If you (or someone) has a faster scope try spreading out X scale to see if each 1-6 ms pulse is composed of >10 MHz modulation. Even an unamplified photodiode should output a few tens or hundreds of millivolts, enough to measure something.
I'd do it but a) I gave my old scope to erco, and b) they accidentally sent my samples to another Gordon McComb that's in their database! Arggh!
It's not the scope; it's the response time of the photodiode.
That is, indeed, a curious waveform. 'Not at all what I would've expected, given my modulation theory -- but not entirely inconsistent with it, assuming there's HF stuff buried in the pulses.
I'm not so sure phase detection is at play here. ST claim their devices are based on *patented* "flightsense"(tm) tech. Finding the patent(s) should really help
You may be right that he device doesn't (solely) rely on phase correlation. The following appears to be the relevant ST patent, or at least one of them:
Like any good modern patent, though, there's a good amount of obfuscation in it. It does mention time-to-digital conversion as a process, but it's not listed as part of the claims, and the patent doesn't delve into the actual mechanics of measuring the time difference.
In the background discussion, they do provide formulas for calculation with ToF, but it also includes "The time shift component (=‘t’) due to the photon TOF, is dependent on the modulation frequency and phase shift magnitude of the waveform." It is somewhat unclear what this refers to, but this sounds like fancy footwork to me, and that modulation and phase differences are still part of the process.
The TI patent application is interesting, though they talk about the modulation/phase aspect in regards to incoherent light, which would make that aspect not applicable to laser-based sensors. But it's telling that even with modulation/phase they refer to these as ToF. In my reading on the subject over the last few years, I've seen quite a bit of latitude in the definition of "time of flight." The typical obfuscation of modern patents doesn't help things.
Understanding how these work is interesting, of course, but I think we're missing the forest for the trees. The RF device is said to support 3D mapping. For a $5 retail part. LaserPing is only the beginning...
What kind of engineers are we without seeking to understand how things work, that is to say, taking them apart or pushing their limits? So much "is said to" in technology buzz about features that are pipe dreams or hardly near working. Yes, though, always only the beginning...
Too many advanced mixed signal products try to hide their principles of operation (aka IP) in a black box. Fine when working straight from the data sheet and demo apps. But try to push the envelope, we are left guessing, having to test on the bench to weasel out even basic performance data, not to mention the possible limits.
This is something I have a keen interest in, especially if it can be done in color. My app is this: a floating autonomous vehicle that takes pictures of the bottom for, say, eelgrass mapping. The problems faced are these: even in direct sunlight, the bottom might be too dim to get a good, non-fuzzy picture of. But if you try to shine a bright light directly down from the floating platform, you get strong reflections from the "flock" suspended in the water.
So what if you could record only those photons that arrive in a certain time slot, for example from depths near the bottom? There would be no flock in the photo!
Uh yeah the 3D mapping claims are a bit of a stretch for these sensors. The ST presentation goes into what might be achieved, but they're using 2D arrays of the sensor placement to decode some of those gestures.
For something like you're talking about Phil I think I would be trying something like Intel Realsense ToF cameras
A couple of years ago I was at some expo and somebody with a stand was showing off a vehicle traffic location and speed detection system. It could detect the position and speed of vehicles along many kilometers of highway.
Talking to the young guys on the stand I got the this amazing story...
Their original problem to solve was that when there is a lot of copper cable laid along highways people tend to come and dig it up. Copper is worth money.
They had a solution. Put a fibre optic cable in the conduits with the copper cables. By means of some kind of time domain reflectometry they could detect when anyone along the kilometers of cable was trying to dig it up. The fibre would detect the vibration.
When they tried it out it looked like it would not work. There was too much noise coming back down the fibre.
Until one bright spark said "That is not noise. What we see there is the motion of vehicles on the highway"
Boom they had a whole new product idea!
I conclude that, even without the fibre, with enough processing of a returned signal one can separate "flock" and other debris from the bottom of a lake.
What kind of engineers are we without seeking to understand how things work, that is to say, taking them apart or pushing their limits?
I get you, but there's no taking one of these apart. The discussion of exactly how it works is elementary at best, and doesn't advance the actual use of the device.
But, it does have an I2C bus on it. So far I haven't read anyone talking experimenting with it *that* way, which to me offers the most fun possibilities, and 3D mapping (for things like edge and outline detection). True, LaserPing doesn't export the I2C pins off the board, but Sparkfun's breakout board does. We could use that to play with the potential of the chip until Parallax releases a version with I2C exposed.
... LaserPing doesn't export the I2C pins off the board ...
Actually, it does. They're just pads, but they're eminently connectable. I think you would have to hold the micro in reset to free the I2C pins from interference. But the reset pad is available, too, so I think it's totally hackable.
Good to know about the pins. I have to wait for new stock before I can get mine.
I want to reiterate that as engineers we should be interested in how things work, but not at the expense of also investigating the practical applications. I know I want to mount a Ping and LaserPing side by side and take measurements from many kinds of surfaces, looking for any benefits of fusing the results, rejecting out-of-bounds readings, etc. For example, how is LaserPing aiming straight at a glass window? Can it handle it?
