Video progress report!

[sutty]

Thanks Tom, good info there, for when I do attempt the OSD, which at this rate will be soon.

Not worried about the crash, especially now I know what it was, I'd flicked, by accident, that damned IOC switch. I am going to have to lock that down at a single value in the radio so it can't happen again. I cannot fly it with that on, it's impossible, completely impossible. I couldn't even work out that this is what was going on, no time. I only know now because just a moment ago, before going for a test flight, I checked my switches, and wouldn't you know it, that iffy broken switch, which is hard to tell where it's at, was set for IOC.

You know me though, can't stand to have a bird down. Fixed it right up in a jiffy. All that I had wrong in total was I had almost snapped the end of one arm off, lost a rubber foot and bent two props, one a little bit, the other worse. The plastic of the arm all married up prefectly and it was kind of long shears down each side, in different positions. It's the kind of plastic that CA melts, which was perfect. I smeared CA all over it until it melted into place, and then I taped it all up. I gave it two hours, and it was strong enough for phase two. I mixed some epoxy, and went over the whole joints inside and out with a fairly thick layer, thining it out as I went away from the joint. The beauty of the cracks not matching up, means that each side of the damage is supported with unbroken plastic.

Obviously I had to do a lot of disassembly, and reassembly, but its all done now and a twist test of the arms shows it to be as strong as the others. I've order a frame and props, and some other goodies, so I should be okay for next time.

Oh, the props that were bent. One a tiny bit, and one worse, I used the undamaged ones as a template, and bent them back perfectly, teasing them very gently until they matched. Pretty sure the structural integrity is good, as there is no sign of where it was bent, and no additional flex on that blade in hand tests. If it gives out on me I'm sure I'll notice as it plummets from the sky.

After that, I just left the epoxy to set for two hours, full set in 24 hours, and then re-assembled.

Went outside just now, everything went fine. Took off perfectly, hovered with little to no drift. Flew forward 20 feet, engaged loiter, and it was locked as normal. Flicked RTL, just to be sure, and slowly up she went, which is what I'd hoped would happen, since I was well below the RTL altitude. Once at the correct height she flew back the 20 feet or so and descended nicely for a soft landing.

Happy as a clam. Got my toy back. Must remember to fit lost model alarm. Stupid that I didn't do that whilst it was all in bits.

Cheers

Sutty

[jperkosk]

Originally Posted by Tomstoy2

The way most osd's work, they hop along the right audio signal.
You do not want any microphones attached on right audio, such as, if your fpv camera has a mike, most do, ( disconnect the right audio pin from the connector ), And the video transmitter,

You keep saying that, Tom, but my former OSD (a POS HK system that filled the screen with hard-to-read useless info) as well as the DJI iOSD system that I'm installing now during the hexa upgrade both just take a video input (output from the camera), overlay on it OSD info, and output a video signal with OSD info on it that gets connected to VTx. Audio in on VTx I simply leave unconnected. Why would you need a separate channel for OSD?

@Andy: I checked and that's what I have in my APM 2.5 kit for OSD, quite simple and inexpensive, as wel as untested by be, unfortunately :

https://store.3drobotics.com/product...nimosd-rev-1-1


http://store.3drobotics.com/products/ftdi-cable-3-3v

[sutty]

Good god, do I have to solder it? That will go well with my hands, not. I'll have to find one with the pins in.

I'll see what I can find out this week, as to what works with the 350Pro, as it doesn't look like an APM, even though it must somehow be based on it. If people have used that, then that's what I'll get.

Cheers

Sutty

[jperkosk]

Yep, you'd need to solder the pins, they came separately, I wonder why?

Sounds like a plan to me

My hexa is a couple of days away from being ready for re-maiden, will post in its own thread. OSD is incredibly good by comparison with the old HK system that I had, but APM's OSD beats it by quite a bit.

The info option set of what can be displayed is impressive:

[sutty]

Very cool. Just found out the telemetry isn't routed off the board to a connector in the side anyway. Just pads on the main board.

