Monday, June 25, 2012

Summary of the ROV1 project.

After this weekend of test diving I feel that i can conclude my first attempt to make a simple ROV.

Here is a summary of the process: (Links lead to blog posts/photos about each topic)

ROV1 homemade by DZL, from oz2cpu on Vimeo.

Future of the project:

After using several cameras it has become clear that a single fixed focus wide angle camera gives the best result. I plan to hack one of the popular GoPro cameras so it can be remote controlled and provide a live video feed.

The camera will be mounted on a pan/tilt unit inside the ROV for better coverage.

An IMU and pressure sensor will be installed for closed loop control and navigation.
Side thrusters may be installed for side to side control and drift compensation.

Many more posts will appear here over the summer as the project progresses.

Dive day, midsummer at Illutron


Added a whole bunch of new stuff to the ROV: Switch mode power supply, remote controlled camera, landing gear and new firmware.
I realized that i needed some feed back from the ROV in terms of power levels and moisture levels so I added a LCD display to the control panel.
This display will eventually display data from several attached sensors.
Here is an underwater video from the dive, shot by Thomas Scherrer.

Saturday, June 16, 2012

ROV testing at Illutron


ROV testing at Illutron, a set on Flickr.
Today the ROV1 got the first live test dives in the water at Illutron.
Right from the start everything worked great.

ROV1 tank test

Tested the ROV1 in a glass tank at Illutron. All went well and everything worked right away. The controls are accurate and all seals are good.
Short video of tank test
Less crappy videos:
Sail off 1
Sail off 2

Tuesday, June 12, 2012

ROV1 ready for 1. test.


ROV1 ready for 1. test., a set on Flickr.
Did final assembly of the ROV1 and control box.
Installed micro controller and leakage detector.
ROV1 is connected to the control box on the surface via a single CAT5 cable. This cable carries high voltage power (240V), bi-directional data and video.
A test camera and rudimentary LED lighting is installed for the first tests.
All wiring is implemented in a way so that any short between the high voltage and other wires will immediately trigger a safety relay.

Monday, June 11, 2012

The Scream Detector and digital photo frame

We were invited to participate in a gallery opening in Sweden. The theme was a 20 by 20 cm canvas interpretation of Edvard Munchs "The Scream".
I hacked a digital photo frame I found. Put a picture of The Scream on it and mounted the screen and electronics in a 20 by 20 cm. frame.
I located the wires for the backlight of the LCD and put a FET on it.
I used the song detector project as the basis to make a scream detector that in term controlled the backlight via the FET.
This way you have to scream to see the scream...

Energy harvester/perpetual sensor

Energy harvesterEnergy harvesterEnergy harvesterEnergy harvesterEnergy harvesterEnergy harvester
Energy harvesterIMG_20120611_180752IMG_20120611_180623IMG_20120611_180636

Some time ago I was presented with the challenge to make a device that would collect information such as temperature, light, radiation and similar without ever being serviced or recharged. The device could possibly even be integrated into a wall or a rock for all eternity.
I started experimenting with running micro controllers in really low power configurations and developing algorithms for extreme power consumption.
The images above show a couple of these devices. They are all powered from a single 2.5V/100µA solar cell and use super-capacitors for power storage.
One have a simple LCD that displays the power stats for the device. This allowed me to get an idea of what kind of power an computation power was available.
The others uses a power booster circuit to blink a white LED that need higher voltage than 2.5V.
The LED blinks once per second. The duration of the blink is determined by the amount of stored energy allowing the blinking to go on during low energy times. Once per hour, if the stored power allows for it, the blink is modulated in a way that allows collected data to be transmitted to an external device. The average current drawn by these devices is around 6.2µA with a bus voltage of about 2.5V.
During energy starvation the bus voltage may drop as low as 1.3V. If this happen the firmware will put the device into 8 second inactivity cycles until better times.

Sealing ROV1


Sealing ROV1, a set on Flickr.
After initial tests of the ROV1 it became clear that it was really leaking.
I has used cheap foam gaskets for attaching the windows and no seal at all around the bushings for the motor and surface cables.
Christian of Illutron gave me some soft 3mm. vinyl sheet. This stuff seem ideal except it's really nasty to laser cut.
Adding an extra O-ring seal ensures that all packing surfaces are facory finished.
After testing with a vacuum pump I am now much more confident that this little ROV will stay dry inside.