Phase detection of RF carriers

expect to sub-contract the design work but looking for an opinion first about practicality.

range, and report their relative phase with at least 50 nsec accuracy.

advance and are assumed to be perfectly stable. Nominal C/No >50 dB/Hz each carrier.

want to know is how much one leads or lags the other in time, like measuring the relative zero-crossing times of the two sinusoids. For measurement accuracy it is acceptable to time-average each measurement over as long as 100 msec.

assuming this concept is practical. Any recommendations for someone to do the design are appreciated. We are located in Southern Calif.

curve fit of the difference tone amplitude against time? How do we get relative phase of the two RF carriers from the curve fit?

And there is never enough of either. Sometimes I wish the day would have

30h, and I bet the FCC wished that UHF had another dozen or so 100MHz slivers they could hock for a ton of money :-)

Nowadays we can say it's easy to measure but that was different when we were young. Back then the only way to measure frequency (somewhat) precisely was to beat it with known signals, hoping the radio stations were tuned well. Which was sometimes a dicey bet. Same with time, the most precise instrument we had was our grandparent's regulator clock. Then at age 18 or so I had enough dough from my army pay that I could buy a 1MHz crystal and some TTL chips to build a more precise timing device.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg
Loading thread data ...

Joerg schrieb:

Hello,

a good pendulum clock was more precise than a cheap crystal oscillator without oven or temperature compensation.

Bye

Reply to
Uwe Hercksen

That adds delay!

The beauty of the timestamper is that it reports the edge times of every RF cycle with no obligation to think about what that may actually mean.

--
John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   
 Click to see the full signature
Reply to
John Larkin

Yes, but there can be minor inconveniences when you have to measure whether a time interval is 112.3msec or 112.4msec and there is only one event.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

expect to sub-contract the design work but looking for an opinion first about practicality.

range, and report their relative phase with at least 50 nsec accuracy.

advance and are assumed to be perfectly stable. Nominal C/No >50 dB/Hz each carrier.

want to know is how much one leads or lags the other in time, like measuring the relative zero-crossing times of the two sinusoids. For measurement accuracy it is acceptable to time-average each measurement over as long as 100 msec.

assuming this concept is practical. Any recommendations for someone to do the design are appreciated. We are located in Southern Calif.

curve fit of the difference tone amplitude against time? How do we get relative phase of the two RF carriers from the curve fit?

I once visited ONI (Offshore Navigation Inc), who did Raydist (vaguely loran-like LF hyperbolic) navigation services for oil rigs in the Gulf of Mexico. They had a wall full of racks of GR gear that was, essentially, a huge frequency counter. There were ovenized oscillators, WWV receivers, LC wavemeters, interpolation oscillators, lissajous oscilloscopes, all sorts of wild stuff. It used to be hard to measure frequency.

An oil company would hire them to put their gear, and an operator, on a boat to locate and relocate well heads and such. It was CW RF, all incremental tracking of hyperbolic lanes, so if there was a power glitch or something they lost the position, and they'd have to motor back to a known reference point.

--
John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   
 Click to see the full signature
Reply to
John Larkin

Yes, but a known delay.

The band above AM radio is full of static, distant thunderstorm pulses, crackling, switching power supply noise, marine transmitters, PLC crud, DSL leakage and whatnot. It will be tough to detect things like zero-crossings with any amount of reliability in a wideband setting.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

On Wednesday, August 8, 2012 10:06:59 AM UTC-7, Joerg wrote: [ about making phase determinations on circa 2 MHz radio]

So, the receiver will mainly be a PLL locked to the input signal. If you use a PLL that generates in-phase (A_sine) and quadrature (A_cosine) signals (really easy to do), then you can do start-stop timing on edges of the two phases of signal A against a single phase of signal B.

When phase is zero, a flipflop set on A_sine and reset on B_sine has zero pulse width, and increases linearly with phase until 360 degrees (when it goes back to zero).

