Gimbal Not Centered

[User Name]

You're changing the goal posts now, since when was this conversation about gimbals 10 degrees off or more off? It never was.

If a gimbal is way off, do the level calibration and compass calibration to make sure the gimbal IMU and Solo's IMUs are in sync then retry. Then contract support. 10 degrees is a significant amount and doesn't tie in with typical examples. Unless of course you're plucking figures like 10 degrees out of thin air just to continue an argument.

I measured mine at 13 degrees, so nothing is being plucked out of thin air. Calibration does nothing with regard to the issue. Mine may indeed be an extreme example, but the issues remain - it may be less noticeable on less extreme examples, but they ARE there. No goal posts are being changed, Solo is simply not functioning as you suggest it is or should.

My screen-record example and Joe's video disproves several contentions:

-His balloon IS at the center of the orbit point because it remains stationary throughout the video. You've simply picked camera frame center, NOT orbit center. The ground continually moves beneath your suggested orbit point, proving it's not the actual orbit point.

-Solo is NOT compensating for an offset camera rotation.

-You can NOT set an orbit point offset from a subject and keep the subject stationary throughout an orbit.

3DR may indeed intentionally rotate the camera to keep the left leg out of the camera's view, but they have NOT compensated for this in the flight controller and/or during smart shots.

 

[Ian [P13]]

You are moving the goal posts and you're talking out of your backside to suggest all Solos are 10-13 degrees. Attached is a typical Solo gimbal offset and it's nowhere near 13 degrees, to actually be pointing at 13 degrees the right leg would be in shot.

I've attached an image of 13 degrees. I look forward to a photo of your solo in flight from below showing a 13 degree offset.

 

[User Name]

You are moving the goal posts and you're talking out of your backside to suggest all Solos are 10-13 degrees.

First of all, I have never suggested ALL Solos are 10-13 degrees off. And I'm not talking out of my backside. I've posted a video of mine demonstrating this several times and I'll do it again. Please acknowledge the following so we can all be assured you aren't simply ignoring this.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

Second - speaking of backsides, you have made several claims that have been proven false:

-You have claimed Joe's balloon was not the center of the orbit in his video. This is demonstrably false. I've suggested (several times) you hover your mouse over your suggested center as the video plays. The ground will move beneath your mouse. Please do this now and either explain how a moving point can be the center or acknowledge your mistake.

-You have said that Solo orients itself based on the gimbal orientation - "cablecam and orbit etc is driven by the gimbal's orientation, so it's flying the direction of the gimbal and the Solo is aligning itself to the gimbal". By remaining stationary throughout the shot, the balloon in Joe's shot is clearly at the center of the orbit. If Solo oriented itself based on the gimbal orientation, the balloon would be center frame. It clearly is not, so this claim is also demonstrably false. Please acknowledge.

I've acknowledged several times that my 13 degree off gimbal is extreme, but if compensation exists, it shouldn't matter how off a gimbal is. And since extreme examples aren't compensated for, on what do you base the existence of any compensation at all? What I'm saying is that on lesser gimbal rotations, the effects are simply not as noticeable, but they ARE there.

I'm not looking to argue just to argue, I'd simply like to get to the bottom of the issue. But you're repeatedly making claims that are demonstrably not true.

 

[Ian [P13]]

Your gimbal isn't typical, this thread is not about your gimbal you've simply hijacked the thread to make it so. The only time I've seen a gimbal that far off has been with an install where the hdmi is pulling. So check your install and if all is correctly installed contact support.

Your video pre-dates the last firmware, the gimbal code has been having continuous tweaks, so I'd also suggest doing at least a level calibration with the current firmware to see if that helps.

I'm not getting into an argument without logs on a distant subject, I don't think that balloon looks like the point of focus and the only proof either way is the logs.

 

[User Name]

Your gimbal isn't typical, this thread is not about your gimbal you've simply hijacked the thread to make it so. The only time I've seen a gimbal that far off has been with an install where the hdmi is pulling. So check your install and if all is correctly installed contact support.

You have an unfortunate tendency to overlook and/or refuse to acknowledge the obvious. Look at my gimbal when the power is off - it happily sits facing directly ahead proving the issue is not any pressure on the mount.

