Measuring TV Video Carrier Frequencies to 1 Hz Precision

[VK2SWL1]

VHF TV video carrier offsets are measured to 1 Hz resolution in order that specific transmitter locations can be assigned to geographic areas. This is of great benefit to six metre hams and TV DXers who need to know what geographic windows exist for VHF propagation.

Basic Setup:

Icom R-8500 fitted with high stability CR-293 PLL crystal unit.

5-el 45-70 MHz horizontal polarized yagi.

Spectrum Lab [1].

15625 Hz TV-Derived Frequency Reference Generator (outputs in 10 KHz steps over HF to UHF).
(see "Electronics Australia" Oct 93). Accuracy is specified to 0.001 ppm.

The 15625 Hz TV-derived reference unit is connected to a standard VCR via the video 1v p-p line-out. The VCR is tuned to TCN9 Sydney in order to receive 15625 Hz sync pulses from a Rubidium transmitter source.

The first step is to ensure that your receiving rig exhibits at least relatively stable BFO frequency response. A stock Icom R-8500 did not fit this requirement, hence it was necessary to fit an Icom CR-293 PLL crystal unit. Assuming the receiver exhibits poor frequency stability, you will see diagonal traces on Spectrum Lab (a most undesirable situation when attempting to measure weak AM carriers).

Download Spectrum Lab and set all parameters to default. Then make the following adjustments:

Select: Options, FFT settings.

FFT input size: 16384 Hz
Internal FFT Average: 30

Connect the REC out (audio line-out socket on receiver front panel) to the line-in audio socket on the back panel of the PC tower. Adjust the PC audio mixer controls until the optimum line-in level is obtained.

See SL screen showing 45.249985 MHz Wellington next to 45.250005 MHz Hedgehope, NZ via Es. SL indicates 1474.73 Hz as the peak signal (Wellington):



Inserting the RF Reference Carrier

Option 1: use a 2-way high isolation RF switch for switching between the reference and DX carriers for rapid comparison.

Option 2: inject the reference carrier so that it it mixes with the displayed DX carrier.

Calculations

With the Icom R-8500 tuned to 1 KHz below the nominal RF frequency (in this case, 45.249 MHz USB), the RF reference carrier is indicated as 1000 Hz on Spectrum Lab. The DX carrier as received on the Icom R-8500 indicates 975 Hz on Spectrum Lab.

1000 Hz - 975 Hz = 25 Hz low (or 0.000025 MHz).

45.25 minus 0.000025 = 45.249975 MHz (Wellington, NZ).

After each TV video carrier has been assigned to a location, the database is updated [3].

Example Practical Applications

When looking for multi-Es openings out to the Pacific, the Icom R-8500 is tuned to 55.249 MHz USB. A trace around 1600 Hz indicates Wharite, ZL2, while a 1550 Hz trace indicates KVZK-2 Pago Pago, KH8.

When conditions open up to DU, the most common trace is from Davao (55.249740 MHz). Hence with the R-8500 tuned to 55.249 MHz USB, Davao shows up as 1325 Hz on SL.

Conclusion

The settings currently used on SL are the same I use for detecting UHF TV carriers off the moon [2], hence the sensitivity is more than adequate for terrestrial DX work.

Todd Emslie

1. http://freenet-homepage.de/dl4yhf/spectra1.html

2. http://home.iprimus.com.au/toddemslie/s ... -tvdx.html

3. http://home.iprimus.com.au/toddemslie/Tevhftvlist.html

[VK3AUU]

Thanks for that Todd.

I have been using a simpler method which is probably good to 1 Hz.
I am just using an Icom 706 MkIIG with a large muffin fan under it keeping the oscillator compartment at a fairly steady temperature.
Using Speclab, I set the RX to 14.999 MHz on WWV and get the reading of say 1001.45 Hz averaged over a minute. (it might vary .05)
The correction at say 45.25 MHz is then 1.45 x 45.25/15 = -4.37 Hz
For Hedgehope, I then set the Rx at 45.249000 MHz and averaged over a minute I get a reading of 1004.9. This give a frequency of 45249000 + 1004.9 - 4.37 = 45250000.5
which agrees with what you get.

If you can't get WWV, you will need to find some other source and multiply the frequencie accordingly to calculate the correction needed.

At 15 MHz the 706 drifts about 7 Hz higher as the ambient temperature goes from 20 to 30 degrees C. It has the standard Oscillator, no TCXO. Without the muffin fan it drifts terribly.

David

[VK3HZ]

Hi Todd,
A couple of comments regarding this approach.