Comments
I'd do it but a) I gave my old scope to erco, and b) they accidentally sent my samples to another Gordon McComb that's in their database! Arggh!
That is, indeed, a curious waveform. 'Not at all what I would've expected, given my modulation theory -- but not entirely inconsistent with it, assuming there's HF stuff buried in the pulses.
-Phil
Doing a bit of a hunt, here's a presentation detailing the use of SPADs (single photo avalanche diodes) to do the timing detection
http://www.st.com/resource/en/product_presentation/time_of_flight_from_1dto3d_vl6180x.pdf
It refers to SPAD arrays but less on how they work or are used
Here's some decapitated chip pics showing the SPAD arrays
http://www.techinsights.com/about-techinsights/overview/blog/stmicroelectronics-time-of-flight-sensors-and-the-starship-enterprise/
Here's one relevant ST patent. Its kind of hard to know whether this is the most applicable one, or perhaps there are others, but this is pretty close
http://www.techinsights.com/about-techinsights/overview/blog/stmicroelectronics-time-of-flight-sensors-and-the-starship-enterprise/
I realise this is all about the ST device, but seems very similar to the RF one
https://www.google.com/patents/US8610043
Like any good modern patent, though, there's a good amount of obfuscation in it. It does mention time-to-digital conversion as a process, but it's not listed as part of the claims, and the patent doesn't delve into the actual mechanics of measuring the time difference.
In the background discussion, they do provide formulas for calculation with ToF, but it also includes "The time shift component (=‘t’) due to the photon TOF, is dependent on the modulation frequency and phase shift magnitude of the waveform." It is somewhat unclear what this refers to, but this sounds like fancy footwork to me, and that modulation and phase differences are still part of the process.
https://www.google.com/patents/US20140152974
Here's an ST Micro patent that discusses SPAD arrays in such sensors:
https://www.google.com/patents/US8610043
-Phil
Understanding how these work is interesting, of course, but I think we're missing the forest for the trees. The RF device is said to support 3D mapping. For a $5 retail part. LaserPing is only the beginning...
Too many advanced mixed signal products try to hide their principles of operation (aka IP) in a black box. Fine when working straight from the data sheet and demo apps. But try to push the envelope, we are left guessing, having to test on the bench to weasel out even basic performance data, not to mention the possible limits.
So what if you could record only those photons that arrive in a certain time slot, for example from depths near the bottom? There would be no flock in the photo!
-Phil
For something like you're talking about Phil I think I would be trying something like Intel Realsense ToF cameras
A couple of years ago I was at some expo and somebody with a stand was showing off a vehicle traffic location and speed detection system. It could detect the position and speed of vehicles along many kilometers of highway.
Talking to the young guys on the stand I got the this amazing story...
Their original problem to solve was that when there is a lot of copper cable laid along highways people tend to come and dig it up. Copper is worth money.
They had a solution. Put a fibre optic cable in the conduits with the copper cables. By means of some kind of time domain reflectometry they could detect when anyone along the kilometers of cable was trying to dig it up. The fibre would detect the vibration.
When they tried it out it looked like it would not work. There was too much noise coming back down the fibre.
Until one bright spark said "That is not noise. What we see there is the motion of vehicles on the highway"
Boom they had a whole new product idea!
I conclude that, even without the fibre, with enough processing of a returned signal one can separate "flock" and other debris from the bottom of a lake.
Sorry I have no idea how
I get you, but there's no taking one of these apart. The discussion of exactly how it works is elementary at best, and doesn't advance the actual use of the device.
But, it does have an I2C bus on it. So far I haven't read anyone talking experimenting with it *that* way, which to me offers the most fun possibilities, and 3D mapping (for things like edge and outline detection). True, LaserPing doesn't export the I2C pins off the board, but Sparkfun's breakout board does. We could use that to play with the potential of the chip until Parallax releases a version with I2C exposed.
Actually, it does. They're just pads, but they're eminently connectable. I think you would have to hold the micro in reset to free the I2C pins from interference. But the reset pad is available, too, so I think it's totally hackable.
-Phil
I want to reiterate that as engineers we should be interested in how things work, but not at the expense of also investigating the practical applications. I know I want to mount a Ping and LaserPing side by side and take measurements from many kinds of surfaces, looking for any benefits of fusing the results, rejecting out-of-bounds readings, etc. For example, how is LaserPing aiming straight at a glass window? Can it handle it?
Digikey have 50 in stock. Well 49 in a few minutes....
Samples per second clearly depends upon target distance.
Here is the rough experiment reported in terms of rough distance and the period in clocks(at 80Mz)... done using P2v.vers31:
With the LaserPing aimed at the ceiling (well beyond 2m) the period maxes out at around 950_000 clocks at about 74 hz.
With a black piece of plastic covering the LaserPing sensor and emitter... no distance
Period= 26250 at 255 hz
with target at a couple of inches... about 19800 clocks at 530Hz
and about an inch farther 28500 at 497 Hz
when pointed at object just beyond arms length...
about 297000 clks at 184 hz
Pretty snazzy.
Seems like a lot of room to play in though:)