Using GPS to synchronise transmit frequencies

[k4kyv]

One thing that exacerbates the skywave interference problem on the AM broadcast band is that weaker stations in the background that share the same frequency may be off frequency a few Hertz, causing a fluttering effect.  If all stations could be synchronised exactly, so there would be no frequency error and therefore no beat frequency product, the interference would be less distracting and would tremendously improve the quality of nighttime AM reception, or even daytime reception when there is co-channel interference.

This is now feasable using the GPS satellite signals. Using GPS, the system acquires a signal from three or more satellites, providing a highly accurate time source, which is used as a reference point.  Using the GPS signal allows the frequencies from multiple transmitters at widely separated locations to be synchronized, and stay phase-locked indefinitely.

This technology could also be used with amateur transmitters, as the ultimate in frequency synthesis.  For example, two stations could agree to meet on 3885 kHz, then each would dial up 3885 on their DDS VFO's, and the carriers would, by default, be locked exactly on frequency.  Receivers and transceivers could be equipped as well. The cost shouldn't be much more than with existing equipment, using similar technology to what is currently employed in even the cheapest cell phones.  This technology could most easily be incorporated in SDR systems such as the FlexRadio.

http://www.freepatentsonline.com/7103374.html

[Steve - WB3HUZ]

Rather old idea. That freq sync with GPR stuff has been done since the 1980s.

[w1vtp]

Rather old idea. That freq sync with GPR stuff has been done since the 1980s.

Yes, it's an old idea, a but the practical application is "today."   I'm not sure it's necessary but certainly since the amateur radio tradition is excellence -- why not implement it.

Al

[Steve - WB3HUZ]

I told that to some SSBers probably 20 years ago. Then it would have been a good way to get everyone exactly on frequency for better sounding audio. With synthesized and usually very accurate VFOs in the newer rigs, it's probably not required.

I think some of the TAPR guys were looking at using GPR to sync up some DSSS and FH type systems. That's probably still a good use.

[W3LSN]

It's already being done by those AM stations that have adopted HD Radio. The HD exciters typically come with a GPS receiver and antenna.

73, Jim
WA2AJM/3

[Mike/W8BAC]

I use the 10 MHz GPS reference here Don for my Flex Radio SDR 1000's and 5 other pieces of test equipment. I picked up a Brandywine GPS4 time and frequency standard and a Symmetricom 58532 LNB on eBay. I pipe the 10 MHz standard through a kit built ham bone W8ZR receiver multicoupler to the SDR's (2) and test equipment. It's nice to know I'm within a few parts per million of the indicated frequency on the transmitter/receiver. Better yet, my test gear is on the same master oscillatory.

Mike

[Mike/W8BAC]

Since my last post I realized I left something out. Don, I wonder about time shift due to atmospheric conditions. If two stations separated by, say, 1500 miles are exactly on the same frequency and you hear the flutter you describe, could this be a path distortion and not frequency offset issue?

My other point of interest is Clear Channel 780 KHz WBBM in Chicago has lately had interference on it's upper sideband. It is (no doubt to me) IBUZZ but I cannot find the carrier/culprit in the vicinity. If any of you are interested and happen to figure this out, let me know please.

73
Mike

[k4kyv]

I use the 10 MHz GPS reference here Don for my Flex Radio SDR 1000's and 5 other pieces of test equipment. I picked up a Brandywine GPS4 time and frequency standard and a Symmetricom 58532 LNB on eBay. I pipe the 10 MHz standard through a kit built ham bone W8ZR receiver multicoupler to the SDR's (2) and test equipment. It's nice to know I'm within a few parts per million of the indicated frequency on the transmitter/receiver. Better yet, my test gear is on the same master oscillatory.

It should be much better than "a few parts per million".  One part per million would be approximately 4 Hz on 75m. You should be able to sync to within a fraction of a cycle another station using a similar setup on any amateur band. Atmospheric perturbations might shift the phase back and forth a little, but at MF and HF I doubt if it would be enough to make an audible beat note, except perhaps during an aurora disturbance. It should be possible to make slopbucket sound almost as good as AM if the transmitter and receiver are in sync. Where the ESSB boys are missing the boat is running completely suppressed carrier and then manually tuning the other stations in.  With some voices it is easy to tell when there is zero frequency error, but with others it is more  difficult.  I find it almost impossible with female voices.

Another alternative for ESSB would be to run a pilot carrier about 20 dB down from p.e.p. and use a PLL type synchronous detector to lock the BFO onto the pilot carrier.  AT&T did that with their transoceanic radiotelephone circuits in the 40's and 50's, before the advent of satellites.

