Low thrust in yaw direction

Blue Robotics Software Cockpit
1

I’ve been working with the BlueROV for many years. For about two months now, I’ve been using Cockpit. I’ve noticed that I have very little power when yaw. I have full thrust when needed for vertical and horizontal movement, but only about 20% for yaw. I’ve already checked all the settings. With QGroundControl, I had significantly more power. I also don’t have any restrictions like expo or anything similar in my joystick settings. Have I overlooked something?

2

Hi @Lukasz, welcome to the forum :slight_smile:

This sounds very weird - there shouldn’t be a difference between Cockpit and QGC’s default configuration of the inputs. I’m curious how you’re measuring the difference - is it in response time / acceleration of the yaw, or maximum rotation speed, and are you measuring that by looking at the vehicle, or at the telemetry data / video feed? Are you doing back-to-back comparisons of the two control station software options, or is this based off your memory of how things felt when controlling with QGC?

Some ideas off the top of my head:

  • The yaw axis of your joystick is not registering its full range of motion, or has been configured to a reduced range
    • Can you confirm in the table view of joystick configuration that your R axis “preview” column changes between -1 and 1, and has output values between ±1000?
  • Your system is having latency issues, which are more immediately obvious with rotation controls than translational ones
    • Which Cockpit, BlueOS, and ArduSub versions are you using? I’m curious whether the latest Cockpit v1.17 beta resolves the problem
    • Are you able to confirm the telemetry adjusts with minimal delay when you grab the vehicle and rotate it?
    • Can you run a network test in BlueOS?
  • You are testing in a flight mode where the yaw rate is affected by autopilot control system tuning parameters, like acro, guided, or auto mode
    • Can you confirm the issue is present in manual mode?
    • And also in depth hold and/or stabilise mode?
2 Likes
3

Thanks for the quick reply. It’s clearly noticeable both on dry land and in water. Maximum thrust is possible in all directions, but not in yaw. This is noticeable in our test tank, as I can move quickly in all directions. But if I want to turn around, it’s slow. It was also noticeable inside a wind turbine, where eddies are to be expected. My ROV could no longer maintain its heading and slowly rotated because the drive power wasn’t sufficient. Maintaining position wasn’t a problem, as I’m using a DVL. This was not a problem in QGround.

Yes, I can confirm this.

I have no latency.

Cockpit: The problem occurs in the following versions: 1.16 and 1.17-beta 7 and 8
Blue OS: 1.4.2
ArduSub: 4.5.6

Yes. Video and other telemetry data arrive without any noticeable latency.

Tether Cable 100m: Download:65 Mbps / Upload: 63 Mbps

The problem occurs in all flight modes: Manual, Sbilize, Depth Hold and POS Hold.

4

@Lukasz that’s indeed a very strange issue.

Could you open Menu>Tools>MAVLink, search for the MANUAL_CONTROL message, actuate your yaw stick to both extremes and tell us what are the maximum and minimum values that appear for the R axis?

A screen recording actuating all axes would be even better, but as we are only seeing problems in the Yaw the max/min for R should be enough.

1 Like
5

It would also be helpful if you can check the SERVO_OUTPUT_RAW message outputs, to see whether the motors are actually not being commanded to full thrust (since that may be due to autopilot configuration rather than the inputs being received).

Have you tried QGC recently? It’s still not clear to me whether you’re doing back to back testing (in which case it most likely is a Cockpit issue), or if you’re just basing this off memory (in which case it might be an autopilot or motor/hardware issue).

2 Likes
6

I will check this in the next few days and get back to you.

7

I’ll check it out!

1 Like