Jump to content
Daniel Wee

First few flights - test run FPV

Recommended Posts

Today, I finally managed to get a proper test-run with my rig. I cannibalized a no-name China made video extender (2.4GHz 500mW) for this test. The unit was a bit bulkier than the AWM modules but it worked and I managed to get it into the plane - a Cessna trainer with a 39 inch wingspan. The camera is fixed pointing forward just behind the cowling which I sprayed black for contrast purposes.

I had a video camera setup to record the flights and brought the plane up high and started looking at the camera to fly a bit, then back to visual. I was afraid I might lose sight of the plane if I looked down too long.

Some of my initial thoughts:-

1. Video link is very critical, obviously. Even momentary fade-outs can be very stressful as it can be hard to spot the plane. I have goggles but have not flown with them as my video link is suffering some problems. I have some Wi-Fi interference and I think the receiver or transmitter is drifting.

2. The plane is very twitchy in the response. I may need to add some expo to the controls but what appears like a smooth flight from the ground may be a very bumpy ride and this is made worse with strong wind.

3. It is quite hard to tell if you are going up or down (or slipping sideways) without extra telemetry. I'm not sure how you guys are dealing with this aspect. How do you guys maintain good control in strong winds?

4. Cloudy days are good for visibility, and it seems to work best when the sun is high up in the sky.

5. Overall, it isn't as easy as it may sound, to fly FPV, at first, but it is very exciting. I think I need a more stable platform as the little Cessna is having a hard time in the wind. Maybe with a bigger wingspan, and more expo in the controls, will make it fly more like my Flight Sim games.

I'm still editing the video of my bumpy ride.

Daniel

Share this post


Link to post
Share on other sites

Oh boy! Awesome another video! I can't wait to see your video Daniel! I never had success with those Chinese made A/V RX/TX, so I would love to see your results after you put it through your surgery room! ;)

Share this post


Link to post
Share on other sites

Okay, here's the video - hope it works.

FPV Test Flight

There some tearing in the video - looks like interference but I'm not sure what's causing it. I'm guessing it might be the PLL losing lock, due to oscillator drift in either the transmitter or the receiver. I need to do more testing but I think the horizontal green lines are due to Wi-Fi (lots of that around here).

Daniel

Share this post


Link to post
Share on other sites

OK, let me give it a guess... you have a KX131 and a panasonic camcorder, possibly GS400?

Share this post


Link to post
Share on other sites

Actually no - the camera is another nameless AVC566SP - I don't even know which company makes it.

It's specified at 12V +-10% but I'm running it off a 2S 850mAh Li-Po. Gets a bit warm but I'm surprised it worked at so low a voltage.

As for the camera - actually I am using a DVX102B - a bit overkill but it's what I have here at the moment. But you're pretty close with the PANASONIC guess!

If it wasn't for the compression artifacts and the interference, the picture is actually pretty good. The camera seems to run as low as 5V. I have another one here - the AVC666SP - a SONY 1/3 inch camera. The picture quality is pretty good - wider FOV and better (more saturated) color, but there's some color noise. The main problem is that the camera is a bit bigger and heavier.

Daniel

Edited by Daniel Wee

Share this post


Link to post
Share on other sites

OK. Why I said that is because I had that combination (KX131 and NV-GS400) and had the exact same red/green lines problem. The conclusion was that they didn't like each other... I could use the GS400 to record anything BUT the KX131's image, and I could record the KX131 with anything BUT the NV-GS400. Just the 2 together would not work. Maybe try recording (or at least viewing) with another device. That might fix it.

I have another older Panasonic (NV-DS38) which doesn't have that problem, but I haven't tried newer ones. Maybe it's a general issue with Panasonic and some sources?

Edited by Kilrah

Share this post


Link to post
Share on other sites

Ahhh... how very interesting. Could be a voltage level standard thing. Since we know the Panasonic part of it, perhaps my camera is a rebranded KX-131?

I'll have to look for a non Panasonic recording device but I like the mini-DV resolution. It is possible that the video tearing is also due to that compatibility because when I first tested it with the goggles, there was no problem.

