This documentation is no longer maintained. For documentation of the current version of emc2, please see http://www.linuxcnc.org/docview/html

Table of Contents

List of figures

1 HAL Examples

Most of these examples assume you are starting with a Stepconf based configuration and have two threads base-thread and servo-thread. The Stepconf wizard will create an empty custom.hal and a custom_postgui.hal file. The custom.hal file will be loaded after the configuration hal file and the custom_postgui.hal file is loaded after the GUI has been loaded.

1.1 Manual Toolchange

In this example it is assumed that your "rolling your own" configuration and wish to add the HAL Manual Toolchange window. The HAL Manual Toolchange is primarily useful if you have presettable tools and you store the offsets in the tool table. If you need to touch off for each tool change then it is best just to split up your g code. To use the HAL Manual Toolchange window you basically have to load the hal manualtoolchange component then send the iocontrol "tool change" to the hal manualtoolchange "change" and send the hal manualtoolchange "changed" back to the iocontrol "tool changed".

This is an example of with the HAL Manual Toolchange from the Stepconf wizard

loadusr -W hal_manualtoolchange
net tool-change iocontrol.0.tool-change => hal_manualtoolchange.change
net tool-changed iocontrol.0.tool-changed <= hal_manualtoolchange.changed
net tool-number iocontrol.0.tool-prep-number => hal_manualtoolchange.number
net tool-prepare-loopback iocontrol.0.tool-prepare => iocontrol.0.tool-prepared

This is an example of without the HAL Manual Toolchange from the Stepconf wizard

net tool-number <= iocontrol.0.tool-prep-number
net tool-change-loopback iocontrol.0.tool-change => iocontrol.0.tool-changed
net tool-prepare-loopback iocontrol.0.tool-prepare => iocontrol.0.tool-prepared

1.2 Compute Velocity

This example uses "ddt", "mult2" and "abs" to compute the velocity of a single axis. For more information on the real time components see the man pages or the Realtime Components section ([->]).

The first thing is to check your configuration to make sure you are not using any of the real time components all ready. You can do this by opening up the HAL Configuration window and look for the components in the pin section. If you are then find the .hal file that they are being loaded in and increase the counts and adjust the instance to the correct value. Add the following to your custom.hal file.

Load the realtime components.

loadrt ddt count=1
loadrt mult2 count =1
loadrt abs count=1

Add the functions to a thread so it will get updated.

addf ddt.0 servo-thread
addf mult2.0 servo-thread
addf abs.0 servo-thread

Make the connections.

setp mult2.in1 60
net xpos-cmd ddt.0.in
net X-IPS mult2.0.in0 <= ddt.0.out
net X-ABS abs.0.in <= mult2.0.out
net X-IPM abs.0.out

In this last section we are setting the mult2.0.in1 to 60 to convert the inch per second to inch per minute that we get from the ddt.0.out.

The xpos-cmd sends the commanded position to the ddt.0.in. The ddt computes the derivative of the change of the input.

The ddt2.0.out is multiplied by 60 to give IPM.

The mult2.0.out is sent to the abs to get the absolute value.

The following figure shows the result when the X axis is moving at 15 IPM in the minus direction. Notice that we can get the absolute value from either the abs.0.out pin or the X-IPM signal.

Figure: Velocity Example

1.3 Soft Start

This example shows how the HAL components "lowpass", "limit2" or "limit3" can be used to limit how fast a signal changes.

In this example we have a servo motor driving a lathe spindle. If we just used the commanded spindle speeds on the servo it will try and go from present speed to commanded speed as fast as it can. This could cause a problem or damage the drive. To slow the rate of change we can send the motion.spindle-speed-out through a limiter before the PID, so that the PID command value varies slowly.

Three built-in components that limit a signal are:

limit2
limits the range and first derivative of a signal.
limit3
limits the range, first and second derivatives of a signal.
lowpass
uses an exponentially-weighted moving average to track an input signal.

To find more information on these HAL components check the man pages.

