Update on 2024-Apr-13 -- Removed Gitlab references

Github makes it easier for me to provide binary package releases along with my source code.

Update on 2024-Mar-23 -- Moving to Github

Development has moved to Github.

The first thing I realized when I started to use tcladu was the need for convenience functions. For example, combinations of write_device and read_device should be combined into a query command for convenience. And it would be nice for those kind of commands to be in the same tcladu package. So I started to write some Tcl, and I realized I didn't really know how to manage the package version with this high-level addition.

The tcladu.so binary already provides the tcladu namespace and package version. I set this version in a makefile that also drives testing, so I'd really like the makefile/binary package version to be the source of truth. So I need to extract that package version in the convenience code. It turns out that there's a package command for this. These links are to Tcl 9.0 documentation, but my usage doesn't care.

Adding these lines

load ./tcladu.so
set version [package present tcladu]
package provide tcladu $version

...to the tcladu.tcl source file synchronizes the Tcl and binary source versions. I can then package this up with

pkg_mkIndex -verbose . tcladu.so tcladu.tcl

...using the pkg_mkIndex command. This produces the pkgIndex.tcl script distributed with the package source files:

package ifneeded tcladu 1.1.0 [list load [file join $dir tcladu.so]]\n[list source [file join $dir tcladu.tcl]]

...where you can see the combination of sourcing and loading done to make the package available. These convenience functions are coming in tcladu version 1.1.0.

Image by Freepik

Links

• A nice description of creating a Tcl package
• Turning a SWIG module into a Tcl package

Tcladu is a Tcl package providing an interface to the ADU100 from Ontrak Control Systems.

One of my projects will be powered by a battery, and will sleep for long periods while it waits for user input. My battery life will depend on this sleep current, which might be about 10 µA. My handheld meter will measure this, but it suffers from mediocre common-mode rejection when I use it to measure high-side current. I really wanted to probe my sleep current and send it directly to an oscilloscope. These INA169 breakout boards seem like a convenient way to make this probe, but using them raises calibration, noise, and bandwidth questions. I want to describe how I put down a simple calibration circuit to give myself some confidence in this measurement.

Wiring the sensor

Figure 1 shows the current sensor, which is built around TI’s IN169 current sensor IC. Adafruit sells the breakout board I used, which builds in the IC, the current sensing resistor R_shunt, the current-to-voltage resistor R_load, and a bypass capacitor. It comes configured with a convenient gain of 1 V/A, and sells for about 10 USD. Sparkfun and others make similar breakout boards. These boards can be modified for different “burden voltages” and power configurations.

Xfig's depth settings can create some headaches when you use figure libraries. Library figures can use depths that conflict with your current drawing. Say I have this drawing of the Teensy Audio Board and a red circle,

...and then I move the red circle over the board:

This is clearly not what I want. The circle has a depth of 50, and the board has depths between 10 and 60. The circle ends up on top of some objects and not others. If you were using PowerPoint, you would just select Bring to Front or Send to Back to arrange the objects. But Xfig requires setting the depth of each member object individually, which you should really do with a script like figdepth.

Using figdepth

Figdepth has some options, the most important of which is the minimum depth argument

...which sets the depth of the topmost object in a fig file. The script simply increments all depths by some number to make this minimum depth to be your setting. You will, of course, need to have everything you want to modify in its own fig file. You can then run figdepth like

...to write your new depth-adjusted figure. Use Xfig's merge command to bring the new figure into your old drawing. My new drawing looks like this:

...with all objects in the Teensy Audio Board at a deeper depth than the red circle.

How does it work?

The Fig file format is really simple, which is one reason why I like Xfig so much. You can figure out what kind of object each line describes by looking at the first character. You then figure out which number in the line is the object's depth and increment it. The switch statement below shows how one line is processed.

