My virtual playground. Meer info

Recent additions

Texas decadence?

Last friday I ordered 5 samples of a tiny smd op-amp from Texas Instruments. Very easy, no big fuss like the guys at Microchip are doing lately. I was very happy to see that they were shipping them with fed-ex and it would arrive on monday! They didn’t need to as samples are free, but well that was the only option :-)
They even put it in a huge box before they shipped it with fed-ex:

CO2 tax anybody?

(Oh yes, what did I order? A replacement for the AD8552 dual op-amp used in my thermophile sensor module. Analog didn’t give samples in the SOIC8 package. Texas Instruments did: the OPA2335
Note to self: + is Vss and – is Gnd ;-) )

6 August 2007, 00:28 | Link | Comments

How to desolder a chip

This weekend, I made a bad PCB. The only option that was left, was creating a new one. But my PIC was already soldered on! After a search on the net, I didn’t find any approach with simple tools to desolder an SMD chip (in a way that it will still work afterwards!).
This and this are brilliant examples of how not to do it :-)

So I came up with my own solution:

  • Put a lot of solder on the 4 sides of the chip.
  • Turn an iron (to iron your clothes) upside down on maximum temperature.
  • Put your PCB on it (no solder on the bottom).
  • Now quickly heat up the 4 soldered sides with your soldering iron.
  • The heat of the iron will make sure the solder remains liquid for a few seconds.
  • You can use your soldering iron to push the chip off the PCB when you’re desoldering the last side.
  • Do it quick so your chip won’t be burnt!

Good luck :-)

30 July 2007, 16:46 | Link | Comments

Autopilot phase 1 (stabilization & servo controller)

I’ve been slowly working on my own autopilot system for more than half a year. I spent most of my time gathering knowledge and components. The last few weeks I worked a bit more intense on the project and it paid off: the stabilisation and servo controller are finished. They have been tested successfully and proved to be reliable!


Now, what can my module do:

  • Calibration on startup (defining min and max values for the sensors, and the neutral point)
  • Manual mode: Pass transmitter signals to the servo’s
  • Stabilized mode: Read the desired roll- and pitch angle from the transmitter sticks and stabilize the plane in those desired angles. This stabilization works with thermophiles. They sense the temperature difference between the ground and the sky.
  • (mode is determined by a slider on the RC-transmitter)

Other features:

  • Delta mixing is done in the microcontroller so it’s not required on the transmitter (makes it easier to read pitch and roll input signals seperately)
  • DSP filtering. The PPM signal is taken from the receiver before any DSP-filtering is done, so we need to do it ourselves to eliminiate glitches. Features:
    • Checks if every PPM frame has the same number of channels.
    • Checks if every pulse in the PPM frame is within a valid range (1ms and 2ms).
    • When a bad frame is received: keep the last valid positions. Go to failsafe positions after 2 seconds.

Planned features:

  • Read input from UART instead of RC-transmitter (interfacing to other microcontroller or PC)

I tested the stabilization unit on a small delta wing (40cm span) because I see no use in testing it in an easystar :-)

Module on the bottom of the wing:

Sensors on top of the nose, slightly pointing up:

Now some showing off :-)
The stabilization works so well I can launch the plane with the transmitter on the ground! When I give full left, the plane understands this as “go as much left al long as you can still see the horizon”. Considering the IR-sensor lens of 100 degrees, this will probably be about 40-50 degrees.

15 July 2007, 13:34 | Link | Comments [10]

New inertial sensor

Analog devices has created a new inertial sensor, the ADIS16355. It includes a 3 axis high precision gyroscope, and 3 accelerometers. On top of that, it includes digital I/O with some on-board basic sensor filtering. It’s target market are dead-reckoning applications such as car navigation when a GPS signal is lost.

Being a cube with a side of 23mm, it seems kind of big. Unfortunately no weight specs on the datasheet yet. For the very small MAV’s this might be too big, but for most autopilot systems this sensor will improve navigation performance!
Samples will cost $350, +1000 will be $250 per unit.

4 July 2007, 16:23 | Link | Comments [2]

Connecting a GPS (EB-85) to a PIC microcontroller

Last month I finally bought a GPS module. I decided on the EB-85 because it’s cheap (thanks to mr. rc-cam!) and it has a 5Hz update rate!

However, this module has an UART that operates at 2.8V. My PIC runs at 5V. After some investigations I figures out that the TTL pins of my PIC would interprete the 2.8V as logical high, but the ST-pins (Schmitt-trigger) won’t. Because the hardware UART or PIC uses ST, I was left with 2 options: Use the TTL-pins and a software UART or continue looking for another solution :-)

What I need is a TTL chip that has en open-collector output. TTL accepts 2.8V as a logical high, and the open-collector port allows me to use any voltage as logical out if I use a pull-up resistor.
My solution uses the 7407 IC, which is a “Hex Buffers/Drivers With Open-Collector High-Voltage Outputs”.

This is the schematic:

On the left you see the PIC, on the right you see the connector to my GPS. In the center-top you see the pull-up resistor to 5V, in the center-bottom you see the pull-up resistor to 2.8V that consists of a voltage divider that yields 2.8V.

On my test-board it looks as follows:

This is what my test-program does:

  1. Startup
  2. Send a message to the GPS so only the RMC (Recommanded minimum) messages come through:
  1. Every time a byte is received on the UART, an interrupt is launched and the received data is decoded.
  2. The main loop prints the data in the LCD

The overhead of the UART and GPS decoding is very low on my 4Mhz PIC. Perfect!

Here are some nice sites for electronics newbies like myself:

There was a slight error in my schematic: I played the voltage divider before the buffer instead of after it. It is corrected now.
This should also work (but my 7407 chip has 6 buffers so I didn’t care):

30 June 2007, 15:25 | Link | Comments [11]

Extremely basic autopilot

I found an interesting link while surfing the net (well, doesn’t it always happen like this ;-) )

It’s a rudder-driven autopilot with waypoints hard coded, but works with a PIC16F877
Very basic, but comes with pic-code and a very simple circuit.
Interesting to read.

23 June 2007, 07:08 | Link | Comments


After months of building, my yoda is finally finished!
This high aspect ratio wing is perfect for thermalling. It has a rather high speed, but it can make very small circles!
Flying it is easy, but it’s very sensitive on it’s pitch. Also, you need to stay very focussed because it’s hard to see how it is oriented, you need to know it! Thas why it’s hard to fly it in the evening (after work :-) ) when it get’s a bit darker.

The maiden was perfect, first a glide test, and then the real liftoff!

For more info on the Yoda, see:

23 June 2007, 07:06 | Link | Comments [1]

Older articles | Newer articles