Multi rotor camera mounts, the good, the bad, and the ugly

[RTRyder]

ken,
put the props under the arms. that's my plan for Son-of-Okto and I'd love to see how it helps a standard frame with airframe vibrations. it's an easy mod for standard frames, I just haven't had the time to pull my stuff apart when I'm still trying to get rev 1.0 done and flying.
good luck,
bart

Flipping the motors would actually be a very easy thing to do on the Droidworx frame, remove one bolt and rotate the arm 180 degrees, reinsert bolt. Problem there is I'd now have the motor and prop in full view of whatever I have mounted on the camera platform. As it is right now the end of the arm is either in view or dips into view occasionally depending on what resolution I have the GoPro set for. I wonder what the end result would be if I just flipped the motors on arms 2,3 and 5,6? Interesting to think about though, when flipped the motors would be pushing up on the airframe rather than pulling it up as it does now.

I stripped the Hexa down early this AM, takes all of about 10 minutes the way I have it built now with bullet connectors on the motor leads. Tried some rubber between the center plates and one of the arm mounts, doesn't seem to have much of a deadening effect using my high frequency testing methods. I took the motor off the mounting pad and was rotating it and noticed that just the action of the motor turning on the bearings and the magnets passing the windings can generate a fair amount of vibration by itself. I don't know that its possible to eliminate all the vibrations just due to the fact that there's six motors all buzzing away at the same time and continually changing RPMs, even if everything were precision balanced. I think the only real solution to the video buzz from vibration is down the route that PhotoShip one took with the mkTR rubber mounting system to stop the vibes from getting to the mount though that can present its own unique set of problems if the mounts aren't exactly the right stiffness.

At this point I'm thinking of digging out the stock frame and reassembling the electronics into it just to see the difference between it and a complete carbon fiber frame with the same lower unit and mounting system. If there's no significant change in the video with the Canon Vixia mounted then there's no point in trying to squash the vibes in the frame itself and time would be better spent devising a working isolation system between the upper and lower units. If there is a large difference then the problem is the carbon fiber frame and at that point I'd have to rethink what parts are going to be used to build the best video platform with what I have available. Judging by the video I took with the stock Hexa and a couple different camera mounts, I don't think the carbon fiber is to blame though, I think it's just the nature of the beast.

More experimentation to come...

Ken

 

[Bartman]

ken
you're right about the props whirring just in front of the camera lens. if you flip the motors to the bottoms you have to reestablish the correct order of things as to which props are turning which way and make sure it's all correct for the FC board. the esc's can turn the motor either way and the props just have to get flipped over to push down which isn't a problem if you have the 5mm adapters on the motors. i'm really curious is the props on the bottoms would make much of a difference. have you ever tried to scoot into a piston single airplane with the engine running? it's a load of force blowing from the prop even at idle. at full power the wash from the prop is giving the airframe a heck of a beating. i'd think a lot of that would be reduced if the props were below the arms. it's part of the reason some aircraft designers put engines in back and use pusher props, it eliminates a lot of airframe vibration in addition to a lot of cabin noise.
good luck with your experiments.
bart

 

[Xaircraft USA]

does the vibration issues tend to affect CCD as-well as CMOS Cameras Equally ??

 

[RTRyder]

does the vibration issues tend to affect CCD as-well as CMOS Cameras Equally ??

The short answer is no, though both are affected by vibration a CCD sensor isn't bothered by it as much as a CMOS. CMOS sensors get what is commonly known as the jello effect when subjected to vibration making the video appear as though you are looking at the scene through a bowl of jello as it's wiggling around. The proper term is rolling shutter artifacts and has to do with the difference in how CMOS and CCD sensors work in the camera. It's a bit of a long explanation and a quick google search will get you a number of pages with a full explanation.

Ken

 

[matwelli]

sorry to hijack the thread, the prop under the arms makes sense, will give it a go to reduce vibration.
I also guess efficiency may go up.
I have an eagle tree logger, planed to compare SF vs DD props (watts required to hover) and will do the same test over vs under frame motors.

 

[RTRyder]

So, much experimentation and tweaking has taken place in the last week and I think I finally have things pretty much where I want them. I swapped out the prop adaptors from the homemade Delrin inserts to a set of the correct 5 mm metal units with balanced Graupner props on them. Two of the Delrin inserts were drilled a tiny bit off center as it turns out and causing a fair amount of vibration, the first flight with the new adaptors and props was noticeably smoother. Just for the sake of experimentation I put the Canon Vixia camcorder on the mount and did a test flight to see if the vibration factor is any better and it was, considerably. Still not able to get usable video out of it, but the image isn't a total blur like it was before so I know I'm getting closer.

