[time-nuts] Lady Heather's Tbolt oscillator auto-tune function
holrum at hotmail.com
Sat Sep 10 21:03:44 EDT 2016
Here's a little info on Lady Heather's oscillator autotune function for the Thunderbolt GPSDO:
The autotune function tries to optimize the settings for the oscillator disciplining parameters, antenna signal mask angle, and the signal level amplitude mask beyond what the default setting (which are more for telecom timing applications) are.
To use the auto-tune function the receiver should be warmed up and stable. Manually set the antenna mask angle to a low value (say 0-5 degrees) with the FE keyboard command and set the signal level mask to a low value (1 AMU, 20-30 dBc) with the FL command. Clear the signal level log with the CS command. Let the receiver collect signal level data for several hours (overnight is good). The data collected is used to determine where your signal levels begin to drop vs the satellite elevation angles.
Since the unit should be locked and stable, the current DAC setting is where the oscillator is at 10.0000000 MHz and will be set in EEPROM as the initial DAC setting. The tbolt uses this value to speed up locking the oscillator when powering up.
Then issue the autotune command (&a). This will put the DAC into manual control mode and step the DAC control voltage 5 mV high and then low and measure how the oscillator frequency changes. This takes a few minutes to complete. This is used to determine the oscillator voltage gain. The calculated value seems to be quite accurate. Tbolts set up for external oscillators can be used to determine unknown 10 MHz oscillator EFC characteristics.
Next, Lady Heather sets the loop time constant to 500 seconds and the damping to 1.0 These value seem to be good conservative general purpose values for a typical unit and should not cause any loop stability issues.
Finally, Lady Heather sets the antenna mask angle setting to where the signal level starts to drop off rapidly and the signal level mask value to 30 dBc (or 1 AMU unit). After the auto-tune completes, the antenna mask angle setting might need to be manually set to a lower value... it tends to find that the signal level falls off at a higher angle than one might expect. You can check the signal level vs elevation plot with the SAE keyboard command (or ZE in the next release will show the plot zoomed to full screen). High antenna mask angles cause more satellite constellation changes which is not good. Low antenna mask angles subjects the receiver to multi-path effects which is also not good. You are damned if you do and damned if you don't...
More information about the time-nuts