My problem with all this is that I have effectively been caught, prosecuted and convicted by a machine.
I am entitled to know that the machine was working properly at the time.
The official response appears to be "Don't worry, its a very clever machine"
|
Yes but surely Lawman the point is that you seem to accept thats its entirely possible you were toddling along at a speed somewhat in excess of the prevailing limit. It also has to be accepted that had you seen the van you would have not traveled in exess of the limit. - So all in all slap wrist and take it like a man.
------------------------------
TourVanMan TM < Ex RF >
|
I accept that it is entirely possible that I might have been pressing on a bit. It is also entirely possible that I wasn't doing more than 70. I don't remember every second of every journey that I make, so I am forced in this situation to "take the word" of the machine,.
Am i right in thinking that hand held machines have to be calibrated daily?
|
So, RF, if the machine self-calibrates... why does it need calibrating annually?
I don't accept what you write. There must be something in the system that relies on being calibrated - otherwise how does it know what frequency it sends out, or receives back?
|
|
|
I don't remember every second of every journey
Clearly, just as you didnt see the camera, which after all is the real driving mistake here.
------------------------------
TourVanMan TM < Ex RF >
|
|
|
|
|
My problem with all this is that I have effectively been caught prosecuted and convicted by a machine.
Brave new world,I`m afraid. Seems we can soon look forward to toasters that feel guilt if they burn the toast! (Yesterday`s Today programme)
|
|
|
|
|
I cannot believe that a simple one point calibration like this would be acceptable in a court, especially as it is a zero point. I occasionally review scientific papers for a respected journal and if I came across a submission where such a one point process had been used I would reject it out of hand without even bothering to read the rest. In order to be any where near acceptable a 3 point calibraiton is needed: 2 points to set the range and 1 additional to fix that range in the correct place. If you were to calibrate a digital tehermometer using melting ice all you could be sure of is that is reads temperatures of 0 correctly. In the radar example all this calibration has done is ensure that the machine can identify a non moving object. Correct calibration is really difficult if performed properly and there has to be more to it than this.
|
|
All that self calibrating does in these cases is to establish the accuracy of speed detection on stationary objects !!
|
Lots of machines self-calibrate these days. For example the crash sensors in your airbag system self-check and self-calibrate every time you switch on.
To self-calibrate a doppler radar is relatively simple. You would either be using something which physically vibrates (e.g. an oscillator driving a piezo element) or you could excite some component which displays large impedance variations (e.g. a varactor diode) which would modulate the reflected radar signal. Any uncertaintainty is then in the frequency of the test oscillator, but realisitically its going to be extremely reliable.
|
|
|
All that self calibrating does in these cases is to establish the accuracy of speed detection on stationary objects !!
No, I'm sure that's not true.
|
It's calibrating against a stationary object so the only speed it's accurately calibrated to is zero.
It's accuracy at other speeds is dependent on design and correct operation NOT on calibration.
It's like saying that if a car speedo is accurate at zero then it must be accurate at 70mph - but we all know that doesn't happen.
|
It's calibrating against a stationary object so the only speed it's accurately calibrated to is zero. It's accuracy at other speeds is dependent on design and correct operation NOT on calibration. It's like saying that if a car speedo is accurate at zero then it must be accurate at 70mph - but we all know that doesn't happen.
No, you misunderstand. The Doppler radar system works by determination of the beat frequency between the outgoing radar signal and the reflected (and Doppler-shifted) signal from the car. At 0mph there is no output. What you need to check and calibrate is the frequency of the radar oscillator and whatever oscillator/counter is being used to measure the difference frequency. Usually these oscillators are very stable, but do show some drift with temperature and with time.
The normal way to calibrate these, as I said above, is to use a tuning fork or some device which can exhibit big impedance changes (like a varactor diode) - these effectively and very accurately provide a simulated return signal so that the oscillator accuracy can be checked and calibrated. You only really need to use a real car to check the aiming of the system and camera timing etc. Most handheld police units come with a tuning fork in the case for periodic checking and calibration.
|
|
CAMIC machines also self-calibrate, if I understand correctly.
|
Radar speed measuring guns (not talking lasers here) emit a continuous wave, typically at a microwave frequency (Sorry TVM but you are way way off in your description of how these radars work.)
Just to cut down on the irritating wasted verbiage that seems to be trotted out attacking most of the serious technical contributions made in these HJ columns, I have spent several years in microwave design and have expired patents to prove it and was at one time an expert witness on this gun subject.
The wave bounces off the target vehicle, from the surfaces that are normal (at right angles) to the line going to the vehicle and back to the radar. There the return signal is added to a sample of what is being transmitted. (the text books do it the difficult way -- or did -- via a doppler explanation.)
