The eventual data for the US to turn off the L1CA code transmission will depend on when the entire constellation is fully replaced with Block 3 and 3F satellites. As of Jan 2021 only 5 block3 satellites are in orbit.

Not that long ago Deere/NavCom cut its contract with the JPL lab for use of the Gypsy software that had been producing the corrections for all NavCom Starfire systems since the company started. Deere/NavCom developed their own software for calculating the corrections and began transmitting those corrections in a new message format. At this time all SF3000 (saphire GNSS processor) and newer receive this new message data from the SF system. They did not or could not change the system on older Navcom 2000/2100D processors (Deere gen2 and iTC) to this message system. At the end of 2020 Deere ceased any broadcast of the data that originated from JPL Gypsy software so SF1 and SF2 data can no longer be decoded by the old receivers.

I believe the issue in the iTC came about because the iTC is "seeing" SF satellite broadcast but can't decode the correction data. I think the internal receiver firmware was jumping out of a WAAS fix when it was able to track and see a strong signal for SF1 but since it couldn't decode the data it got stuck in 3D fix mode and would not use WAAS so long as the SF signal was strong.
I think what the "fix" does is simply stop the receiver from even caring if there is an SF satellite signal so it will always default down to using a WAAS fix.

That's what I think is happening based on what I have seen.