But this is beside the point because I most certainly am NOT hijacking the thread -
the issues here effect ALL gimbals, it's simply less noticeable on gimbals with less rotation. My gimbal is in the process of being replaced, but rest assured, if the new one comes back only offset 5 or 6 degrees, it'll demonstrate these same exact issues, just less so.

I'm not getting into an argument without logs on a distant subject, I don't think that balloon looks like the point of focus and the only proof either way is the logs.

I hope Joe does post his logs to prove to your satisfaction the exact location of the orbit point. But this really is unnecessary and your refusal to acknowledge this is telling. Solo simply CAN NOT orbit a moving point. The spot you've chosen as the orbit center is moving, therefore it is not the center. Fact. Case closed. End of story, logs or no logs.

The problem is that you have a basic misunderstanding of how Solo works. Since there is no manual control over gimbal yaw, in fact no control over yaw at all outside stabilization, it's meaningless to say that Solo aligns itself to the gimbal. Beyond minor rotations necessary for stabilization, the gimbal is effectively fixed in yaw - of course they're aligned, they can't NOT be. Solo DOES NOT orient itself to the gimbal orientation. What it does do is allow the user to freely yaw during smart shots and it automatically offsets the flight controller's orientation relative to the airframe in real-time. This allows it to fly a fixed path regardless of airframe orientation.

Put more simply, it decouples airframe-front from flight controller-front. This has nothing to do with gimbal orientation which, again, is fixed, and only effects cable cam and follow modes, not orbit, since you can not freely yaw in orbit. So gimbal orientation is doubly irrelevant in orbit mode.

I know you'll refuse to acknowledge this, and I suspect you'll even refuse to read it, but thems are the the facts. If you have a direct line to support, I guarantee they'll back this up.

 

[Ian [P13]]

I also pointed out, your demo is not on the current firmware and you'd likely not run the level calibration afterwards, but you clearly ignore that as you don't seem to want any help at all.

I never said the gimbal itself was yawed in a move, I went so far as to clarify I wasn't saying that.

I've said the gimbal has its own IMU, the Solo has its own IMUs, that the SoloLink controls both, with the gimbal driving the moves, i.e. when you store a position for example in cablecam, the position including the camera is stored, at all times the SoloLink is using the information it has on the gimbal relative to the legs to keep them out of shot. The fact they have separate IMUs is why there's no harm whatsoever in doing a level calibration.

I am querying with 3DR if it'll be possible to manually recalibrate the gimbal position in a future firmware as with handheld gimbals such as the Ronin you can as IMUs aren't perfect...

 

[User Name]

I also pointed out, your demo is not on the current firmware

I had the latest firmware on my Solo and gimbal as of the date it was removed to send back for a replacement, which was January 1st. It exhibited the same issue, so unless they fixed it in a firmware released in the past 18 days, this is irrelevant.

and you'd likely not run the level calibration afterwards

Please explain, specifically and explicitly, how you determine the likelihood of my not having run level calibrations after a firmware update? For someone who talks of pulling information out of one's backside, you do a hell of a lot of that yourself.

but you clearly ignore that as you don't want any help at all, you simply want an argument.

I've ignored nothing.

You even attempt to put words into my mouth, which is your misunderstanding of points I've said.

Your misunderstanding of the way Solo works is so profound you can't even see how little sense your explanation makes. As pointless as I'm sure this will be, here goes:

"at all times the SoloLink is using the information it has on the gimbal relative to the legs to keep them out of shot"

There is no relative information between the gimbal and the airframe, because the two are effectively fixed. Your explanation directly implies that the gimbal can yaw, but when I point this out you say "I never said the gimbal yaws".

Have you ever built a multirotor? Flight controllers have a set "forward" orientation. If the flight controller is oriented in the airframe with it's "forward" offset from that of the desired airframe forward, you need to configure an offset to compensate for this. If the flight controller is oriented 90 degrees to the right, you'll configure a yaw offset of -90 degrees so the flight controller knows to treat it's left side as front, it's right side as back and so on.

