Findings and questions

[daedalus]

Hi everyone,

Bench testing
Power using FTDI cable with an laptop was not enough with anything more than the GPS. Lots of brownouts.
Our biggest mistake was to incorrectly apply power. We have tried batt-ESC-receiver-gluon-servos (with SJ2 and SJ3 soldered) and did not work stable enough - brownouts. In the wiki is clearly stated that the receiver should be powered through gluonpilot yet we managed to do it the other way.
Now we have a setup with the ESC powering the servos and an UBEC powering the gluonpilot and the receiver (unsolder SJ2 and SJ3). Appears to be stable but my concern is regarding the output voltage from UBEC - 5.25V.

Stabilized mode
Working very good and interesting (first flights with an stabilizer system). It gives you confidence. Because we are flying with and Arcus glider the adverse yaw is quite strong. Next days we will try to couple the rudder servo on the second gluonpilot channel to reduce that adverse yaw.
The reduced authority of commands in this mode makes landing challenging and incorrect (too much speed). Could it be possible to limit the pilot inputs only after exceeding defined "max pitch" and "max roll"?

Autopilot mode
Strange behavior (same heading). Tried it in two flights with two different command sets (actually the first and the second example from wiki).
Sometimes strongly oscillates in pitch, yet in stabilized mode it was quite stable. Maybe it was the wind (30km/h).

Camera trigger
How can we use the two servo_set and servo_trigger commands?

Other questions
How is the engine managed in autopilot mode?
Could it be possible to connect an airspeed sensor (maybe using this http://store.diydrones.com/product_p/br-0004-03.htm)?

[Tom]

Thanks for your quick review

Bench testing:
The brownouts really depend on the 5V USB supply. Powering servos on USB is indeed tricky.
Don't worry about the 5.25V BEC: no component on the gluopilot module uses the 5V supply rail directly.

Stabilized mode:
A glider will always have less authority than e.g. a funjet. I would propose you to increase the P-values (e.g. x 1.5) for pitch and roll in the PID-tab. Maximum pilot input = max pitch/roll as defined in the configuration (so you will need more stick input than in manual mode)

Autopilot mode:
What does the barometric altitude show in the datalog? Maybe a flow of air confuses the pressure sensor. Using the checkbox in "Control" you can decide how altitude control will work in stabilized & autopilot mode.


SERVO_SET allows you to set one PWM output (micoseconds). Of course you should not set output 1-4 because the setting will directly be overwritten with the Control loop.
SERVO_TRIGGER allows you to set the output of a servo, and it will move back to the original position after a number of configurable milliseconds.
A good way to experiment with this is by using attaching a servo to the board.
e.g
1. servo_set(5, 1200)
2. servo_trigger(5, 1800, 1)
3. servo_trigger(5, 1200, 1)
4. goto 2
(this will move the servo on channel 5 every second).

I guess you will need development code & gluonconfig to do this.


The engine is always manually controlled. In the future this will be changed. An airspeed will also be added, somewhere in the future (using ADC or SPI/I2C). If you are a developer yourself: feel free to experimant with the module and the code!

[magellan]

Good afternoon Tom,

I received my Gluon 2 days ago. I have put it together for testing. Everything works brilliantly.
GPS locks rapidly, even indoors. Everything seems to be normal.

After a few rounds of bench testing and fine tuning, I will go ahead and install it in this little plane:
http://www3.towerhobbies.com/cgi-bin/WT ... &search=Go
I am removing the ducted fan and using a propeller on a pod instead.

I have ordered and airspeed sensor, I have an HMC5843 Mag as well, I plan to implement both. Next, an OSD.

I have installed my XBee pros and everyhting works well too. Great stuff.

I'll keep you posted.

Regards,

[Tom]

Good to see everything is going good! Just be carefull of the downward force a pylon-mounted motor triggers (it's a kind of leverage).

Tom

[magellan]

Yes, will make the pod adjustable so I can change the angle to counter-act the leverage effect until I get the perfect angle.
Today, I received an MPXV7002DP airspeed sensor from DIYdrones.

I gotta figure out how to implement that in the Gluon at some point.

I'll keep you posted as things move forward.

Cheers,

[Tom]

Great!

For the airspeed sensor: You could use the AD-input on the gluon board, but even better would be to use an I2C ADC

[magellan]

Hi Tom,

As it is, the MPXV7002DP has an analog output. How would you go about making this an I2C device?
Is there something at Sprakfun I can buy to make it happen? I will gladly try it out if you know a way.

By the way, where are the AD inputs on the Gluon?

Thanks!

[Tom]

You could use the analog input (pin hole on the bottom left of ic number 16 in this drawing: http://www.gluonpilot.com/wiki/Hardware_v0.1n)

or use a chip like this: http://www.sparkfun.com/products/8636 (I2C would be better than SPI)

[marc]

Dear,

Some comments:

- The MPXV7002DP requires a CLEAN 5V power supply which is not directlty available on the Gluon pilot... the easier is to use a sensor requireing 3V3 with a regulator to get that clean power supply..I suggest a MP3V5004 DP (free sample)

- a speed of 5 km/h gives a dynamic pressure of 1,18 Pa.... meaning 1...2 mV on the pressure sensor (see the attached .xls sheet).... therefore a 10-12 bit ADC is not accurate enough....

- having analog cable going out from the Gluon will get noise ... while we are measuring mV, according to me, it is not a good idea...

- I thing we have to work with an external ADC directly connected on the pressure sensor (ADC like ADS1100)...

Attached you will find a "dirty" prototype of Pitot tube I made...
Feel free to comment and cross-check everything...



Rgds

MArc

[magellan]

Bonjour Marc,

Nice work, very interesting setup. I agree with the 3.3 v business too. Makes sense.
Did you test this with your Gluon module yet?

I ordered some samples from freescale. I wonder how much it would cost to get some boards made with that setup. It looks like the best way to go.

Thanks for the info, much appreciated!