Gordon Nelson first introduced precision frequency measurement of AM carriers in the late 1960s to facilitate identification of weak medium wave DX signals.

Ian Roberts, ZS6BTE, did all the research for precision frequency measurement of DX TV video frequencies, and to the best of my knowledge, there are still only a small number of DXers who can measure TV carriers to a resolution of 1 Hz.


Over the years I used different radios for early detection of long distance TV video and audio. My first receiver was a Realistic Patrolman 50 multi-band radio. For the first time I could at least hear the vision buzz of DX-TV carriers. Because the Patrolman-50 only had analog tuning, I couldn't accurately measure the frequency. However, I could at least tell what television channel was being received, for example, chE2, chR1, or ch0 etc.

It eventually became obvious that a radio with digital read out was needed, since I couldn't even tell if the DX video carrier was on -10, 0, or +10 KHz offset.

Anthony Mann (Perth, Australia) purchased an ICOM R7000 receiver in 1989. I was impressed how he was able to measure TV carrier frequencies to 100 Hz resolution. For example, 48.2396 MHz or 46.1718 MHz, etc.

I purchased an Icom R7000 back in 1993, and have never looked back. My VHF-DX reception logs were greatly improved by virtue of being able to detect and measure DX TV carriers at extremely weak levels.

It eventually became obvious that more precision was needed. For example, I would often receive more than one video carrier on 48.25 MHz. Since the offsets were typically less than 100 Hz apart, I could not measure the video offsets, and thus could not verify the origin of these signals.

In 1998, Ian Roberts (South Africa) developed various techniques for accurately measuring television (or any other relatively stable AM carriers) to a resolution of only 1 Hz.

I was also impressed on how Ian was able to separate and identify three chE2 TV transmitters (Germany, Thailand, and West Malaysia), that were all within ~ 30 Hz of each other! A DXer using a receiver with 100 Hz resolution, would find it impossible to measure and identify TV video carriers that are typically only 10-50 Hz apart.

Ian's method was to use the WWV reference signal on 15 MHz and calibrate the offset to TV carriers. Since then he has refined the technique, and now uses SpectraLab software for visual display of TV video carriers on a PC. Click here for more details of Ian's measuring techniques.

The benefits of this approach are now obvious; it is possible to accurately assign TV video carrier offsets to specific transmitter locations. This is of great benefit to six metre hams that want to know what geographic windows exist for 50 MHz propagation, and TV DXers who are trying for specific transmitters.

The end result of all this research has been a database of band 1 TV video carrier frequencies. For more details go to: 45-55 MHz TV database.

High stability frequency references

Accurate frequency reference signals are one of the main requirements for measuring carriers. Access to one or more of the following examples is essential:
15 MHz WWV short wave frequency standard.
Local rubidium based TV video carrier.
Local rubidium based 6m ham beacon (for example, the UK Buxton beacon).
Crystal reference generator.
Harmonics generated by a TV-derived reference unit.
GPS reference standard.

Audio filter

To facilitate measurements of extremely weak video carriers, I am currently using a home made 16 Hz audio bandpass filter, which can be tuned from ~ 400-4000 Hz. Narrow band filtering and amplification enables extremely weak carriers to be measured on a multimeter. Also, because the video carrier has been narrowed down to 16 Hz, signal to noise ratio is improved. Narrow band audio filters are standard equipment for EME DX. By using the audio filter, it is possible to measure TV carriers via tropospheric scatter on a daily basis, out to 500 miles distance.

My Icom receivers use a 2.4 KHz bandwidth for SSB. I have programmed all TV video carriers between 45-108 MHz, using USB on -/0/+ offsets.

The results are excellent, for example, it is now relatively easy to identify and separate 48.239580 MHz Genting Sempah, Malaysia, from 48.239590 MHz Nakhon Ratchasima, Thailand.

