Speed gun

[GTBecker]

Has anyone built or hacked a speed gun?

[dlh]

I haven't but this guy has.

http://www.make-digital.com/make/vol10/?pg=150

[GTBecker]

Yeah, there are a few projects like that online, no more than repackaging and repowering the Hot Wheels unit. I'll need more range than dozens of feet and I need to get the data from it.

Bushnell has an inexpensive one http://www.bushnell.com/products/other-products/speed-gun/101901/ that I'd trust more than Mattel's and, discounted, isn't much more costly. In either case, it appears I'd need to get at the LCD segments for the data - unless there's a magic circuit board land.

Is sampling LCD driver lines much of a problem, I wonder?

[dlh]

I think most character based LCDs have RAM that is readable so the difficulty probably depends on the particular LCD or, more specifically, on the controller used with the LCD.

[GTBecker]

We'll see; I ordered a Bushnell.

I expect there will be just a single DSP part that handles ranging and the display, too, so getting at the data bus of the LCD driver might not be possible. I'll report when I have the thing ripped open.

[dlh]

It may have a custom LCD but these have become so standardized and commodified that I suspect it will use the Hitachi 44780 (or equiivalent) controller. Here's a site with tons of info on it.

http://home.iae.nl/users/pouweha/lcd/lcd.shtml

[GTBecker]

Here is the Bushnell Velocity speed gun. Apparently operating in K-Band (suggested by my car's radar detector, perhaps 24.150 GHz), it looks like it's a Gunn oscillator module at the vertex of a ~4"-long cast brass horn with an RF lens at the mouth. Much more rugged than the Mattel unit, which is reported to operate in X-Band (probably 10.525 GHz).

Three circuit boards - signal, RF preamp and digital processing/display; the OEM appears to be Progress Instruments. Looks like a magic connector and some unusually fat through-holes on the top edge of the display board. Very good range for autos (it claims 1500 feet but it read a car 2000' away in my first test). It can read the wood blade tips of an overhead fan (17MPH on low). My app is for boats, so I expect the range will be less than cars for most fiberglass hulls.

More and detailed shots at http://rightime.com/images/Speed/Bushnell_Velocity_1.rar and http://rightime.com/images/Speed/Bushnell_Velocity_2.rar .

[GTBecker]

The large through-holes are apparently for coax connectors. They are labeled "Aud", presumably audio of the radar return beat note, and "Comp". Composite? Compensation? There are a few unmounted parts and other unused connector positions, perhaps for a higher-cost version of the product. The Gunn module is a MA-COM, processor is a TI M430F413 8MHz 16bit RISC which drives the LCD directly. J6, the promising edge connector, at least in part, is JTAG and power. http://focus.ti.com/docs/prod/folders/print/msp430f413.html Amazing; an American product.

A few more photos at http://rightime.com/images/Speed/Bushnell_Velocity_3.rar .

[GTBecker]

Well, I can find no signal that appears to be a digital representation of speed in the Bushnell Velocity gun, and J6 seems to be only JTAG and power (although there are two pins on that connector that I can't identify). While I think I can interpret the LCD segments differentially against its backplane, that would take some level conversion and, perhaps, synchronous latching. The Aud point, though, is promising.

The attached scope image is what an overhead fan blade looks like on the Aud point when the the display shows ~15mph; this signal is apparently the filtered beat note of the sweeping transmitted RF frequency and its return. Beat note cycle periods are a little less than a millisecond at that speed and correspondingly shorter at higher speeds. The signal voltage is clipped at about 2.7vpp, and it looks like the signal is highpassed so that speeds below about 10mph yield a small signal. Speeds above ~12mph produce a signal that is fully clipped.

A second test shows ~675uS at an indicated 22mph and ~1027us at 14mph so the formula appears to be Speed(mph)=0.015/t.

Next step is to feed that to a processor for timing.

[GTBecker]

Here is a 60-second MP3 of the Aud signal audio. Initially, the fan is at medium speed then slower and faster. Note both rate and frequency changes. Since the forum doesn't appear to accept audio files, copy and open this link in a player. http://rightime.com/images/Misc/Doppler_FanBlade_MP3.asf

I also took the thing out this morning and found that it can read the water velocity of lawn sprinklers. And, like an automatic door opener sensor, it senses body and hand motion - but much more selectively since the beam is pretty well shaped by the RF lens. I can envision a hand-wave door opener with a twist: with the beam crossing a path to a door, simply approaching it will do nothing - but waving a hand sideways (parallel to the beam, perpendicular to the path) while in the beam will be detectable.

I think I'll add a small amp and speaker to the device to better investigate the outside world of Doppler.

[GTBecker]

FWIW, the speed gun Aud output yields genuine audio - if the gun is pointed at a moving loudspeaker cone. In this recording the gun is about sixteen inches away from a small radio speaker, aimed a little off-axis to reduce the reflected signal below clipping. I placed my hand in the beam twice, which mutes the audio.

I'm not sure what this says about the demodulation scheme since the moving cone produces FM'd RF back to the gun, which should sound like broadband noise, I think, not like the audio that moved the cone.

If this link doesn't play directly, copy and open the URL in a player: http://rightime.com/images/Misc/Speaker_Cone_MP3.asf

Any guess about why this works?

Another interesting effect is that the beam seems to be very well reflected by the arc in a fluorescent tube. More experimentation is called for.

[GTBecker]

Ah, the return beam is not FM'd; it is PM'd by motion. That is, the return beam phase, not frequency, is advanced or retarded by the relative target motion. In the case of an approaching vehicle, the phase advance is continuous until the vehicle passes the gun or otherwise leaves the beam. Continuous phase advance is the same as increasing the frequency according to speed. In the case of the speaker cone, though, the return beam phase cycles through advance and retard for each cone motion cycle, while the average frequency is fixed.

My guess, then, is that the Aud point is the output of a phase comparator. In a Gunn diode mixer I think that is inherent, so this is a natural product of the device. Sweeping the diode frequency wouldn't be necessary except perhaps to measure distance, I imagine.

[spamiam]

Tom,

Wow, this is a great project! The discussion of phase madulation etc. is WAY over my head, but I get the jist of it. Very interesting. I never thought that the radar guns could read such fine detail as a water spray, or a speaker cone!

What does the other connector next to the "AUD" do?

-Tony

[GTBecker]

The Comp connector's function is unclear, and the signal on it is complex. It is, it appears, the input to a lowpass filter that produces the Aud output, so I assume it means Composite, the sum/difference (or product) of the outgoing and incoming RF signals.

I, too, was surprised that the device can detect the motion of a paper cone, but things are different at 24GHz. We had a hard rain a few days ago so I took it outside and pointed it up into large droplets. It indicated 15mph - about 22fps - which I suspect is correct. It can see steam from a tea kettle.

I've added a small amplifier, switch and speaker to it. Beyond simply hearing the returns, that aids in aiming - and understanding how closer or larger objects dominate the signal. [And, although surely $1000 guns used by police are better devices, it shows that trees and other obstacles between a moving gun and a target make measurements of anything but the tree difficult or impossible, and, among a cluster of cars, it's pretty difficult to determine which might be speeding even if the gun can measure the higher speed.]

My wife was concerned when I showed her the closing velocities of oncoming vehicles on a rural two-lane road we often use. 135mph is not uncommon. It fires radar detectors in other vehicles, too. Shooting at a car ahead of us turns on the brakelights on occasion! Ha!