Jump to content

camera movement follow headset movement


Recommended Posts

  • Replies 286
  • Created
  • Last Reply

Top Posters In This Topic

Not a bad price on the headset, I'd like to know how they send position information out. I'm going to guess USB, possibly serial, but more likely USB.

I was shocked and scared at the prices they wanted for some of their "tracker" devices.

Quick question about the signal into the student port, PPM or PCM? I have input my own PPM signal into the student channel before to use one of the older flight yokes for a computer with the pedels for rudder. Worked great. I just didn't take the time to take apart the PCM signal and do it that way. With what I know now and the help I can get with writing C microcontroller code it wouldn't be any problem.

I remember 1 quirk with the 9C I was using was that it only had 8 out of the 9 channels in PPM mode, you had to be in PCM to get all 9 channels. Still hold true?

Thanks,

Matt klarich

Link to post
Share on other sites
Quick question about the signal into the student port, PPM or PCM?

FWIW, I'm not aware of any R/C Tx's that send PCM out of the trainer port. From what I have seen, the PCM encoding is performed by the master Tx and not from the trainer Tx.

Link to post
Share on other sites

"I remember 1 quirk with the 9C I was using was that it only had 8 out of the 9 channels in PPM mode, you had to be in PCM to get all 9 channels. Still hold true?"

yes, the channel 9 is only available on PCM on my 9C

Link to post
Share on other sites

Futaba TX do output the same signal on the port than on RF. BUT the input will only accept PPM. PCM output is used when you want to use the DSC function, that is link your transmitter via cable to the receiver with RF off on the field. Then the pulse setting must be the type the RX expects.

There would be no point in generating PCM, as these codings are proprietary and thus different for each manufacturer. Why bother when the good'ole PPM is universal :)

On Futaba, CH9 is only a switching channel, so one single bit in the PCM frame. Would take too much space in PPM, a framerate reduction would be necessary to send it.

Yeah, that thing seems USB-connected and -powered. Could be nice but a little too PC-oriented ;)

Link to post
Share on other sites
  • 3 months later...

Wow, that's a great news. I'm very happy to see that you made it, and it look good.

The second version is better, I agree. Do you already plug it in the trainer port of your Futaba 9Z or you are wire on the pot. inside the Tx?

That's great to see it's work on the two axis, this is the cherry on the sunday. But for my glow planes, I will need to make a very strong support for the cam to be able to move in the two axis. I have a Cap 232 (Cap maniac from BME aircraft) still in the box, I will build this plane in the summer with a two axis cam, but for now, my Funtana 40 is ready for the test. I will try the left/ right movement only, it will be enough for the training, it will be a lot different to do aerobatic, I was using two cam, one looking straight forward and one looking straight down, it was not realistic but it was working, now it will be like in my most beautifull dream :)

To modify the sevo for 180 deg. we can add two resistors on the pot, it's an easy mod. and accessible to everyone.

Me too on my side I was working on that project. I made the prototype suggested by Temporary insanity, and also make a mechanical prototype from my own. I don't know what to say about the infrared version, on the bench it was working, but I must stay straight, it's not linear, a little movement, and it goes rapidly to the side. Also I must say, when I will use this outside, where I sit, the sun will partially come through the trees, and I think it will affect the photodiode. But overall it work.

I made test with a mechanical prototype and it was looking good, so I make a final version. A pot. is fix on the side of the helmet and an arm will pivot when the head will turn. At the end of the arm, a spring and a magnet are installed. On my shoulder, I install another magnet that will stay in place using a bandage, They are under my t-shirt. See photo

post-6-1141005270_thumb.jpg

Edited by Vrflyer
Link to post
Share on other sites

The arm can swing up to the side of the helmet, a metal plate is glued there, so the magnet stay there when I remove the headset. A switch as been added on the RC tx, the movement of the cam will be transfert to the pot ch.6 before removing the headset.

post-6-1141006885_thumb.jpg

Edited by Vrflyer
Link to post
Share on other sites

It must be plugged directly inside the tx on the pot. The movement of the arm is small, the servo was moving only a little bit, so I modified the servo by adding two resistors on the pot like I said before to be able to reach 180 deg.

The movement of the cam is linear when I move the head. I can rotate all the way from left to right, the magnet stay togheter, but sometime, they will disconect, the arm will come down due to gravity, so the cam come center, I center my head and the magnets reconnect toghether without help. It work well in my house, but it will need adjustement while using it like all mechanical design.