The PLL loop filter will limit how nimble the frequency-phase changes in the input signals are represented, but also allows easy filtering

Reply to
whit3rd

pulse width,

zero).

That won't be very efficient. All it takes is a nice big tropical thunderstorm in the vicinity and it'll all go off kilter. Or a big fat noisy transmitter a couple of kHz higher. It'll cause the PLL to veer.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

expect to sub-contract the design work but looking for an opinion first about practicality.

MHz range, and report their relative phase with at least 50 nsec accuracy.

advance and are assumed to be perfectly stable. Nominal C/No >50 dB/Hz each carrier.

want to know is how much one leads or lags the other in time, like measuring the relative zero-crossing times of the two sinusoids. For measurement accuracy it is acceptable to time-average each measurement over as long as 100 msec.

assuming this concept is practical. Any recommendations for someone to do the design are appreciated. We are located in Southern Calif.

curve fit of the difference tone amplitude against time? How do we get relative phase of the two RF carriers from the curve fit?

Back in the 1970s, NBS Boulder was building laser frequency chains (colour centre laser locked to the 5th harmonic of a CO2 laser locked to the Nth harmonic of a methane laser....locked to a caesium clock). In the 1960s, people were doing Mossbauer effect measurements looking at the differential tuning of X-ray emission and absorption due to general relativity effects between the basement and the attic of a building.

You and I were just stuck with cut-price toys, is all. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
 Click to see the full signature
Reply to
Phil Hobbs

AT-cut crystals can do 2 ppm/K tempcos barefoot. Good luck getting that on a pendulum.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
 Click to see the full signature
Reply to
Phil Hobbs

expect to sub-contract the design work but looking for an opinion first about practicality.

MHz range, and report their relative phase with at least 50 nsec accuracy.

advance and are assumed to be perfectly stable. Nominal C/No >50 dB/Hz each carrier.

want to know is how much one leads or lags the other in time, like measuring the relative zero-crossing times of the two sinusoids. For measurement accuracy it is acceptable to time-average each measurement over as long as 100 msec.

assuming this concept is practical. Any recommendations for someone to do the design are appreciated. We are located in Southern Calif.

curve fit of the difference tone amplitude against time? How do we get relative phase of the two RF carriers from the curve fit?

I remember jumping for joy in the 70's when there was a sale on crystal analog alarm clocks, five Deutschmarks. So I grabbed my bicycle, threw it into 10th gear and floored it. About five miles, got there just in time, only a few clocks left. Two became mine and they were pressed into service for all kinds of experiments. Strangely, neither became butchered and they still run fine. One is now my barbie timer clock so I can tell my wife when dinner is going to be ready on weekends.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

.

o pulse width,

zero).

Yep. PLLs are the best flywheels ever known. The external noise in that f= req range can be 60+ dB above kTB. In fact, external noise is so dominant = that receiver G/T doesn't matter much at all. Lossy ferrite antennas with = their poor receiver G/T are a don't care because they lose as much signal a= s noise ... :o) =20

Reply to
George

This really sounds like RAYDIST, except the carriers are farther apart. Raydist evidently used resolver (or similar) driven mechanical counters.

Steve

Reply to
Owen Roberts

It was Raydist. And it did drive resolver-type motors and mechanical counters to track the path through the hyperbolic lanes.

Google has a lot of stuff on Raydist.

--
John Larkin         Highland Technology, Inc

jlarkin at highlandtechnology dot com
 Click to see the full signature
Reply to
John Larkin

Well, you DO need to add a GPS disciplining circuit to the pendulum first...

Reply to
Ralph Barone

If the frequencies are known precisely, as George indicated, then it is indeed possible to determine their relative phase. You just convert them to baseband and take the phase difference with atan2(Q0,I0) - atan2(Q1,I1). This is how modern phase noise analyzers work, in which the reference and input sources don't need to be at the same frequency.