When a TVDX video carrier is strong enough to produce a picture, there is usually sideband frequencies present as well. These sidebands have a raspy sound, that distinguishes them from a genuine TV video carrier. The mathematics is simple:

48.2396 MHz (genuine) + .015625 MHz = 48.25522 MHz (sideband).
48.2396 MHz (genuine) - .015625 MHz = 48.223975 MHz (sideband).
48.2604 MHz (genuine) + .015625 MHz = 48.276025 MHz (sideband).
48.2604 MHz (genuine) - .015625 MHz = 48.24477 MHz (sideband).

Spectrum analyzer programs

Visual display of AM DX-TV carriers on a audio spectrum analyzer, is the optimum method of detecting and measuring extremely weak signals.

Ian Roberts uses SpectraLab. This spectrum analyzer is not freeware. Other audio frequency spectrum analyzer programs include SpectrumLab, which is freeware.

My PC, especially the monitor, produces high level RF interference, hence I often prefer to make a good quality 2 minute audio tape recording of a DX opening. Just connect the line-out of your SSB scanner, and connect to the input of a tape deck. Once the DX opening is over, I connect the output of the tape deck into the line input on the back of the PC.

The bandwidth of a scanner's SSB filter is typically around 2.5 KHz. This will be the audio frequency limit for frequency response on the spectrum analyzer.

Tony Mann sent me a video, describing his measuring method. A high stability crystal reference source, with outputs at 2.5, 5, and 10 KHz, is mixed with the DX signal. Using a quality coupler (splitter in reverse), the DX and reference signals are mixed. Variable attenuation is needed to reduce the output level of the reference signal. A 0-20dB Tandy overload attenuator could be used. Tony's Icom R-7000 receiver is tuned to 48.249 or 49.749 MHz, etc, (1KHz below the reference carrier). We now see a main spike (reference carrier), which of course is stronger than the other adjacent minor spikes (DX carriers). By positioning the cursor on the reference spike, and then on a adjacent DX carrier, the frequency difference between the two signals is automatically displayed on the screen. For example, one carrier is 106 Hz higher than the 48.25 reference. This is 48.250106 MHz Dubai.

ICDX correspondence

Below are some comments regarding my method of frequency measurements. These comments originally appeared in the ICDX Australian/Pacific TV-FM DX e-mail discussion group archives.

From: Todd Emslie
Subject: New method of frequency measurement


I have recently started to measure TV video carriers to a resolution of 1 Hz. Below is a brief explanation of my method.

Terrestrial TV channels have their line frequency locked to a standard. The TV line frequency is 15625 Hz, ie 1 MHz divided by 64. If there is a local TV station that has it's 15625 Hz line frequency locked to a rubidium reference, you have access to a extremely accurate reference standard.

Equipment required: Digital multimeter which can measure audio frequencies to 3.75 digits, and ideally at least 0.1% accuracy. I am currently using a Fluke 19 digital multimeter (100 Hz-20 KHz frequency range).

TV-derived 15625 Hz line frequency reference unit (home constructed),

A receiver which can receive upper side band (USB), and has digital readout, eg Icom R7000, R7100, 8500, AOR-R5000 etc.

A TV or VCR that has a 1v p-p base-band video output.

A calculator for the primary school mathematics.


My Toshiba C-531 TV is tuned to my local TCN ch 9 (196.25 MHz video) TV channel. A VCR could also be used.

TCN9's line frequency (15625 Hz), is locked to a rubidium frequency standard, hence the line frequency is stable to 0.01 to 0.001 ppm. My Toshiba's TV video line out (or VCR), 1v p-p composite video signal is connected to a low cost TV- derived frequency reference unit. The TV-derived frequency reference unit generates accurate harmonics at 10 KHz intervals, hence I have an extremely accurate and stable frequency reference.

Actual worked example:
I recently received a TVDX video carrier on approximately 46.24 MHz. I tuned the ICOM R7000 receiver to 46.239 MHz USB. I usually tune approximately 1-2KHz below the DX video carrier. The SSB audio output on the ICOM R7000 seems to peak around 1200 Hz. I connect a lead from the Icom's headphone or line out jack to a digital multimeter. I then measured the DX signal's audio frequency and obtained a reading of 1600 Hz. Immediately, I switched to the TV-derived reference signal, and obtained a reading of 1564 Hz.

