Really wierd Deviation Issue - Bug or "Feature"?

Background: I am running Deviation 5.0.0 on a new (white) model Devo-10, with Neewar Black nRF24L01 module added. I've done a ton of flying with it, and never had a problem until now.

So yesterday, just before I had some friends visit to do some flying, I cleaned up the model file I had previously made for my Syma X8C quads. Just minor stuff, some GUI tweaks for positions of toggles, a little bit of whitespace editing in the .INI file to make it a little more readable, adjusted one Yaw setting to make it easier for a beginner to control on FMOD 0 "Easy" setting... stuff like that.

Along with these minor "tweaks", I also changed the name of the first channel mixer section from [channel1] to [channel11] in order to match the rest of the channel headings (example: my Ch3 Throttle mixer is [channel13] and so on).

The flying day was a DISASTER!

Completely uncontrollable, pitch and roll seemed to be reversed and more powerful in one direction than the other, and my throttle control seemed to be all over the place! Crashed one X8C into the side of a large metal silo at about 40MPH after trying to re-gain control of it following a quick takeoff and bank (which happened the wrong direction from what I commanded). WTH? I thought maybe the X8C control board had gone bad in that quad, so I brought out my other X8C (with video link installed) only to have a violent "faceplant" into the ground after dropping from about 60 feet, after having it rocket straight up (I am using a 3S Lipo pack on that one because of the extra video Tx weight) because the throttle seemed to suddenly lose its mind! Crinckled one arm almost to breaking failure, very irritating (and embarassing with my friends there all watching the whole fiasco! )

I could tell that something was HUGELY wrong... so I spent 4 hours today slowly "undoing" all of my .INI changes one by one (comparing to my previous version as I went), then flight testing in my enclosed shop area with one of my newly rebuilt X8C aircraft trying to get things back to normal. Finally changed the [channel11] heading back to [channel1] as a last ditch attempt after going back to the original code on just about everything else, and suddenly everything was magically working again! I then restored my model file back to yesterday's "disaster" version, then changed ONLY that [channel11] line back to [channel1] and it ALL worked perfectly!

Wierd. Does Deviation absolutely REQUIRE that a [channel1] heading be present in the .INI file in order to work properly?

I don't have X8C or any addon NRF HF modules...

Sorry, I have somehow no clue what you are talking about
Why don't you upload both model.ini files?

Why is CH3 mixer channel11?
The mixer is associated to CH11, not CH3 if the mixer section is below channel11.
It only works if you assign CH3 the input of CH11 (which the mixer controls)?!?!?

What channels does the H8C expect?

Have you checked your advanced mixer GUI what happens to your channel1 vs channel11 vs your mixer which is now associated only to
[channel1] means channel one.
The [mixer] section below [channel1] means the mixer is associated to CH1.
So this is CH1 and not CH11.

I am almost sure that if your protocol is set to use 11 or 12 channels that there need to be [channel1], [channel2]... sections first and you can not just simply delete it.

Maybe a Deviation core developer or a NRF / H8C protocol dev can tell you more.


After directly editing the model.ini file you may want to check in the mixer GUI if the file could still successfully read and all channels and mixer stuff works.
Try your channel monitor! What value is sent on CH1, especially if you remove the channel from the ini?

For sure your CH1 mixer setup won't send on channel 1 anymore...so depending on what your quad does when it sees no CH1 signal...

Personally I do believe that you should not do a complete re-write of the the model.ini from scratch without further testing and double-checking channel values.

Too bad that there is no way for your H8C to check the FC what signal values are displayed for all required+optional channels.

Thomas,

I read your post, and I think you are confusing actual RC channels (such as Ch1, Etc.) with mixer channel headings (such as [channel1], etc.)

My problem is that by changing the [channel1] mixer heading in the .INI model file to [channel11], it caused HUGE chaos to the whole system. This is very strange, because whoever originally wrote the .INI model file (it was written for X5C-2 quads) had every mixer channel name increased by 10 (example: Ch3 RC channel was handled by mixer heading [channel13] and so on.) All I was doing was making the [channel1] match up with the rest of them. Apparently, though, Deviation does not like this, and seems to need a [channel1] mixer heading for some reason!

Any insights from you or others? This is a pretty strange behavior for the software...

FDR wrote: I don't think it supposed to work with renamed headers.
Only the virtual channels can be renamed...

If that's the case, then why have I seen model files that have [channelX] mixer heading numbers ALL OVER THE MAP? Some are sequential and in order, some skip numbers... there seems to be no defined structure to it.

We're still waiting for a specific example - can you attach your config file here?

victzh wrote: We're still waiting for a specific example - can you attach your config file here?

OK.
This works:

[channel1]
subtrim=65
template=complex
[mixer]
src=AIL
dest=Ch1
scalar=50 blah blah blah blah

This does not work:

[channel11]
subtrim=65
template=complex
[mixer]
src=AIL
dest=Ch1
scalar=50 blah blah blah


The only difference is line 1.

PhracturedBlue wrote: can you please post an entire ini file?

the # after 'channel' must be less than the number of output channels defined.
Also, using channel # that are not sequential is very innefeient. These are not random names you can define as you like. The # after the word 'channel' defines which slot the channel fits into in the table. basically, this is not something you can play around with however you like. In almost every case, these should be sequential starting from '1', possibly skipping a few channels that are not used by the Rx. Also, the 'dest' value of the mixer should match the Channel # usually. changing the header without changing the 'dest' is certain to break things

Aha!

That would explain things... finally. An answer that makes sense!

I cannot post the entire INI file, I'm posting from an iPad, and the only way would be to copy/paste the entire file into the post... long and inefficient.

Anyway, it sounds like you have answered my question completely... I now know what to avoid in the future!

Thanks.