When Solo is in a smartshot where the operator has free control of camera yaw, you're not actually controlling camera yaw, nor are you "flying the camera" with the Solo updating the airframe to match the gimbal, keeping the legs out of the frame. All that's happening is that you're yawing SOLO and the yaw offset between the flight controller and the airframe is being updated in real time to counteract your yaw - Solo is flying along it's programmed path with the airframe pointed ahead and you yaw to the left, Solo's flight controller gets updated in real time to change it's yaw offset to the right. That's it.

The gimbal has IMUs to read and report orientation so the motors can counteract unintended movements of the camera, but this information is not reported to the flight controller to control the airframe itself in any way.

I'm virtually certain this is the way it works, and I've e-mailed support to verify. If I'm mistaken, I'll freely admit it, unlike you, who refuses to acknowledge when your assertions are directly contradicted by observation.

 

[Joe Horn]

I hope Joe does post his logs to prove to your satisfaction the exact location of the orbit point. But this really is unnecessary and your refusal to acknowledge this is telling. Solo simply CAN NOT orbit a moving point. The spot you've chosen as the orbit center is moving, therefore it is not the center. Fact. Case closed. End of story, logs or no logs.

That balloon flight was almost a month ago. I've flown many times a day, almost every day, since then. So its log went into the bit bucket many flights ago. But as you say, the manifest evidence of the video itself suffices to prove that the balloon was in fact the center ("target") of the Orbit, and that the gimbal was aiming the GoPro off-center.

Somebody asked what that balloon is for. It's for balloon rides. It's one of the largest tethered helium (not hot air) balloons and the first of its kind in the United States. Standing 118 feet tall with a volume of 210,000 cubic feet, the Great Park Balloon can hold up to 30 passengers in its 1,810-pound gondola. Flying at an altitude up to 400 feet, it is always tethered to the ground by a steel cable with the strength of 99,000 pounds, making it a safe and fun experience for everyone. But I digress. Back to off-center gimbals.

 

[Ian [P13]]

@User Name

The gimbal is Mavlink and reports back to the SoloLink, the data from the gimbal is also logged. The gimbal has its own EKF algorithm and is more complex than you give it credit for.

The SoloLink can control the gimbal and Pixhawk using information from both.

The gimbal can dynamically change its own torque and plenty of other tricks, just glance the code it is anything but a dumb slave to the frame:

ardupilot-solo/libraries/AP_Mount at master · 3drobotics/ardupilot-solo · GitHub

A browse of the commits shows how much work is continually going into improving the gimbal logic:

History for libraries/AP_Mount - 3drobotics/ardupilot-solo · GitHub

 

[User Name]

The gimbal is Mavlink and reports back to the SoloLink, the data from the gimbal is also logged. The gimbal has its own EKF algorithm and is more complex than you give it credit for.

The SoloLink can control the gimbal and Pixhawk using information from both.

The gimbal can dynamically change its own torque and plenty of other tricks, just glance the code it is anything but a dumb slave to the frame:

Never implied it was dumb, but none of that signifies that it works as you suggest. Of course it's controlled through Mavlink over SoloLink - that's how pitch and camera functions are controlled. EKF is common in position estimation, which the gimbal does in order to stabilize the camera. Same goes for dynamic torque changes of the motors. Again, none of this means that it's yaw orientation is specifically referenced by the flight controller in order to keep parts of the airframe out of the camera's view.

And just as a reminder, you still haven't explained how you came to the conclusion that I didn't recalibrate after firmware updates.

 

[Ian [P13]]

You said the gimbal didn't report data back, you said the airframe and gimbal were simply fixed which is saying the gimbal is dumb.

This code is only Arducopter side too, the SoloLink code isn't on GitHub and plays a key part in Solo. Python code is used on the SoloLink to interface with the Pixhawk and gimbal.

There is no relative information between the gimbal and the airframe, because the two are effectively fixed.... The gimbal has IMUs to read and report orientation so the motors can counteract unintended movements of the camera, but this information is not reported to the flight controller to control the airframe itself in any way.

Then after I pointed out the Mavlink interface you immediately go 180 degrees...

Never implied it was dumb,

 

[User Name]

You said the gimbal didn't report data back, you said the airframe and gimbal were simply fixed which is saying the gimbal is dumb.