I use a Tandy high isolation switch for immediate switching between the reference signal and the DX signal.

Because the audio frequency (1600 Hz) of the DX signal was higher than the audio frequency (1564 Hz) reference signal, I already know that the DX signal is 36 Hz higher than 46.24 MHz.

Since the DX signal is 0.000036 MHz (36 Hz) higher than the 46.240000 MHz reference signal, we add 0.000036 MHz to 46.240000 MHz, which equals 46.240036 MHz.

Thus the TVDX video signal has been measured as 46.240036 MHz. With practice, it takes less than a minute to obtain a measurement.

My next step was to build a narrow band (~16Hz) audio filter, to facilitate measurements of extremely weak video carriers. The audio filter is also useful for measuring video, when multiple carriers are present. The filter circuit is based around a LM348 OP IC amp. Weak video carrier signals were initially unreadable on my meter, hence the need for additional amplification and filtering.


Todd Emslie.
Sydney, Australia

From: Tony Mann

Subject: Re: And now there is 4.

The method Todd is using is correct, but his description of it may be confusing.

If we have any scanner with a SSB facility we can use it to measure the frequency of a TV carrier, provided there is a (rubidium-type stability) reference carrier nearby. Mathematically it's a lot easier to see.

fb = the audio beat frequency

fl = the ( Icom R7000) scanner internal oscillator frequency
fc = the frequency of the am carrier we wish to measure
Then for usb mode we have:
fb = fl - fc

Taking the example of 46.24 MHz:
for the reference at 46.240000, we have
fl - 46240000 = 1600 equation (1)

while for the tv carrier, call it fx, we have
fl - fx = 1564 equation (2)
Subtracting equation (2) from equation (1) we find:
fx - 46240000 = 1600 - 1564 = 36 Hz
fx = 46240036 MHz.
Note that fl has cancelled out. i.e. This method is independent of the scanner internal local oscillator frequency (fl), provided you make the two beat frequency measurements fast enough that fl does not drift appreciably between them.
Instead of the frequency meter, as Ian has already outlined, one can use the sound card of a computer (& the right software), to give you a spectrum of the DX carriers, typically 3 kHz wide on the R7000. You still have to know/measure the frequency of one of them to get the frequencies of the others. This is where the "TV-derived frequency standard" is useful.

The unit Todd is using is based on a circuit that was an Electronics Australia project kit in 1993. Its performance was verified against atomic standards at our National Measurement Laboratory in Sydney (see "Electronics Australia" Oct 93), and from memory it achieved stability's of 0.01 to 0.001 ppm, which is adequate for DXing!

The rubidium reference is in the timing of the TV station sync pulses, not the RF carrier. It seems that only the ABC and Channel 9 TV networks in Australia use a rubidium reference for the sync. The others use quartz which can be up to 0.5 ppm off.

Basically the circuit extracts the sync pulses (& corrects for the blanking interval) and divides by 2 to give a 7.8125 kHz reference. This is compared to the 7.8125 kHz obtained from a 10 MHz VCO divided by 128. The error signal from the phase comparator locks the VCO.

I modified the Electronics Australia circuit to deliver harmonics every 250 kHz and also every 10 kHz (by adding another divider circuit).

Todd is using the first Electronics Australia prototype I made, and I have compared its frequency against my high quality quartz setup, and am satisfied the unit is performing correctly. If I ever get any significant spare time I may make up a few more, but I am so busy at the moment.


Tony Mann,

Perth, Western Australia.
(32 S, 116 E)

Message: 4
Date: Tue, 31 Oct 2000 11:16:45 +0200
From: Ian Roberts
Subject: Re: And now there is 4.

Tony, thanks for explaining Todd's method.

There are some points interleaved in this thread below, which may not be obvious to others. At the end of the day we don't want a bunch of TV stations identified when they do not in fact exist, but are measurement errors.

