so ive been messing around with an arduino mega for the past few weeks trying to build something that'll automatically switch my antenna tuner settings when i change bands on the radio. the idea is pretty simple in theory — read the band data output from my IC-7300 (it puts out CI-V over USB) and then drive some relays to switch in different capacitor/inductor values on a homebrew L-network tuner i built last year.

the CI-V parsing part wasnt too bad, i found a library that handles most of the serial decoding. the part thats killing me is the relay timing. when the radio does a band change theres like a 200-300ms window before it actually starts transmitting and i cant figure out if my relay switching is finishing inside that window or if im occasionally catching the tail end of a transmission on the wrong antenna config. i dont have a scope right now which makes this basically guesswork.

has anyone done something similar or at least messed with CI-V on the 7300? wondering if there's a more reliable way to get a band change event that i can actually sync to rather than just watching the frequency change and inferring the band.

yeah ive done almost exactly this but with a Pi instead of an arduino, mostly because i already had one sitting around doing nothing. the CI-V thing is honestly pretty solid once you get the timing sorted. what i ended up doing was adding a small delay after i detect the band change before firing the relays — like 50ms — and then i also watch for the radio to stop being in transmit state before i allow any switching at all. the 7300 will tell you transmit state over CI-V too so you can poll that. its not the cleanest solution but its been working for about 8 months without a hiccup.

the other thing worth checking is whether your relays are actually rated for what you need. i fried a ULN2003 driver chip once because i didnt account for the inductive kick from the relay coils. flyback diodes on every coil, dont skip that step.

not sure about the CI-V specifics but for the relay timing issue have you looked at using a hardware interrupt instead of polling? might get you tighter response times. also the raspberry pi 4 has way more headroom if the mega starts feeling cramped, ive been running a whole logging and digital mode pipeline on one and its barely breaking a sweat. though for something this timing-sensitive maybe stick with the arduino since you dont have an OS in the way messing with latency.