While I was tinkering with the electronics I upgraded the firmware to .82b, wiping the eprom on the F/C in the process to make sure nothing remained of any of the previous settings. Did a complete recalibration before flying and only changed a few parameters back to where I had them set previously, I wanted to be as close to out of the box settings as I could get to see if that made any difference in the way it flys, can't say I really notice anything different.

Next piece was to install the new FPV TX and wire everything up for live video out of the GoPro. The TX I was using last season found its way onto the Hoverfly quad and from there to the parts bin after the Hoverfly did a series of 5 aileron flips while falling from roughly 100 feet up. Amazingly it only broke one prop and smashed the FPV TX, nice of the TX to sacrifice itself to save the quad...

Anyway it was pretty obvious the TX wasn't going back on the Hexa so I ordered up a pair of new ones, one to live on the Hoverfly and the other to go on the Hexa, now that job is done.

While all this was going on I was also doing testing and experimentation with getting the best video results I could using the GoPro and trying out several editing programs for post process. It became pretty clear that it was either learn to live with what I can get from the GoPro, or dig deep and spend a LOT more $ to get a better camera that will work on the existing camera mount without being done in by the vibration that doesn't seem to at all bother the little GoPro. I'm still playing around with the editing programs, seems the ones that really do all the things I would like are silly amounts of $ and I'm not spending many hundreds up to a thousand or more just to be able to make the videos look good, I'll get the best quality raw video I can so it doesn't need major work on the backend. Part of that solution is dialing in the flight control settings a bit better and practicing smoother flying. Another part is some modifications to the GoPro itself and I'll have more to say on that subject soon.

Ken

 

[Bartman]

ken,
a separate thread on gyro settings would be a very welcome addition to the site. in return for your generosity i'll work on getting a better selection of smileys for us.
bart

 

[RTRyder]

I'll put something together on flight controller settings soon, still have some Germanglish to translate into readable text.

Things are coming together nicely, I'm down to fine tuning of the settings to get everything perfectly smooth. Here's the last test flight done yesterday afternoon, other than conversion to WIndows format then posting to Youtube, this is untouched raw video from the GoPro @ 720P 60FPS...

http://www.youtube.com/watch?v=-zNbXcvRE3k

Ken

 

[Bartman]

this thread has taken a few different directions but I'd like to offer another angle on the issue of vibration. at first glance the multi-rotor copter seems like an ideal platform for AP as it has positioning capabilities, gyro stabilization, altitude hold, loads of power to loft heavy cameras into the air...but... it is also a lifting device with absolutely no ability to maneuver without tilting itself and creating a horizontal component of lift. it has to roll or nick to give itself the ability to move horizontally across the sky and each one of those little tilts and un-tilts (for every tilt there's got to be an un-tilt, no?) has to be accommodated by the camera mount. i think that that is asking a lot and i think that that is why large scale RC helicopters continue to be the gold standard. correct me if i'm wrong but helicopters have their rotors uncoupled from the airframe via the pivoting mast they use for control. the rotors may be constantly pivoting but the helicopter's fuselage remains hanging below the rotor with lots of mass to absorb whatever vibration there is. so, to me, it looks like a traditional helicopter may put some of the rolls and nick for control into the camera mount but not nearly as much as a multi-rotor will.
i've got an idea to deal with this and it's outlined in the beginning of my Son-of-Okto thread so i won't get into it here, but i'd love to hear denny's response to this as he's the one with the real experience.
i think a lot of the solution will come when we un-couple the mount from the millions of little control inputs the multi's make just to remain under control.
just my 2 cents.
bart

 

[RTRyder]

this thread has taken a few different directions but I'd like to offer another angle on the issue of vibration. at first glance the multi-rotor copter seems like an ideal platform for AP as it has positioning capabilities, gyro stabilization, altitude hold, loads of power to loft heavy cameras into the air...but... it is also a lifting device with absolutely no ability to maneuver without tilting itself and creating a horizontal component of lift. it has to roll or nick to give itself the ability to move horizontally across the sky and each one of those little tilts and un-tilts (for every tilt there's got to be an un-tilt, no?) has to be accommodated by the camera mount. i think that that is asking a lot and i think that that is why large scale RC helicopters continue to be the gold standard. correct me if i'm wrong but helicopters have their rotors uncoupled from the airframe via the pivoting mast they use for control. the rotors may be constantly pivoting but the helicopter's fuselage remains hanging below the rotor with lots of mass to absorb whatever vibration there is. so, to me, it looks like a traditional helicopter may put some of the rolls and nick for control into the camera mount but not nearly as much as a multi-rotor will.
i've got an idea to deal with this and it's outlined in the beginning of my Son-of-Okto thread so i won't get into it here, but i'd love to hear denny's response to this as he's the one with the real experience.
i think a lot of the solution will come when we un-couple the mount from the millions of little control inputs the multi's make just to remain under control.
just my 2 cents.
bart