Tony Mann wrote:

> Mathematically it's a lot easier to see.
> Let
> fb = the audio beat frequency
> fl = the (R7000) scanner internal oscillator frequency

A better description would be "the SSB receiver's USB carrier frequency". The internal (LO) oscillator frequency in a receiver is typically between 10 and 30.2 MHz and is not the frequency to be used here.

> fc = the frequency of the am carrier we wish to measure
> Then for usb mode we have:

> fb = fl - fc

This should read:
fb = f1 +/-fc

the DX TV carrier could be below or above the audio beat frequency produced by the frequency reference, I think.
If the reference produces a beat of 1600 Hz and the DX TV produces a beat of 1700 Hz, the DX TV frequency is below the carrier produced from the reference and 100 Hz must be deducted.
Example: the rig is tuned to 48.238 USB, the reference produces a beat of 1600 Hz: a frequency of 48.239.600. The DX TV produces a beat of 1700 Hz. The DX TV's frequency is 48.239.600 - 100 = 48.239.500.
Example: The DX TV produces a beat of 1500 Hz. The DX TV's frequency is 48.239.600 + 100 = 48.239.700.
All strong TV signals produce 50 Hz sideboards spread symmetrically each side of the carrier, at reducing amplitude, and strong line sync bunches at 15625 Hz high and low frequency of the carrier (for 625-line TV). The numbers or 60 Hz and 15750 for 525-line TV (NTSC).

Ian Roberts.

From: Ian Roberts

Hi Todd,

There are problems when using the AOR range of scanners. Udo Deutcher, in recent correspondence regarding measuring TV frequencies with an AOR, noticed anomalies and serious errors. Seems that the synthesizer is not what one hopes it should be... I don't recall which AOR he was using. Comparing it to his ICOM 706 ham transceiver, he obtained differences in the apparent frequencies. Also, the AOR gave different answers depending on whether USB or LSB was used, a certain indicator that the method used was incorrect, or that the synthesizer is no good. He also obtained anomalies when using my method you copied in this thread when using the AOR, but not with the ICOM. He is lucky that the German Biedenkopf TV transmitter is within range and was on frequency at the time (not anymore, it is now 4 Hz high). The method is fine, the rig's frequency control is not up to snuff.

This being the case, then your advice regarding bfo-ing to the same audio beat at a specified offset relative to WWV may be bad advice.

Fortunately, this business of synthesizer steps/frequency control tends to fall away when a round synthesizer step such as 1 kHz is used. Then one simply reads off the audio beat frequency and either adds it to the USB RF frequency, or deducts it from the LSB RF frequency.

If a sound card is used as a audio frequency counter in asymmetric mode (for example only USB is used, or LSB) then it is vital to understand how to correct the apparent error. Very few persons got this right the first time...

Also, if one examines the frequency synthesis of rigs such as the ICOM 756 transceiver, one may be lead to believe that there is a standing error in the frequency synthesis plan; this rig has 5 oscillators with DDS (Direct digital synthesis). Don Graham researched this and produced a correction equation for the IC756. This worried me to the extent that I re-investigated my own methods and results on the two rigs I have, the IC-R8500 and IC746. Both produced results within an absolute error of .5 Hz at 50 MHz, which is about as good as one can get when using WWV/WWVH. Don, I believe, had not considered the digital phase control where the "lock" can be slipped (offset) in software to compensate for the fixed step errors. It does not matter whether the rig's DSP is on or off. I have no doubt that the IC756 also produces exact synthesizer steps, due to the necessary corrections being built into the software.

Todd Emslie wrote:

> To enable non-technical VHF DXers to contribute video offset measurements, > simplified methods are needed for relatively accurate measurements to 4 > decimals. Do you have any further suggestions?

Yes, turn off RIT, BFO, clarifier or other RX offsets, get the receiver onto frequency, then add the audio beat to the USB frequency.

To sum up:
All rigs are not equal.
Determine the frequency of a known AM carrier to confirm the rig's accuracy. The rig does not have to have 100/10/1 Hz steps, 1 kHz is just as easy and may be more accurate on the cheaper receivers. If one uses an asymmetric mode such as a sound card in USB only, the system is not self calibrating and the user must make sure the the results are interpreted correctly.

