Tuesday, June 14, 2016

Terry & ROS

After a number of adventures I finally got a ROS stack setup so that move_base, amcl, and my robot base all like each other well enough for navigation to function. Luckily I added some structural support to the physical base as the self driving control is a little snappier than I normally tend to drive the robot by hand.

There was an upgrade from Indigo to Kinetic in the mix and the coupled update to Ubuntu Xenial to match the ROS platform update. I found a bunch of ROS packages that I used are not currently available for Kinetic, so had an expanding catkin ws for self compiled system packages to complete the update. Really cool stuff like rosserial wasn't available. Then I found that a timeout there caused a bunch of error messages about mismatched read sizes. I downgrade to the indigo version of rosserial and the error was still there, so I assume it relates to the various serial drivers in the Linux kernel doing different timing than they did before. Still, one would have hoped that rosserial was more resilient to multiple partial packet delivery. But with a timeout bump all works again. FWIW I've seen similar in boost, you try to read 60 bytes and get 43 then need to get that remaining 17 and stuff the excess in a readback buffer for the next packet read attempt. The boost one hit me going from 6 to 10 channel io to a rc receiver-to-uart arduino I created. The "joy" of low level io.

I found that the issues stopping navigation from working for me out of the box on Indigo were still there in Kinetic.  So I now have a very cool bit of knowledge to tell if somebody has navigation working or is just assuming that what one reads equals what will work out of the box.

Probably the next ROS thing will be trying to get a moveit stack for the mearm. I've got one of these cut and so will soon have it built. It seems like an ideal thing to work on MoveIt for because its a simple low cost arm that anybody can cut out and servo up. I've long wanted a simple tutorial on MoveIt for affordable arms. It might be that I'm the one writing that tutorial rather than just reading it.

Video and other goodness to follow. As usual, persistence it the key^TM.

Wednesday, June 8, 2016

libferris 2.0

A new libferris is coming. For a while I've been chipping away at porting libferris and it's tree over to using boost instead of the loki and sigc++ libraries. This has been a little difficult in that it is a major undertaking and that you need to get it working or things segv in wonderful ways.

Luckily there are tests for things like stldb4 so I could see that things were in decent shape along the way. I have also started to bring back the dejagnu test suite for libferris into the main tree. This has given me some degree of happiness that libferris is working ok with the new boost port.

As part of that I've been working on allowing libferris to store it's settings in a configurable location. It's a chicken and egg problem how to set that configuration, as you need to be able to load a configuration in order to be able to set the setting. At the moment it is using an environment variable. I think I'll expand that to allow a longer list of default locations to be searched. So for example on OSX libferris can check /Applications/libferris.app/whatever as a fallback so you can just install and run the ferris suite without any need to do move setup than a simple drag and drop.

For those interested, this is all pushed up to github so you can grab and use right now. Once I have expanded the test suite more I will likely make an announced 2.0 release with tarballs and possibly deb/rpm/dmg distributions.

New filesystems that I've had planned are for mounting MQTT, ROS, and YAML.

Tuesday, May 10, 2016

Through hole PCB Making -- Same Day

I initially thought that removing the multiple week wait for a board would be the true joy of making PCBs locally. It turns out that quick iteration is the best part. Version 2 and 3 of a board flows quickly and you end up with something unexpected after only a few days of tinkering.

I'm still at the level of making through hole stuff. Hopefully I can refine the process to allow some of the larger SMT stuff too. Throwing some caps, leds, resistors, dc jacks, and regulators on for a first cook round will cut down on the soldering phase.



My hello world PCB was an ESP8266 carrier with an mcp23017 muxer and a bunch of buttons. This is an MQTT emission device which I will be using to assist in the controlling of a 3d printer. While web interfaces are flexible, some tend to put buttons too close and you can fairly easily crash the bed by clicking down instead of up in some cases.






Today's iteration is an esp8266 breadboarder. This allows 3v3 intake, has a TTL serial header (on the left) and a resistor + led combo on pin 14 for blink testing. The button at top right toggles into flashing mode and the bottom of the board breaks out 7 gpios onto the breadboard. The 3v3 and ground also have a header under the hot glue to the power rails on the breadboard. Very handy for testing a breadboard layout before designing the next PCB to have an ESP8266 pressed into it.



The breadboard side needs a little trimming back. Turns out the older breadboard I used to measure was wider than this one :o

Monday, April 18, 2016

Making PCB with a hobby CNC machine

One of the main goals I had in mind when getting a CNC "engraving" machine was to make PCB at home. It's sort of full circle to the '70s I guess. Only instead of using nasty chemicals I just have the engraver scratch off an isolation path between traces. Or so the plan goes.


My "hello world" board is the above controller for a 3d printer. This is a follow up to the similar board I made to help use the CNC itself. For a 3d printer I added buttons to set Z=0.1 height and a higher Z height to aid in homing. The breakout headers on the bottom right are for the ESP8266 daughter board. The middle chip is an MCP32017 gpio extender. I've had good experiences using TWI on the ESP8266 and the MCP overcomes the pin limitations quite nicely. It also gives all the buttons a nice central place to go :)

The 3v3 regulator makes the whole show a plug in the AA pack and go type board. The on/off switch is the physical connection to an external battery.

One step closer to the design in the morning, physically create in the afternoon, and use in the evening goal.

Thursday, January 28, 2016

CNC Control with MQTT

I recently upgraded a 3040 CNC machine by replacing the parallel port driven driver board with a smoothieboard. This runs a 100Mhz Cortex-M mcu and has USB and ethernet interfaces, much more modern. This all lead me to coming up with a new controller to move the cutting head, all without needing to update the controller box or recompile or touch the smoothieboard firmware.



I built a small controller box with 12 buttons on it and shoved an esp8266 into that box with a MCP23017 chip to allow access to 16 gpio over TWI from the esp mcu. The firmware is fairly simple on the esp, it enables the internal pull ups on all gpio pins on the 23017 chip and sends an MQTT message when each button is pressed and released. The time since MCU boot in milliseconds is sent as the MQTT payload. This way, one can work out if this is a short or longer button press and move the cutting head a proportional distance.

The web interface for smoothie provides a pronterface like interface for manipulating where the cutting head is on the board and the height it is at. So lucky that it's open source firmware so I can see the non obfuscated javascript that the web interface uses. Then work out the correct POST method to send gcode commands directly to the smoothieboard on the CNC.

The interesting design here is using software on the server to make the controller box meet the smoothieboard. On the server MQTT messages are turned into POST requests using mqtt-launcher. The massive benefit here is that I can change what each button does on the CNC without needing to reprogram the controller or modify the cnc firmware. Just change the mqtt-launcher config file and all is well. So far MQTT is the best "IoT" tech I've had the privilege to use.



I'll probably build another controller for controlling 3d printers. Although most 3d printers just home each axis there is sometimes some pesky commands that must be run at startup, to help home z-axis for example. Having physical buttons to move the x axis down by 4mm, 1mm and 0.1mm makes it so much less likely to fat finger the web interface and accidentally crash the bed by initiating a larger z-axis movement than one had hoped for.

Thursday, December 17, 2015

17 Segments are the new Red.

While 7 Segment displays are quite common, the slightly more complex cousin the 17 segment display allows you to show the A-Z range from English and also some additional symbols due to the extra segments.

The unfortunate part of the above kit which I got from Akizukidenshi is that the panel behind the 17 segger effectively treats the display as a 7 segger. So you get some large 7 segment digits but can never display an "A" for example. Although this suits the clock that the kit builds just fine, there is no way I could abide wasting such a nice display by not being able to display text. With the esp8266 and other wifethernet solutions around at a low price point it is handy to be able to display the wind speed, target "feels like" temperature etc as well as just the time.

With that in mind I have 3 of these 17 seggers breadboarded with a two transistor highside and custom lowside using an mxp23017 pin muxer driving two 2803 current sinks which are attached using 8 up 330 ohm resistors in IC blocks. This lowside is very useful because with a little care it can be setup on a compact piece of stripboard. All the controlling MCU needs is I2C and it can switch all the cathodes just fine.

While experimenting I found some nice highside driver ICs. I now have custom PCB on their way which can each drive 2 displays and can chain left and right to allow an arbitrary number of displays to be connected into a single line display. More info an photos to follow, I just noticed that I hadn't blogged in a while so thought I'd drop this terse post up knowing that more pictures and video are to come. I also have the digits changing through a fade effect. It is much less distracting to go from time to temp and back if you don't jump the digits in one go.

Monday, October 26, 2015

ESP8266 and a few pins

The new Arduino 1.6.x IDE makes it fairly simple to use the ESP8266 modules. I have been meaning to play around with a some open window detectors for a while now. I notice two dedicated GPIO pins on the ESP8266, which is one more than I really need. So I threw in an led which turns on when the window is open. Nothing like local, direct feedback that the device has detected the state of affairs. The reed switch is attached on an interrupt so as soon as the magnet gets too far away the light shines.


I will probably fold and make the interrupt set a flag so that the main loop can perform an http GET to tell the server as soon as it knows when a state has changed.

Probably the main annoying thing I've still got is that during boot it seems the state of both the gpio pins matters. So if the reed switch is closed when you first spply power then the esp goes into some stall state.

It will be interesting to see how easy OTA firmware updates are for the device.