When I run the stock voxl_mpa_to_ros node on the VOXL2 and subscribe to the topic for stereo_disparity, the stereo interface throws the following error

Stereo interface received non-stereo frame, exiting stereo

I checked out the code and noticed that the problem is here in the stereo_interface.cpp at line 111

if(meta.format != IMAGE_FORMAT_STEREO_RAW8){
       printf("Stereo interface received non-stereo frame, exiting stereo\n");
       interface->StopPublishing();
       interface->StopAdvertising();
   }

I added a printf to see what format is being received using a printf statement and it is an integer with value 0 or IMAGE_FORMAT_RAW8
got format 0

So i modified the node to be the following

if(meta.format != IMAGE_FORMAT_STEREO_RAW8){// && meta.format != IMAGE_FORMAT_RAW8){
       printf("Stereo interface received non-stereo frame, exiting stereo\n");
       printf("got format %d\n", meta.format);
       interface->StopPublishing();
       interface->StopAdvertising();
   }

The node crashes once more and this time, without any error. When running with GDB i see that it crashes at line 152, which is

memcpy(&(imgR.data[0]), &frame[dataSize], dataSize)

Commenting this line allows the node to work, but now i cant access the right stereo images via ROS

I am running system image 1.5.5-M0054-14.1a-perf with voxl-suite: 0.9.5

Seems like there should be a quick fix for this, can someone help me out here ?

Fixed the problem, the mavlink stream name is SERVO_OUTPUT_RAW_0.

changing that to 200Hz gives us the RPM feedback at 200Hz on mavros

So when i try px4-listener it says command not found.

The command that works is
listener esc_status
This produces the following resutls
never published

But after some investigation, the uorb topic with the rpm information is
actuator_outputs

I can read the output

This topic is being published at 500Hz

I tried to increase the rate of the following mavlink topics

mavlink stream -u 15446 -s actuator_output -r 100
mavlink stream -u 15446 -s ACTUATOR_OUTPUT -r 100
mavlink stream -u 15446 -s actuator_outputs -r 100
mavlink stream -u 15446 -s ACTUATOR_OUTPUTS -r 100

Which results in the following error message
ERROR [mavlink] stream ACTUATOR_OUTPUTS not found (same for all other)

Hi @modaltb
Thanks for sharing that information. It happened today and changing the power cable fixed it.

I had this weird problem with the VOXL2 board I'm working with. When I first connect the battery to the platform, the VOXL2 does not boot up, but instead gets stuck in a boot loop (the fan powers down for a second and starts back up, a green led turns off and on on the bottom of the board) and we need to keep reconnecting the battery until it stops boot looping and starts working again. Sometimes I need to put it in fastboot mode and reboot right after for the board to work again. Does anyone here know whats happening ?

Hey, thanks for that. I was able to pull the voxl-dev branch and got it working on the VOXL2, I can see the feedback of the RPM on the mavlink topic SERVO_OUTPUT_RAW. But the frequency is at 10Hz.
When I try to use the following command

mavlink stream -u "port_number" -s SERVO_OUTPUT_RAW -r 100

I get an error stream SERVO_OUTPUT_RAW not found

From the uart_esc code in px4-firmware I see that the uorb topic thats being used for RPM information is ESC_STATUS. And doing the following commands yields no result. The frequency is still 10Hz

mavlink stream -u "port_number" -s ESC_STATUS -r 100

mavlink stream -u "port_number" -s ESC_INFO -r 100

Is there anything I can do to increase the frequency of the message ?

@Daniel-Mellinger
I am still facing the issue with the remote, but i believe its only with the remote as I was able to get it flying in offboard mode using the VIO onboard. Its flying and is quite stable at the moment. Due to the nature of our research, we dont fly with the manual remote (we only use that for initial tests).
But from my experience im not sure if the remote controller could be at fault, because

1. I have used the same RC / receiver pair on the VOXL (version 1) and its able to fly in manual mode without any problems
2. If you look at the Manual Control input plot in the px4 review mentioned in the 1st post, you will see that there are no discontinuities present in the signal. The setpoint thrust and pitch curves look fine. Whereas the setpoint pitch and thrust going into the attitude controller is quite discontinuous

In summary, we are just using the remote to change modes and trigger emergency kill switch and thus this bug is not affecting us anymore

Im using the spektrum DX8 radio with the satellite receiver

Hi Chad,
Im not sure if the problem lies with PID tuning, as the pitch and thrust setpoints (directly taken from the RC) are the ones that create the problem in the first place. If you have a look at the PX4 flight review, you will see the following

Between time 21:26 and 21:36 the RC input received on the pitch is 0 but the pitch setpoint oscillates between -35 and 0 (degrees).
In the same time my thrust setpoint at the RC is between 0.9 and 1.0, whereas the thrust setpoint is also oscillating between 0.08 and 1.0 or 0.9

This behavior cant be explained by a bad PID tune as the thrust and pitch setpoints should follow the RC directly.

Another thing i see is that in the link provided, I see that voxl2 has PX4 v1.12 but the github repo has v1.4.24 as the latest version. Is there a dev build for v1.12 that i could access somewhere else ?

I have a VOXL2 using the following configuration

1. Frame -> dragonfly 230 (check this image https://cdn.shopify.com/s/files/1/0101/1977/4259/files/3mav_desk_1_2048x2048.png?v=1616628087)
2. ESC -> VOXL ESC v2
3. Flight Controller -> VOXL2 using PX4

I was able to setup the VOXL2 and run PX4 on it. I did the calibration of all sensors and could get it to arm, but when i fly it I see a very weird jittery motion. When I check the ulog file on PX4 flight review I see the following on the pitch, roll and thrust setpoints where it oscillates between the setpoint provided using the RC and another setpoint which is different

pitch setpoints oscillates between the value set by the RC and -35 degrees (max tilit)
thrust setpoint oscillates between the value set by the RC and 0

The review for ulog is here

https://review.px4.io/plot_app?log=444738f8-fc52-4c52-a53d-831342feeba3

I have also tried different PX4 versions that are available here https://github.com/modalai/px4-firmware/ by manually compiling for the rb5 platform and all of them have the same issue.

Is there anyone here who has had a similar experience and was able to solve this issue ?

Regards
Vivek