Category Archives: HowTo’s

#VSCP interfaces howto

Something that may confuse new users of VSCP is the GUID of interfaces. Looking at the interface above the CAN4VSCP driver have GUID set to


which means nodes connected to this interface will come in with GUID’s


If we look at the GUID’s if this machine we see that they all start with


As of  the spec we know that this is a GUID constructed from an Ethernet address. In this case only four of the six MAC digits is used. The


is the interface id as set by the VSCP daemon and the last two digits are the nickname id for a connected node.

The problem that can occur here is that the interface digits can be different at different runs of the VSCP daemon. The number is just set when the interface is set up and from time to time this setup can happen in different order. So sometime


can be


or even


One can’t tell beforehand.

This is a problem if one want to use the GUID to identify a node. Problematic as this is just what we want in most cases. To trigger on an event from a specific node in a decision matrix the GUID is the item to filter on. Just as in this case

Here we store a temperature measurement in a variable if it comes from a node with GUID


and measurement index = 1

If the interface ordinal  becomes something else like


we are in trouble here as the action (store measurement value in variable) will not be triggered.

If we look at the current driver for this setup it is set to


<!-- The can4vscp driver -->
<driver enable="false" >

The “problem” is solved by changing the <guid> tag from an all zero value (or not defined) to a valid GUID.  If not defined or all zero the VSCP daemon will assign a GUID as of above. But if a valid GUID is set this value will always be used.

In my case I can use


as I have some assigned GUIDs.  The two LSB’s is still used for the nicknames.

After that change the “problem” is no more and I can filter on the new GUID instead which alway will be the same.



#ESP8266 #VSCP Development Intro – part 1

Serial port

Connect the USB cable to the board.  You will get a new CDC serial port. You can check which one you get with

ls /dev/tty*

Normally you get /dev/ttyUSB0 or /dev/ttyUSB1 or the like. To make them usable for you you should add yourself to the dialout group

addgroup user dialout

where “user” is your username. You can also do this manually in the /etc/group file or

chmod a+rw /dev/ttyUSB0

if you like brute force.

Loading code to the ESP8266

The ESP8266 CPU can be booted in three different ways:

  • Flash Mode: default booting mode. Firmware is read and executed from the flash memory. Pins need to be set to: GPIO15=0, GPIO0=0, GPIO2=1
  • UART Mode: used to program our board (through a Serial-to-Usb adapter). Pins need to be set to:  GPIO15=0, GPIO0=1, GPIO2=1
  • SDIO Mode: loads firmware from an SDIO card? Pins need to be set to:  GPIO15=1, GPIO0=0/1, GPIO2=0/1

The boot process is described here.

Switching from a mode to another requires to reset the module while the pin states listed is set.

With the nodemcu board and esptool flash uploader all this is handle automatically but if you have another board you must handle this yourself.


You find it here.

Fetch it

sudo git clone

cd esp-open-sdk

Install it

apt-get install make unrar autoconf automake libtool libtool-bin gcc g++ gperf flex bison texinfo gawk ncurses-dev libexpat-dev python sed

git clone --recursive

cd esp-open-sdk/

Error when building?  See this post
This builds the standalone version of the SDK (Non FreeRTOS)
You should add the bin directory to he path
echo 'PATH=$PATH:~/development/esp8266/esp-open-sdk/xtensa-lx106-elf/bin' >> ~/.profile

echo 'PATH=$PATH:~/development/esp8266/esp-open-sdk/esptool' >> ~/.profile


I have installed the toolchain in


so change paths above for your installation folders.

Update it (when updates are available)

make clean
git pull
git submodule update

Firmware upload tool

You need the to upliad firmware to the module, you find it here
Install with
pip install esptool

First code (blinky)

You find it here

cd ~/development/esp8266/esp-open-sdk/examples/blinky



to build it. Remember that the paths above must have been set. You may get some warnings. No problem.

For your information. If you build from source code should be loaded like this:

  • bin/0x00000.bin to 0x00000
  • bin/0x10000.bin to 0x10000

That is just what we do here. Upload to your module with --port /dev/ttyUSB1 write_flash 0x00000 blinky-0x00000.bin 0x10000 blinky-0x10000.bin

No need to press any buttons during flashing. It is handled automatically by the USB DTR circuitry.

