X8 build

[RTRyder]

Made a lot of progress in the last 24 hours, the ESC's I needed arrived in yesterday's mail so I spent last evening getting the necessary connectors soldered on, setting the programming options, then installing on the frame and connecting the motors. Around 9:00 AM this morning I caught a break in the rain showers we're having and managed to get outside for a test flight, had some issues with it wanting to yaw slowly to the right after it got in the air so I landed and brought it back in to check things out. Found I had somehow managed to reverse most of the top/bottom motor connections while I had the ESC's disconnected for programming, fixed that and tried again only this time it wanted to yaw in the other direction, hmmm...

Back in the house and hook it up to laptop for a look at the setup and had one of those DUH moments, the mixer table was still setup for a quad! Fixed that and finally had it flying normally. Initial flights were done with a 3S 2200 mAh pack and naturally the 8 motors sucked that down pretty quick so I landed it after less than 5 minutes before it ran out of juice. Rain started back up right about then so I decided to tackle the few things I still needed to do on the camera mount and that wound up taking 4 or 5 hours to machine the pieces and get it all reassembled and back on the frame. By the time I had all that done the rain had been stopped for a while so I went out for another test flight, this time with a 4S 5000 mAh pack.

Flew pretty good though I didn't see a lot of difference on the bigger battery, still has plenty of power and hovers at a bit below half throttle.The control response seems a lot better than what it had as a quad, and it's amazingly quiet, a lot quieter than either of the Hexa's so I would guess that has a lot to do with the stacked counter rotating motors. Still had the gain settings from the quad setup so I figured it wouldn't hurt to try some changes to see if it would make a difference so I came back in and hooked it up to the laptop again. While I was at it I decided to upgrade the firmware to the newest version, didn't like the results when I had upgraded to it in the quad config but a few people have said it works well with the X8 so I decided to give it another try. They were right it works a lot better on an X8 than it did on the quad, still has a lot of annoying quirks in it though and they messed up the tilt and roll compensation trying to fix the roll servo jittering from the last version, now it slows to a crawl as it nears level and takes way too long to settle out. If things are happening quickly then it tends to overshoot level and has to back up though the Avertical mount tries really hard to deal with it.

Went out for one last flight with the GoPro on the mount just to get a feel for what the mount was doing and it actually turned out fairly well other than the platform level was off. Yet another annoying quirk of the Hoverfly code, it changes the level setting once the board is fully armed which it can't be when doing the setup so you have to kind of eyeball it and do a lot of trial and error until you find the correct offset between the set point in setup and the set point after it arms so that the platform winds up actually level when the board is ready to fly. Did I mention the firmware has a lot of annoying quirks?

Anyway, I flew for a timed 5 minutes to get an idea of what the power usage is going to be like with the GoPro recording during the flight. Somehow managed to get the firmware gain setting pretty close for taking a shot in the dark and it flew well, auto level sort of worked, it only wanted to drift slightly rearward and a tiny bit to the right, so overall a very successful flight. I didn't tempt fate by trying altitude hold, it barely worked in this firmware version on the quad, didn't feel like I needed the excitement of watching it either drop 10 feet or shoot skyward when I flipped the switch so I leave that for another day.

So far I really like the way it flys, its quiet, has plenty of power, control response is outstanding, stability is very good, now all I need is a flight control board that works the way its supposed to. As soon as the I2C to PWM convertor I ordered last week gets here I am definitely going to swap in one of the MK flight controller boards and see how it flys with that, already have an X8 mixer table for it, just need the last piece of the puzzle to make it work.

I'm going to try and get over to a ball field near the house in the morning for a test flight with the video cam on the mount if the rain holds off for a while, I'm really curious to see how it does in a large area where I can actually fly it around.

Oh yeah, power consumption appears to be about the same as the AD-6 heavy lifter, the battery will hit 80% discharge at around 7 1/2 to 8 minutes, so I think I'll be flying a pair of the 4S 5000 packs in parallel and that should give me about 12 minutes flight time.

