Ideally someone in this group who has some microwave expertise would
> like to work with me on this project. All open source, of course.
>
Surely it would be *much* simpler to use two lasers, suitably spaced apart, positioned so that their light is interrupted when passed by the approaching car. You have them aligned so they hit a couple of photo diodes on the opposite side of the strip. Write a script to calculate the speed from the time interval between diode 1 and 2's blackout and the known distance between the lasers.
--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.
I do some RF design that includes some products that operate at GHz frequencies. A few points that may be useful.
If you're doing microwave radar, you'll probably end up in the
24Ghz band or higher. Other available frequencies are X, K, Ka, and Ku bands. 24GHz can be done on PCB boards, but not common G10/FR4 PCB's. There are quite a few options in microwave PCB's available all rather expensive: The common method is to split the PCB into a small area that requires low loss and well behaved PCB material for the microwave stuff, and a larger FR4 section for the lower frequency stuff.
If you're doing RF layout on a PCB for your initial prototype, you might consider another way. Microwave modules and sub-assemblies are quite common. In general, they're housed in milled aluminum housings and interconnected with SMA or smaller connectors. Make a block diagram of your RF section, build it using modules, see how it plays, and then think about combining the boxes into a single PCB. MCL is a favorite vendor of such modules: For example, here's a general purpose transceiver implemented totally in MCL modules:
There are also web sites full of RF design aids and data. My favorites extracted from my messy bookmarks collection: and magazines:
RF and microwaves are all about systems design. The calculations revolve around transmit power, coax/waveguide losses, transmit antenna gain, path loss, reflection efficiency, receive antenna gain, more coax/waveguide losses, receiver sensitivity, system noise figure/factor, and demodulator efficiency. Also, receiver dynamic range, intermodulation problems, FCC incidental radiation compliance, protecting other users of your chosen frequency, and whatever else I've forgotten to list. For the RF part, try starting with Appcad 4.0 to give you an idea of what's involved: See 3 examples under Signals and Systems -> NoiseCalc section.
Gunn diode transceivers are still manufactured but are scarce: No clue on price or availability.
I can't offer much on test equipment required. I built up my collection more by what I could afford than what I needed. Much of it is totally obsolete by todays standards. Here's the stuff I use, mostly from the 1980's and 1990's. There's about 3 times as much stuff stored off camera: For RF and radar, you'll need a spectrum analyzer that covers all the frequencies involved (RF and IF), a signal source capable of simulating the transmitter, a very fast scope capable of seeing nanosecond rise times (to see the reflected pulse), a directional coupler for impedance matching, and a sweep generator for tuning filters. I think you can live without a network analyzer. You'll also need a pile of SMA and N adapters and pigtails, maybe some waveguide components, low loss patch cables, RF terminations, attenuators, step attenuators, diode RF detectors, calibrated test antennas, and associated hardware.
After you select an operating frequency, I can perhaps be more specific. If you use modules instead of designing an RF PCB and assume the modules meet spec, you can probably get away with just the fast scope and some of the RF hardware. However, if you need to do any manner of RF troubleshooting or optimization, you'll need the most expensive part of the puzzle, the spectrum analyzer. When you start on the PCB, you'll need most everything I listed.
I have two old Kustom Signals Falcon 24GHz radar guns. Something like these: One works the last time I checked, the other has something wrong with it. You're welcome to borrow one or both if you think it will help. Send email.
Good luck.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
Those Bird RF wattmeters are becoming somewhat collectable these days, Jeff. Do you have the complete range of inserts for it? They fetch about a hundred bucks a piece last time I looked (the inserts). Got one myself as it happens, plus over a dozen inserts. Never seem to use it so should stick it on ebay I guess. :-/
--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.
No. When I was in the 2way radio biz, everyone had a Bird 43 wattmeter and collection of elements (slugs). Later, I got into engineering and decided to test a typical Bird 43 for accuracy and the effects of impedance mismatch. It sucked. Basic accuracy is +/-2% of full scale on the meter and +/-5% on most of the slugs. So, if you're measuring a 3 watt transmitter output with a 10 watt slug, the error can be +/-7% of 10 watts = +/-0.7 watts. That means the measured power can be anywhere from 2.3 watts to 3.7 watts. Not the greatest, but the best that could be done at the time for a reasonable price.
The grey Bird meter in the photo is a Bird Termaline wattmeter. It doesn't use elements (slugs) and has a built in 200 watt terminator. I've blown the diodes out of it a few times, so I won't claim it's accurate. The black Bird meter in the photo has a built in diode detector and is calibrated for military frequencies between
200-400MHz. I made a calibration table so that it's usable on 150 and
450MHz. Again, no elements are used.