Now we are ready to do some real work for our VSCP system with the ESP8266.  If you rather prefer Arduino this howto may be the one you should go for instead of this one. Life is much simpler in the Arduino world. But the degrees of freedom is better if you do it all by yourself. The penalty for freedom is more problems of course. As always.

Another intro is here
esp8266 wiki is here.

Part 2 will follow.

Howto: Minimum VSCP wifi Level II node (VSCP PID detector)

In this  howto we will build a simple VSCP level II node that with the help of a PIR motion sensor detects motion and sens this as a VSCP event wireless to a VSCP daemon.

We use the famous ESP8266  for this  and work in the even more famous Arduino environment.  The ESP8266 comes in many flavors but a convenient form factor is the NodeMCU board.  You can buy it on Aliexpress or Ebay at a low-cost (USD 2.2).

There is also a baseboard available for the NodeMCU and I use it as it is convenient when working with a new design. It to is available on Aliexpress and Ebay at a low-cost (USD 1.54).

The last item you need is the PIR sensor, they to are available at low-cost. I bought mine from Ebay (USD 0.99).

If you haven’t set up your Arduino IDE for work with the ESP8266 you need to do that now.  Here is an excellent tutorial on how to do this. Make the blink example work before you move on.

I find the Arduino IDE a bit crude to work with so I use Visual Studio Code instead.  An excellent programming editor that works on all the major operation systems. I tend to use it more and more. But there are other options to such as Atom. Anyway instructions on how to get it working for Arduino development is here.  Make sure to set the “output” tag to a valid location in the .vscode/arduino.json file otherwise upload will be VERY slow.

You also need a VSCP daemon/server running for this example. All info about it is here.

To connect the sensor to the NodeMCU we ned to know about he port mapping for the module which is a bit different from for other Arduino devices. It looks like this

static const uint8_t D0   = 16;
static const uint8_t D1   = 5;
static const uint8_t D2   = 4;
static const uint8_t D3   = 0;
static const uint8_t D4   = 2;
static const uint8_t D5   = 14;
static const uint8_t D6   = 12;
static const uint8_t D7   = 13;
static const uint8_t D8   = 15;
static const uint8_t D9   = 3;
static const uint8_t D10  = 1;

D0 has a blue LED connected to it so it is an obvious candidate for status. D1 can be used to connect to the PIR

The PIR device I have outputs a 3V signal even when powered with +5V so it is OK to connect directly to D1. So the three pins on the PIR

1     +5V
2     Output to D1 (High on detect)
3     GND

The LED on D0 light up when a LOW is written to it. This is opposite to the PIR output. We take care of this in software.

We implement a dump VSCP device here. This is a device that does not handle any of the register reads that “normal” VSCP requires and which don’t have a MDF file either. A dumb device has been 14 set in the header.

We send two events. The heartbeat event (CLASS1.INFORMATION, Type=9) every minute. This event is recommended for all nodes as it is used for node discovery and detection. The other event we send is the detect event (CLASS1.INFORMATION, Type=49 detect) when an object is detected.

The test setup looks like this

and you can find the complete code is here

The heartbeat coming into the VSCP daemon (using VSCP works)

and the detect event

With the event in the VSCP daemon it is easy to add a DM row that for instance light up the lamps in a room. You can also use the rest or the websocket interface to do give visual feedback.

We will follow-up this howto with a post where we use the Expressif SDK instead of the Arduino and use a plain ESP8266 board. But also a ost where we implement a full Level II node that have registers and a MDF and show the advantage we get with a node like that over a dumb one.

An alternative wifi lib is documented here.


#VSCP update process #howto

In previous versions of VSCP whenever you did an install all configuration files were replaced with the latest version. This is not true anymore.  Now the new version is instead written as a copy with the date of the install appended to it.  So if you after a “make install” or a “dpkg -i vscpd” want the latest configs you have to copy the backup to the actual config file and restart the VSCP server.

Files that is handled in this way is


Also if you are on unstable code you should remove the databases before you start the updated server . Use

rm /srv/vscp/*.sqlite3

for this.

Also note that the web sample code is not installed in the install/update process no more. The process to get this subsystem installed with be described later.

#VSCP HOWTO: DM Using timers

Timers. One can wonder what they do in the decision matrix? They even have their own events defined in the CLASS2.VSCP class.

