I'm currently counting the T1CLK pin for 1ms, then doing a div4, to average it out, this gives 16kHz steps. The frequency to memory lookup takes about 500 clks (so 250us).
This is done twice.? If these are the same, then I do a relay set, and an SWR voltage read/SWR calc.
I'm planning on using two SWR targets, one is the good target, to store (maybe 1:1.3), the other an? OK target, to use (maybe 1:1.5).
I'm using 14-bit program memory, with the top bits initialised to 1's.? When I store anything I clear the top bits, so these can be used to tell if there is a valid store value.
In the future, I might use the top bits, to store the SWR.? We have 12-bits, so 8-bits for the SWR, would leave enough for other flags.
So read the main memory, if that's got top bits set, then step -1 or +1, then -2 / +2 to find some settings.? If any are found, then the relay are set to the memory, and the SWR is read.
If the SWR is Good, then it's used.? If it's OK, then it does a fine_tune() to try to improve things.
If this gets it down to Good, then it stores it, if it gets it down to OK, then it's used.
If there is no stored or if the SWR is above the OK level then we do a tune(), or Tom's tune2(), if the SWR is below the Good level, then the L/C/CSW get stored.
I've put some more pictures, and the full schematic on by website:
This includes my horrible soldering where I swapped the pin 11 and pin 23.
Pin 11 is the T1CKI, is was the Cap SW relay, but I cut that connection, and used a wire from the 74AC14/74HC393.
Pin 23 was the Bypass button, I #ifdef'd the code for the button, and used that pin for the Cap switch
RogerW/GW6HRU