Hi,
seems to be a bug in the Linux Kernel / DT-overlays.
I have opended an issue. Now we have to wayt until the developer fix it. It isn´t an issue in the OP-App or the board.
I cannot do anything at the moment. Hope the developer fix it fastly... ;-).
Until this happened, do not update the Kernel over 5.4 or if you have done this downgrade to under 5.4... ;-)
https://github.com/raspberrypi/linux/issues/3765
Thomas
I had the same issue a few weeks back, solved by reinstall of MCS app.
It broke when I updated Moitessier HAT and crashed a lot of i2c stuff...
Thomas;
Would you expect the Seatalk1 input to work better with a PullUp connected to it?
I'm not sure what's on the MCS board so would it make a difference you beleive?
Also since you completed the magic with ST1 input, are you considering making an output as well?
Hi flybob,
this issue cant be solve by reinstall the MCS App. Its an linux issue in the raspberry firmware.
I have testet now many st1 devices without any issues. You find the wiring in the st1 git folder. You can try to use a 10KOhm Resistor in series to the MCS Input.
At the Moment i have not planed an output because there is no software site availible at the Moment. The Hardware is easy, but you need a script witch translate all the sentences....
(2020-08-06, 08:11 AM)glowntiger Wrote: [ -> ]Hi flybob,
this issue cant be solve by reinstall the MCS App. Its an linux issue in the raspberry firmware.
I have testet now many st1 devices without any issues. You find the wiring in the st1 git folder. You can try to use a 10KOhm Resistor in series to the MCS Input.
At the Moment i have not planed an output because there is no software site availible at the Moment. The Hardware is easy, but you need a script witch translate all the sentences....
I'm thinking about $STALK sentences only, into ST1. So not creating all the sentences but enable other scripts to create $STALK that are created into ST1.
That way it would be possible for many available scripts to communicate.
Example:
https://opencpn.org/wiki/dokuwiki/doku.p...ilot_rm_pi
Hi flybob,
at the Moment i have several ideas to integrate this.
1. Also a bitbanging output on the Pi (pigpio did not support this at the Moment)
2. A microcontroller as an add on for the MCS which directly translate st1 <=> nmea0183
3. The Sc16is752 is able to work in 9bit mode. But the actualy dt-overlay for the pi did not support this. There are changes on the driver nedable.
At the Moment I prefer 3. But I have to less time to work on this because its much software work. The hardwere for testing I have here.. ;-)
https://www.nxp.com/docs/en/data-sheet/S...50_760.pdf
Ok,
I'm looking at Marco Bergmans code and wondering if it can help, as he manages sending (but not generic $STALK though), using only regular GPIO.
https://github.com/marcobergman/
Code:
def write_seatalk (xx, yy):
# thanks to http://www.thomasknauf.de/seatalk.htm
with serial.Serial() as ser:
ser.baudrate = 4800
ser.port = '/dev/serial0'
ser.stopbits=serial.STOPBITS_ONE
ser.bytesize=serial.EIGHTBITS
ser.open()
ser.parity = serial.PARITY_MARK
ser.write(b'\x86')
ser.parity = serial.PARITY_SPACE
ser.write(b'\x11' + chr(int(xx, 16)) + chr(int(yy, 16)))
ser.close()
There is a HW solution
http://www.thomasknauf.de/rap/seatalk3.htm#Bid which I think will work connected direct to a GPIO out pin, but the program must handle the inverted output.
Another way is to put a transistor driving an optocoupler if You want the added safety.
And just a thought, the program must handle collisions......