using RF to measure distance 1

[shamone]

Hi, I am trying to measure the distance between a stationary point and a moving point up to at most 1200ft. I think RF is the best(or only way) to do it. Wondering if anybody knows an effective, cost efficient, method of doing this. The distance is not always line of sight either. Thanks.

 

[Warpspeed]

O/k fair enough. Do a bit of a search into telurometers and how they work. These are surveying instruments that can measure over very large line of sight distances to extreme accuracy.

The basic idea is you modulate a microwave (or optical) carrier with a much lower frequency, and measure the phase of the modulation at the other end. The lowest modulating frequency may have an effective wavelength of a mile or more. That would tell you very roughly where you are within a whole mile.

You then switch the modulating frequency to a shorter wavelength. Perhaps a tenth of a mile, and measure the phase of that. You just keep switching ranges down to successively finer resolution.

It is like reading a clock where you have hour, minute and second sweep hands. Each clock hand just by itself is pretty useless, but with all three readings you can easily get down to a few seconds resolution in 24 hours in one quick glance.

Telurometers are a bit more complex, but that is the basic concept in its crudest form.

 

[zeusfaber]

You don't specify requirements for endurance, data rate, latency, resolution, area over which the system has to operate, rate of movement of the moving object, or any exclusion rules.

If none of these is demanding, you might get away with borrowing a GPS and some string to help you paint a few range lines on the field, then paying some kid with a phone to stand beside the moving machine and tell you where it is while you type the figures in at the keyboard (or even just write it down for later analysis)

Only make it complicated if it needs to be complicated?

A.

 

[shamone]

Thanks for the input zeusfaber. Yes the reason why there are no requirements cuz i am new to all of this so i don't know what to look out for but i am sure those wil come oup during the design process. Here are some more details: The object that is moving will be stationary when the measurement is taken place. THe key here is to make it much more feasabel than GPS. That is why i am seeing if there any other ways to do it.

Cheers

 

[zeusfaber]

You have an interesting problem here. You're trying to measure an inconvenient distance to better than half a percent accuracy.

Half a percent is quite feasible, but is well into the territory where most measuring devices are going to have to be made first, then calibrated.

If you buy, borrow or steal proper surveying gear, a laser rangefinder or a GPS, someone else will already have done this calibration for you.

If you make your own kit, you'll need to calibrate it yourself - for which you'll need to obtain an accurate means of measuring distance to start with, taking you right back to one of the bought-in options. You've got to ask yourself what added benefit the homebrew gadget will bring to outweigh the added error, hassle and expense.

This benefit could be real - for instance, you may only need to hire the expensive gear for one day instead of the whole trial period, and once calibrated, the new toy might be quicker and easier to use - but all these depend on exactly what you want to do.

"Systems" thinking preaches the discipline of understanding your requirement really thoroughly before you get stuck into design work (the politicians who spoil my day job reckon we ought to be spending up to 15% of the project budget before we even start development). Might be a bit doctrinaire, but it helps keep you and your money out of the blind alleys.

A.

 

[shamone]

Perhaps i am missing the big picture here. I was planning on using to transceiver chips. To measure the distance i was just going to measure(with a uP) the time it takes to send and recieve a signal and put that into d=ct equationg to get the distance. Could you maybe explain the need for calibration in this matter? Thanks

 

[zeusfaber]

VE1BLL hinted at this earlier on: The linearly distance-variant bit (about 2.5 microsec) is only part of the cycle time. There will also be the time it takes the transponder to discriminate the incoming signal from the background noise, decide to reply, and build up transmitter power.

Some of these things are likely to vary with distance (for instance, the further you are away, the weaker the received signal is and the longer it is likely to take for the receiver to recognise that it is there. If you just use d=ct, you will overestimate distance - probably increasingly so as your target gets further away.

The way round this is to calibrate the device by recording output at a variety of accurately known distances.

As IRStuff has already pointed out, you'll need to measure your cycle time to within 10 ns. I'm not sure how easy that's going to be just using a uP.

A.

 

[shamone]

Cool, I understand what you are saying now. Thanks so much. Well I found a transceiver chip designed for uP control. It has a CTS(clear-to-send) pin that goes low whenever an incoming signal is received. That is what i am going to use the counter on cuz soon as that bit goes low, i know that the signal is receivee however i will make sure that what is received is what i sent out. Anyway i am gonna give it a try. It is experience for me which i really need so i can't lose either way(new grad from engineering). Thanks for all the help guys and i will keep you posted.

 

[IRstuff]

I can't think of a single processor in the entire universe that can respond reliably and repeatably within a time calibratible to 10 ns.