I'm not sure there is anyway I can solder to those little pads without screwing it up and ruining my Devo module. Wouldn't be so bad to ruin a telemetry module trying, but ruining the main controller would seriously cheese me off.

Also seemed to be a difference of opinion in what two people did, but if I decided to do it I could consult with them, or you, or others, to see what is needed. One suggested changing a surface mounted resistor for some reason, and I can promise that would be a complete impossibility for me, and they used a different ground pin on the OSD. One used the one marked black, and one used the one marked ground. I think they might be on two isolated different power sections on the board, but why they chose different options, I don't know.

I guess they were both making it up as they went along to a certain extent.

Little bit upset now.

[Tomstoy2]

You are correct, Jerry. A few of the new osd's carry both signals, ( osd ), across both right audio and the video channel itself.

The main purpose of this is for antenna tracking. Though, the opnly tracker that I am aware of that tracks thru audio at present is FlyMyDream, ( the more I study this tracker, the more I like it.

Of course, DJI, being a proprietary system, does things their own way and can't be adapted to fit other manufacturers without an electrical PHD.

[jperkosk]

Sort of makes sense

Only got master's in electronic engineering, no PHD, so I guess I won't be able to incorporate antenna tracking into DJI system

Anyhow, my whole new system is just bunch of improvements and corrections addressing weak points of my original hexa build, so I stayed away from long-distance stuff, and am still using the 5.8GHz link with circular polarization only. I want to have a relatively high-end mid-class multirotor AV/FPV platform with ~1km range for now, anything longer range will have to be a next project, and as I don't think it can feasibly be based on multi-rotors, I will probably have change to large helis as a platform; I still don't understand planks.

Looking forward to the results of your longer-range experiments, though, may come in handy some time next year

[Tomstoy2]

Hey, just like fbl, I'll be your huckleberry.

That's exactly what propelled me back towards planks.

Get yourself that ET Guardian I have. Foam flys like a block of foam until you put one of these on it. Cheap, too, and super easy to set up.
I swear, if it wasn't for it I would have hung up the Belt Penguin long ago. Now, I'm adding a pan & tilt for both camera's and converting to 1.3ghz!

I still haven't looked at my Hexa, still pissed at myself!

[sutty]

Hi Guys

Some pottering about over the last couple of days. First up some clips from flights that on occasion were quite windy, but it still coped pretty well. I think at 2.25 there's a buzzard, at least I think it's a buzzard, that takes off from under the quad, and flies off in front of it, as I then try to set off and follow.

Second a video with a few clips with calm weather. Longest trip so far, as mentioned about 250m. Yep just checked it on Google Earth and it was 247m, so not a bad estimate. Then same video, ending with my crash, and a shot of the barley, lol. You will see it sets off at a pretty good clip as I move out and forward to be able to see it around the trees, but my view is still not good. Reviewing the video I seem to have it at one point, but even though it looks okay on the video, it was still moving away, and I just couldn't take the risk of it going over the house with the red roof, which is in that direction, or anywhere else for that matter, plus I still couldn't see it that well as I was rushing out to follow it. I think I had no choice but to ditch it.

All things considered, since it's flying again, not too bad an outcome. Glad the tall barley was there to absorb much of the impact.





Cheers

Sutty

[Tomstoy2]

F***in switch!
It's a sinking feeling when stuff like that happens.
Can't you inhibit that switch in you tx set-up?

[copterboy]

First time I seen anyone use this hobby to chase some birds

[sutty]

Yes Tom, you can pretty much do anything with switch configs in the DX18. Just not looked at it yet to figure it out.

3 things to do before next flight.

1 Label it with phone number and reward.

2 Change the switch config so IOS cannot be selected no matter how much I try to screw up.

3 Fit a lost model alarm. I know they aren't fool proof, or long range, but it might help for just such similar circumstances.

Yes Denis, it's never normally worked for me either, not that I try, but can you imagine, "Do you have any interesting hobbies?" "Why yes, I fly toy helicopters, almost to the point of obsession."

I chased a little wagtail around the lawn the other day in one of the videos too, so I seem to be making a habit of it.