First let us define what a VSCP timer is.

A VSCP timer is a free running 32-bit timer with millisecond resolution.  That is they can hold  0xffffffff = 4294967295 milliseconds which mens they will roll over in about 50 days.

There are actions defined to

There is the following  internal events defined related to timers

To create and start a timer

<row enable="true" groupid="timers" >

    Create timer

  <mask priority="0"
 GUID="00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00" />

  <filter priority="0"
 GUID="00:01:02:03:04:05:06:07:08:09:0A:0B:0C:0D:0E:0F" />



Here a timer with id = 1 is created. The timer has an initial time set to 10 seconds. When this timer elapses it will set the variable //timer1flag// to true. The last argument is the reload flag. Here it is set to true so when the time has elapsed the initial value will be loaded again and the timer will start again.

In the example above we could have added “;4” at the end of the parameter which would have the effect that the reload would stop after four runs. Default is thus forever.

When the timer is started a CLASS2.VSCPD, Type = 25 (0x0019) Timer started event is generated. When the ten seconds has gone and the timer elapses the  CLASS2.VSCPD, Type = 29 (0x001D) Timer Elapsed event is generated.

We can use either one of these two generated events to do any action periodically like this.

<row enable="true" groupid="timers" >

    Handle timer elapsed

  <mask priority="0"
 GUID="00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00" />

  <filter priority="0"
 GUID="00:01:02:03:04:05:06:07:08:09:0A:0B:0C:0D:0E:0F" />

    /srv/vscp/timefile;1;%isoboth: Timer with[0] elapsed %lf


<row enable="true" groupid="timers" > 

   <comment> Handle timer elapsed </comment> 
   <mask priority="0" 
         type="0xFFFF" GUID="00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00" /> 
    <filter priority="0" 
            type="25" GUID="00:01:02:03:04:05:06:07:08:09:0A:0B:0C:0D:0E:0F" /> 

      /srv/vscp/timefile;1;%isoboth: Timer with[0] elapsed[0]ms %lf </param> </row>

Here some info is just written to a file when the timer is started and when it elapses.   But if you want to send an event periodically instead or do other actions this is the way to do it.

You don’t have to give a variable nor a reload flag when you start a timer. If no variable is given (use ;;) it is just ignored. The reload value will be set to false as default  value, that is the timer will run only once. As the last parameter you can set the number of times the timer should reload before it should stop.  The full documentation is here and here.

One useful use of timers is to handle resend of events. The working is like this

Send the event you expect a reply event from another node.

  1. Create/start a timer that have  period equal to resend intervals for the event you want to send. The timer should  have true for autoloading and the number of autoloads set to the number of resend that are allowed.  You can trigger the creation of the timer  by make a DM entry that triggering on one of the reserved events for example,  CLASS1.LOCAL  or CLASS1.LABORATORY so that when you send this local event  the actual event will be sent (see next point)
  2. In the timer started event send  the event you want to send.
  3. When/If the expected reply  is received pause the timer.
  4. Now if the timer the timer stopped event is detected the reply wait timeout has expired so do timeout handling there by sending another local event or CLASS1.ERROR, Type = 32 Time out.

Can be used for much more of course. Just useful and simple.

#VSCP HOWTO: DM write/append to file

The Level I logger and the Level II logger are great tools for logging events in a VSCP based system.  Useful also for debugging etc. Another method is to use the execute external program action and execute a script and write to a file there. We have seen this method being used in other howtos.  Also while running a JavaScript or a Lua script files can be written.


file writing is also available as an action.  It is documented here. This action allows writing or appending a string with output to a named file that is created if it does not exist. With the VSCP escapes a lot of dynamic information can go into this file.

Suppose that we want to  log events of a certain type to a file, here CLASS1.DATA, Type=2 A/D values form a specific device. We then write a decision matrix (DM) row

<row enable="true" groupid="" >

 Collect A/D values from node X

<mask priority="0"

<filter priority="0"
 GUID="00:01:02:03:04:05:06:07:08:09:0A:0B:0C:0D:0E:0F" />

 /tmp/addata;1;%isoboth: %measurement.string %lf 


will generate output content like this

With the other VSCP escapes and literals you have many options to generate meaningful output.