Ken

 

[matwelli]

well done, so do you like the x8 config ? is it stable, yet manuverable at the same time ?
Man im a glutton for punishment, have two X8 builds on the go, one with KKuk and one with Ardupilot !

 

[RTRyder]

well done, so do you like the x8 config ? is it stable, yet manuverable at the same time ?
Man im a glutton for punishment, have two X8 builds on the go, one with KKuk and one with Ardupilot !

So far it appears to be a good platform but I haven't flown it enough to really be able to get a feel for what it can do. The big issue at the moment is trying to get the right gain settings on the Hoverfly board to really lock it in, but so far it's been at least as stable as a quad and surprisingly manuverable for something this size. That's the one thing that takes some getting used to, this thing is BIG, the motor to motor distance is a couple inches larger than the Droidworx AD-6 and it's also a couple inches taller.

Late yesterday a bit before dark when the winds had died down I went out front and flew it up and down the street with the GoPro on the platform, first time I've had enough room to actually fly straight lines for any kind of distance. After the flight the video looked good veiwing it on the LCD screen, I was going to put it up on youtube but the GoPro somehow managed to scramble the file beyond recovery between when I watched it and put the memory card in my laptop. May not have been a fault of the GoPro, I might have absent mindedly pulled the card while the camera was still accessing the file thanks to being distracted by our dogs fighting in the other room I'll try again later today if the weather holds.

My biggest issue has nothing to do with the X8 itself, the problem is the buggy Hoverfly firmware and the way it works, or I should say doesn't work, with the camera mount. I've got it dialed in as well as the software will permit but it overshoots level and at times seems to want to do something completely different from what the frame is doing. The mount worked much better in the previous release but that one was beta for the X8 code, so I can have it fly crappy and compensate the camera platform well, or fly well and take whatever I get on video...

I'm hoping the I2C convertor arrives in the next day or two, I want to try the X8 with an MK controller which I know works well with this mount from the limited time I had it on the stock Hexa. Unfortunately the stock Hexa doesn't have enough power to fly the mount with a camera on it, the X8 on the other hand does quite well with it.

Ken

 

[Macsgrafs]

Hi Ken,

I was thinking of the hoverfly pro for a AD8HLE, but reading your posts maybe I should go for something different, especialy if as you said the GPS costs as much as the H/F pro!!!!
I'm a total newb to this game, so I'm going to be bugging you for a while

Ross

 

[RTRyder]

Had another go at getting the X8 in the air and stable enough to get a proper evaluation of the Avertical mount. I finally got a combination of props, firmware, and gain settings that seem to work fairly well but the Hoverfly board can drive a person crazy trying to make it do what it should. I really think it would be so much better if the end user could configure parameters in the software ala Mikrokopter rather than relying on just a gain setting to do the job. If I had the ability to tune PID settings I probably would have had this thing dialed in a while back, instead I've been futzing around with end point adjustments on the TX trying to find the closest gain setting that will impart some level of stability to my particular setup, ARRRGGGHHH...

In any case, I wound up with Grauper 11 x 5 props on the top motors and 10 x 5 on the lowers to get a speed differential between top and bottom and it appears to work much better than the same prop top and bottom. I did a quick eyeball calibration of the Avertical mount platform after reflashing the H/F firmware and then got outside for a test flight in front of the house between thundershowers. This is raw video direct from the GoPro memory card to Youtube, you'll see some odd flickering especially when the camera is turning, apparently Youtube has a hard time digesting 60 FPS video streams...

You'll also notice some twitchiness on the roll axis, this is a known firmware problem with the Hoverfly board, supposedly fixed in the next release, (and supposedly fixed in the last release as well, better but not fixed!) but overall discounting the software glitches this is superb tilt and roll compensation and exactly what I expected this mount to be capable of, now I just need a flight controller that can handle flying and camera mount compensation at the same time. For that I have just ordered up a Quadrino board with baro and compass plus camera servo ports, would be rather ironic if a $145 open source board can do the job better than the $450 Hoverfly whizbang board but I wouldn't be at all surprised if it does work better.