15625 Hz TV-Derived Frequency Reference Generator (outputs in 10 KHz steps over HF to UHF). (see "Electronics Australia" Oct 93). Accuracy is specified to 0.001 ppm.
The 15625 Hz TV-derived reference unit is connected to a standard VCR via the video 1v p-p line-out. The VCR is tuned to TCN9 Sydney in order to receive 15625 Hz sync pulses from a Rubidium transmitter source.

How current is this information (i.e. Rubidium source)? My understanding is that, since the advent of Frame Stores, the Rubidium's are no longer necessary. Therefore, as the TV station equipment has been updated (e.g. for 16:9), the Rubidium may have been "updated" to a simple crystal oscillator.

My other comment is that, with a 10 kHz "picket fence" generator, you could have up to 5 kHz difference between your reference and the TV carrier. PC soundcards can be off frequency by significant amounts - I've measured one as 6.6 Hz off at 1kHz. The error also varies with sample rate (BTW, usually very close at 48 kHz sampling). Therefore, trying to measure to better than 1 Hz using an uncalibrated soundcard may not be possible.

As a frequency reference, you may be better sourcing a 10 MHz GPS Reference such as a Trimble Thunderbolt - readily (and fairly cheaply) available on eB*y at the moment. Build up a divider to get 10kHz and 1kHz outputs. Graft the 10kHz into the back end of your TV reference to get an accurate RF "picket fence" and use the 1kHz to set the calibration for your soundcard in Spectrum Lab.

Regards,
Dave.

[VK2SWL1]

Hi Dave,

With reference to ABHN5A Newcastle, I get 138.275961 MHz. This is at variance with VK7MO's 138.276025 MHz reading obtained in past years. Of course, ABHN5A could have drifted since that time. If not, I've got some work to do!

The GPS source you suggest on eBay [1] sounds a viable alternative to 15625 Hz TV-derived reference units. The limitation with TV-derived units is that only one of my local TV stations (TCN9) provides accurate timing in the video sync pulses. The other TV stations result in increased errors to varying degrees.

The accuracy of the reference generator can be tested on either 57.25 ABNS1 Port Pirie, 57.25 ABGS1 Mt Gambier, or 57.26 ABEV1 Bendigo. Using my current system, I obtain measurements of 57.250000 (ABNS1 and ABGS1) and 57.260000 (ABEV1).

I take your point re sound card freq response accuracy. For many years I used a Fluke digital audio multimeter to measure AM carriers via the Icom R-8500 RX. A useful exercise would be to compare the reading obtained on Spectrum Lab vs Fluke multimeter. Alas, the 9V battery is currently flat in the Fluke meter, hence comparisons will be done later. Assuming the sound card is giving incorrect readings, I suppose a correction factor could be applied. I can't yet speak for the accuracy at ch5A (I look forward to your ABHN5A reading). Nevertheless, there is a chE2 TV station in 9M2 West Malysia that is received in Aust/NZ via multi-Es during the afternoons. My measurement (48.239598 MHz) matches everyone else to 1 Hz, hence the PC sound card and/or TCN9 sync pulse source can't be too far off.

Todd

1. http://cgi.ebay.com.au/TRIMBLE-GPS-RECE ... 8345QQcmdZ

[VK3HZ]

Hi Todd,

Sorry, for the sake of anyone else reading this, I deleted the Ch5A question after I found your email of a few years ago where you got the same result ar Rex VK7MO when he was using GPS-locking.

However, for the sake of the exercise, I'll measure it again and see what I get. Back shortly!

Regards,
Dave.

[VK2SWL1]

Hi David,

Yes, 15 MHz WWV is a potential reference source providing a good HF antenna is available. My R-8500 features separate HF and VHF RF inputs, so it is easy to compare the WWV tone up against the TV carrier. This was the method I used while listening at Darwin during solar cycle 23 peak. The challenge with WWV is the doppler shift experienced with HF F2 propagation. This explains why you average the measurement. You can also model your receiver's error response by measuring the slope (gradient). Your receiver's BFO oscillator response can be modelled in Y = mx + b form.

It is easy to very the validity of your measuring system by tuning to either 57.26 ABEV1 Bendigo or 57.25 ABGS1 Mt Gambier. If the measurements are not right on 57.250000 or 57.260000, this indicates reference source error.

Regards,

Todd

[VK3HZ]

I measure it as 138,276,011.0 Hz
I think it's one of the stations suffering Rubidium withdrawal

I might see if Rex VK7MO can also measure it to add to the confusion

Yes, that eB*y link is the sort of unit I was talking about. There seem to be a lot coming out of China at the moment too - perhaps they've just dismantled a CDMA network of something.