As with instrumentation applications, band-limiting is vital for successful phase-difference recovery in the presence of noise. Some discussion of measurement bandwidth would be appropriate before any sure-fire answers can be given.

-- john (recently working on this stuff)

Reply to
John Miles, KE5FX

Not after the signal has been "processed" by the ionosphere. For example:

Incidentally, that's why Loran used 100KHz. No skywave.

The ARRL still conducts frequency measurement contests every year.

The results are currently in thousandths of a Hz.

Not bad for a 1 minute measurement window.

There's a Yahoo group for frequency measurement nuts:

Of course, there are also time nuts:

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
 Click to see the full signature
Reply to
Jeff Liebermann

(Also, the frequencies don't even have to be known precisely, as long as the LOs used for baseband downconversion are known. No need to downconvert all the way to DC, in other words -- you just need to end up inside the measurement bandwidth.)

Reply to
John Miles, KE5FX

The OP said that the SNR was at least 50 dB. I wouldn't have thought that sky waves ever got that good--do they?

And it isn't time and frequency that are hard, it's that hams and WWv fans are trying to correct for the spurious modulation due to a fluctuating soupy atmosphere. _Measuring_ the modulation is easy, it's only _correcting_ for it that's hard.

You guys are way overthinking this, unless the OP has other specifications that he hasn't given us.

Two signals at different frequencies have a perfectly well-defined phase relationship at any instant, provided:

  1. The SNR is high enough that you don't lose cycles in the measurement; say 20 dB for practical purposes.
  2. You pick some instant to be t=0.

Given those things, the phase difference is

Delta phi(t) = phi(0) + 2*pi* integral (0 to t) of (f2(t') - f1(t')) dt'

If one signal is known to have a higher frequency than the other, and the signal frequency variations aren't too fast, the required apparatus is:

one (1) mixer, e.g. an NE602, a Mini-Circuits diode bridge, or even a

74HC4016 plus a couple of inverters;

two (2) amplifiers, to get the signals up to the right level for the mixer;

one (1) simple lowpass filter to get rid of the sum frequency;

one (1) ADC to measure the filter output voltage; and

one (1) suitably programmed microcontroller to keep track of timestamps, do the curve fitting, and stream the data out over USB or Ethernet.

For a fancy version, the OP can use an SA614 ($2), which has a limiting IF and a mixer at the output, as well as a nice logarithmic signal strength indicator. A calibration signal would calibrate out the AM-PM conversion in the limiter and any nonsinusoidal behaviour in the mixer.

A second SA614 would provide a nice limiter output for driving the other port of the mixer, and with a bit of thought, one could easily get I and Q by using a phase shift network on one port of one of the mixers.

This can be done easily on a 60 mm square board, for a parts cost of probably $20 on a bad day.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
 Click to see the full signature
Reply to
Phil Hobbs

50dB SNR is quite unsustainable in that frequency range unless you are right next to the transmitters or each of them has its own nuclear power station.

Don't want to rain on your parade but let me throw in a few comments:

Measuring is difficult. How do you measure the deviations caused by a tropical thunderstorm going on 30 miles to the south? Lightning is totally unpredictable and it makes you amplitude and thus phase thresholds bounce like popcorn.

He wrote 50nsec which is 10% of a cycle at 2MHz. Very feasible if you have enough SNR, but that's where issues can sneak in.

A NE602 in the marine band will be like a Fiat 500 with a five-ton trailer :-)

You need at least something in the above +10dBm class from Mini-Circuits.

Amp before the NE602? That's like a Fiat 500 with a 30-ton trailer :-)

An IF limiter can work well here but never without narrowband pre-filtering. All it takes is Captain Gump calling shore to tell them he's returning from his shrimping trip and the whole system will keel over.

Wot's a meemaid'r? :-))

I'd concur, but I'd put it more in he $50 range, on account of better mixers, more LO power and most of all some nice filters. But only if you can have the crystals almost made for free somewhere in Asia. Out here you be looking at >$10, per crystal.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

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.