proc increase_object_depth { xfig_line depth_increase} {
    # Return a new xfig line with the depth increased
    #
    # Arguments:
    #   xfig_line -- Single line from an xfig file
    #   depth_increase -- Amount to increase the depth
    set entry_list [split $xfig_line]
    set first_character [lindex $entry_list 0]
    switch $first_character {
    "0" {
        # Color object -- nothing to do
        return $xfig_line
    }
    "1" {
        # Ellipse
        set depth_index 6
        set old_depth [get_object_depth $xfig_line]
        set new_depth [expr $old_depth + $depth_increase]
        set new_entry_list [lset entry_list $depth_index $new_depth]
        set new_line [join $new_entry_list]
        return $new_line
    }
    "2" {
        # Polyline (also imported picture bounding boxes)
        set depth_index 6
        set old_depth [get_object_depth $xfig_line]
        set new_depth [expr $old_depth + $depth_increase]
        set new_entry_list [lset entry_list $depth_index $new_depth]
        set new_line [join $new_entry_list]
        return $new_line
    }
    "3" {
        # Spline
        set depth_index 6
        set old_depth [get_object_depth $xfig_line]
        set new_depth [expr $old_depth + $depth_increase]
        set new_entry_list [lset entry_list $depth_index $new_depth]
        set new_line [join $new_entry_list]
        return $new_line
    }
    "4" {
        # Text
        set depth_index 3
        set old_depth [get_object_depth $xfig_line]
        set new_depth [expr $old_depth + $depth_increase]
        set new_entry_list [lset entry_list $depth_index $new_depth]
        set new_line [join $new_entry_list]
        return $new_line
    }
    "5" {
        # Arc
        set depth_index 6
        set old_depth [get_object_depth $xfig_line]
        set new_depth [expr $old_depth + $depth_increase]
        set new_entry_list [lset entry_list $depth_index $new_depth]
        set new_line [join $new_entry_list]
        return $new_line
    }
    "6" {
        # Compound -- nothing to do
        return $xfig_line
    }
    default {
        # Some non-object -- nothing to do
        return $xfig_line
    }
  }
}

Having trouble with cmdline?

Figdepth uses the cmdline package from tcllib to handle command line arguments. I usually run scripts out of Eshell in Emacs, so I have this in my .emacs:

...to set the TCLLIBPATH environment variable. This tells Tcl where to find Tcllib's packages.

Where is the figdepth script?

Figdepth is part of my xfigart repository.

I still love Xfig when I need to make a sketch of something with roughly-correct diminsions. I've been neglecting my Xfigart repository, but I'm going to try to be better. I'll start with this drawing of the CES-571423-28PM speaker from CUI.

The siunitx package for LaTeX makes it easy to handle units in both normal and math modes. I end up needing units like V_rms when talking about electronics, so I add this in my preamble:

\usepackage{siunitx}

\DeclareSIUnit \vrms {\ensuremath{\mathrm{V_{rms}}}}
\DeclareSIUnit \vpp {\ensuremath{\mathrm{V_{pp}}}}

...to define these units to siunitx. I can then invoke the units with something like

free air.  I drove them to about \SI{9}{\vpp}.

where I know siunitx will then make the proper space between the number and the unit. This all renders like this:

...in the PDF output.

I've been wanting to experiment with audio signal processing on the Teensy, so I bought a Teensy 4.0 and an Audio Adaptor Board. I had to do a bit of research to find the right headers to plug one board into the other. I ended up with the headers listed below.

I used these headers in the photo shown below.

I've been interested in fieldbusses lately, and MODBUS in particular is used by some pretty cheap hardware modules. It also offers "driverless" communication — you don't need a .dll or .so file to talk to your hardware. The hardware I've been experiementing with is shown below:

...and can be bought from eBay for around 50 USD. I found some very useful code on the Tcl wiki and was able to write and read from registers with a USB/RS-485 adaptor. The RTU-307C user's manual implies that you can change the RS-485 baud from 9600, but I wasn't able to figure out how that works.

After some more help from The Tcler's Wiki, I put together a GUI to demonstrate MODBUS communication with the RTU-307C module. The animation below shows the GUI next to an oscilloscope (yelloscope) display. The wiki helped me get the mouse wheel to move those sliders.

Links

Source: Repository on Gitlab

The ATmega328P is used on the Arduino UNO, the Nano, and on many other mass-produced microcontroller boards. But it has only 2048 bytes of SRAM, which is shared between the stack and storage for things like local variables. Bizarre behavior can happen if the SRAM needed for variable storage overlaps that needed for the stack, but it's hard to understand how much space these variables will use until the program is actually running. This is where Michael McTernan's stack monitor is useful.

The stack monitor "paints" SRAM storage registers with a magic value before application code executes. You can then check for unused stack memory during execution by counting up these magic values. Here's some debug output from my Diamond firmware running on the EVL-510:

The unused SRAM is initially 200 bytes, but falls to 179 after turning the volume knob. Changing the volume calls a few functions that allocate local variables in SRAM. These get cleaned up when the function exits, but the stack monitor can see that the memory was used. The idea is to exercise as much of the system as you can and then make sure you still have some unused memory.

I call the function with a scheduler made from avr-simple-scheduler. The code block below shows adding the stack reporting function to the scheduler with a period of 5s. I could have made this even longer, since I only need the report to run after I've excercised as much as I can.

if (strcmp( LOG_LEVEL, "debug" ) == 0) {
    // Start the task to monitor free memory
    OS_TaskCreate(&system_report_stack_free, 5000, BLOCKED);
}

This desktop speaker turned out well, and I asked Jack at Prime Images Photography to take some real product photos. There's lots of interesting stuff here, including the volume knob's state machine and some handy prototyping tools from Mikroe. I'll get into these in some later posts.