By the way, I don't seem to be getting bored of the scenery. It isn't really the point, it is just so cool to fly like this.

Cheers

Sutty

[sutty]

Phone number reward applied to the side. Switch config changed to have only one value by mixing it with itself.

Only thing left to do is to fit a lost model alarm. Sadly, the ones I had don't work. I knew that, but I'd forgotten. I ordered 3 very cheaply from China, on ebay, but they didn't work with Spektrum, so that is why they were sitting there.

Had the whole quad stripped down again to fit it, but of course it didn't work. I've ordered another one that works, like I have on the 250. Only £4 including carriage, so can't be bad. Should be here for Tuesday.

Cheers

Sutty

[sutty]

Not a terrific video, but it isn't to show off anything other than a few things I learned from GoPro Studio. It is a very cool and easy to use piece of software.

First thing I did was change the camera to use spot lighting, which I thought sounded like a good idea. Instead of using the average lighting over a large portion of the entire frame it uses the lighting in the centre of the frame to set the exposure. I thought this was a good idea because we are rarely interested in the edges, and the centre of the frame is where we look, and where the lighting should be set correctly. Just a trial, and not sure how much it helped, but I thought it sounded good.

Next, in GoPro Studio, I discovered that there is anti-fisheye. I also discovered that if you make a protune recording you can apply automatic protune colour correction, which looks very good. Needed for professional movie editing to get perfect matching between clips from different cameras, apparently.

Next I discovered that you can rotate and then zoom to crop the distorted edges of the rotate. Better to zoom so you keep the same aspect ratio.

Then export it with the automatic default best settings for compatibility with you tube, to save wasting time uploading a massive file.

Here it is. Fist minute raw, second minute, fisheye removed, protune colour correction, levelled, and zoomed a fraction to crop the skewed edges of the rotate, and all with my crappy PC. With most other editors my PC dies when I try to manipulate these videos.

Big thumbs up from me. Pity there wasn't more fisheye to make it more obvious, it was on narrow anyway, but you can still see it.



Cheers

Sutty

[jperkosk]

I wanted to see the difference but the whole clip is 42 sec, what happened?

[sutty]

Sorry Jerry, my bad. I thought it was longer that's all. You just need to watch the change from before 21 seconds to after 21 seconds. Two 21 second clips, and I just thought they were longer.

First half unprocessed, second half, same thing, with the adjustments.

Cheers

Sutty

[jperkosk]

Hehehe now that I knew where transition was it was easy to find. Quite a bit of a difference. May give it a shot and record in Protune on my hexa's re-maiden soon.

[sutty]

The bit that pleased me the most was rotate. You could do it with extreme precision.

[Tomstoy2]

Found a nice spot, about 20 minutes away, to fly.

Early morning, absolutely no wind. This is all line of sight, had a slight cable issue for the camera, so no fpv.
That's ok, as this was mainly to test out the new pan & tilt. It's a heavy beast, well built.
It needs to be, as landing into tall grass is like 15 mph to 0 in 1 second.
You don't so much as slide it in, she will tip stall if too slow. You more or less pick your spot and fly it into it.

Had a lot of fun on this flight. This is the total flight, so it's a long vid, I suggest you thumb thru it,

https://www.youtube.com/watch?featur...&v=I0JLgToUs90

[sutty]

Couldn't skip through Tom, too interested. It's the first time we've had chance to see an extended flight. I thought it was flying very well, or at least it felt like it for th emajority of it from the onboard perspective, but of course you'll know better from direct visual feedback. The only odd thing I thought I noticed was maybe a little high speed instability a few minutes in. Did you notice that, or was it nothing to worry about. As I watched I was wondering if you had a stability controller, and maybe it needed a tad more tuning?

Terrific looking spot, and when you do get to try it out FPV it looks like there might be some interesting terrain to over fly and inspect as you stretch your legs.

I managed only one pack yesterday, as by the time the wind had dropped it left me really short of time. The flight was really just a battery capacity test, so I'll know how long I have with the 5200 packs, so nothing much to post up, but it is on film, just not reviewed yet. If there's anything there, I'll post it up.

Cheers

Sutty