Hack: wireless live-view for polecam v2–$50

Photo – A&C

Back in July, I posted about my new Looking Down series of photos. What I didn’t show are the tools I built to do this series. This is the polecam: a camera mounted on the end of a 20-foot extendable pole, that is more conventionally used for window washing and painting. The camera, another 2nd hand purchase, is running the Canon Hacker Development Kit (CHDK) alternative firmware, which allows me to run scripts on the camera to automate it. In this application, I’m using a simple intervalometer to snap a pic every 2 seconds. I built this simple, manually-adjustable pan and tilt camera mount of out sheet aluminum. Version 1 of the polecam gave some great serendipitous results, shooting blind, but I also had a few shoots that were total misses where the camera quit shooting from some reason or I wasn’t framing the shots the way I had hoped. What I really need was to be able to see what the camera is seeing from…

20 feet up in the air.

Here are the components of my $50 solution, that allows me…

to see from the ground exactly what the camera sees up in the air and…

do a far better job of framing the photos and tracking progress of the shoot on location. Technical details follow for those that are interested.

Ah, version 1. In the make-o-sphere, there is this phenomenon of inertia due to perfectionism. Makers or almost-makers sometimes get hung up *thinking* about their project instead of actually *making* their project. They may spend endless hours googling how dozens of people on the net have already done this project to a freakishly pro level. That freezes them up or heavily discourages them or pushes them to endless design and redesign cycles. I’m speaking from experience and recovering. To hell with that!

Striving to be a guerrilla maker, thinking with my hands (see Tim Brown’s talk on creativity and play), I just deferred ideas and implementations of motorized pan and tilt mechanisms, remote view and shutter mechanisms, and wireless image streaming, and got *something* done. Here, I added the white straws to the camera deck to get a general idea of where the camera was pointing.

I am a total sucker for cheap surplus gear with high utility.

When I saw this automotive wireless rear view camera in the XS Cargo flyer for $50, I snapped it up, powered it with 12V cordless drill batteries on both ends to prove the concept, and then left it under the bench for a couple of years.

Considering the component parts that include camera, wireless link, and display, purchased separately would cost 6x as much, I knew the DG018 Wireless Rear View Camera would some day come in handy. The challenges I was having on the Looking Down series prompted me to pull this gear out again.

My initial thought was to simply mount the wireless backup camera above my Canon PowerShot to get a general sense of the framing. I tested that and it worked… ok. Then James Bastow suggested piping the composite video out on the PowerShot right into this backup camera unit for transmission down to the display. Hmmm, good idea. So I cracked the camera unit open…

and happily discovered two mostly-independent printed circuit boards: the camera board (left), and the RF transmitter board (right).

Manufacturers, like Sunway in China who produced this device, often make or source separate daughter boards for wireless communication like this board and incorporate them into a variety of larger solutions. That lowers both their material costs as well as testing and certification costs for the RF link.

I faced two primary challenges: 1. how to power the pole-mounted camera unit with something smaller than a 12V drill battery, and 2. how to get video out of the point-and-shoot camera and into the aforementioned RF transmitter.

I *love* reverse engineering consumer electronics. It’s like a great puzzle. Recognizing that the 12V input from this automotive application was likely *way* higher than this camera needed, I was hoping I could bypass the 12V input with something lower. I powered the unit from 12V (red and black wires at right) and then used a multimeter to measure voltages at various points on the board. I found these five header pins serving up 5V and 3.3V which are very common operating voltages. Considering they were right beside the big black inductors on the board, I reckoned these were the outputs of a buck converter switched mode power supply that took 12V in and stepped it down to more useful voltages without generating the heat you’d get from a linear regulator. Through some guessing and experimentation, I discovered that 3.3V powers the camera and 5V powers the RF board. I tried putting 3V (two AA batteries) across the 5V pin, and that gave some static in the monitor, so I knew I was on the right track. I bumped it up to 3 AA batteries for 4.5V and the first problem was solved. Well…

solved after I made this lightweight dummy AA battery replacement out of plastic tubing and screws. I only had a 4-way battery holder, so this worked a charm. Power problem solved.

I attached my hacked up 4.5V supply via the orange wire shown and cut off the board’s original 12V input wire. Ground remained the same, of course.

The problem of getting video into the RF transmitter was easier and harder. Easier in that the manufacturer had silkscreened “VIDEO OUT” on the PCB. Harder in that when I tried to pull the camera and RF boards apart, I totally ripped a crucial (and microscopic) trace off the PCB. That required board repair surgery that necessitated using a cheap microscope, my soldering iron, and The Force to fix.

In the course of that work, I was able to successfully pull the pin connecting the video-out from the camera to the video-in of the RF transmitter. Wanting the flexibility to either use the integrated camera or an external video feed, I took a page from last year’s Project Loud Monkey and installed a male/female RCA connector pair. This photo shows the integrated camera looped back into the transmitter. Trial runs confirmed that, happily, the video out of the camera was indeed composite video as James had suggested and I had hoped.

Reckoning that the bulk of this board’s 230mA operating current goes to the RF portion, I thought it wise to not risk further board damage trying to cut out the camera.

Including this pic in case the notes help someone else along. Bad form here by not using a my proper lab notebook.

With such success on the transmitting end, I cracked open the receiver, hoping to also bypass the 12V step down and use smaller batteries here too. However, this end was more complex, likely due to the LCD display and requiring 5.2V, 5V, 3.3V, and 1.8V, to name a few. Too risky, but…

this SDA, GND, SCL pinout looks intriguing for future research. Smells like I2C access to the receiver.

With the composite video RF transmission figure out I gathered the pieces to make something a bit more robust for field work. From top left: Canon PowerShot point-and-shoot, pole-top battery pack, power wires, power LED, 2.5 LCD display, transmitter board, composite video out cable for the PowerShot.

The camera’s comp video out 3.5mm port takes that cable in…

and the yellow RCA plug on the other end goes into the red video-in cable I installed on the transmitter. The chain is point-and-shoot –> video out cable –> RF transmitter board. I was delighted to discover that comp video out shows the live view of what the camera sees in shooting mode, with all the icons, grid, indicators, etc that you would typically see on the camera’s built-in LCD.

I whipped up a little custom cardboard box and packed it full of the transmitter board, batteries, a power switch, status LED, and antenna (taped to lid).