Actually helicopters are not at all unlike a multicopter in how they maneuver, they just do it with a single rotor. On a traditional heli the swash plate is the component that tilts and causes directional shift, the rotor shaft and rotor head assembly are fixed in relation to the fuselage of the heli. The swash plate can be tilted fore/aft and side to side plus any angle in between those positions. What happens is as the swash tilts it increases the pitch of the blades slightly as they sweep past the position opposite where you want to go and decrease the pitch slightly as they sweep past in the direction you want to move in.

The end result is the amount of lift varies from one side to another just like a multicopter and the heli moves in the direction of the lesser amount of lift. The fuselage does tilt that way also though the amount of tilt may be very slight compared to the amount a multicopter might tilt for the same degree of movement. That is one of the reasons why a heli is generally a better platform for A/P work, it can make small movements with less off axis motion than a multicopter can and teh variable pitch tail rotor allows it to rotate about the main axis with less positional disturbance than on a multicopter. Another big reason is the degree of stability generated by the large spinning single rotor vs. 4, 6, or 8 smaller rotors and also the ability to autorotate safely to the ground in the event of a engine failure, making it a slightly safer way to put expensive camera gear in the air.

Bottom line is that no matter the platform, heli, multicopter, airplane, or even a balloon, there is always going to be angular variations of the flight path due to the 3D nature of flight as well as variations of the wind, so whatever method is used to carry camera gear it is going to have to contend with those changes.

Ken

 

[RTRyder]

this thread has taken a few different directions but I'd like to offer another angle on the issue of vibration. at first glance the multi-rotor copter seems like an ideal platform for AP as it has positioning capabilities, gyro stabilization, altitude hold, loads of power to loft heavy cameras into the air...but... it is also a lifting device with absolutely no ability to maneuver without tilting itself and creating a horizontal component of lift. it has to roll or nick to give itself the ability to move horizontally across the sky and each one of those little tilts and un-tilts (for every tilt there's got to be an un-tilt, no?) has to be accommodated by the camera mount. i think that that is asking a lot and i think that that is why large scale RC helicopters continue to be the gold standard. correct me if i'm wrong but helicopters have their rotors uncoupled from the airframe via the pivoting mast they use for control. the rotors may be constantly pivoting but the helicopter's fuselage remains hanging below the rotor with lots of mass to absorb whatever vibration there is. so, to me, it looks like a traditional helicopter may put some of the rolls and nick for control into the camera mount but not nearly as much as a multi-rotor will.
i've got an idea to deal with this and it's outlined in the beginning of my Son-of-Okto thread so i won't get into it here, but i'd love to hear denny's response to this as he's the one with the real experience.
i think a lot of the solution will come when we un-couple the mount from the millions of little control inputs the multi's make just to remain under control.
just my 2 cents.
bart

Actually helicopters are not at all unlike a multicopter in how they maneuver, they just do it with a single rotor. On a traditional heli the swash plate is the component that tilts and causes directional shift, the rotor shaft and rotor head assembly are fixed in relation to the fuselage of the heli. The swash plate can be tilted fore/aft and side to side plus any angle in between those positions. What happens is as the swash tilts it increases the pitch of the blades slightly as they sweep past the position opposite where you want to go and decrease the pitch slightly as they sweep past in the diection you want to move in.

The end result is the amount of lift varies from one side to another just like a multicopter and the heli moves in the direction of the lesser amount of lift. The fuselage does tilt that way also though the amount of tilt may be very slight compared to the amount a multicopter might tilt for the same degree of movement. That is one of the reasons why a heli is generally a better platform for A/P work, it can make small movements with less off axis motion than a multicopter can and the variable pitch tail rotor allows rotation about the main axis while remaining in position better than a multicopter can. Another big reason is the degree of stability generated by the large spinning single rotor vs. 4, 6, or 8 smaller rotors and also the ability to autorotate safely to the ground in the event of a engine failure, making it a slightly safer way to put expensive camera gear in the air.

Bottom line is that no matter the platform, heli, multicopter, airplane, or even a balloon, there is always going to be angular variations of the flight path due to the 3D nature of flight as well as variations of the wind, so whatever method is used to carry camera gear it is going to have to contend with those changes.

