(state)
if state==1 then
      GoToWorkZero()
end
end,
Note that this is an entry in the signal library table and the second 'end' statement concludes the in line function definition
and the trailing ',' (comma) concludes the table entry. Don't forget the trailing comma!

Craig

2940

Mach4 General Discussion / Re: CNC drive AXBB-E

« on: March 09, 2019, 07:54:24 PM »

Hi,
yes 'potentially' is indeed the keyword.

Take an AC Step/Direction servo. A stream of pulses, let say 10,000 pulses per second are applied to a servo with a
10,000 count per revolution encoder. The servo will turn one revolution per second. Should the servo be loaded
it will slow somewhat, but the servo drive recognizes that its slipping behind and drives the servo to catch up.

Modern servos can maintain angular position and/or angular speed to within very small fractions of 1%, in the case of the
servo I described above to within 2 parts in 10,000 or so, or 0.02%. More modern servos with 20 bit encoders are very
much better again.

There are limits of course, if the load exceeds the torque the servo can produce it will slip behind and cannot catch up,
a servo would fault 'following error'.

Excluding the situation where the servo is overloaded the ability of the servo to accurately follow its commanded position
(or speed) is determined by the frequency at which it can correct itself. Modern servos (and VFDs) have typically 'refresh
rates' of 5-10kHz. This results in a closed loop servo bandwidth of 500Hz to 1kHz. Such a servo can respond and correct
a fluctuation in position (or speed) with 1/500^th-1/1000^th of a second, pretty damn good.

The ESS PID refresh rate is approximately 1kHz for a servo bandwidth of 100Hz. About 10^th thats available by
an AC servo but still INFINITELY better than open loop. That is what ESS spindle PID does, it makes a feedback servo
of an ordinary spindle motor, albeit with somewhat reduced bandwidth.

You will certainly recall conversations on the forum where the concept of Mach enacting feedback has been discussed.
The communication delays preclude Mach itself from doing a good job. External motion controllers like the Hicon and CSMIO/A
can do so by adding a PID controlled analogue voltage to an analogue servo drive. They add that ability at considerable
cost. The ESS has enacted a similar feature but with one motor only ( the spindle) at no extra cost.

Thus the 'potential' is in fact 'ACTUAL' within the exception of overload, for which PID feedback cant fix, nor any other means either.

Craig