G4EML wrote: ↑Wed Mar 31, 2021 9:35 am
Don’t forget that we have seen before that some radios had EEPROMs that were slow to write. We had to add the code to wait until the write was completed. Maybe speeding up the bus has caused the write to fail again.
Colin.
I did some testing and there appears to be a few problems.
The system we use to wait for the EEPROM to finish its write, doesn't appear to work correctly.
The data sheet says that it takes 5 milliseconds to write any data and during that period the EEPROM will not respond to other requests.
We poll the EEPROM waiting for it to respond at 1 millisecond internals, but this doesn't seem to work if the bus speed is 400kHz.
Unfortunately changing the timing of this, causes another problem with the CPS, as the CPS won't work correctly if the firmware doesn't respond quickly.
I noticed this problem with the CPS over a month ago, mainly on Widows 10, which seems to behave slightly differently to Windows 7.
But to fix the CPS, the entire COM's transfer system needs to be rewritten to change it from polling for data from the COM port, to being event driven.
The code in the CPS is based on Kai's original COM's test code, which worked fine for his original testing, but is now becoming unusable.
But to rewrite the entire COMMS system is non trivial and will probably take several days of solid effort to complete, and at the moment I'm not keen to give up complete weekend's for things which are working, but not working in an optimal fashion.