Ian Roberts, ZS6BTE
South Africa

From: Todd Emslie

Hi Ian,

It seems that all VHF DXers who are able to measure TV carriers to a few Hz accuracy, have technical backgrounds in electronics.

These same DXers (Tony Mann, Ian Roberts, Adam Maurer, Don Graham, and myself) are using different methods to achieve similar results. For example, they have access to Digipan, Cool edit Pro software, digital multimeters, 15625 Hz line TV-derived reference units, and tunable crystal reference units, etc.

It is apparent that a large percentage of VHF DXers do not have a technical background or access to the above equipment. Some do not use a VHF receiver that covers the 15 MHz range and the VHF range.

To enable non-technical VHF DXers to contribute video offset measurements, simplified methods are needed for relatively accurate measurements to 4 decimals. Do you have any further suggestions?

On my R7000, if the reference frequency is too far from the DX frequency, e.g. using ABC ch2 TV on 64.2500, you get an offset. From memory this is approximately 50 Hz at 48.25. The offset gets smaller as the reference and DX carrier frequencies get closer. This is why I prefer to use my 15625 Hz TV-derived reference unit. Hopefully due to the superior design of the R8500, the offset between 15 MHz WWV and the DX video carrier is not significant.


Todd Emslie

From: Ian Roberts.

Hello Todd,

Some comments on your interesting suggestions, as requested:

Here is what I found below regarding your specific questions:

Method 1 - Simple audio BFO pitch comparison of WWV to DX vision carrier.

1. Tune the IC-R8500 to 14.999 MHz USB (1 KHz below WWV), place this in a memory channel.
2. Using 10 Hz steps on the R8500, tune the DX video carrier, until you obtain a similar BFO audio beat to 14.999 MHz WWV. Then add 1 KHz to this measurement.
3. Take note of the difference between the apparent and actual frequency measurement.

Using Kenya last night I obtained an error of 36 Hz. In 10 Hz steps this could be around 46 Hz, which is probably not acceptable.

Method 2 - Simple audio BFO pitch comparison to a known band1 DX reference carrier.

1. Tune the IC-R8500 to 1KHz below a known reference carrier, for example,
48.2474 MHz Biendenkopf.
2. Place 48.2464 MHz in a memory channel.
3. Using 10 Hz steps on the R8500, tune the DX video carrier, until you obtain a similar BFO audio beat to 48.2464 MHz. Then add 1KHz to this measurement.
4. Take note of the difference between the apparent and actual frequency measurement.

This method is OK provided both frequencies are close together. The maximum error amounts to the synthesiser steps used. For 10 Hz steps this is a bit more than 10 Hz, depending on how far the frequencies are apart.

Method 3 - Simple offset correction to a known band1 DX reference carrier. This method is only suitable for 4 decimal measurements.

1. Tune the IC-R8500 until the DX video carrier's beat note is no longer audible, ("zero-beat")
2. Use a reference carrier, for example, 48.2500 MHz, etc, to correct for any rx offset.
3. Subtract or add the difference.

This method is suitable for 3 decimal measurements. The error I obtained averaged 100 Hz, but this is with a rig on frequency. Uncalibrated rigs might double the error to around 200 Hz due to the user (by definition) being unable to hear the zero beat frequency.

The attached scan shows the pass band of the IC-R8500, with convenient markers from local 50 Hz multiples at 50,100,150 and 250 Hz. The user will not hear anything below about 150 Hz unless the volume is max. If another carrier is also at the offset then the zero beat will be inaudible. So one can obtain 48.2493 for example, plus or minus the typical 200 Hz on top.

Ian Roberts

Measuring TV carriers on a Icom R-8500 by Todd Emslie

I recently bought a new Icom R-8500 and the CR-293 high stability PLL crystal unit.

It occurred to me that I could use my TV-derived reference to measure the incremental frequency errors on my Icom R-8500. By measuring the USB audio frequency of 14.999 MHz WWV, I could compare the USB audio frequency of other band 1 TV carriers between 45-108 MHz.

