I've tried delays before :( This is the exact type of thinking that will solve my problem though so I appreciate it. Please provide any other suggestions :)

The issue with the time delays is that they are limited by the MX-900 editor (lowest is .1 seconds or 1 second can't remember without looking). When I use the time delays along with a trigger, the responsiveness of the MX-900 seemed really sluggish and almost unusable for things like brightness increase (based on the price of the MX-900 and MSC-400). I think this is because the MX-900 has some sort of extra pause it must do for macros (and this pause is undocumented of course).

A macro with a trigger and something else seemed to always exhibit this extra undocumented pause too... and the pause was substantial although I ddin't measure it. However, this did reduce the errors, but limited the number of user presses substantially. So if you pressed the button really fast, the MX-900 would just ignore your press until the macro finished...

Any time delays added to the macro inside of the MSC-400 did nothing to eliminate the reliability problem (only adding delays to the macro within the MX-900 editor worked). I'm not sure why this is...

It's interesting to hear you've encountered the same issue. Take a peak at the CCP file I posted (use the google doc link a few posts up). I've been using my work around (which is a PITA to do 100 times, but it works), and it still works flawlessly and is very responsive. I'd be interested if my method helps your issue too.

I've also added a new method to my invert variable work around that works for press/hold buttons (I can post this too if anyone is interested). Basically, I use the press and hold functions to press buttons (on a hidden device page) that use the invert variable work around for each button. In effect, this allows me to use a variable macro for the press and a different variable macro for hold.

I still don't know why changing the RS232 trigger in between presses fixes the issue entirely. It makes me think it's something to do with how either the MSC-400 or MX-900 is handling denouncing or it's some method to keep from overflowing the buffer of whatever device is attached to the MSC-400. None of the other HA systems I've played with do this kind of thing though. They just work.

I'm experiencing a very similar problem (which I've posted here: [Link: remotecentral.com]).

When my RS232 command is the only command in a macro (on the MX-880) it works nearly flawless.

However, when I use the RS232 command in a more complex macro it only works 20% of the time. I'd tried adding a 3.0 second delay to my complex macro before it issues the RS-232 command but it doesn't help at all.

I know this is a pretty old thread but I have this issue and am looking to see if it is the case for anyone else still or if it has been fixed and I am doing something wrong.

I use a mx1200 and a mx900 with a msc400. I only use 1 rs232 out that dumps commands to a Server that then processes the commands and sends out IP commands to the rest of my devices. Everything works fine except for the fact that sometimes the smart macros won't fire and so it may take multiple button presses to get the channel to go up for example.

Each button has a smart macro assigned to it to send a type ascii string to the listening server. There is also only one command in each macro as the server will process the command and execute the appropriate macro command (I did this to take the load off the msc400). I record each button press which is how I am able to tell if a command is not sent to the listening server.

I wasn't able to find any recent information so either I am doing something wrong or everyone just deals with it.

Any suggestions are greatly appreciated.

Sounds like the RFX250 antenna is receiving RF noise. Look at the RF light on it and see if it's flickering. Try relocating the antenna to improve noise rejection.

That's what a MSC does when there's noise.