There will be a much bigger limitation soon regarding the TV-based reference - when do the Sydney analogue stations shut down?!

Regards,
Dave.

[VK3HZ]

Rex VK7MO measured it as 138.276 010 719 MHz with a confidence of about 0.05 Hz.

So, I wonder if TCN9 is still Rubidium locked??

Regards,
Dave.

[VK2SWL1]

VK3HZ wrote:

<Rex VK7MO measured it as 138.276 010 719 MHz with a confidence of about 0.05 Hz.>

<So, I wonder if TCN9 is still Rubidium locked?>

For some unknown reason, my initial ABHN5A measurement was in error. The VCR is pre-set to ch2, 7, 9, and 10. Using a good quality metro VHF/UHF yagi aimed at Artarmon/Gore hill, RF is fed into the VCR. The 1v p-p RCA output is connected to the 15625 Hz TV-derived ref unit, which then outputs RF in 10 KHz intervals up to the low microwave region. When different channels are selected on the VCR (2, 7, 9, 10), the resultant RF output from the 15625 Hz ref unit varies:

ABHN5A Newcastle measurement on SL: 1505.53 Hz (+/- 0.02 Hz)

ABN2 source for TV-derived reference: 1494.4 - 1494.7 Hz
ATN7 source for TV-derived reference: ~ 1856 Hz
TCN9 source for TV-derived reference: 1494.4 - 1494.7 Hz
TEN10 source for TV-derived reference: ~ 1398 Hz

Conclusion

ABN2 and TCN9 are both equally effective as accurate reference sources for the 15625 Hz TV-derived unit.

TEN10 (and especially) ATN7 are totally useless as reference sources for the 15625 Hz TV-derived unit.

ABN2 and ATN7 references (average) = 1494.65 Hz on SL.
ABHN5A (average) = 1505.53 Hz on SL.

This corresponds to 138.276 010 880 MHz.

This exercise also adequately demonstrates that certain PC sound cards do not introduce appreciable audio frequency error. The variation from the TV-derived reference source was small (+/- 0.15 Hz), however, the variation from ABHN5A was even smaller (+/- 0.02 Hz).

Compared to VK7MO's measurement, my ABHN5A reading (when using either ABN2 or TCN9 as 15625 Hz reference sources) is ~ 0.000000161 MHz higher. VK7MO's measuring setup is probably one of the best in VK/ZL. With this context in mind, my + 0.161 Hz error is a satisfactory result.

Regards,

Todd

[VK7MO]

One of the problems in checking frequency accuracy off NewCastle 5a to less than one Hz is that it does drift around. The drift seems to relate to termperature so I doubt it is now locked to Rubidium as it seemed to be 3 years ago. Over the last week I have measured it varying between 10.7 Hz and 10.960 Hiz high on 138.276 MHz. The higher frequencies relate to evenings and cooler termperatures at Newcastle and the lower frequency to daytime and hotter termperatures. This morning ie 20th of January I measure it on 138.276 010950 or 10.950 above a nominal 138.276 MHz.

I think therefore the differences measured between Todd and I could well be just the fact that our measurements were at different times. If Todd would like to make a measurement at a specific time I will do also for comparision.

The variation of around 300 mHz at 138 MHz is only about 2 parts in 10^9 which is much better than I would expect for a good quality cystal oven yet probably less than one would expect for Rubidium or GPS locked system so it would be interesting to know what it actually is.

Rex VK7MO

[VK2GOM]

The correction at say 45.25 MHz is then 1.45 x 45.25/15 = -4.37 Hz
For Hedgehope, I then set the Rx at 45.249000 MHz and averaged over a minute I get a reading of 1004.9. This give a frequency of 45249000 + 1004.9 - 4.37 = 45250000.5
which agrees with what you get.
David

David,

Just wondering what the divide by 15 relates to in your formula? I'll probably say "D'oh!" to myself when you tell me

73 - Rob VK2GOM / G0MOH

PS. Hang on - I might have figured it out? 15 as in 15MHz - the actual WWV freq?

[VK3AUU]

That's the frequency of WWV, the standard frequency station that I use as a frequency reference.

David
VK3AUU

[VK2GOM]

Thanks David,

My GPS disciplined rig shows 14.999MHz as reading between 999.8Hz and 1000.2Hz on WWV in Spectrum Lab.

That should put me in a reasonable position to measure TV carrier frequencies as is.

73 - Rob VK2GOM / G0MOH