With all the software defined stuff out there these days, why aren't we hearing more talk in amateur circles about this kind of experimentation?

And why aren't all AM broadcast stations synchronised on each channel, whether running hybrid digital or just plain analogue?

...and they have the gall to tell us that we are outmoded because we run full carrier DSB AM!

[N4LTA]

GPS is the frequency standard used by most all of the POP digital service providers.

I did the engineering on quite a few back 8 -10 years ago. One of the electrical engineering tasks was to provide a GPS roof mounted antenna. We used 9913 to route from the antenna to the GPS clock.

The GPS clocks are so accurate that they have to have periodic correction for the relativistic effects of the earths gravity slightly warping space time which makes them run different than clocks on earth. This is one of the effects that accurately tested Einstein's theory. If the corrections are not done, the accuracy would drift off to errors of kMs in a few weeks.

Pat
N4LTA

[WA1GFZ]

TAPR just sold a pile of Thimble Thunderbolt modules used in cell sites as a GPS clock source. They put out 10 MHz and 1 PPS. I plan to use it as a reference in my shack. Not sure if it means much on 75 AM.

[Ed/KB1HYS]

I always thought that the human ear could detect low frequency beat notes pretty well.  When I've tuned guitars and such starting off with a known pitch, the rest of the strings I had tuned by ear were right on when checked with a Korg tuner.

The point is, I think a properly Zero beat signal would not be off by more than a very few cycles anyway, you can easily hear a 5 cycle beat note. How much more accurate do you need to be? 

[w3jn]

Dave, W3NP has the Fluke sig gen he uses for a VFO GPS synced, as I recall.

[k4kyv]

I always thought that the human ear could detect low frequency beat notes pretty well.  When I've tuned guitars and such starting off with a known pitch, the rest of the strings I had tuned by ear were right on when checked with a Korg tuner.

The point is, I think a properly Zero beat signal would not be off by more than a very few cycles anyway, you can easily hear a 5 cycle beat note. How much more accurate do you need to be? 

If your AM receiver has good low frequency response you can hear the beat note.  Typical "communications quality" audio rolls off at 200~ or even higher, so the beat note just disappears as you approach zero-beat.

With slopbucket, there is no carrier to zero-beat against.  You have to adjust the sound of the voice by ear.  If the frequency error is more than a few cycles off, the voice sounds unnatural, particularly if the lower frequencies are preserved, as the ESSB gang tries to do.

I have carefully tuned in some ESSB signals using the synch detector in manual mode.  If you get it just right, some of them sound very good, near broadcast quality.  But where they are missing the boat is not having some means of easily synchronising the re-inserted carrier to the suppressed one.  The easiest way to do that would be to transmit a weak pilot carrier, enough for the PLL in a synch detector to lock onto.

The GPS synch-lock system would make the pilot carrier unnecessary.

By "ESSB", I don't mean slopbucketeers who run out and buy a load of audio processing gear and then try to force-feed broadcast quality audio through the stock filter on a ricebox by using an equaliser to turn the highs and  lows, that fall outside the passband of the filter, way up.

[Ian VK3KRI]

I run a rubidium clock that I compare long term against a GPS, currently its about 7.5 x 10e-11 off frequency  - thats 0.00075 Hz @ 10 MHz . I guess its time to tweak it again. 

Why bother ? - because I can!

                                                       Ian VK3KRI

[WU2D]

Can I frequency lock my Type 10 Hartley with that?

[W2ZE]

My other point of interest is Clear Channel 780 KHz WBBM in Chicago has lately had interference on it's upper sideband. It is (no doubt to me) IBUZZ but I cannot find the carrier/culprit in the vicinity. If any of you are interested and happen to figure this out, let me know please.

The only real powerhouse on 790 is KABC in LA. There are a lot of other stations on 790, but not strong enough to interfere with BBM. if your hearing solid white noise on 790, its WBBM's upper IBOC carriers. If your hearing a frying egg sound on WBBM, then its self interference from its secondary and tertiary carriers, possibly its core (primary) carriers if the rx is wide enough.

[Bill, KD0HG]

Don and all:

There is a real problem with frequency/phase synchronized AM transmitters.

A bit of background: The idea is an old one. Goes back to the 1930s. The idea then as now was that frequency-locking two broadcast stations sharing the same frequency would eliminate the beat note and thusly minimize interference with each other, as Don suggested.