I was going to try the goggles alone today but forgot to bring the batteries for it. I'll see about testing it tomorrow but thanks for the heads-up. First time I'm reading about this.

Daniel

Share this post


Link to post
Share on other sites

Well, I've never heard of anyone else having that problem up to now ;)

The video tearing could be due to a slightly too high video level either on the TX or RX, but it doesn't seem that significant. Or maybe to the undervoltage, with some part of the circuit not having enough swing and staurating.

I had tried adjusting the level to help solve the line problem but it didn't change a thing. I just don't have an explanation, all the levels are good, nothing odd on the signal (and it works on any other monitor/recorder)... :unsure:

Share this post


Link to post
Share on other sites

Hi Terry, thanks for the input. I am especially interested in the 0.2V bias shift in the light level changes in the picture. I think this may be due either to a design problem, or a power problem.

Here's the theory - IF there is a regulator in the camera, and we are using the SAME basic modules (re-branded perhaps), then either the smoothing capacitor POST internal regulator may not be up to the job. Furthermore, I think the video cameras may be more sensitive to this kind of bias shifts than the goggles and so they don't really show up in the goggles. Either that or the goggles have some kind of flywheel design for the H-sync and thereby avoids the problem.

It's too bad I'm missing my DSO (scope) now or I would have been able to investigate this further. I'll try to see if I can arrange something. Let me know if you come up with any solutions.

Daniel

Share this post


Link to post
Share on other sites

Okay, I've done some further testing. I had two problems - one was a gradual loss of signal which I think I've traced to my 78L05 regulator for the receiver, which might have been underrated. It was getting pretty hot and I think it was either shutting down (thermal shutdown) or the voltage was drifting sufficiently to cause the PLL trouble. I've now replaced that with a regular 7805 and it seems to be working better. One down (I hope).

Video Tearing

The video tearing - I've ascertained that this is not due to the camera itself. A different camera yields the same problem. The issue occurs primarily during a transition from very bright to very dark in the horizontal. So a bright sky and dark ground, for example, would result in the "tearing". The camera itself does not have this problem, only when fed through the video link. The tear, occurs for a short while, usually, before it sorts itself out. It would appear that this phenomena affects not still scenes but scenes involving changing brightness (very bright) in the horizontal. This seens almost certainly to be a capacitor related issue somewhere in the circuitry.

Possible causes

If my observations are correct, and if we assume that this problem is occuring primarily at the transmitter end of things, then we can make some guesses as to what might be causing this video tearing - which is basically a loss of the negative going H-sync signal. I think such bright/dark transitions result in a sudden voltage swing that is big enough to throw off one of the biasing/coupling capacitors. If this is the case, then maybe, just maybe, by placing another capacitor across the supply rail closer to the signal generation circuitry may alleviate the problem somewhat. I think I will try this.

Terry, perhaps since you have this problem and apparent have a access to a scope, you can check to see where the 0.2V bias is occurring. Is this happening at the transmit side or the receive side?

Another possible solution might be to put an ND (neutral density) filter in front of the camera to bring overall brightness down, possibly closer into the design parameters. It may be that the camera was not designed to operate in such bright light and we pushing the circuitry. This is also quite easy to test. [Edit] Ok. Scratch that idea. I just tested it and it does not help at all.

I'm off to do more testing and will report back here if I get good results. If you guys have other suggestions, I'd be most happy to hear it.

PROBLEM SOLVED

Wahoo! Problem solved. As it turned out, some of my guesses were quite close. Firstly, this is something on the transmitter side. In fact, I had another problem being that the color of the camera was a little washed out - which pointed to voltage levels being a little high. In any case, I started tracing the video signal from the camera through the modulation stage and bumped into a small variable resistor on board. It appeared to be used to adjust the signal going into the FM modulator and so I thought to myself, this could make a difference, better than putting an ND in front of the camera (I used my wife's Oakley's for the test earlier - didn't work). So I started lowering the setting and immediately saw an increase in saturation, higher contrast and slightly darker picture. So, I'm thinking, "Hey, this is nice. I should test the transitions again." Lo and behold, the problem was gone. I have now adjusted it so that I am just where the problem doesn't occur anymore. The good side effect is that the picture looks much better now and closer to what it would look like if I fed the signal direct from the camera to the goggles.