You won't get $100/ea for the elements (slugs). The running price is about $30/ea for the common elements and maybe $60 for the low power and broadcast elements:
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
So I've heard. My chief technical advisor here (Phd in electronics and still sharp as a razor at 78 years old) told me about the accuracy issue some time ago. He has one, too. I was surprised I admit, as this meter has acquired something of a mystical, legendary aura about it (at least here in Yurp) which may account for why they're so sought after and cost more here. I won't miss it if it goes as I've always preferred scopes for power measurement (two-tone test and all that).
--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.
What's the item immediately to the left of the Birds, just out of curiosity?
-- This message may be freely reproduced without limit or charge only via the Usenet protocol. Reproduction in whole or part through other protocols, whether for profit or not, is conditional upon a charge of GBP10.00 per reproduction. Publication in this manner via non-Usenet protocols constitutes acceptance of this condition.
I usually ignore one line comments because they usually have little of value to offer, but I'll make an exception this time.
I can't design anything that's even close to the state of the art. I'm much better at finding and fixing other peoples problems. However, if you want to pay me to educate myself in material science, high speed logic, semiconductor processes, exotic materials, patent evasion, and industrial espionage, I could probably produce something worth testing in about 10 years.
As for a faster CPU, I don't follow the rumors and press releases, but I did manage to find something interesting to help inspire additional topic drift:
Incidentally, Hittite (now part of Analog Devices) has a few high speed dividers in their product line-up. Near the bottom of the page: Maximum frequency is 18GHz, which makes then suitable for satellite TV local oscillators. I'll Start with a simple divide by 8 device: which dissipated 0.5 watts and probably has 20 gates in the divider section. An Intel i7 CPU: which has about 2 billion gates, a CPU using the same inGaP GaAs HBT (heterojunction bipolar transistor) technology as the Hittite device should dissipate: 2*10^9 / 20 * 0.5 watts = 50 million watts However, all the gates aren't simultaneously running at 18GHz. Assuming 0.1% are, that's still 50,000 watts. Did I mention that I'm also good at business plans, science fiction, and ruining everyone's day?
Prediction: The next generation of CPU will have enough cores for half of them to run the OS and applications, and the other half to keep an eye on the OS and application for spyware, tracking software, privacy violations, and system security. If there are any CPU's left, they can be configured as a Hyper-Snooper which keeps an eye on everything else.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
IFR-1500 service monitor circa about 1985. IFR is now Aeroflex: It currently has a sick power supply that refuses to fix itself:
On the shelf, under the pile of 4 assorted HP counters, is a similar SSI/Wavetek 3000 service monitor circa about 1978. More of the same with a pile of HP sweep generators on the left:
Both are sometimes referred to in the 2way radio biz as a "shop in a box". They contain everything you need to tune, align, measure, tweak, test, and bless a 2way radio. These became largely obsolete with the introduction of cellular and digital radios which require elaborate and specialized digital modulation testers and generators. I still deal with conventional 2way FM radios, so they remain useful.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
I just turned 70 and am rather dull. However, I've taken up knife making and sharpening as a hobby, and after some practice, hope to sharpen my wit.
Analog meter accuracy is often specified as some percent of full scale which means you should do your measuring in the upper 1/4th of the meter scale.
Most everyone in the radio biz has a Bird Model 43 and a box of slugs stolen from a previous employer. I've lost count of how many have disappeared from various factories and radio shops. There's nothing mystic about them. They're not particularly accurate, but since Bird didn't make any changes to the basic design since 1960, everyone's meter (including the FCC) had meters that read equally bad with the errors all in the same direction. If you do something wrong often and long enough, it becomes a standard.
So, how accurate is your scope for measuring RF power? Hint: Power = Voltage^2 / 50 ohms. So, if you make a small error in voltage measurement, that creates a much larger error in the power calculation when it's squared. Two-tone tests should be done with a spectrum analyzer, not a oscilloscope.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
The system bandwidth of a doppler speed meter is in the low KHz range. A time-of-flight distance measuring rig will need system bandwidth in the 100 MHz sort of range. So there is a s/n penalty in the rough order of 60 dB, favoring doppler.
The Radar Equation is cruel.
And it's easy to measure a doppler beat. It's more work to measure time of flight to nanosecond resolution.
If multiple vehicles are in the antenna cone, the doppler tones will be superimposed. It isn't hard to do filtering or FFT to separate the tones, in particular to nail the fastest driver. With TOF, all the pulse echoes will be confusing.
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
Can't say I've ever had any problems with the accuracy of the scopes I've used (got over a dozen of them here; most of them fully working and a couple recently calibrated). Should I choose to switch to doing this measurement with a SA, I have plenty of those to choose from too. But I'll stick with the scopes. 'Better the devil you know' and all that.
--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.
Interesting. The bulk of my stuff is from exactly the same era as yours. I got some Marconi sig gens that currently await my attention (they'll be waiting a while yet). Marconi also got swallowed up by Aeroflex IIRC. So much to do; so little time to do it all. Sigh....
--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.