The problem was, there was no way to do that with 1930s technology. No GPS or atomic clocks back then, crystals in ovens was the best they could do. So the beat went on and the idea was abandoned for many years.

So now we can lock transmitters on exactly the same frequency, even in phase. That gives us another problem. There are going to be deep nulls and peaks in the field between two separated phase-locked transmitters. Every half-wave between then you'll get a deep fade where the signals are 180 out of phase, move on a bit and there will be a peak when they're exactly in phase. A peak or null could stay in the same physical area for weeks or months. And it does exist.

The same effect happens on skywave as the height of the reflecting layer changes. Big time fading effects, because the transmission medium isn't stable and the two signals phase in and out, even though they're locked on precisely the same frequency.

The guy that hired me for this job is now the CE of KKOB-AM, AM 770 in Albuquerque. He operates one of the most famous synchronized AM stations in the country. 50 KW, but they have to put a deep null to the NE at night to protect WABC. That null, unfortunately, is right over the top of Santa Fe, the state capitol and a major market.

So for many years, KKOB has used a synchronized low-power transmitter near Santa Fe at night to solve the problem. But even though they can now precisely phase-lock both transmitters together, they don't.

Because of the null-peak effects between both transmitters as I described. So now, KKOB has a deliberate frequency offset between their transmitters of a few HZ, which physically moves the nulls around a geographic area. It's mainly audible in the overlap zone where both signals are roughly equal strength, but it is an annoyance.

I believe that WLLH 1400 in Boston has a similar setup with the same issues. A transmitter in Lowell and one in Lawrence with a reported mush zone in between, but I'm not familiar with that area.

[Mike/W8BAC]

Hi Mike,

What I'm hearing, and seeing is what looks, and sounds like somebody's AM digital signal that is appearing in BBM's analog upper sideband space. Looking at the new broadcasts you see the carrier in the center with analog audio on either side out to about 10 or so KHz and a digital hump above and below the analog audio. Usually you can tell who's who with the spectrum display but I haven't figured out who's digital signal it is over BBM's upper sideband analog audio. Thanks for your reply and sorry for the hijack.

Interesting story Bill. You put it together for me.

Mike

[k4kyv]

The problem was, there was no way to do that with 1930s technology. No GPS or atomic clocks back then, crystals in ovens was the best they could do. So the beat went on and the idea was abandoned for many years.

I once read in one of the 1930's broadcast rags that a couple of stations accomplished it by  sending the synchronising signal over a phone line.  I think this was also done by the BBC for national coverage on a single broadcast channel.

So now we can lock transmitters on exactly the same frequency, even in phase. That gives us another problem. There are going to be deep nulls and peaks in the field between two separated phase-locked transmitters. Every half-wave between then you'll get a deep fade where the signals are 180 out of phase, move on a bit and there will be a peak when they're exactly in phase. A peak or null could stay in the same physical area for weeks or months. And it does exist.

I think you are talking about where the same program is transmitted from both synchronised transmitters, and they are trying to attain solid coverage.  The peaks and nulls would be a problem where the signals are close to equal in strength, but it shouldn't be a problem when the interfering signal is much weaker, as in the case of the AM broadcast band when the desired signal often has a weak distant co-channel station in the background with a different program, causing interference and an audible flutter.

The guy that hired me for this job is now the CE of KKOB-AM, AM 770 in Albuquerque. He operates one of the most famous synchronized AM stations in the country. 50 KW, but they have to put a deep null to the NE at night to protect WABC. That null, unfortunately, is right over the top of Santa Fe, the state capitol and a major market.

So for many years, KKOB has used a synchronized low-power transmitter near Santa Fe at night to solve the problem. But even though they can now precisely phase-lock both transmitters together, they don't.

Because of the null-peak effects between both transmitters as I described. So now, KKOB has a deliberate frequency offset between their transmitters of a few HZ, which physically moves the nulls around a geographic area. It's mainly audible in the overlap zone where both signals are roughly equal strength, but it is an annoyance.

The same effect happens on skywave as the height of the reflecting layer changes. Big time fading effects, because the transmission medium isn't stable and the two signals phase in and out, even though they're locked on precisely the same frequency.

It would seem that the scheme would be OK only where the directional null fell right smack in a heavily populated area that is surrounded by a large rural area. The station would sacrifice some coverage in the sparsely populated  rural area while regaining coverage in the highly populated area, using the  low power transmitter.  I would think the pulsating beat effect of the two stations slightly off frequency from one another would be just as devastating as the fixed nulls and peaks from synchronised carriers.  The only solution where the signals are close to the same strength would be a third station on a different frequency channel altogether.  Maybe that would be a use for the new FM translators the FCC has authorised for AM stations.