Another thing I noticed was that the video monitor that I was using handled the tears better than the goggles did. It was an analog unit and seem to have better tolerances for out of specification signals.

So, Terry, hopefully this may give you a clue to solving your problem - it seems to be a pretty easy fix even if you don't have a built in variable resistor. You just need to trace the path of the video signal to the modulator. Now, the wind has picked up and I can't do any flight testing with this, will have to wait for the evening before giving this a go. Two down!

Daniel

Edited by Daniel Wee

Share this post


Link to post
Share on other sites

I'm glad you found the answer Daniel, I adjusted the same pot but found it was not a complete cure.

I have changed cameras now and the problem is gone, the odd thing is the old camera works fine when connected direct to a monitor or VCR.

Terry

Share this post


Link to post
Share on other sites
Another possible solution might be to put an ND (neutral density) filter in front of the camera to bring overall brightness down, possibly closer into the design parameters. It may be that the camera was not designed to operate in such bright light and we pushing the circuitry. This is also quite easy to test. [Edit] Ok. Scratch that idea. I just tested it and it does not help at all.

I'd have said it wouldn't work too. If you put a filter on the cam, it will see a darker image and its gain control will simply increase exposure to find the same level again, which for it is the right one.

PROBLEM SOLVED

I think I have to be sorry on that one, I just read I simply didn't give enough details... When I said "The video tearing could be due to a slightly too high video level either on the TX or RX, but it doesn't seem that significant", I meant in my head "Try to adjust the video level pots in the TX or RX". I just didn't write the sentence :(

All transmitters I've seen have an adjustment, and that's what we usually check first when there's an image problem...

Share this post


Link to post
Share on other sites

Nothing to apologize about Kilrah - I understood what you meant even though you didn't write it out :) We're all just experimenting and learning here so even having a problem like this is good as long as we can find a solution. Now I need to figure out what is causing those green lines on my Panasonic DVX102B.

I'm going to test this thing in flight later on and will report on the effects of the changes. Maybe I can record another clip and post it here too if anyone is interested.

Daniel

Share this post


Link to post
Share on other sites

I just got back from flight testing the changes I made and am happy to report that all the problems are solved - good video, strong signal lock. What a thrill to be flying that plane on video. The only thing that made it less fun was the strong winds which made me work to keep control of the plane.

Now, on to the telemetry side of things.

Daniel

Share this post


Link to post
Share on other sites

Just some updates on my ongoing project.

The changes I made to the camera in order to prevent video tearing in very bright situations where there is sudden big change in horizontal contrast seems to be dependent on actual amount of light available. After I made the changes yesterday, I flew the setup in bright hot afternoon sun and everything worked great. However, this morning at first light, I was testing the setup again and to my surprise, the problem was now occurring in the darker areas - sort of like the opposite of my initial problem. Once the sun got higher and there was more light, the problem went away.

My conclusion, therefore, is that given the dynamic range of the transmitter input range, or perhaps due to it's weak design at the modulator, it can be optimized to work in bright OR dark situations, but not both. I am wondering if there is a design change I can make to completely eliminate this problem - perhaps by scaling down the camera Vp-p range (compressing). I probably will lose contrast that way though. I'll see if I can live with this because it only affects it when really dark and it was still flyable - just needs a bit more light.

Apart from that, I have now clocked about 30-minutes or more of FPV or partial-FPV flight and am slowly started to relax and get used to navigating through the video feed. I still have a problem with determining if I am climbing or descending and I think the wide angle lense (2.6mm) I am using is not making this part easier. The WA lense does make overall navigation easier though, and gives me better SA (Situational Awareness). In fact, at this point I don't really feel the need to pan/tilt the camera, maybe when I get the hang of this and relax even more I will want to look around a bit more.