But it can not work in the winter simply due to the fact I need to wear a jacket very thick. With a smaller arm, it will not work, not enough space. I need to wait summer before trying it...

So we have two working unit, but I prefer far more your design :)

I intend to try your design with my Futaba 9C, it can be plug in the trainer port without internal modification. I can remove my setup easily to try your's, I was expecting doing that so the headset is ready.

Edited by Vrflyer
Link to post
Share on other sites

The weather is still offering me everything between snow, rain and wind :(

However I've just managed to do a short 5-min flight under a light snow shower... <_<

I hadn't made the 2-axis mount yet, so that was pan only. That was very nice, but the most difficult thing is actually to get used to that again :(

This summer when I seriously started doing a lot of FPV flight I was trying to move the camera with my head, which obviously didn't work. So I added that rudder -> camera mix which was really very nice. The bad point about it is that with the many flight hours I've done with it I've got used to it way too much :(

That is when I flew with the headtracker the other day it got everything but natural, had to concentrate about turning the head instead of giving rudder <_<

So at the moment it's pretty weird, I'm actually too concentrated to enjoy it, if you can understand my point. I'll have to accumulate some flight time to get rid of that bad habit. Nothing wrong with the system though, it's just me :(

Regarding the device I'll definitely have to rewrite everything in assembly. That's not very complicated as I already have most of the functions available from other projects, just a matter of time. But I want to be able to put quite a number of configuration options, like 2-4 travel settings on each axis, possibility to choose what channels to use, servo end point setting (for use with TXs that don't allow using the trainer TX as a source for its own mixes, like my 9Z), and hopefully the pass-through feature to enable chaining it between 2 TXs.

Regarding performance I'm very optimistic. For now it has a slight drift which is not even very disturbing, but I've made no compensation at all yet. Remembering my first try which was totally unusable at the beginning (servo would escape to end point in 1 second if I brought my hand close to the sensor...) and which got pretty good after some correction there should be no big issue in getting rid of that here.

Yes the device of course already uses the trainer port. I run a 5-wire cable to the input, and have a little cable going out of the plug again, with a little push switch at the end. I velcro this one on the TX so I can remove it easily. The goal of making it as discrete as possible is achieved. I didn't want more cables and things everywhere as we're already pretty heavily loaded on the field.

For my flight test I was standing, with video RX in one pocket, goggles controller/battery in the other, and TX with neck strap. Sensor was fixed to the goggles with double-sided tape. That means I could move freely with everything on me. Pretty nice :)

Another cool thing is I've been able to integrate the sensor and electronics into another set of video goggles (the ones that were mentioned in another thread, 640x480 version), which have quite some empty space inside. Even better, then the only thing added is a plug for the TX :)

Now with this weather I've taken the "opportunity" to make a little pan/tilt mount for my small camera, and a friend made one for the bigger one. Some photos:

P1040601_s.jpg

P1040589_s.jpg

The bigger one needed a servo to be reversed, so I quickly made a reversing module for that purpose.

The small one works very well at least on that plane, I haven't seen any vibrations, but well it's electric... the big one will have to be tested with a gas engine...

I've modified my 4MP digital camera with an integrated R/C shutter I've made as well (not much space in there, but the unused MS slot enables to stow the plug when not used :P ), so many new things ready for when the weather gets better :)

P1040557_s.jpg

P1040571_s.jpg

Link to post
Share on other sites

Good Stuff !

Just for info you said "The bigger one needed a servo to be reversed, so I quickly made a reversing module for that purpose" you can reverse the servo by swapping the motor and pot wires.

Terry

Link to post
Share on other sites

Yeah I know, but I like to avoid mechanical hacks when I can, plus they have put glue on all the solder points...

Actually that was quicker to write 6 assembly instructions than even to disassemble the servo ;)

Link to post
Share on other sites

'" That was very nice, but the most difficult thing is actually to get used to that again"

Exactly what I was anticipated. On my Funtana, I will kept the cam in the wing when I will add the cockpit cam, so I will toggle between them if I don't feel confortable or something wrong happen with the cockpit cam.

I'm very happy to know it work fine, or close to. I don't understand " like 2-4 travel settings on each axis" explain it more.

"That means I could move freely with everything on me." That's cool, it's allow to turn yourself in the direction off the planes to kept good video reception while being able to turn your head and you will recover the planes rapidly if it land too far, you don't need to pack everything :)

I can hardly imagine you wearing your t-shirt on your head if it was a sunny day, you need peripherical view to maintain your balance. even if you are concentrate on the video, your brain continue to see around. Without this view, you will not be able to stay balance. Try to stay on one feet with eye close.

