General problem with MultiWii dev's

[eXc]

Hello,

the last time i have tested several dev's after the stable 2.1 version which is working fine in my case. But i wanted to use HC-SOR4 Sonar which is by now not implemented in this stable version.

While testing failsafe mode i recognized a problem with my rx to be seen in the linked video. When i move one of my sticks to absolute 100% negative position, the failsafecounter increases, as if there is no obvious signal from this channel. Normal moving or 100% positive movement doesn't matter.

I have set the failsafeCnt on debug[1].

Perhaps someone can give me a hint on how i can remove this behaviour.

Best regards
Christian

http://www.myexcalibur.de/mutiwii_failsafecnt_error.wmv

[Hamburger]

It is most likely not a software problem but I will give the usual hint is to check endpoints to be not below 1000 nor above 2000.

[eXc]

Is it possible to set this endpoints in multiwii software or must i check and adjust that with my rx? Ask this, because my rx just gives -100% to +100%.

Christian

[Mis]

You must decrease endpoints in your TX.

[jevermeister]

Had the same issue aome time ago.
I believe Hamburger ran into that a few weeks too.

[Hamburger]

yes. The new failsafe detection code is much better than previous, but unfortunately also way more sensitive to over extended endpoints. It is an issue that comes up frequently, now.
Should add a note to the wiki, probably.

[Hamburger]

done that.

[jevermeister]

hi, extended endpoint are out of spec so it is okay for failsafe to kick in in my opinion...

[Hamburger]

right. But leaves users like me clueless.

[eXc]

Changing the endpoints worked for me. Thx.

[rbirdie001]

Hi all!
I didn't find better thread for such question so I'm trying this one:

I built second tricopter for the fun and wanted to use for it some older sensors from the drawer.
I wanted to use ITG3205 + standalone NUNCHAK as I2C device (I remember this config worked sometimes at 1.8 or 1.9 versions well for me) but I can't make it work now with R1240. Things like configuration, 3.3V supply, pullups and LLC should be OK.
I told myself, OK, I have also analog ACC chip so I'll connect it via Arduino ADC and it will be nice "old school".
I did so and it works and flies, but now I have some I2C errors - lets say ~10 per minute - but no complete hangup or visible flight problems. When I disable analog ACC, no I2C errors anymore.
(OK, I CAN do more testing around, go back to 1.8, 1.9, 2.0 etc, but already spent three evenings with it and I'm about to order new "all in one" FC board. )
Does anyone have some experience about any of those problems or at least can test it on another hardware to confirm or uproot it?
Thanks!
Roman

[alll]

It is most likely not a software problem but I will give the usual hint is to check endpoints to be not below 1000 nor above 2000.

And another reason to implement a rx-calibration procedure?...

manu
PS: can somebody explain the main program structure / blocks? Thanks.

[rbirdie001]

Hi Guys,
I loaded r1311 to the tricopter with the Crius SE board and regular Spectrum RX (connected with 6 lines to separate RC inputs D2, D4, D5...) With 2.1 it worked perfectly but with r1311 is reading of RC signals some way lagy, sometimes stucks for a while and especially throttle channnel reacts very strange. If I set full throttle and then go fast to throttle off, throttle bar in the GUI stucks somewhere about 1500 and never goes down. Slower I'm moving the throttle from the top to down, better the bar follows the stick and I can get bar under 1000 but anyway I can't arm. Even if GUI shows trottle under 1000 and I set YAW stick to right, LED just makes one short blink and remains off, ARM field in the GUI is still inactive, motors off. Sensors are read OK, no I2C errors, cycle time between 3100 and 3500. There should be nothing special in the config.h (see attached).
When I load back 2.1, everything works normally.
Any ideas, please?
Tomorrow I'll continue troubleshooting it and if it persists, I'll try to make video.
Thanks for any advice!
Roman

[rbirdie001]

Seems to be really bug in failsafe function of r1311!
I found that this problem is caused when I uncomment failsafe. If commented, RC pulses are readed OK.
On 2.1 it worked fine. For descriprion please see my yesterday's post.
Can please someone check that? If it's really a bug, it should be corrected for future 2.2.
Thanks!
Roman

Hi Guys,
I loaded r1311 to the tricopter with the Crius SE board and regular Spectrum RX (connected with 6 lines to separate RC inputs D2, D4, D5...) With 2.1 it worked perfectly but with r1311 is reading of RC signals some way lagy, sometimes stucks for a while and especially throttle channnel reacts very strange. If I set full throttle and then go fast to throttle off, throttle bar in the GUI stucks somewhere about 1500 and never goes down. Slower I'm moving the throttle from the top to down, better the bar follows the stick and I can get bar under 1000 but anyway I can't arm. Even if GUI shows trottle under 1000 and I set YAW stick to right, LED just makes one short blink and remains off, ARM field in the GUI is still inactive, motors off. Sensors are read OK, no I2C errors, cycle time between 3100 and 3500. There should be nothing special in the config.h (see attached).
When I load back 2.1, everything works normally.
Any ideas, please?
Tomorrow I'll continue troubleshooting it and if it persists, I'll try to make video.
Thanks for any advice!
Roman

[Hamburger]

most often problems with the vastly improved failsafe stem from endpoints not set properly in the tx. Unwanted values below 1000 or above 2000 are detected way better than before; requires user setting endpoints though.

[rbirdie001]

most often problems with the vastly improved failsafe stem from endpoints not set properly in the tx. Unwanted values below 1000 or above 2000 are detected way better than before; requires user setting endpoints though.

You were right, Hamburger, thanks!
I didn't know that now it's too sensitive now .
I readjusted endpoints and everything works, just it's good to know for everyone, that when failsafe is activated, last valid value of the channel remains displayed in the GUI which was in my case confusing. If I moved throttle stick down slowly, there remained about 1050 but when very fast, there persisted value like 1500.
Generally it would be usefull, if there is some kind of obvious signalisation that failsafe is active (e.g. box in the GUI or special blinking pattern of the LED.) Just for dumb people like me who aren't reading carefully
Roman