While I was flying and looking at some forests, I nearly flew into the treetops! Good thing I looked at the plane visually. Also in today's flight (alone) I lost visibility of the plane a few times, forcing me to fly back to familiar landmarks - which by then I had fortunately started to identify. Twice this happened but I am getting more confident in finding my way around. As I said, I just need some way of determining how high I am so the GPS/OSD thing will have to be done soon. I already have the 5Hz GPS module and am waiting to get the BOB-4 (which I hear has a problem with it's harmonics interefering with GPS reception thus requiring modification to it's oscillator frequency) OSD. I will be using my own controller and software to manage the data.

Also, the Cessna I've been flying has been a great workhorse but is starting to show signs of wear and tear. It has never been crashed and every landing has been perfect. The foam is not holding up well though so I guess I should start work on the Easy Star soon, or buy an Easy Glider.

Finally, the transmitter/receiver is working remarkably well, losing signal only when I am flying above the receiver (both using the supplied whip antennas). I have flown out as far as 400m, maybe more, and the signal has been rock solid. The best part is that I only paid about US$60 for the 500mW combination. That's less than a third of the price of the systems out there (albeit they're smaller). I will be looking forward to comparing their performance when I receive those (500mW and 200mW Lawmates) in a few weeks time. In fact, I saw a 2W set in the shop as well but I didn't want to fry my brains or worry about power supply issues so I went with the 500mW solution. Both have been running cool - the TX (and RX) runs off 5V through a 7805 regulator which I happened to have lying around. It also comes with cables and wall-plug adapters to power both units as a bonus. The TX is powered using a 2S Li-Po which is running surprisingly well and long on a small (850mAh) battery. Very little heat is being generated by the combo.

That's it for now. I'd be happy to take questions on the setup. I am thinking of trying another camera today. My wife and kid comes back from their shopping holiday tonight so my free time will be less after today :)

Daniel

Share this post


Link to post
Share on other sites

More updates and camera tests

Okay, I've done more testing and here are the results.

Firstly, more about the video "tearing" - it seems that the problem is still there but is much reduced in it's impact, so while there is some video offsetting, the picture is largely viewable.

Secondly (and not unexpectedly), I found that using a whip is fine if you're not going to fly really high and up close. Basically once you get high up, it's easy to move outside of the donut-shaped lobe of the whip antenna. This is not so much a problem when flying low and out front. I'm wondering how you guys normally aim the 8dBi patch? Do you have to keep moving it all the time or do you fly inside of the cone?

Lastly, I mounted my second camera today - a 1/3-inch CCD using a SONY chipset. The fact that it is a 1/3-inch CCD means that the FOV (field of view) is naturally wider than that of a 1/4-inch CCD with the same lense. This allowed me to dispense with the WA lense I had been using with the smaller sensor. I could immediately see the improvement, better resolution, better detail, and better colours! Wow, I can actually see the grass and other ground detail now which was lost to me earlier. I think I'm going to stick with this camera since the extra details really help with the depth perception besides making it a lot more fun to fly. The AGC behavior is similar to that of the other camera, if not a bit better. The only down side is the slightly bigger size and added weight. I'll do more tests with this camera - and may have to re-adjust the TX modulator to match the behavior of this camera.

And here's the video of the test - I have made some color corrections that makes things look darker than they really are, and needless to say, the compression takes a way a lot of details visible in the original footage (mini-DV). The horizontal green lines are due, apparently, to some incompatibility with the recording device.

FPV Test 2 (new camera)

Daniel

Share this post


Link to post
Share on other sites

I made further adjustments to the TX modulation level and with today's test flight, the video is perfect in bright sunlight - not a hint of video tearing throughout the flight and in all orientations. I will try again this evening in poorer light and then confirm that the video recorder doesn't have problems either. So far it looks very very good and I had an enjoyable FPv flight today.

Daniel

Share this post


Link to post
Share on other sites

Glad to see you are making progress Daniel. Originally I was going to use a dual board Sony 420tvl 1/3" CCD camera with a vari-focal lens. Who knows I might in the future, but watching you do it successfully makes me feel more comfortable of going in that direction. I do like my KX131 for it's low power consumption, light weight, but it lacks fine detail in the distance. It's a trade off I guess, so if I do use my larger camera (BTW I can buy IR lens filter/polarized lens for it... maybe great for sunny days when the sun is not in it's highest position. ) I better shave off weight somewhere else on the plane.