If it could fit, you will need tint acrylic sheet, I made mine by cuting a rectangle bigger than neccessary and using a heat gun to bend the corner. I used a rotary tool to shape it to fit my google.....

post-6-1141518717_thumb.jpg

Edited by Vrflyer
Link to post
Share on other sites

"Yes the device of course already uses the trainer port. I run a 5-wire cable to the input, and have a little cable going out of the plug again, with a little push switch at the end"

I have no problem to modify my Futaba9C a liitle bit, if it inside only, and if I can reverse the mod with minimum trace. I will buy a trainer cable and use it to make something look like this:

post-6-1141521492_thumb.jpg

Edited by Vrflyer
Link to post
Share on other sites

The trainer cable will run directly to the headset to connect your gyro. I will disconnect the momentary switch inside my Futaba, and use it like your push button for the reset. If I understand well, it will leave a free wire on the trainer port, if it so, I will disconnect this wire from the trainer port and connect the battery 9.6Vdc on it and feed this to power the google. I donot want a separate battery for the google, I will kept a second freshly charged batt in my car.

The receiver is mount on top off the headset, it have a second video/audio output. I connect a cable on it and record using a camcorder.

Link to post
Share on other sites
like 2-4 travel settings on each axis

I mean how much you have to turn the head to go from one end to the other. For now I'm using the throw doubler module on the Pan servo to get 180°, and have set the sensor so that it will need around 180° as well to wipe all the throw. That makes a pretty real impresson. But, if the user doesn't want to use the 180° servo you'll have to move the head 180° to get 90° out of it. Here's where a gain setting is needed.

The next step is to have only to move the head half of the desired servo movement. Can be nice too, requires less movement. Haven't actually tested that yet.

I'll offer the same options for the tilt axis, even if it turns out that people will most likely use a 2x greater gain on this one than on the other axis. It's not really useful to move on 180° here, apart from aerobatics. And, our head will also tilt only around 90° in a comfortable position.

It seems to me now that the best settings are 180° head - 180° servo for pan, and 90°-90° fot tilt. Maybe 90°-180° on tilt for aerobatics.

I've found a servo that will move nearly over 200°, add to this the 70° FOV of the camera... that allows you to see far back!

I can hardly imagine you wearing your t-shirt on your head if it was a sunny day, you need peripherical view to maintain your balance.

Nope, already flown standing with the t-shirt... last time was 4 days ago with the head tracker :P

I had some doublts about this, but there's no problem :)

I have no problem to modify my Futaba9C a liitle bit, if it inside only, and if I can reverse the mod with minimum trace.

The only thing you have to do is replace the momentary trainer switch with a standard ON-ON one. Otherwise you'll have to hold the switch as long as you want to use the sensor. The T9C doesn't allow choosing what switch to use for that purpose, it's fixed on the momentary one...

Edited by Kilrah
Link to post
Share on other sites

I agree for 180deg-180deg on pan and 90deg-90deg on tilt. 180deg on tilt, it's too much, I don't want that. I agree about 90deg-180deg. on pan, it could be nice too.

I have one servo that have more than 180deg, and like you said, panning with a wide angle lens, I can see the tail.

Nope, already flown standing with the t-shirt... last time was 4 days ago with the head tracker

Oh :huh: I'm surprise to heard that, I was thinking the contrary, but if you do it ...

The 5 wires are power, ground, reset, PPM out and a provision for PPM in. You can use the PPM In one if you don't need a second pilot with TX, or share the power wire

I was thinking the power wire was 5V, but if it's come directly from the power switch,, I don't need another wire, I will use it to power the gyro and the headset. I will check on my trainer port to know.

Edited by Vrflyer
Link to post
Share on other sites

There's a pin that gives out the TX's battery voltage for powering the student TX through the trainer port, 9.6V for Futaba. That's what I wanted to use, thus the onboard regulator.

Link to post
Share on other sites
  • 3 weeks later...

Ahh, I found where the party's at! :) Active head tracking is something I'm very excited about incorporating into my setup. I've been doing some research on sensors and such, but it sounds like some of you may be a bit ahead of me. Anyway, if you guys are building this as open source, I would love to help out in the development. Here are a bunch of random thoughts on this subject. If this is old news stuff, I appologize. I'm just trying to get some of my thoughts down on paper (err, cyber paper I guess. :D ).

