Photographing the invisible - adding a photo scanner to the GUI
Progress and a mystery.
RF camera and radio telescope - Table of Contents
Since we’ve got all the furniture from the living room stored in the garage while we’re having the living room painted and new flooring put in, I couldn’t really work on any of the hardware projects I’ve got going on - all the dust and grunge from sawing and filing would get all over the furniture.
Instead, I spent a couple of evenings working on Grote, my RF camera software.
I had last planned to work on the temperature measurement and calibration. What I did was to work on the photo scanner function. I’m more interested in seeing what things look like than I am in measuring the temperature.
The photo scanner looks like this:
| Photo scanner |
|---|
 |
You can pick one of the two RF bands that the LNB has, as well as one of the two polarisation options. You can also select which sampling rate to use. Lower rates (or higher averaging) gives a clearer (less noisy) picture but it takes longer for a scan to complete. A typical scan takes around fifteen minutes.
When I had most of the functionality implemented, I had to try it out. I didn’t pay any attention to where the dish was pointed, I just made sure it wouldn’t hit anything in my hobby room.
This is the first picture I made with Grote’s photo scanner:
| First photo |
|---|
 |
There’s a surprise - there’s a “cold spot” there in the middle.
It took a little experimentation to figure out exactly what it was looking at. The satellite dish I am using is an offset dish - it “looks upwards” at some angle so that “straight ahead” is actually “up there somewhere.”
The dish was looking at this corner:
| Where the dish was looking |
|---|
 |
The cold spot is at the junction of those three walls. It is most definitely not the monitor or the desk the monitor is on.
At first I thought it was something to do with the angles of the walls, so I made a wooden model of the corner and setup things up in the living room to make some comparisons.
| Test setup |
|---|
 |
The wooden model is there on the floor in front of the music stand I was going to use to hold it up.
Being the paranoid critter I am, I moved the model and the music stand out of the way and made a check to be sure that the blank wall was really blank - and got another surprise.
The blank wall in the living room has a cold spot just like the one in my hobby room. It isn’t a malfunction of the scanner or the hardware - the cold spot stays put when I move the scanner to a different place or turn it to point in a different direction.
I could believe that the corner in my hobby room is cold - it is on an outside wall and at least part of it is sheetrock and insulation. The wall in the living room is an inside wall and it is solid. It shouldn’t be cold at all.
There’s something going on here that does not depend on the temperature and that has nothing to do with the corner as such. I did carry out the experiment with the wooden model of the corner, but it was invisible. The wood had the same temperature as the wall and it radiates about like the plaster and wallpaper so that there’s no visible (to the microwaves) difference between the wood and the wall.
I really don’t know quite what is going on, but I do have some “maybes” that I’ll be checking out.
- Maybe there’s something under the wallpaper that radiates differently than the rest of the plaster.
- Maybe the laptop is radiating at microwave frequencies and the dish is simply seeing its own shadow on the wall.
That last one occurred to me while I was writing this up. I’ll have a look at it in the next few days - I’ve had to pack my toys away again because the flooring guys will be back tomorrow to finish up.
That fifteen minute scan time really drags. I implemented a bidirectional scan to speed it up, but had to take it back out again. The scanner mechanism isn’t robust enough to scan up and down - it wobbles something awful on the down strokes. Maybe I ought to make it do the bidirectional scans left and right instead of up and down. Another time.
Here’s what a vertical bidirectional scan looked like:
| Zig-zag |
|---|
 |
Because of the wobbling, alternate lines don’t line up as they should.
The Arduino Uno in the scanner seems to have croaked - I’ll have to repair or replace it before I can do any further experiments.