In 2002, Ian Roberts, ZS6BTE, South Africa claimed reception of the Kenya 48.25 MHz chE2 tx via EME. For more information, go to: Using EME to Locate VHF-SHF Transmitters.

Extreme DXing - Determining the Longitude of a DX TX from the EME echo.

A high power DX transmitter's position can be established from its EME reflection. As the zero elevation point of the setting moon moves across the position of the DX TX so that the EME Window closes, the "moonbounce" signature disappears; this provides the basic longitude of the TX. Under certain positions of the DX TX, the moon, and the RX station, the latitude can also be established. If the moon is approaching, the onset of signal determines the position of the TX.

On Wednesday 1/May 2002, I did such an exercise on the ch E2 TX from western Kenya, which is about 3000 km north of my location, using the morning setting moon.

The attached shot Instanttrack graph.jpg shows the zero elevation point to the moon LOS (loss of signal), the white line, moving from east to west across Kenya with the elevation angle to the moon at 3.2 degrees. Note that the line is not north/south, as the moon is at -24 deg latitude over southern Brazil. The point is, would the estimated ERP of the Kenyan TX, ~160 kW, produce a response in my receiving system? According to VK3UM's EME Planner (screen shot in EME budget.jpg) the signal-to-noise ratio would be around -19 dB ("station B as received by station A"). Additional factors come into play: ground gain at low elevation angles in the TX and RX directions would add several dB to the signal-to-noise ratio, but this would have to occur simultaneously, and software with considerable DSP enhancement using a 64k FFT would "look" into the noise by around 20 dB. Also, the "pol" (polarization) error (see EME timing.jpg) of 21 degrees would degrade the recovered signal to some extent.

I monitored and recorded the entire EME Window from ~8h55 local time down to the zero elevation time in Kenya and ran the 24min 15 sec WAV file into a display store which integrated the display, together with a "peak hold" function to catch amplitude peaks. This is attached as Kenya_raw_Wav.jpg showing the result between 500 and 2500 Hz (500 Hz vertical divisions). The peak at ~500 Hz is local equipment noise. Two thirds of the way across the scan is another peak at 1834 Hz shown in the zoomed image Kenyaemezoomed.jpg, complete with a 50 Hz sideband. This largest response was monitored visually - the 50 Hz sideband occurred a short time later, both happened when it was estimated best ground gain/elevation conditions were occurring for this particular path. The WAV file was subsequently further processed in another package after zooming in to a three minute sequence bearing the relevant echoes and placed through another 64k FFT/filtering algorithm - this is attached as Post_processed.jpg and makes the visual effect more striking: - to my amazement additional 50 Hz sidebands appeared. This is because the "peak hold" mentioned previously had also held noise at peak value, which was then removed. The signal-to-noise ratio at the carrier is about 4 db after all this processing (the horizontal divisions are 6 dB spacing). The signal-to-noise ratio before processing appears to be around -25 dB, somewhat worse than the -19 dB predicted in the EME budget.jpg illustration. Possibly I have overestimated the ERP of the Kenyan TX, a source places it (ERP?) at only 15 kW. The carriers on the left are local mains 50 Hz, plus harmonics thereof.

Arithmetic - measured in USB mode at 48.248000 MHz: Uncorrected response peak frequency: 48248000+1834 = 48.249834 RX error: 14.13 Hz low Doppler shift: -145 Hz at 50 MHz (HM-HM in EME timing.jpg). Since the measurement frequency is 48.25 MHz this needs to be apportioned a bit: 48.25/50*145 = 139.9 Hz Freq is thus: 48249834-14.13+139.9 = 48249959.7 Hz.

This is in astonishing agreement with the frequency 48.249960 MHz I measured a few days ago 20/4/2002 for this TX (see the debate about the long path reception of Kenya on ch E2!).

An attempt to perform the same measurement on the rising moon the previous evening was unsuccessful due to the band still being open and very noisy, but had it worked, a target zone of some accuracy could have been defined by overlapping the AOS (acquisition of signal) with the LOS graphs.

Summary: The approximate longitude of the TX lies just west of the white line shown in Instanttrack graph.jpg, and the frequency was accurately recorded to about 1 Hz.

A copy of the relevant WAV file sequence of ~3min (520k as zipped Wav, 1.7 MB when unzipped) is available if wanted.

Equipment used:

Antenna: 50 MHz yagi operating below optimum gain frequency at 15 m above ground level.
~20m Andrews 1/2" low-loss cable.
ICOM IC-R8500 receiver in USB mode.
Creative Labs PCI 128 on-board sound chip using 8kHz sampling/8 bit/mono, and 800 MHz Celeron CPU PC, suitable display/processing software.

Minimum receiving system requirements

This report seemed to generate some interest, but the principle is not very state-of-the art.

The basic requirements are as follows:

1. The antenna must be pointed in the direction of the common EME window (software for predicting this is VK3UM's EME Planner, free, ex I-net), with an accuracy of about 10 degrees.

2. Most DX antennas will struggle to achieve 10 dBi of gain, so a preamp will be necessary to overcome cheap coax cable. ANY preamp is good enough (gain about 15-20 dB max) because the antenna noise temperature varies between 1500 and 6000k under quiet terrestrial conditions - forget about doing this during a tropo, TEP, spE or F2 VHF opening where the antenna temperature could go to 12000k...

3. The EME Planner software calculates the vital POL (depolarisation error), which should be kept below +/- 45 degrees if loss is not to exceed 3 dB. Ideally look for a time when POL is below about 20 degrees, and the elevation is low (below 20 degrees) if the antenna does not have azimuth/elevation control.

4. The TX should be on!

5. In most cases the echo will be inaudible due to the fact that it is not switched on/off, but appears at a constant frequency.

6. Plenty of DSP is required to recover the signal.

7. A calibrated SSB receiver is required to gain full benefit from the exercise. Forget AM/FM. Allow for Doppler in setting the receiver's offset.

8. This works at any frequency, use the correct antenna and preamp, push the simulation through the EME Planner's nearest ham frequency, and make sure the signal-to-noise ratio remains above about -25 dB (add some loss for depolarisation as necessary).

9. To capture the audio, use an audio capturing device WITHOUT compression, then apply the DSP to this (WAV) file.

A superb piece for this sound card-based task is CoolEdit 2000 available from www.syntrillium.com which I think has a 30 day trial period. This implements DSP/FFTs/noise reduction as required for the task of recovering weak signals, probably below noise most of the time.

I am pretty sure I have detected UHF TV carriers from the USA via moonbounce here in Perth, Australia.

With my UHF TV Yagi pointed at the moon, weak carriers with frequency drift (~ -1.5 Hz/minute) were noted on 26, 27 & 28 May:

Date time (UTC) freq (MHz)* freq drift predicted Doppler (Hz/min) (Hz/min)

26th 1021-1028 483.250537 -2.5 -0.9
26th 1115-1122 501.248339 -1.5 -1.4
27th 1147-1206 501.248312 -1.4 -1.4
28th 1235-1258 501.248292 -1.26 -1.28

(* at fade-out.)

The timing is in excellent agreement with a moon scheduling program (such as GM4JJJ's) for moonset at the US TV tx sites of:

501.25 KWBT-19 Muskogee, OK 35.75N, 95.8W.
483.25 WNDU-16 South Bend, IN 41.6N, 86.2W.

Both of these txs radiate 5000 kW erp horozontal. polarisation., with an omni-directional pattern and have no vertical beam tilt. (Information on these stations may be found at the FCC TV database http://www.fcc.gov/mb/video/tvq.html#START

Most of the 5000 kW US UHF TV tx's are directional. So, for a DXer to the west, it's essential to select only txs that have maximum ERP in the direction(s) of the setting moon (230-310 degrees). Rather few high power txs that I surveyed had power in the optimum directions. There were some with a maximum at 300 degrees and not 240, or vice versa, so they are not available over the full cycle of the moon.

On all 3 days the frequency drift of the 501.25 MHz signal is in excellent agreement with the predicted rate of Doppler shift. The -2Hz/minute drift of the 483.25 MHz tx is about 1.6 Hz/min faster than prediction. Maybe the tx frequency itself is drifting, which is why I didn't find it after the first night.

Sky noise is getting worse: predicted 145K 28/5 vs 60K for 27/5 and 30K for 26/5, as the moon moves east in front of noise sources in the Milky Way.

The signal-to-noise ratio was several dB, as can be seen from spectrum analyser scans of 501.25 MHz KWBT-19 Muskogee, Oklahoma, and WNDU-16 South Bend, IN, at http://www.physics.uwa.edu.au/~agm/eme1.JPG These scans were done every 40 seconds and the bandwidth (spectrum analyser bin width) is about 2 Hz. One could watch progress of signal in near real time: the signal on 501.25 on 27 & 28 May fluctuated by at least 7 dB (in/out) several times and was strongest a few minutes before fade-out.

The antenna is a 22 director Yagi (in Australia, Jaycar's "91" element fringe antenna, model LT3182) with nominal 15dB gain at 500 MHz (15-19 dBi gain from 470 to 862MHz). It was mounted (on a camera tripod) at 1.5 meters (5ft) above ground level, with manual adjustment of the elevation and azimuth to track the moon. Horizontal polarisation is used. It was mostly elevated at about 22 degrees (optimum for 501.25 MHz, KWBT).

I can't yet determine the amount of Faraday rotation. I've been sitting on horozontal polarisation the whole time with a single Yagi. Faraday rotation is supposed to be a problem at UHF. The tx end always sees the maximum amount of it. Certainly some days I don't get signals.

The (1st) preamp is a 2 dB noise figure GaAsFET unit, with 20 dB gain, mounted at the antenna. It is a JIM model M-75, made in Japan (once sold in Australia by Dick Smith Electronics). Its RF bandwidth is 225-1500 MHz.

The receiver is an Icom R7000 (USB mode), plus PC (Mac) based audio spectrum analyser sampling the receiver's audio through the sound card.

To measure frequency accurately I use harmonics of a frequency divider chain which provides combs every 10 kHz from a very stable 5 MHz crystal oscillator. The 5 MHz reference is a high performance Vectron ovenised quartz crystal (circa 1989 vintage), with a frequency drift of better than 0.001 ppm/day.

A simple calculation of the signal-to-noise ratio [= PG/LkTB, where P = TX erp, G = RX ant gain, L = path loss (264 dB at 500 MHz), k = 1.38 x 10^-23, T = noise temperature., B = bandwidth (Hz)] gives about 15 dB for a 3 Hz bandwidth and 300 K noise temperature., which is in the right ballpark.

There are many more UHF transmitters to investigate via EME. Here in Australia the task is made easier by the lack of TV allocations below 526 MHz.

One main requirement is to pick a US TV channel that's in a quiet part of the spectrum where you have decent antenna gain. It could even be in Band 5, as antenna gain compensates for the increased path loss. Preference goes to the few 5000 kW omnidirectional transmitters. For example, some listed at http://www.w9wi.com/articles/uhflow.html

Last night I simulated an EME signal and looked at it on my spectrum analyser.

When the signal to noise ratio was 15 dB (bandwidth ~ 2-3 Hz) I could still hear it at 1 kHz (although very weak) through my 20 Hz bandwidth audio filter. However when the SNR dropped to 8 dB (the most I'm getting from EME) I could not hear it with the audio filter. I think therefore that you must implement a PC spectrum analsyser if you want to do EME DX.

Low noise UHF TV preamplifiers

Masthead mounting of low noise preamplifiers is essential at UHF. If a UHF TV yagi is used in the backyard, a low noise preamp fitted directly to the antenna dipole terminals is essential. Finding low noise UHF TV preamps can be a challenge.

Most commercial UHF TV preamps have typical noise figures between 2-3 dB. Unfortunately, these noise figures are too high for UHF EME. What are the alternatives? Narrow band 0.5 dB GaAsFET preamps can be built. However, the disadvantage is the relatively narrow bandwidth, and with re-peaking the tuned circuits, hence only a few UHF TV channels can be covered.

A more practical method is to use a 1.5 dB n/f wideband UHF preamp, thus no re-peaking is required.

So far, Tony has found that the Alcad BR-105 low noise 1.5 dB UHF balun preamp is equal to the M-75 GaAsFET wideband scanner preamp.

I find I have to use both M-75 GaAsFET preamps to minimise intermodulation, and hence maximise the number of usable UHF TV channels. I can't even use a lesser quality bipolar amp as the second preamp (to overcome the 8 metre cable run through the house). I have a 1 dB NF preamp (bought in the USA in 1985) but I can't use it because it overloads worse than anything else.

Why is a second line UHF amp sometimes necessary? Consider the following:

Masthead preamp gain = 20dB.
UHF receiver or tuner noise figure = 10dB.
Gain needed to establish new receive system noise figure = 10dB + 3dB = 13dB.
Gain available for cable loss without affecting new system noise figure = 7dB.