Drift

Drift is going to be an issue no matter what you do, unless you use sensors that have absolute measurments. Accelerometers, gyros, etc. all sense relative motions. So you are going to have problems eventually with drift causing registration issues. Eg: head centered, but camera is offset to one side. I guess the ultimate would be to use an electronic compass for pan, and an analog inclinometer for pitch. But these are harder to find, and especially harder to find for cheap. As has been mentioned, the compass needs to be tilt compensated or it won't read correctly. A cheaper option is of course to use accelerometers or gyros. But then you have the issue of drift. Gyros seem to be the best option to me. An accelerometer output has to be integrated twice (instead of once for the gyro) to give position data, which will cause noise and drift to be more of a problem. You also have the issue of the accelerometer giving a linear motion. What we need is rotational data. So this also has to be converted somehow into meaningful units. So to me, it seems like you get the best "bang for the buck" by using two gyros. And it sounds like that's what some of you guys are doing. You still have the issue of drift, but there may be other ways around this too. A simple analog RC filter could be used to filter out drift. This would tend to slowly creep the servos back to the zero point. The disadvantage to this of course, would be if you wanted to look out over the wing (for instance) for a long period of time. The servo would slowly pan back to zero. The speed at which this occurrs would be controlled by the charateristics of the filter. If very low drift is detected in the gyro system, then the filter could be built accordingly, and this would be a very slow effect.

Reset Switch

I think this is just about mandatory. As I mentioned, drift will cause discrepancies between head "view" and camera view to creep in. And at some point it will become noticeable and need to be zero'd out. So moving your head back to straight forward and pressing a reset switch is needed. It's also necessary for other reasons. Say you want to change your orientation at the field. For instance, if glare from the sun is causing viewing issues with the HMD, you may want to rotate your body 45 degrees so the viewing is better. Well, this will cause a 45 degree change in the camera pan. So you will want to reset it once you are re-adjusted. I also hand launch some of my planes. And I know there's quite a lot of movement while doing this. So a reset is needed there too. Speaking of which a hold or lock toggle switch could be used during this part of the flight to keep the camera perfectly straight, until everything is situated. Although I guess just flipping off the trainer switch (disabling the camera pan) could be done too. If there is some sort of signal sent out the trainer port when the trainer function is enabled, that could be useful. That could be used as another reset switch. This way, anytime you turn on the head tracking feature, the system automatically resets back to zero.

continued in next post....

Link to post
Share on other sites

Sorry, the board software told me my reply was too long. :P

continued....

Couple other random things.... I do not think it is wise to power the HMD via the transmitter. The goggles probably take at least 500mA to power them, and I don't know if the transmitter has the capability to safely provide that much power through the trainer port. Even if it does, this will also drain the transmitter battery much quicker. I would hate to try this, only to find the transmitter looses power when I have the plane 1/4 mile away!

Also, Kilrah, what kind of connector are you using to plug into the trainer port? I have a Futaba 9C which I believe is the same as your 9Z (square connector). I have a plug that I bought for my AeroFly Pro simulator that plugs in, and has a mono 1/8" jack on the other end. But I don't recall being able to open up the plug to get more than two wires out of it. I'm curious what you use that allows you access to the other pins on the trainer port.

As for angles, I agree with pan being about 180 degrees. I would like a bit more than 90 on pitch though. I plan on building an AP platform where the camera(s) are mounted in the nose of the plane. I want the cameras to be able to look straight down for vertical view images. Plus I will also need 30-45 degrees of up tilt for good viewing during banked turns. So I'm thinking 120-135 degrees total tilt would be good to have.

Well, that's about it for now I think. Looking forward to more discussions on this!

Brian

Link to post
Share on other sites
  • 1 month later...
Couple other random things.... I do not think it is wise to power the HMD via the transmitter. The goggles probably take at least 500mA to power them, and I don't know if the transmitter has the capability to safely provide that much power through the trainer port. Even if it does, this will also drain the transmitter battery much quicker. I would hate to try this, only to find the transmitter looses power when I have the plane 1/4 mile away!

Sorry, I miss this one...

I'm agree with you, but I replaced the battery with a high capacity batt. (2600 maH).

I already have a vumeter between the two eyes to monitor the voltage. Also the Futaba 9C have a voltage indicator and an alarm. If the headset fail and short circuit the transmiter battery, a fuse will blow and leave power to the 9C. The total power consumption is around 600 mA, so I should have 4 hours of run time. I done it 3 weeks ago, and it work well. The trainer port was modified to let the 9.6V (swithed) goes through it.

For the headset tracker, no news.

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...