Making the sensor head
We are going to do some serious amplification to get images out of a pinhole-focused light spread over a good portion of a square meter.
In fact, it's sub-picoampere current sensing for the pinhole, and nanoamperes if you put a lens in it.
So, if you put the amplifier on the side, and lead a wire to the PD on the caret, like i tried initially:
Then all you will get is noise, noise, noise.
There are two parts to this current noise.
1. The mains hum and other external EMF.
2. The stepper motors noise and other internal EMF.
First is solved by wrapping the box into tin foil, and connecting that to the common ground inside.
The second would require additional shielding at the scan head itself and some careful design of the scan system.
Same scene above, with the amplifier mounted at the caret like so:
Would look much better, but still with a lot of noise:
With the amp on the caret and shielded on it's own, the noise would go down to almost nothing.
Shown here with InAs photodiode. The piece on the right is the 6.6GOhm resistor assembly for silicon photodiodes.
To be connected in place of the single tiny resistor the InAs uses.
The shielding should connect to the ground near the PD.
The only holes there should be are the PD and the power/signal wire.
With all that the current noise table goes below my best sensitvity, allowing for much clearer images even indoors.
But if you look closer, you'll see a ripple - that's likely voltage noise on the signal wire. A minor problem yet to be solved.
That was all concerning general noise reduction, silicon photodiodes and pinhole optics (for UV down to NIR).
MWIR and InAs photodiodes have their own unique problems.
About them, to be continued.
(Note: Much better write-up of this is now available at Ribbonfarm: https://www.ribbonfarm.com/2016/05/12/artem-vs-predator/)