[Bill, KD0HG]

So now we can lock transmitters on exactly the same frequency, even in phase. That gives us another problem. There are going to be deep nulls and peaks in the field between two separated phase-locked transmitters. Every half-wave between then you'll get a deep fade where the signals are 180 out of phase, move on a bit and there will be a peak when they're exactly in phase. A peak or null could stay in the same physical area for weeks or months. And it does exist.

I think you are talking about where the same program is transmitted from both synchronised transmitters, and they are trying to attain solid coverage.  The peaks and nulls would be a problem where the signals are close to equal in strength, but it shouldn't be a problem when the interfering signal is much weaker, as in the case of the AM broadcast band when the desired signal often has a weak distant co-channel station in the background with a different program, causing interference and an audible flutter.

Don, you're not quite correct.

Assume two AM transmitters locked on a particular frequency; one in Chicago and the other in Nashville. They're running different programming. Someone east of the Mississippi is listening on channel via sky wave.

What happens as the Ionosphere shifts?

The phase between the two received carriers changes. They NEVER can be in constant phase on skywave, other than for accidental moments.

Sure, you eliminate a low frequency growl between the two transmitters, but they'll phase in and out with each other in any particular spot as the Ionosphere shifts.

Another way to think about it is if you had one transmitter feeding two separate antennas (maybe a phased array of separated verticals). If you randomly changed the length of the feedlines of the antennas, what would be the net effect on received signal strength in any given spot?

I'm trying to say that, yes, you are correct about the annoying beat note between two AM stations, but operating transmitters phase-locked creates it's own set of problems, some of which are alleviated by unlocking the two stations, back to square one.

You're trading an audible beat note for signal cancellation effects.

I did synchronized VHF FM two way (police, fire) and paging transmitters years ago. The signal level flutter as a mobile passed through signal peaks and valleys was awful.

(On VHF the peaks and nulls were a few feet apart, on HF they'd be dozens of feet apart, to dozens of yards apart on the BCB))

There were only two ways to fix the issues- Deliberately offset the two transmitter's RF carrier frequencies by a few HZ, or deliberately jack around with the audio delay between both of the transmitters. Yes, it sounded like cats fighting when both transmitters first keyed up, but the adverse effect on intelligibility was less than when multiple transmitters were all locked together and the RF carriers went in and out of phase because of receiver movement.

That annoying low freq beat between stations is probably the least harmful to intelligibility.

[k4kyv]

I think you are still talking about listening to two signals of approximately the same signal strength. The locked carriers would improve local ground wave reception, for example, when there is weak sky wave interference in the background - a common problem on the  local "graveyard" channels.  I am not talking about someone in Little Rock trying to listen to a station in Chicago when there is a similar station in Atlanta transmitting on the same frequency at about the same power.

The phasing problem exists only if the interfering signal is substantially close to the desired signal in strength.  If the interfering signal is a marginal case of, say -20 dB from the desired signal, the fluttering beat note that modulates the interference from the sidebands is in the background, but audible and irritating. The fluttering effect of the carriers themselves is audible, too.  Kind of like rapid QSB.  If the carriers are precisely locked, the interfering sidebands are still there, but the fluttering effect is gone.  The fluttering of the carriers is gone. The phase cancellations between the carriers of the desired and undesired stations would have negligible effect to the listener.  It would be roughly like a QSB varying the carrier amplitude +/- 10%.

If the interfering signal is close enough in strength to the desired signal for the phase cancellations of the carriers to cause a problem with the listener, the interference from the undesired station will be too much for satisfactory reception of the desired station regardless of how the carriers are phased.  It then becomes more like listening to a signal on 75 with heavy QRM conditions - the signal is readable but not too pleasant to listen to.

QRM between two AM signals is less damaging if the stations are zero beat than if they are 500~ off frequency from each other.  A tuner-upper does less damage if he carefully zero-beats the QSO than if he is audibly off frequency.

[w1vtp]

Don

Would not the phasing problem between two stations that are phase locked using GPS be resolved using Sync detection?

Al

[WA1GFZ]

If both signals traveled different distances there would be a phase shift. Say signal A bounced back at 100 miles up and signal B bounced back at 101 miles up. I don't see how you could ever adjust for that variable on AM