Share this post


Link to post
Share on other sites

Hi JMS, I have to say that going from the 1/4-inch to the 1/3-inch CCD made a huge difference in detail and brings a whole new dimension to the flying experience. I can quickly identify specific ground details - for example - what looked like a small road on the 1/4-inch, turns out to be a railroad track on the 1/3-inch. Especially with the problem of the video tearing (hopefully) solved, I no longer have to worry about the video performance. I originally planned to order one of the 480TVL or 520TVL cameras but this is working out so well right now that I am thinking of putting that off till later.

The one other aspect is the color renditions. I've not really compared the two but on my initial tests, the 1/3-inch CCD had more saturation colors, which I like.

Tomorrow morning I will test the setup in dawn light conditions and I will know how well it really performs. When I get the results, I will continue to post the updates here. I will see if I can also record something so you can get some idea of what it will look like.

Daniel

Share this post


Link to post
Share on other sites
While I was flying and looking at some forests, I nearly flew into the treetops!

I find that I need to fly a bit higher than that so as I can recognise where I am. From low level every field or patch of trees looks the same.

I'm wondering how you guys normally aim the 8dBi patch? Do you have to keep moving it all the time or do you fly inside of the cone?

To start with most guys use a spotter to point the aerial.

Terry

Edited by Terry

Share this post


Link to post
Share on other sites

Just a minor update - I have tested the adjustment now in early morning conditions and there is not a single trace of tearing as far as the goggles go now, under any kind of light conditions. The camera works beautifully and the link is rock solid (except when I fly directly overhead but that's understandable). I now bring a chair with me to the field :) and had a nice relaxing flight today. Sorry there's no video because I forgot to bring the video camera. I'll try to do that on my next trip. I'm starting to feel that the field is a little too small (1km by 400m) and am thinking of venturing a little further, or maybe try a different field. For the moment, I'm regarding this as my training hours to get good at FPV handling. I still have not done an FPV landing though, and am wondering how tricky it will be.

I'm also trying to get started on the OSD and other telemetric aspects of the project.

Daniel

Share this post


Link to post
Share on other sites

Great Daniel! I'm looking forward to your samples because you are doing something I had thought about but never got around to try. I know your method has better potential of a more satisfying result! In the mean time I am putting together a mobile camera system (spare KX 131, rangevideo A/V TX) to mount on my reconnaissance plane. I'm going to fly anyways in the same field where my Twinstar went down, might as well try and hunt it down by air; besides I think it would make a funny video if I succeed! :lol:

John

Share this post


Link to post
Share on other sites

Finding a plane with video is VERY hard due to the low res, you usually can't tell the difference between a stone and your plane...

I've put my digital cam on my moviestar once to find a friend's plane, and flown it in FPV over the zone he thought it was around, flying zigzag to cover all the place from several different altitudes, and came back with about 700 pics. We tried looking at the pics at the field bit no luck. Even with this it took me about 1h30 analyzing the pics calmly on my PC later at home, isolating all pics that had spots that could be it, but not all were, overlaying them over each other and some google earth images to have all levels of detail. We finally found it :)

The problem is that when you have a pic on which you can see it well enough to be sure it is it, it's so low you only see about 5m each way so even if you see the plane, you still have no clue of where the place is, and you need different levels in order to be able to use other references to step by step get to a general view.

http://i18.photobucket.com/albums/b105/Kil...ilrah/avion.jpg

The plane is in the red circle. But the 2 spots below have mislead us for a LONG time. We had seen both photos, the one with the plane that is clear enough to recognise it in full size, and the one with the spots. So we directly went to check those 2 spots, but didn't find a thing (didn't even find out what those spots actually were BTW). We certainly were at ~20m of the plane at that point, but forfeited there. When watching in detail, the photo that has the spots has a little thing where the plane actually is, but just barely noticeable unlike the spots so we totally missed it until I tried to match the field patterns more precisely... Hard work :(

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×