I am pretty sure you didn't make a mistake, since you posted the Function=0 command.

I think if IRPanels created 256 buttons, then you probably did everything right to the ccf file. If you have made ccf files work before with MXEditor, then I am at a loss for a simple explanation.

If the commands that I emailed still don't work, then my bet would be some sort of framing error probably caused by the translation program. Not that I am saying it is necessarily defective, just that the real command translated to Pronto hex and that translated to MX800 mave have a cumulative timing error that causes the IR recever to reject the command.

I have had some experience with this sort of thing in the NEC1 protocol with a video switcher and tinkering with the time base in MakeHex ultimately got recognition. In this case someone actually put a scope across the demodulated output of the IR receiver and we knew that the timing was skewed and roughly how much when compared to the OEM remote. If you really want to trouble shoot this, by any chance do you have a JP1 OFA remote? The JP1 software can very accurately display the on and off timing of learned signals. You could compare the learned OEM signal to the same "generated" command from the MX

I went back to the original learned signals from the J1 group. Every properly learned command (and there were 16 of them) all had EXACTLY a frame length of 108,548 uS. The null repeat frames were also exactly 108,534 uS and that is nearly perfect classic NEC1 timing. So if the OEM remote is that accurate (and they usually aren't), then it may have a pretty selective receiver. John, any thoughts?