Re: [Paparazzi-devel] Quadrotor with Lia 1.1 - a sudden loss control authori

[Michal Podhradsky]

Hi Felix,

we are using a Spectrum satellite receiver with a customized antenna (higher gain than the original one) + Spectrum DX7s transmitter , and we had some radio issues before. So it might have been a radio problem.

Although this happened on a very short range, around 20-30m from the transmitter. 

The config file is here: http://pastebin.com/f1a8xx1a

I checked the sys_mon log and everything seems to be fine (a constant performance during whole flight, and CPU rate around 50%).

Regards
Michal

On Tue, Nov 13, 2012 at 10:50 AM, Felix Ruess <address@hidden> wrote:

Hi Michal,

the horrible frame rate might indicate a problem with your RC. The baro is of no concern in RC_DIRECT mode..

I think it should still work at 10Hz update rate (assuming the channels otherwise contain correct values)... you just get some more lag.

The rc commands are only updated when you actually receive a valid frame, otherwise the previous one is used.

Never had that problem so far though...

Were you at the range limit when this happened?

What RC system are you using?

Can you maybe post your airframe config? Best put it on pastebin or something like that.

It would probably help if you could do some range tests, log the RADIO_CONTROL message as well.
Maybe add the sys_mon module to check if the overall timing is still ok.

Cheers, Felix

On Mon, Nov 12, 2012 at 8:41 PM, Michal Podhradsky <address@hidden> wrote:

Hi Felix,

yes, so that is why I am surprised that the quad did not respond for the throttle commands from the RC pilot.
We are using Lia board from Quadshot (http://thequadshot.com/products/lia) with Aspirin IMU (actually we are using only the MS5611 barometer for pressure readings, not the gyros etc.) and GX3 IMU for attitude.

Unfortunately RC messages are not logged.
I checked the suggested logs:
1) 3-baro_abs-est_z.png shows the barometer absolute readings vs. estimated altitude. The baro is calibrated and performs better than BMP085
2) 4-frame_rate-motors_on-in_flight.png - Motors are on, we are in flight - so these are correct. The frame rate is scaled 1/10
3) 5-ap_mode-frame_rate.png - We can see that shortly after going from Z_HOLD to ATT the frame rate suddenly drops (from 45 to 10 fps), but AP didn't go to FAILSAFE, so I assume even 10 fps is enough, right? Or can this be the problem (Radio related)?

Regards
Michal

On Sat, Nov 10, 2012 at 7:28 AM, Felix Ruess <address@hidden> wrote:

Hi Michal,

this is very strange...

Whether you use the adaptive hover throttle or set a fixed NOMINAL_HOVER_THROTTLE makes no difference in the DIRECT vertical mode, it just takes the RC input and directly applies that to the thrust command: https://github.com/paparazzi/paparazzi/blob/v4.0/sw/airborne/firmwares/rotorcraft/guidance/guidance_v.c#L177

So that should be fine..

From the plot, it doesn't look like the thrust command from the RC was put to 100%, which is very weird. 

I guess you didn't log the RADIO_CONTROL or RC messages in that flight?

Did you take a look at rc_status, frame_rate, ap_in_flight and ap_motors_on in the ROTORCRAFT_STATUS message? (Just to make really sure these are ok)

Could you please post your full config and flight logs?

Cheers, Felix

On Fri, Nov 9, 2012 at 10:30 PM, Michal Podhradsky <address@hidden> wrote:

Hi Loïc,

I am attaching the logs:
1) AP mode and Vertical mode vs. thrust command (scaled to fit). It is not obvious for the first sight, but after going from Z_HOLD(8) to ATT(3) our RC pilot was pulling throttle even up to 100% trying to keep the plane in the air, but it is not reflected in the plot until the last moment - right before the landing (t=239s)

2)  Vertical loop - est_z, est_zd, est_zdd and AP mode. You can see the quad was climbing rapidly during Z_HOLD, and then started to descend after switching to ATT.

Regards
Michal

On Fri, Nov 9, 2012 at 1:06 AM, Loic Drumettaz <address@hidden> wrote:

Hi,

I never experienced such issue. I don't use adaptative estimator for thrust, but NOMINAL_THROTTLE instead (see Paparazzi vertical loops wiki page).  From my experience, it is sometime difficult to switch from RC_CLIMB mode to DIRECT mode for the pilot, because of the sudden change of throttle required to hover.

Maybe could you post your logs?

Regards

Loïc

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel