Cables should be 1:1 to the peripheral and TX / RX as well as CTS / RTS should be crossed on the peripheral side (e.g. a radio modem would have its RX port on pin 2, TX port on pin 3, RTS on port 4 and CTS on port 5).
If you buy a pre-made cable on the RFdesign store, such as the PIXH2 to RFD900ux cable harness, when you look at the radio documentation, it is connecting CTS of the flight controller (cube in this case) to CTS of the modem.
What seems to add to the confusion is that we get significantly better performance in certain tasks like param downloads through an RFD TXMOD when we add flow control pins, but it doesn’t seem to matter if we connected CTS-CTS/RTS-RTS or CTS-RTS/RTS-CTS. We could not tell a difference between CTS-CTS/RTS-RTS or CTS-RTS/RTS-CTS, but there was a huge difference in speed compared to having no flow control.
Also be aware it would be rare circumstances that actually require the CTS/RTS.
If you do, keep in mind there’s been cases where startup (connecting your battery) and associated noise can cause the radio to go into programming mode - not necessarily the RFD radios, but I have heard of it.
To try and get around that you could set these, in some combination, experiment with values:
BRD_BOOT_DELAY,3000
TELEM_DELAY,2
Or just:
BRD_SERx_RTSCTS,0
and go without RTS/CTS. I never use the Auto option, I prefer to turn it off than wait for some glitch to produce unexpected results - that’s just me.
I’m definitely in the circumstance where flow control does make sense to use. I did separately email RFDesign and this was the response I got:
Sorry for the confusion here. I can confirm that the labeling of the modem pins is as you describe. This mix up stems from a error on the very original RFD900 design and has been carried on for consistency in the models that share the form factor.