In a standard rangefinder, ALL of this is done ONLY in discrete or FPGA logic. Even then, massive temperature calibration is required to assure accuracy, since the detection logic chain is usually multiple gates, which means that your logic timing must be known to better than 1 ns per gate.

TTFN

 

[shamone]

frick, didn't even think about that. This is slowly turning into a lost cause. Ah well, i shall keep reading.

Cheers IR

 

[Warpspeed]

The processor itself may not have to. External hardware would make the actual phase/time measurement, the processor could crunch the numbers produced at a much more sedate pace.

Ten nS time resolution only requires a 100 mHz clock or sampling rate, which is by no means fast these days. But as has been previously pointed out, analog circuit delays may be highly variable with signal strength adding a most worrying uncertainty to the whole process.

 

[IRstuff]

Frankly, when one is faced with a technical obstacle, it behooves you to review your premises.

Your desire to find a "cheaper alternative to GPS" is, I think, based on a fallacy. I've yet to find even the simplest laser rangefinder that's less than $100, while I just bought a GPS with mapping software for $79. The reason is that GPS is ubiquitous and mass produced, while rangefinders are less so. RF rangefinders are generally useless in ground applications, so there is essentially zero market and you'd have to roll your own completely from scratch.

Given the range involved, you could buy two GPS receivers, one as a basestation, RF modem the remote GPS to the base station and brute-force a differential GPS for less than the cost of a decent laser rangefinder with RS-232 output (the cheapest I found was running over $500).

TTFN

 

[Warpspeed]

Have to agree with IRstuff. Off the shelf mass produced hardware is pretty hard to beat for value.

Quite often the whole professionally built system can be manufactured cheaper in Asia than just buying the bare components over the counter to build it yourself here. Plus you get all the packaging, development expertise, software, and reliability for free.

 

[shamone]

I think you gentlemen are right. I was looking for a cheap alternative for GPS but after i did do some research on GPS and saw that it is a reasonable price with tones of cool features which makes it really hard to beat. I will keep research for an education stand point but as far is development, well........maybe not. Once again, appreciated all of the input.

Mark

 

[IRstuff]

It's one of the perversities of engineering that something "simple" such as measuring the time of flight, as was attempted by the ancients would prove to be more complex and expensive than launching 24 satellites with rubidium time standards and using 12-channel all-in-view receivers.

Go figure...

TTFN

 

[shamone]

Ha, too true and quite funny. Well it is nice to have these forums where one can share his ideas and get realistic feedback. I hope in the futur I can do the same as you guys have done for me. Happy engineering!

Mark

 

[SolderingGunslinger]

Such a system was described by Hugo Gernsback in 1912 in his book Ralph 24C41+

It was developed for practical use in the late 1930s

It's called RADAR.

Now, you can get a GPS engine a whole lot less expensively than a radar unit and you can adapt APRS (Check on ARRL.ORG about APRS) to have the remote unit radio its position on a VHF transceiver.

Good Luck with yourexperiment.

I remain,
The Old Soldering Gunslinger

 

[VEBill]

SiRF makes some very interesting GPS chips. (Disclaimer - my cousin works there. But they're still very interesting chips...)

In case you wander back to the pinging architecture:

Multipath shouldn't be a major issue because the FIRST signal to arrive is the right one. The multipath can cause interference, so you need to get your measurements done quickly. I've seen this exact concept done (prototype) to eliminate skywave interference from MF navigation systems - grab the phase BEFORE the skywave arrives later (taking a longer path).

To eliminate other issues you put a tiny flag of some sort into the signal and you detect the flag, not the raw RF that might bring issues as mentioned by zeusfaber.

Before GPS there were all sorts of interesting methods of (for example) guiding an oil rig back onto the well (within inches) where the nearest fixed landmark was hundreds of miles away (over the horizon). Tricky business...

Still - GPS wins hands down for most applications.

 

[shamone]

That is what I was thinking VE1BLL. The first signal has to be the correct one. However with all the other technology out there it is only worth me doing just for the "heck" of it. Those rangerfinders IRstuff reffered me to do just the trick and they are cheap.

Cheers

 

[zeusfaber]

Let us know how you get on with the homebrew.

In the meantime, does anyone fancy a guessing game over the application for this?

A.

 

[analogkid2digitalman]

I would guess a potential application would be controller feedback for an automated crop circle making machine

Wheels within wheels / In a spiral array
A pattern so grand / And complex
Time after time / We lose sight of the way
Our causes can't see / Their effects.