Anyway, here's the video, Hoverfly and Youtube warts included, best viewed in 720P HD...

Ken

 

[matwelli]

good stuff, interesting on the prop combination that works the best, will have to experiment, my motors turned up on friday, so its almost all systems go !

Happy to hear the avertical mount works well, mine shipped on friday

 

[Macsgrafs]

Ken, that looks quite stable to me, I have seen far worse than yours to be honest! Does MK produce a octo control board & if so have you tried it? The only thing that puts me off the HF pro is how does a camera operator control the gimbal if HF is already trying to keep it level?

Ross

 

[matwelli]

The HF has an input from an aux channel on the transmitter, this controls the desired tilt angle, the HF then compensates for the tilt of the multicopter. ( i actually went away and read the HF manual to find that out for ya )

Big problem is it only works on a fixed direction camera, if you YAW the camera in relation to the frame, stabilisation dosent works, and you will need a seperate stabilization system.

Ken, that looks quite stable to me, I have seen far worse than yours to be honest! Does MK produce a octo control board & if so have you tried it? The only thing that puts me off the HF pro is how does a camera operator control the gimbal if HF is already trying to keep it level?

Ross

 

[jes1111]

Grauper 11 x 5 props on the top motors and 10 x 5 on the lowers to get a speed differential between top and bottom and it appears to work much better than the same prop top and bottom.

That combination goes against all the theory, which says you should use a higher pitch on the lower (since your problem is wanting to get more "bite" on the faster moving airflow from the top prop). I understand that a smaller prop has less drag therefore it will run slightly faster, but the smaller swept area means it would have to run a lot faster to produce the same thrust as the upper. (c.17% reduction in swept area). Have you tried with higher pitch 11" on the bottom?

 

[RTRyder]

That combination goes against all the theory, which says you should use a higher pitch on the lower (since your problem is wanting to get more "bite" on the faster moving airflow from the top prop). I understand that a smaller prop has less drag therefore it will run slightly faster, but the smaller swept area means it would have to run a lot faster to produce the same thrust as the upper. (c.17% reduction in swept area). Have you tried with higher pitch 11" on the bottom?

A lot of MK guys run the lower motors 10% faster with the same prop top and bottom, there's a lot of theorys as to what works and what doesn't. What I wound up with is a result of using what I had on hand and unfortunately Graupner doesn't make the E-props with different pitch in both standard and reverse types or I would definitely try it. The concept as I understand it is to move move air at the lower motor either by increasing the speed or using a prop with more pitch, a smaller prop is as close as I can get the current setup without spending a bunch of $ for more props. If I can find the time I may test it with one of my Mk flight controllers using an I2C to PWN convertor and a custom mixer file to spin the lower motors that extra 10%.

Ultimately I may convert this one to a standard octo or just break it down into parts and build something else from all of it. It's a really large machine with somewhat limited usefulness that overlaps capability I already have with my two Hexas. I really built it more as a test bed for the camera mount and to satisfy my curiousity about coax setups than as a multi I intend to keep and fly long term. Its been an interesting experiment and I can see the value of having redundancy but it also opens up another set of problems and complexities that I just don't have time or desire to deal with.

Ken

 

[Crash]

....snip...I just need a flight controller that can handle flying and camera mount compensation at the same time. For that I have just ordered up a Quadrino board with baro and compass plus camera servo ports, would be rather ironic if a $145 open source board can do the job better than the $450 Hoverfly whizbang board but I wouldn't be at all surprised if it does work better.

Ken, I hoping my Quadrino board will be shipped today. The MultiWii firmware is still in very active development though so I'm personally not expecting perfection at this point. I do expect it to be a fun 'basher board'.

