Project Logs

Collapse

• Everything now runs on VideNet protocol

  h4rdc0der • 07/30/2021 at 05:47 • 0 comments

  Great success!

  Few days ago I've finished CAN component that is able to handle VideNet protocol. Adding new request takes only few lines of code.

  Currently supported features are:

  • Ability to turn on / turn off controller.
  • Ability to read / set hot water mode.
  • Ability to read / set central heating mode.
  • Ability to read kettle temperature.
  • Ability to read hot water temperature.
  • Ability to read burner power (%).
  • Ability to read fuel usage sum (kg).

  Special thanks going to:

  • Kem Klattrup 
  • Digiplex from elektroda.pl forum

• Introducing remote control

  h4rdc0der • 07/09/2021 at 07:40 • 0 comments

  I've found Vide-Net protocol specification (VideNet is an original RC module). With that knowledge remote control is now possible!

• Fixed and working!

  h4rdc0der • 03/06/2021 at 10:36 • 0 comments

  After removing the terminating resistor, the problems with the boiler disappeared. The device has been working continuously for several days without any problems.
  
  Final mounting - transparent enclosure, SMA antenna:

  
  

• Bad things happened

  h4rdc0der • 03/03/2021 at 21:31 • 0 comments

  It looks like with 120 ohm resistor, I have introduced excessive termination that led to errors in CAN transmissions. I had to unsolder termination resistor from my board and trim my cables to work properly as CAN stub.

  Enclosure arrived today, so I have mounted everything inside.

  It requires more testing this time, but I hope that issue is fixed.
  

• Testing

  h4rdc0der • 02/28/2021 at 20:44 • 0 comments

  Today I prepared the cables and prototype software to retrieve data, then connected everything to the heating boiler CAN-bus and communication with MQTT worked properly.

  Example dasbhoard:

  Testing connection and current consumption.

  
  What's next?

  I need to route everything properly, get an enclosure and add fuel level sensor. It will take few days as I have to order cable organizers etc.

• Tests done, everything works!

  h4rdc0der • 02/26/2021 at 12:56 • 0 comments

  Before connecting my device to the boiler, I made sure that all interfaces were working. 

  There was a problem with the R1 resistor, which made the CAN bus not working, but it turned out that it was enough to correct the weld. There is working two-way communication between the two PCBs. The fuel level sensor leads have also been tested and everything is fine.

  Now all that's left is programming and installation on the boiler :)
  

• PCBs assembled

  h4rdc0der • 02/24/2021 at 20:02 • 0 comments

  Unfortunately, I had to wait over a month for the package, but it eventually arrived. Soldering was quick, without hot-air gun.
  
  Anyway, a lot of optimizations have been introduced into my iot framework - ksIotFrameworkLib.
  
  I hope to have some time to launch this board at the destination place in my basement, but it will not happen this week :)

• PCBs in production!

  h4rdc0der • 01/03/2021 at 18:33 • 0 comments

  Since all experiments were successful, it's time to order a PCB. 
  All parts were ordered together with the board.

  As usual, I used the JLCPCB service, combining the order with LCSC. Unfortunately, according to the announcement, the option to combine orders will be withdrawn soon.

• Experiment 3 - Better approximation

  h4rdc0der • 01/02/2021 at 19:44 • 0 comments

  I decided to improve the temperature conversion. Here's where the curve fitting software come in.
  One of them is Curve Expert - can be bought at https://www.curveexpert.net/order, but trial is also available.

  After a few tweaks, the polynomial curve fitting method turned out to be the best (5th degree).

  double calctemp(short x)
  {
      double a = 1.445633283634090E+02;
      double b = -1.386947619796149E-01;
      double c = 9.837913934334235E-05;
      double d = -4.486094388070098E-08;
      double e = 1.076415627857750E-11;
      double f = -1.066640689001453E-15;
  
      return a + b * x + c * pow(x, 2) + d * pow(x, 3) + e * pow(x, 4) + f * pow(x, 5);
  }

  It turned out that all temperatures are transferred in the same way, so this function can be used to calculate different temperatures - e.g. domestic hot water.

• Experiment 2 - Hacking all I CAN

  h4rdc0der • 01/01/2021 at 21:20 • 0 comments

  I previously established that my boiler uses the CAN bus to transfer data between the controller and the touch panel. The controller is somewhere inside and the touchpad is on the front as you can see ...

  The CAN port is on the back - it looks like an ARK connector. There are also CANL, CANH, GND and optional 12V, so the external module (Estyma VideNET) does not require a separate power supply.

  So, let's start hacking...

  I quickly realized that I did not have the right tools to analyze CAN traffic, since I had only ESP32 and CAN transceiver module based on SN65HVD230 chip.

  I tried several internet solutions to send data to Wireshark, but the messages were still difficult to analyze.

  I found a program on the Internet that I can use to suit my needs - CAN-Monitor-3000 github repository. I ended up writing a program in Arduino to emulate one of it's supported adapters.

  Visit my Canmon-Sniffer repository!

  Troubles with temperature...

  After searching the data stream, I found no interesting temperature value.
  Then noticed message 0x2D6... it changes when temperature changes!

  I figured out that I could send the crafted data to the bus. I had to add support for sending data to the CAN bus to the sniffer. 

  I was in trouble when I sent the wrong frame. The boiler reported a sensor failure and turned off.

  But eventually after few hours of headache, I ended with this small lookup table:

  Approximate for the win!

  I entered the given values on one of the pages for the function approximation. Then I put the given function into the test program.

  CRAZY!

  After uploading the program to ESP32, I got an approximate temperature reading.

  There is still a lot of work ahead of me, but this adventure was really interesting.
  

View all 11 project logs