Because of fading and phase differences, I initially took the average of the upper and lower audio frequencies of WWV on 14.9990 MHz. I arrived at a figure of 1013 Hz. To verify that this was a correct reading, I then connected the output of the TV-derived reference unit, and obtained a rock-solid measurement of 1013 Hz. I then measured the frequency difference between 14.9990 MHz USB, and various band 1 carriers using USB, 1 KHz below the main video carrier:

Offset error between WWV 14.9990 MHz and 45.2490 MHz = 26.8 Hz
Offset between WWV 14.9990 MHz and 46.2490 MHz = 27.7 Hz
Offset between WWV 14.9990 MHz and 48.2490 MHz = 29.5 Hz
Offset between WWV 14.9990 MHz and 49.7490 MHz = 30.8 Hz
Offset between WWV 14.9990 MHz and 55.2490 MHz = 35.7 Hz
Offset between WWV 14.9990 MHz and 57.2490 MHz = 37.4 Hz
Offset between WWV 14.9990 MHz and 59.2490 MHz = 39.2 Hz
Offset between WWV 14.9990 MHz and 61.2490 MHz = 41.0 Hz
Offset between WWV 14.9990 MHz and 62.2490 MHz = 41.9 Hz
Offset between WWV 14.9990 MHz and 64.2490 MHz = 43.6 Hz
Offset between WWV 14.9990 MHz and 77.2490 MHz = 55.1 Hz

The audio offset error increased by an average of .88595 Hz betwwen 14.9990 and 45.2490 MHz.

These figures are useful as a reference when measuring the video carriers of DX TV signals.

For example, chE2 Nakhon Ratchasima, Thailand is received via TEP.

1. Select a memory pre-set for WWV on 14.9990 MHz USB.
2. Then tune chE2 tx using USB and 10 Hz steps, until you obtain a similar audio pitch to WWV.
3. Take note of the frequency displayed on the R-8500, and add 1 KHz.
4. Then look at your offset reference list, and subtract 29 Hz.
5. You should arrive at a figure between 48.23958 and 48.23960 MHz.

My Icom R-8500 has been fitted with the high stability CR-293 PLL crystal unit. The measuring method described above is not suitable for receivers not fitted with a high stability crystal.

Depending on if your receiver is high or low, compared to 14.999 MHz USB WWV, a table of frequency error corrections will need to be printed out. These corrections will be very likely different to my own table.

The advantage with this method of frequency measurement is that once you have measured the synthesizer steps offsets relative to WWV on 14.999 KHz, a reference unit, although still desirable, is no longer essential. In any case, most DXers will not have access to a quality reference unit, hence 15 MHz WWV is their only option.

Freeware spectrum analyzer programs

DL4YHF's Amateur Radio Software: Audio Spectrum Analyzer (Spectrum Lab)

SR5 Spectrum Analyzer.


Spectra software for windows and soundcard by Nino Porcino, IZ8BLY . Accurate spectrum analyzer in the range 0 - 2500 Hz.

List of available spectrum analyser programs:


(1) Jim Rowe, Low cost TV-derived Frequency reference. Electronics Australia, part 1-2, Oct-Nov 1993.

(2)David McQue, G4NJU, The poor man's caesium clock. Radcom magazine, Jan 1999.


Scanners and receivers used for monitoring of DX TV signals.

The poor man's caesium clock, 15625 Hz TV-derived reference. By G4NJU.

DL4YHF's SpectrumLab FFT software. All you need is a receiver with USB, eg) IC-R7000, IC-R8500 etc, for visual display of video carriers on a PC.

Precision measurement of TV DX video carrier frequencies By Ian Roberts, ZS6BTE.

Precision frequency measurement of DXTV video carriers by Kunihiko NAKANO, JE7IDA (Japan).

Methods of measuring DX-TV video carriers. Interesting summary on how UK DXers use the Icom PCR-1000 PC receiver.

Copyright © 2009 Todd Emslie