I said the gimbal didn't report orientation data for use in controlling the airframe's yaw. And the gimbal's yaw IS fixed outside movements necessary for stabilization. The gimbal's "smarts" are in it's ability to interface with the GoPro and the general stabilization algorithms, not in any ability to dynamically control or influence airframe orientation.

This code is only Arducopter side too, the SoloLink code isn't on GitHub and plays a key part in Solo. Python code is used on the SoloLink to interface with the Pixhawk and gimbal.

Irrelevant, especially since neither I or, I presume, you can read the code and determine it's function.

And you STILL haven't justified your comment regarding calibrations after firmware updates. I'm going to keep pressing you on this, since you're so keen on claiming others are pulling information from their arse.

 

[Ian [P13]]

The gimbal's "smarts" are in it's ability to interface with the GoPro and the general stabilization algorithms, not in any ability to dynamically control or influence airframe orientation.

Irrelevant, especially since neither I or, I presume, you can read the code and determine it's function.

This is a developer released piece of code for a shot in Python, this isn't Arducopter code on the Pixhawk so it's not airframe code but SoloLink code, which in effect adds a layer of smartness to the gimbal.

Whilst I'm no Python expert it's clear the code can get the camera/Solo's current orientation and also control it, here's an example of orientation values being requested:

        # default camPitch 
        self.camPitch = camera.getPitch(self.vehicle)

        # default camYaw to current pointing
        self.camYaw = camera.getYaw(self.vehicle)

Here's them being controlled, with different code here depending on if the Solo is with or without a gimbal:

    def handlePitchYaw(self):
        '''Handle free look'''
        # if we do have a gimbal, use mount_control to set pitch and yaw
        if self.vehicle.mount_status[0] is not None:
            msg = self.vehicle.message_factory.mount_control_encode(
                0, 1,    # target system, target component
                # pitch is in centidegrees
                self.camPitch * 100,
                0.0,  # roll
                # yaw is in centidegrees
                self.camYaw * 100,
                0  # save position
            )
        else:
            # if we don't have a gimbal, just set CONDITION_YAW
            msg = self.vehicle.message_factory.command_long_encode(
                0, 0,    # target system, target component
                mavutil.mavlink.MAV_CMD_CONDITION_YAW,  # command
                0,  # confirmation
                self.camYaw,  # param 1 - target angle
                YAW_SPEED,  # param 2 - yaw speed
                self.camYaw,  # param 3 - direction
                0.0,  # relative offset
                0, 0, 0  # params 5-7 (unused)
            )
        self.vehicle.send_mavlink(msg)

 

[Ian [P13]]

Furthermore, I'd asked for a photo of the gimbal orientation in the air as that's when the code starts to collect data for orientation, so looking at the gimbal on the ground doesn't prove there's any issue, this is from 3DR's Jason Short who's a hardware lead.

This information also proves the gimbal isn't simply acting as a slave to the airframe and is in fact gathering its own data and acting on it. It's interesting to note mag data is used too by the gimbal, so there's no harm in a compass calibration either.

This was in reply to a guy turning his Solo on, on a table and staring at his gimbal.

 

[SurfnSkate81]

You're changing the goal posts now, since when was this conversation about gimbals 10 degrees off or more off? It never was.

If a gimbal is way off, do the level calibration and compass calibration to make sure the gimbal IMU and Solo's IMUs are in sync then retry. Then contract support. 10 degrees is a significant amount and doesn't tie in with typical examples. Unless of course you're plucking figures like 10 degrees out of thin air just to continue an argument.

Sorry Ian, but this conversation is about the gimbal being out by 10+ degrees. I should know I did start this thread and have had my solo and gimbal replaced four times.

And yes before you ask I had done level, compass and stick calibration as instructed by 3DR. They did authorise and give an RA# for replacement, so putting two and two together one would assume that this is not normal otherwise 3DR would not have given the OK for my retailer to replace.

What would be good is to get a definitive answer from 3DR as to what tolerance is acceptable. Is the gimbal supposed to be spot on centred or not to compensate for GoPro lens being closer to Solos leg and I doubt we will never know cause 3DR may have flood of returned gimbals on their hands and some rally bad press to accompany the whole delayed gimbal saga.

