The Deliverator – Wannabee

So open minded, my thoughts fell out…


Posted by Deliverator on February 24th, 2006

I will be heading down to Portland next Wednesday for the Pacific Northwest FIRST Robotics Competition. The robot was shipped this last Tuesday, and other people are handling most of the administrative stuff, so I have been endulging myself in a welcome break from anything that resembles responsibility. I am still working on a few robotics related projects, but they are non-essential, so I can putz with them at my leisure.

One of the things we wanted to do last year was mount a CCD camera on the robot itself and hook it into a well-protected DV camera. We got the camera mounted, but nobody had a DV camera that could capture composite video to tape (a few would capture to memory cards, but at low-resolution), so the project fell through at the last minute. This year, we are trying to anticipate any potential pitfalls.

This year, we are going to be using two hacked CVS cameras to capture on-robot footage. We have already hit our first pitfall. There is a regulation that states that all devices on the robot have to be powered by the main system battery, which is ~12 volts. The cameras run off 2 AA batteries = ~2.8 volts. In short, I have to build a voltage regulator to convert the 12 volts down to something that won’t fry the cameras. I am thinking of using a simple design based on a LM317, a couple of resistors, a diode and two capacitors to dampen transients. Ryan and I are going to head down to Frys and see what we can find in the way of components.

The other project I am working on is a regulator for LED modules, kindly donated to the ROV project by Lumileds. In the case of LEDs, the concern is current regulation. Once an initial forward voltage is surpased, LEDs will try to draw as much current as they possibly can, causing them to get VERY hot and start smoking and exploding and other un-good things. As such, you have to limit the ammount of current the LED can draw upon to something less than the LED’s exploding point. Ryan and I played around with a variable voltage/variable current power supply a few weeks ago and experimentally determined how high we could drive the LEDs with out them totally frying. You can “overdrive” LEDs by a fair ammount if you can get rid of the heat fast enough, although you will reduce the LEDs useful life by a substantial ammount. We ended up testing one module to the edge of destruction in a large glass of ice water. Even with this degree of cooling, we managed to get the LED smoking underwater! A more sensible voltage/current combo with better long-term surviveablity seems to be ~3.4 volts at ~700ma. I wasn’t in the mood to build some of the more complex circuits (many of them requiring microcontrollers) that I have seen online for doing both voltage and current control, so I searched around for something I could buy.

I found exactly what I was looking for at Taskled. They make a wide range of LED driver circuits for the Luxeon Star LEDs. I emailed the company and expained what we were trying to do and George, the owner/operator/1 man band that runs the outfit wrote me back promptly and offered to sell me whatever I wanted at cost (he still assembles, solders and tests all the units by hand). I orded five regulators to test. Five 3 watt LEDs may actually be bright enough given that we have a very sensitive CCD camera, but if not, we can already order some more. I would love to use some of the 18 LED modules. Those would certainly provide enough illumination. Unfortunately, the optimal voltage on these seems to be about 21 volts, which would require a switching power supply to achieve off 12 volts. We could use two 12 volt batteries to get 24 volts and then use a regulator to get the voltage to 21 volts and a resistor to limit the current, but that solution is less than optimal in a lot of ways.