Multimodule problems.

I was having issues with the MM not being detected, as described by btoschi on the CC2500 thread . It had only been happening at power on, so I ignored it until it happened while a model was in flight, dropping it from the sky (a V911 which took no damage).

So I took the MM out, removed the connectors, and shortened the wires. Plugged it in, and got the "Missing module" message for the A7105, etc. Plugged those in. Got a "Missing module" message for the multimodule.

I did the obvious checks: connectivity over the wires, shorts, and that I get 3.3volts. Everything seems ok.

Since I don't have anything I really want to fly with MM modules at this point(*), I put it off, and disabled the MM in hardware.ini. I still get the "Missing module" message for the MM.

I got a logic analyzer for some other work, but have no idea what I'm doing with it. Can someone point me to what I should hook up and the appropriate settings for to capture meaningful data? And maybe what to look for?

*) I've got some older A7105 models that don't fly very often, one waiting on parts from china. I added the nrf24l01 for SLT compatibility but haven't ordered one yet, the CC2500 for frsky stuff but really need sebydocky's telemetry fixes with some of my code as well.

Ran across a little bug the other day that may apply. If the MM is disabled, but another enabled module is not found, then the MM is also reported as missing. And on the 10 (and I presume the 7) the string is so long that missing multimodule is what's left on the display.

If you're using a 7e or 10, try this patch and see if a specific missing module is displayed.

Well, that did clarify the problem. I had all the modules disabled, and now I get a "Missing module CYRF6936. So I guess the next step is to take out the multimodule completely and see if that cleans things up.

Thanks!

I just noticed that this change doesn't seem to be in the team build.

Any chance you can add it? Would you like me to?

Anything else we've missed?

Feel free to add it. I lost the local change to disk issues before I had a chance to commit.

Not sure it's the best fix because PROTOCOL_SetSwitch() also checks if the multimodule is enabled. Didn't remove check that in case it's needed in other cases, but it seems to me the check for module enabled should be separate from an error return from that function.

Pull request created. Still waiting to get that other approval.

While it may not be the best, it's clearly better than what we have no, so I'm all for doing either as is or without the extra test.