 

[Ian [P13]]

I've given the answer, there isn't an issue, the gimbal tunes and aligns in flight, it has an EKF just like the flight controller. On the ground the gimbal is in low torque mode and is not getting any velocity or heading data so won't necessary point forward, in flight getting said data it points forward. You have the technical detail in the prior post, that's information as of the current firmware so is up to date.

The support guys won't always have the latest information on how the code works as it's changing almost by the week, 3DR give the benefit of the doubt and often RMA when in fact there isn't an issue, as they'd rather check and confirm than have an unhappy customer.

Jason Short along with Philip Rowse is amongst one of the most knowledgeable guys on the Solo, i.e. his word is gospel.

 

[User Name]

On the ground the gimbal is in low torque mode and is not getting any velocity or heading data so won't necessary point forward, in flight getting said data it points forward.

Your ability to disregard provided visual proof that contradicts your claims is amazing.

We have footage, from the air, from orbits, demonstrating the camera is not pointing at the point around which the Solo is orbiting. In a perfect example of circular logic, you pick the center of the frame as the orbit point... because the camera HAS to be pointing at the orbit point, despite that point tracing out a circle in the footage. I provided a clear example of both scenarios - camera offset an actual orbit point, and offset orbit point to correct for offset camera. My video exactly backs up the video and demonstrates the effect of a camera not targeting the orbit point.

This is your in-air proof. Yet, ever the lover of claiming others are moving the goal posts, you now move the goal posts yourself asking for in-air footage of the gimbal, no doubt because I've repeatedly stated that my gimbal has been returned for a replacement and therefore unable to provide requested footage.

As for low torque on the ground vs. high torque in the air, first, that's obvious, as on the ground there's no need for much stabilization. And second, none of Jason's post speaks to anything other than standard stabilization.

And in none of this does it apparently raise a question in your mind as to why 3DR would go to the trouble of developing such an advanced and integrated system whereby information from the gimbal is integral in positioning the airframe, and then construct an airframe with legs that enter the camera's field of view when it rotates more than about 3 degrees to either side. Seriously, wouldn't it not have been simpler to employ retractable legs like numerous other platforms, or simply a different leg design/location? But I'm of course forgetting your propensity to allow the conclusion to drive the evidence.

And let me just state for the record, I'm not in any way criticizing the design of the Solo. With their true standout feature - the decoupling of flight controller orientation from air frame orientation, there's no need for independent camera yaw. The system works fine, and the only issue here is a minor manufacturing or calibration one where some gimbals point excessively to one side, for which they're happy to provide replacements. I'm not bashing the Solo in any way.

Lastly, you STILL haven't answered how you determined I likely didn't run calibrations. Come on Ian, be a man and just admit it was an entirely baseless claim pulled straight out of your arse.

 

[Ian [P13]]

@User Name You've shown your ignorance throughout this thread, you're wrong. I've included the detail from one of the most respected technical sources for the Solo within 3DR and still you continue to argue.

You should really do us all a favour and sell your Solo and buy a Phantom.

 

[pyrate]

Wow how many threads have we seen on this.
Just fly the thing, it works itself out
grab a phantom like Ian said, they are not nearly as sophisticated an you can figure that one out
The gimbal on solo is doing things Phantom cannot, and apparently not everyone can understand

 

[User Name]

@User Name You've shown you're ignorance throughout this thread, you're wrong. I've included the detail from one of the most respected technical sources for the Solo within 3DR and still you continue to argue.

You should really do us all a favour and sell your Solo and buy a Phantom.

Most anyone can see straight through you BS.

I'm done, and you're an idiot.

Just fly the thing, it works itself out

I, and others, have flown ours for months. Did every firmware update. It never "worked itself out", and 3DR has, for those of us who contacted them, acknowledged it was a defect. It's absurd to claim there is no problem and/or that it will work itself out.

grab a phantom like Ian said

Had I wanted one, I would have. I don't, and as I said, I'm not criticizing the Solo. This is a simple manufacturing issue that they are addressing. The way the Solo works outside this is perfectly fine.

The gimbal on solo is doing things Phantom cannot, and apparently not everyone can understand

Yes, it can control a GoPro and it's settings.