If the vehicle is stationary, that sum signal shows a steady amplitude because the two signals have a fixed relative phase. If the vehicle is moving the relative phase of the two signals is changing and so the sum signal exhibits an amplitude modulation --- sometimes the two are in phase and add, sometimes they will be in antiphase and subtract. In fact, one cycle of amplitude modulation occurs per half wavelength change in distance. So speed is so many half wavelengths per second.
This amplitude ripple signal is extracted by what is called a mixer (usually when describing Doppler shift) but can also be seen, from the above, as an amplitude extractor. Either way, that ripple frequency is used to derive speed after it has been amplified to make it bigger.
For instance, convert it to pulses at the same frequency as the ripple one and count pulses per second as a measure of speed. Pulses per second are proportional to microwave frequency. So that must stay accurate. Also the indicated result depends on the accuracy of the counting circuit. The gating/counting period is usually derived via a crystal signal so as to be known and stable.
There are other circuit tricks that must be used to verify the authenticity of the speed signal and remove the effects of spurii.
Part of the calibration check of the equipment makes use of a tuning fork to produce a known amplitude ripple frequency so as to check the counting circuit. Do not confuse this process with producing a doppler signal. It is a phase wobble process and can now be understood from the above.
The microwave frequency accuracy is typically about 0.1%, (and hence capable of being part of a very accurate speed measuring process) but microwave oscillation carries a risk of jump mode and hence frequency.
The next most likely mode is 1.5 times up in frequency, depending on the oscillator design. However, the Microwave Associates oscillator that was used in the most popular police speed Gunn in my day, the Muniquip, was of a very reliable design. It is possible to test for jump mode tendency in the lab.
But ultimately, the only way a layman has of proving all is OK (outside of a lab) is to run a car against it at a known speed. Here you are not usually looking for percents of error, but rather to rule out a spurious microwave frequency.
The oscillator frequency is best measured and set in the lab. Indeed it has-to/should meet frequency regulation standards.
|
So how is it, that on a recent TV program, a stationary car was shown to be moving at speed and a car going at a fixed speed was zapped and the radar gun rerturned a speed figure substantially higher ?
Wasn't there also the demonstration that showed a radar gun aimed at a wall and said wall was doing 20mph !
|
You are confusing the laser gun, as shown on TV, with the radar gun which I described.
Lasers measure distance and use the change of distance as a measure of speed. Light has such a short wavelength, typically a million times shorter than what comes out of a microwave gun, that it will reflect back from the fine structure of brickwork. So you just walk the beam spot along the wall during the measurement and get the speed of the spot moving along the wall.
|
|
I tend to agree - I don't see the doppler shift being a big effect with microwave radar - the speed of the car is tiny when compared with the propagation speed of the wave.
Number_Cruncher
|
Slight slip, for | speed along the wall | read changing spot distance -- there is none if the wall is a right angles!
N_C: The shift is of the order of 1 KHz at 30 mph for a 10.587 GHz radar. (as was then in use) What is very nice about this, electronically, is the signal processing is done at audio type frequencies. So very simple amplifiers are used after the mixing.
|
Thanks buzbee,
Just for interest, here are some estimates (ordinary Doppler and relativistic Doppler) of the amount of Doppler shift there is;
v30mph is the vehicle's speed in metres per second,
vwave is the propagation speed of the electromagnetic wave, 3e8m/s
v30mph=vms
v30mph =
13.41083333333333
(vwave/(vwave+v30mph))
ans =
0.99999995529722
sqrt((1-(v30mph/vwave))/(1+(v30mph/vwave)))
ans =
0.99999995529722
i.e., virtually none!
To complete the picture, looking at the phase change, or analgously the interferometric operation, at 10.587GHz, the wavelength is about 28mm.
In 1 second, the vehicle moves 13metres, and the light making a round trip travels an extra 26m
So 26m / 28mm gives approx 1000 wavelengths per second, which tallys with the 1Khz you describe. Although I'm not too sure how the mixer would actually work - does it produce an output for each phase change of 2*pi, or does it work on half cycles?
Number_Cruncher
|
A mixer will produce the sum and difference of input frequencies along with various spurious products that can be minimized with good design.
In this case the 10GJz reference signal is mixed with the 10.000001GHz reflected signal to give a 1KHz signal, a 20.000001GHz signal and various other spurii that are easily filtered out.
The accuracy of the reference signal doesn't make a huge difference to the measured result.
Its certainly easy enough to generate signals with better than 1ppm accuracy, every mobile phone manages this.
--
I read often, only post occasionally
|
|
|
|
|
|
|