I have not used a compensated cam mount with my current MWC setup. I think there may be one problem. I don't think there is any way to have a compensated mount and the ability to move the mount with the radio. I think that the Atmel chips that most use have run out of pins. That could possibly be fixed with a PPM sum receiver and a firmware hack.

Is there a servo signal mixer on the market that could do the job?

 

[RTRyder]

X8 build, the final chapter

I finally found the magic combination that has turned the X8 into what I had hoped for when I started this project. First mistake, using a Hoverfly pro board, just like when I had it on a quad the board proved to be nothing but headaches.

The combination that worked? Well, no big surprise here, a Mikrokopter 2.1 flight controller and a $15 I2C to PWM converter, modify a couple connections here and there, load a custom mixer file into the Mk board and it flys as good if not better than either of the two Hexas I have. With a bit of heft to it, RTF weight in the 7 pound neighborhood, it certainly is more stable than the much lighter box stock MK. I'd like to do a comparison to the Droidworx AD-6 but at the moment the AD-6 is a large pile of carbon fiber parts and electronics on the bench and has donated its F/C board to the X8 project while it awaits the remaining parts to finish making the Droidworx frame into a proper heavy lift version.

The correct way to setup a coax is to spin the lower motors faster than the upper using the same prop on both and I can say with certainty it makes a world of difference. I have the mixer table setup for a 10% difference and the smoothness is unlike anything I was able to achieve using the Hoverfly board. With the H/F I could see the ends of the XL landing legs vibrating at varying speeds as it flew, with the Mk I see no vibration at all so it appears the speed differential is key to making a coax setup smooth and as vibration free as possible.

For the second flight I setup altitude hold and connected the Avertical camera mount to the MK controller, mount looked like it was working really well from the ground if its decent tomorrow I'll put a camera on it and see just how good it is. I think this is going to be a winning combo and its why I built the X8 in the first place to be able to use this particular camera mount. Now I need to clean up some of the wiring and put the rest of the AD-6 electronics stack on it to see how well it works with GPS. Looks like I'll have to sell off some of the excess stuff I have hanging around to finance yet another MK electronics stack so the AD-6 can have its gear back once I get it rebuilt.

Ken

 

[jes1111]

The correct way to setup a coax is to spin the lower motors faster than the upper using the same prop on both and I can say with certainty it makes a world of difference. I have the mixer table setup for a 10% difference

Sounds like you've done comparative tests yourself, or are you quoting a trustworthy source? If this is indeed "the correct way" I shall be very pleased since it is obviously the most "tunable" approach.

 

[RTRyder]

Sounds like you've done comparative tests yourself, or are you quoting a trustworthy source? If this is indeed "the correct way" I shall be very pleased since it is obviously the most "tunable" approach.

Both actually, I did a lot of testing on my X8 in collaboration with another X8 Hoverfly pilot from Texas, and from much reading about coax setups I see most people running the lower motors anywhere from 7 to 12 percent faster. Other strategies are to use a higher pitch prop on the lower motor or some variation of diffrent props top and bottom to regain the efficiency lost due to the coaxial setup. My experience is that the increase in motor speed with the same props top and bottom work best for my setup.

Most of the MK mixer files I gathered from other sources are set to a 10% differential top motors vs bottom so that number generally seems to work the best for MK, other controller boards could fall within the range I've seen others use. I would think that the optimum setting would vary somewhat depending on the platform, motors, props, loading, etc.

Bottom line is the lower motors need to somehow make up for the loss of efficiency of the coaxial setup, spinning the motors faster or using some combination of larger/smaller props or different pitch top to bottom have all been used with varying degrees of success. From my testing I can say that simply spinning both motors at the same speed with the same props will make for a very vibration laden airframe, the speed difference has calmed the vibrations to point of no longer being able to see the landing gear vibrating in flight. If I can catch a break in the rain we've been getting the last few days I'll hang a camera on it and compare to the video taken with the original setup for levels of vibration as seen from the camera platform.

Ken