[time-nuts] KS24361 Ulrich Z83XX software mod works
csteinmetz at yandex.com
Mon Dec 29 03:06:06 EST 2014
Some time ago, there was a thread about whether Z38xx writes to the
GPSDO's NV memory every second, raising the possibility that it could
prematurely wear out the NV memory if it were used continuously to
monitor a GPSDO. Here is what Ulrich said (23 April 2013):
>due to the discussion that evolved concerning Z38XX i looked up the sources
>and found that indeed every time the current time is read using a
>is sent in advance to make the receiver answer in the format that Z38XX
>needs. I have not been aware that this could lead to NVRAM wear out but have
>been thinking this simply sets the format of the next output. I have changed
>the software so that the FORM F2 command is now only written once after
>program start. The software can be downloaded from the usual place but I
>have currently no time to test it.
>My own experiences with NV memory that can only be written for a limited
>number of times (EEPROM on ATMEL processors) indicates that a clever made
>NV writing routine would first compare the NV contents to what shall be
>written on a byte by byte base and write only changed bytes.
Does anyone know for certain whether current vesions of the program
really do not pose this potential risk to the NVRAM?
More information about the time-nuts