Ken

 

[Bartman]

.....The fuselage does tilt that way also though the amount of tilt may be very slight compared to the amount a multicopter might tilt for the same degree of movement. That is one of the reasons why a heli is generally a better platform for A/P work, it can make small movements with less off axis motion than a multicopter can......

that's the point i was trying to make. i have an idea that would decouple the camera mount from the multicopter but it wouldn't do too well, without significant development, if you want to fly fast and maneuver a lot. if i ever get Son-of-Okto built, i'll have a go at it.

 

[Droider]

Just found this. Not sure where it should be posted but its helped me as I was just about to post a question on exactly this subject!

Move it to where you see fit Bart. I am sure it is a useful piece of information. Interesting what Quardrocopter have said about voltage

http://www.droidworx.com.au/camera.html

 

[RTRyder]

Just found this. Not sure where it should be posted but its helped me as I was just about to post a question on exactly this subject!

Move it to where you see fit Bart. I am sure it is a useful piece of information. Interesting what Quardrocopter have said about voltage

http://www.droidworx.com.au/camera.html

I bought one of the 6 volt recoms being offered as a direct replacement for the existing 5 volt unit by our friend Ziggy. Last night I put the Droidworx Mk on the bench and it took roughly 20 minutes to strip the stack down, swap the recom, and put it all back together. When I powered it up it seemed that the mount was responding quicker with the full 6 volts going to the servos, a test flight a few minutes ago pretty much confirms it. I had the mount dialed in nicely before the swap, now the video is appearing a little twitchy so it's time to go back into MKtools and tweak the settings again to damp down the response a bit and get back to where I was.

I've found the servos being used on a mount, and how they respond to changes in pitch and roll can make a world of difference. On the Hoverfly quad I've got a low end HiSight II mount now with a set of HiTec 5065 digital servos that I programmed for travel and centering for the requirements of each axis of the mount. I was quite surprised to find with the careful setup I did afterward, I'm getting results that equal what I can get from the Droidworx with high end mount

Now that has me thinking about a lot of things like Hexa vs. Quad, heavy vs. light, intended use, flying area, etc. One thing about the testing I've been doing is that it's primarliy very low altitude with a lot of significant elevation changes and obstacles in a small area contributing to a lot of ground effect being added into the mix. I think the Quad first of all doesn't generate as much downdraft and consequently doesn't get bounced around as much by ground effect at low altitude. Second, being significantly lighter than the Hexa it doesn't take the Quad as much engine rpm variation to either maintain or change altitude so there's less vibration to deal with. I could go on but I think you get the idea here.

There is a definite application for the larger Hexa, but also for the Quad in places where the larger craft is just too unwieldy to be able to fly it smooth enough to get the results I expect. Bottom line is getting a feel for the capabilities of each and then determining which is the right tool for the job. I know in a large area or where I need to have GPS and or position hold the Hexa is the better choice, but for close to the ground where being able to manuever through and around a lot of obstacles is needed the Quad is a much better choice.

The biggest surprise of the whole exercise is how well the little HiSight II can be made to work when the whole package is setup correctly!

Ken

 

[Droider]

Ahh so which way do I go?? Pro Mini or HighSite II?

I just been out and run my first full on test of my ADX3.. WOW I am going to post it in the build section

 

[RTRyder]

Ahh so which way do I go?? Pro Mini or HighSite II?

I just been out and run my first full on test of my ADX3.. WOW I am going to post it in the build section

Depends on what you intend to do with it. If you're only ever going to shoot stills the HiSight will be fine as would the Pro Mini, but if you're looking for really good video with a minimum of fuss making it look good on the backend, then you want a mount like the Pro Mini.

I'm getting decent results doing video with the modified GoPro on the HiSight, but it will never be professional quality, for pro level results you need a Pro Mini or equivalent and a lot of time spent dialing in the F/C settings to work best with the way you fly.

Ken

 

[Droider]

I need the video quality so it looks like the ProMini. I would be interested in how you have set it up in MKtools

 

[RTRyder]

I need the video quality so it looks like the ProMini. I would be interested in how you have set it up in MKtools

I'll let you know once I get it fully dialed in, it's going to need some tweaking now with 6 volts going to the servos. I was also told by Rob from Droidworx that Kim at Photohigher is spec'ing new servos for the mount that are better than the upgraded ones my mount came with and they will be sending me a set once they finialize which ones they'll be using. Knowing that, I'm not going to spend too much time getting it perfect to only have to repeat the procedure in a few weeks time when the new servos arrive.

I think from now on I'm going to start charging vendors for being what amounts to a Beta tester, it's darn near a full time job keeping all these things working right when they keep changing them...

 

[Droider]

In this world of hi tech stuff, which moves so fast we are ALL Vendor testers!!

 

[Droider]

Ken.. can the highsite II be fitted to the Droidworx landing gear?

I cant seem to get any response from anyone about the avialibiltiy of the ProMini