I have recieved my Railstars RAILbooster in the post today.
It works - currently outputs 14.3VDC from supplied 15VDC, I used a bridge rectifier to enable multimeter measurement of the track voltage. (sourced the idea from dccwiki article on How to convert dcc to dc)
Microduino sending dcc signal to booster which is being isolated in order for the arduino to monitor the dcc packet. Can't have 14.3 VDC going straight to the Arduino that would kill the arduino.
DCC monitor in action - all those numbers (except for the idle packet - 11111111 00000000 11111111) are due to my major speed change.
Wednesday, 16 July 2014
Sunday, 13 July 2014
DCC Controller - revisted
Before I could test out the logic analyser I had to make use of a Microduino-Core to generate the DCC signal, as I only have one Arduino Uno board. The microduino is arduino compatable and uses the ATmega328P same as the UNO.
New Fritzing images using the Microduino, fingers crossed that the CmdrArduino files will work with it. (Updated as incorrectly connected dcc out to ground when it should goto pin 10).
I need to do more testing of the dcc signal to see if CmdrArduino is running correctly on the microduino.
There is limited sampling due to the size of memory on the Uno board, but it is useable, I just need to adjust my triggers.
What we are seeing is the idle packet for all decoders which is in the format of:
11111111111111 0 11111111 0 00000000 0 11111111 1
As shown on line 90 of the NMRA Specification S-9.2 DCC Communication Standard (pdf document).
I have put together the dcc monitor and ran it testing the Microduino running CmdrArduino, but with no result, including no idle packets. Looks like dcc monitor cannot handle positive unipolar signal, which is okay as the dcc monitor was designed for trackside which is a bipolar signal. Just means I need to wait until I have my booster before I can test it out. Correction: CmdrArduino has a differential signal, the example sketch just mentions pin 9, the two pins required are 9 & 10 on Arduino Uno (and microduino).
Update: dcc monitor will work without booster now that I am using pin 10, instead of ground.
Between marker 1 (1.45 ms) and 9 (7.23 ms) is the whole dcc idle packet, with both dcc signals shown - pin 9 in purple and pin 10 in blue.
New Fritzing images using the Microduino, fingers crossed that the CmdrArduino files will work with it. (Updated as incorrectly connected dcc out to ground when it should goto pin 10).
I need to do more testing of the dcc signal to see if CmdrArduino is running correctly on the microduino.
Logic Analyser
The logic analyser works, it supports 4MHz & 2MHz sampling rate. I tried it at 100 kHz sampling rate with the microduino running CmdrArduino, shown below.There is limited sampling due to the size of memory on the Uno board, but it is useable, I just need to adjust my triggers.
What we are seeing is the idle packet for all decoders which is in the format of:
11111111111111 0 11111111 0 00000000 0 11111111 1
As shown on line 90 of the NMRA Specification S-9.2 DCC Communication Standard (pdf document).
DCC Monitor
I have put together the dcc monitor and ran it testing the Microduino running CmdrArduino, but with no result, including no idle packets. Looks like dcc monitor cannot handle positive unipolar signal, which is okay as the dcc monitor was designed for trackside which is a bipolar signal. Just means I need to wait until I have my booster before I can test it out. Correction: CmdrArduino has a differential signal, the example sketch just mentions pin 9, the two pins required are 9 & 10 on Arduino Uno (and microduino).
Update: dcc monitor will work without booster now that I am using pin 10, instead of ground.
Between marker 1 (1.45 ms) and 9 (7.23 ms) is the whole dcc idle packet, with both dcc signals shown - pin 9 in purple and pin 10 in blue.
Subscribe to:
Posts (Atom)