Re: [Paparazzi-devel] Cam4 initial testing report

[nisma]

Hi,
take a look at this page:
http://zipham.free.fr/oe5dxl.htm
This is a simple 9600 modem with serial input/output.
If we implement only the sender, then it's very simple.
Further view this page for explanation (french) 
http://zipham.free.fr/picrct.htm
and compare the output section from this MCU implementation
http://zipham.free.fr/pic9k6.gif to the simple MCU-less circuit.

The decompiled asm sourcecode is really simple. This application is written
using a higher level language, so requesting the sourcecode form the author
can be a solution. Maybe it's written in C.
But the idea is to integrate this into one already used MCU (atmel).
The difference from outputting RS232 or HDL-tx is very small.
Or you can just use the MCU-less rs232 schematic.
Theoretically, the free OP-amplifier can be used for the IR board,
or the modem can be located on this board in order to lower IC count.

For a simple RPM:
This is not linear, but it's ok for measuring the difference between motors.
Instead of the potentiometer, two fixed resistors can be used.

Motor 
-----[>|-+------+----+   
         |      |    R
         -      =    R<---- RRR --- MCU ADC pin
         L      |    R
         +-RRR--+----+---GND

Simple circuit to measure RPM using led,diode,condensator,2xresistor and
poti.
This measures the outgoing PWM.
The led is to consume current and to protect the ADC pin.
The condensator is to integrate the PWM.
The voltage divider adjust the voltage to the MCU ADC input range.
The last resistor is to limit the current for the ADC and to match the ADC
input impedance.

Personally, i have abbandonated the Paparazzi way.
Using more then 10mW TX is illegal without having a radio licence and the
video transmitter transmit in the allowed frequency, this on most countrys.
I have used the maxstream modules (worldwide approval, 19200) for long range
transmission.
But the cost is 500$. Additionally i have used the digital camera (cellular
type),
and a gumstix board (400MH Xscale processor running linux) for the
navigation,
filtering, and mpeg compression.
For manual flight, i sample the trainer port from the RC/TX and transmit it
using
the datalink. 4800 bps is sufficient for that.
Viewing the cost: 500$ radio, 100 Euro GPS, 200 Euro airframe with
accesoires,
200 Euro Gumstix + video (including IR), 
the cost of a 2axis gyro and 2axis inclinometer + 1x altimeter is 100 Euro.
I have opted for it.
I have seen a bid advantages from using DGPS. This adds the precision
significantly.
I have used a small pocket AM radio that i have adapted to this purposes.
IF EGNOS should be active, then this is not necessary.
Annother point is that the wind direction and wind speed is really a matter
for the
position estimation filter and that the GPS filter must be set for areal in
addition
to tweaking the rules for the active SV.
I have recorded the RINOS data for postprocessing and as result, i can say,
that the
accuracy is really good. The important point tweaking the SV settings is,
that the
UAV must have a altimeter onboard. This is essential. The reason is, that if
less then 4 SV are visible and active, the GPS passes to the 2D mode and no
altitude
data is available. On standard settings, more SV are active.
One alternative to this is reconfiguring the GPS unit during flight.
Currently, i think in adding a compass and airspeed measurement in order to
can fly
without valid GPS data. Having a decent computer and a bidirectional link, i
can
time the transmission delay and translate the position to real coordinate.

I have a component cost of 1200 Euro.
This is really tree times more then you'r solution.

On the other side, for a small UAV, the minimal cost is:
Garmin etrex (yellow, 100$) - FMA autopilot (100$), picaxe MCU with
experiment board:
18 Euro:
Programming in basic or flowchart using the autopilot navigation sequences.
Fail proof RS232/servo and other commands are available.
The GPS altitude is used for holding the altitude, the rest of SW can be
used from the rc-autopilot (rcpilot) project, already basic source code, but
this code should
only used as example, because it's not really good.
No power supply  is required, the RC receiver works from 6V.
Using a silicon diode, this dropps 0.7V, allowing a supply voltage of 6.2 V.
Quoting the price for the airframe at 200 Euro (including brushless motor
and batteries and motor controller). Adding the RC transmitter/receiver and
servos, 
The cost is 500 Euro. In this example, no really difference is getting the
FAM 
autopilot or building one for 4x 12 Euro (ir sensor)+ PBC + micro.
I have tried this using the orginal rc-autopilot, without the PSU,
 in addition to a picaxe that have made the altitude holt and camera trigger
on 
programmed waypoint arrival and after every 180 seconds.
I have used a low cost (100Euro) webcam/camera combination with 2Mp.
It have worked as expected.
For a real UAV, it's necessary to add a absolute and relative pressure
sensor in order to be sure to don't stall (constant relative pressure) and
don't lose the
height, if no gps data is available. This eliminate the need to have the
finger always on the RC-TX.

A bit thing, that i don't understood is the IR sensor. 
Why you don't measure the IR sensors on each channel instead of measuring
the difference of two channels. Measuring all channels, gives out
additionally information for calibration, like minimum and maximum value,
remote temperature
and more.

Regards
Chris

-- 
NEU: WLAN-Router für 0,- EUR* - auch für DSL-Wechsler!
GMX DSL = supergünstig & kabellos http://www.gmx.net/de/go/dsl