Place the following in a text file called softstart.hal. If your not familiar with Linux place the file in your home directory.

##############################
loadrt threads period1=1000000 name1=thread
loadrt siggen
loadrt lowpass
loadrt limit2
loadrt limit3
net square siggen.0.square => lowpass.0.in limit2.0.in limit3.0.in
net lowpass <= lowpass.0.out
net limit2 <= limit2.0.out
net limit3 <= limit3.0.out
setp siggen.0.frequency .1
setp lowpass.0.gain .01
setp limit2.0.maxv 2
setp limit3.0.maxv 2
setp limit3.0.maxa 10
addf siggen.0.update thread
addf lowpass.0 thread
addf limit2.0 thread
addf limit3.0 thread
start
loadusr halscope
##############################

Open a terminal window and run the file with the following command.

halrun -I softstart.hal

When the HAL Oscilloscope first starts up click "OK" to accept the default thread.

Next you have to add the signals to the channels. Click on channel 1 then select "square" from the Signals tab. Repeat for channels 2-4 and add lowpass, limit2, and limit3.

Next to set up a trigger signal click on the Source None button and select square. The button will change to Source Chan 1.

Next click on Single in the Run Mode radio buttons box. This will start a run and when it finishes you will see your traces.

To separate the signals so you can see them better click on a channel then use the Pos slider in the Vertical box to set the positions.

Figure: Softstart

To see the effect of changing the set point values of any of the components you can change them in the terminal window. To see what different gain settings do for lowpass just type the following in the terminal window and try different settings.

setp lowpass.0.gain .01

After changing a setting run the oscilloscope again to see the change.

When your finished type "exit" in the terminal window to shut down halrun and close the halscope. Don't close the terminal window with halrun running as it might leave some things in memory that could prevent EMC from loading.

For more information on HalScope see the HAL manual.

1.4 Debounce

The debounce component is used to filter a noisy digital input.

The debounce filter works by incrementing a counter whenever the input is true, and decrementing the counter when it is false. If the counter decrements to zero, the output is set false and the counter ignores further decrements. If the counter increments up to a threshold, the output is set true and the counter ignores further increments. If the counter is between zero and the threshold, the output retains its previous state. The threshold determines the amount of filtering: a threshold of 1 does no filtering at all, and a threshold of N requires a signal to be present for N samples before the output changes state.

The debounce components are created in groups with the size of the group determined by the configuration when loaded. This allows the integrator to set the parameters of a group with one line in the HAL file and to add the group to a thread with one line. In the following example 2 groups are created with 2 filters to show how the addressing works. If your using a Stepconf Wizard generated configuration put the following in your custom.hal file.

# load the debounce component
loadrt debounce cfg=2,2
# add each debounce group to different threads just for an example
addf debounce.0 servo-thread
addf debounce.1 base-thread
# set the debounce delay for group 0 to two cycles of the servo-thread
setp debounce.0.delay 2
# set the debounce delay for group 1 to three cycles of the base-thread
setp debounce.1.delay 3

Now when you run EMC and look in the HAL Configuration window you will see the two debounce groups.

Figure: Debounce Groups


Now lets say we want to debounce an encoder signal from phase Z that is coming into EMC via the parallel port pin 13. First we have to remove the link to the pins then run them through the debounce then out to the component. If your using a Stepconf Wizard generated configuration put the following in your custom.hal file.

# unlink the encoder and parallel port pins
unlinkp encoder.0.phase-Z
unlinkp parport.0.pin-13-in
# create a new signal and link the parallel port to the debounce component
net filter-p-13 <= parport.0.pin-13-in => debounce.1.0.in
# create a new signal and send the filtered output to the encoder input
net filtered-p-13 <= debounce.1.0.out => encoder.0.phase-Z

Now if you have more pins to filter it is lather rinse repeat. If your encoder is on the base-thread then you want to put the debounce on the same thread usually so the timings are the same. You can look in the main hal file to see where a component is added to a function with addf.