IR detector system, biasing of photo diode

Yes, the inner loop must be a lot faster than the control time of the robot, or it will oscillate ;-)

Correct, all though high resolutionn cameraes are very cheap nowadays. And they have a Linux maching running the application SW, so adding a camera does not add much extra cost, just a camera interface IC.

It's a good ride. ;)

The really good thing is that all the relevant noise sources can be calculated really easily from first principles.

If you calculate using just the shot noise and Johnson noise formulas, you'll get the right SNR answer to within a fraction of a dB. So if you figure out what the low- and high-background limits are, you can figure out what the performance will be in a few minutes.

Cheers

Phil Hobbs

Nice

Hard to do without trashing the noise, at least in dim light. (In really bright light the shot noise of the sunlight photocurrent can easily swamp everything.)

Cheers

Phil Hobbs

Or like certain radio navigation systems popular in Western Europe, 80 or so years ago. ;)

That's a good example of Horace Darwin's maxim, "Always try it the other way round", quoted in "Instruments and Experiences" by R. V. Jones.

Jones is one of my technical heroes, and was instrumental in defeating a couple of the aforementioned navigation systems, specifically "Knickebein" and "X-Geraet", as explained in his book, "The Wizard War" (entitled "Most Secret War" in the UK). Horace Darwin was a grandson of Charles Darwin, and a colleague of Jones's at Aberdeen after the war.

Cheers

Phil Hobbs

VFDs put out large amounts of magnetic crap from the hundreds of hertz on up. I saw your VFD EMI filters at your Otis St shop. ;)

Or just do three lines of algebra to pick the right resistor value, AC couple, and be done.

"Well-lit", as in probably 1000 lumens of LED or fluorescent light, which has very little output in the >700 nm region.

There are lots of imponderables there, though. For instance, on account of the slow speed of sound in air, a 1 m/s breeze (2.2 mph) will make the apparent direction of the acoustic source move by 3 mrad.

Cheers

Phil Hobbs

;)

Cheers

Phil Hobbs

Horace Darwin - Charles Darwin's youngest son - founded Cambridge Instruments where I worked from 1982 to 1991. He died in 1928.

I actually knew one of Charles Darwin's grandsons, when he was a professor at Sussex University. If there was a Horace Darwin in his generation, Google doesn't know about it.

If you can afford a little gain in the two pole Sallen-Key filter stage, you can use the gain to tune for a critically damped response with equal value capacitors, which can be from the same batch, which makes procurement a lot easier.

Or spend around a thousand dollars on a really narrow band interference filter that only lets through your laser wavelength, and blocks sunlight.

An expensive solution, but hard to beat.

If one of the chirps has been reflected from a distant hard surface the long delay could interfere with the next pulse. Something like that happened to an early-warning radar sytem when the moon rose and reflected long-delayed pulses.

A slow pulse rate or a variable one would overcome this.

Life would be so much better without customers.

Cheap at twice the price!

It’s a unique book—a collection of Jones’s groundbreaking papers on things like measuring the ether drag on light in a moving dielectric, which would be a niche interest except that each one is paired with an essay of about equal length explaining how the measurement was done.

Those pieces are of lasting value—they’re confidential in tone, including the inventions required, the blind alleys, and very often, how somebody made a big difference with a piece of hard-won advice, like that remark of Darwin’s.

Cheers

Phil Hobbs

That was conducted EMI. 20 volt spikes everywhere on the top floor. Mag fields drop rapidly with distance, 3rd power or something.

It's homing into the mother ship so a breeze will very slightly curve the path.

You're an optics guy, so maybe don't like the sound thing.

Hidebound prejudice is the only possible explanation. ;)

Cheers

Phil Hobbs

(I don’t necessarily dislike the sound idea, but it’s more of a science project than the LED approach. )

Think so? A quick experiment would be easy. Two MEMS mikes would feed an oscilloscope directly. No inductors or TIAs or filters needed. One could trigger the scope from the source and signal average for extra fun.

The phase difference between two microphones would be a great direction sensor. This would be so easy in production, a couple of surface-mout MEMS mikes on a PC board.

You would need more than two microphones to get unambiguous results.

John

It would be better than two photodiodes in black tubes.

Three or four mikes would give 360 degree sensing.

As I understood the problem, the robot is probably pointing at the mother ship, and just needs to be guided the final few feet.

So simple. Some engineers don't like simple.

It sounds as though the problem could be solved easily and reliably with a VLF magnetic field, rather than light. Something along the lines of the Lorenz blind landing beacon. By using dot-dash modulation, the receiver would know which side of the centre line it was on.

The only problem would be convincing your client that it wasn't radio - but that sounds like a much easier problem than all the solutions so far suggested.