In the previous example it is clear that more than 7dB loss between the masthead preamp and the receiver input will degrade the system noise figure. The way around this problem is to use a UHF line amp further down the line near the receiver. The noise figure of the line amp should be well below the receiver noise figure. Consider the following:

Line amp noise figure = 3dB (typical).
Minimum input level to line amp = n/f + 3dB, or 3 + 3 = 6dB.

Tony found a source of M-75 GaAsFET preamps, but they are not cheap: http://www.durhamradio.com/jim-scanner-preamplifiers.htm

The list of USA UHF TV, including one from Australia, received in Perth, Australia via moonbounce so far:

1 WNDU-16 South Bend, IN 483.2505 ? 41.6N, 86.2W 5MW Z H
2 KWBT-19 Muskogee, OK 501.2482 35.8N, 95.8W 5MW Z H
3 WAPT-16 Jackson, MS 483.2510 32.3N, 90.3W 5MW ZdH
4 KUPB-18 Midland, TX 495.2512 31.8N,102.5W 5MW Z H
5 KXTX-39 Dallas, TX 621.2496 32.6N, 97.0W 5MW Z E
6 KPTM-42 Omaha, NE 639.2595 41.1N, 96.2W 5MW + H
7 KPPX-51 Tolleson, AZ 693.2494 33.3N, 112.0W 5MW Z
8 KUVS-19 Modesto, CA 501.2401 38.1N, 120.7W 5MW -d
9 KDTV-14 San Francisco, CA 471.2603 37.5N, 121.9W 5MW +d
10 WFTT-50 Tampa, FL 687.2497 27.8N, 82.3W 4MW Z H
11 KWEX-41 San Antonio, TX 633.2596 29.3N, 95.3W 5MW +d
12 CTC-35 Mt.Ulandra,NSW,Aust. 576.2496 34.8S,147.9E 1.6MW Z H
13 WLTX-19 Columbia, SC 501.2589 34.1N, 80.8W 5MW +dH
14 WXIX-19 Newport, KY 501.2598 39.1N, 84.6W 5MW +dH
15 WAND-17 Decatur, IL 489.2499 40.0N, 88.8W 5MW ZdH
16 KTVG-17 Grand Is., NE 489.2489 40.7N, 98.6W 5MW ZdH
17 KXVO-15 Omaha, NE 477.24995 41.1N, 96.2W 5MW ZdH
18 KXAN-36 Austin, TX 603.24956 30.3N, 97.8W 5MW Z E
19 KTWB-22 Seattle, WA 519.25982 47.6N,122.3W 5MW + H
20 WBBH-20 Ft. Myers, FL 507.25995 26.8N, 81.8W 5MW +dH

Last column is: power (MegaWatts); channel offset (Z=zero, +10 or -10 kHz); d = directional antenna (blank = omnidirectional); polarization, H = horizontal, E = elliptical.

I emphasize that the signal-to-noise ratio is very poor, typically 5-8 dB in a bandwidth of ~ 2Hz. The strongest signal one night was 12 dB from KWBT-19, but I can follow signals down to 2 dB above the noise.

There is a small chance txs 13 and 14 are the other way around - they were received nearly simultaneously.

Tx no 12, at Mt. Ulandra (QTH of ABMN0), was also measured at 1210 UTC on 21 June by Todd Emslie. He obtained 576.249653 MHz, which is very close to my measurement the previous day of 576.249660 MHz. Todd is only 200 miles from this tx and receives it via tropospheric scatter.

tx date time,UTC freq. QTF* drift [Doppler]
(Hz/min) **
1 ? 26-May 1021-1028 483.25049 242 -2.5 [-0.9]
2 26-May 1115-1122 501.24844 244 -1.5 [-1.3]
2 27-May 1147-1206 501.24842 240 -1.4 [-1.3]
2 28-May 1235-1258 501.24842 238 -1.3 [-1.3]
2 29-May 1340-1352 501.24842 238 -1.3 [-1.3]
2 11-Jun 0127-0142 501.24823 299 -1.3 [-1.3]
2 15-Jun 0512-0516+ 501.24818 294 -1.5 [-1.2]
2 16-Jun 0550-0556 501.24814 289 -1.8 [-1.4]
2 25-Jun 1138-1145 501.24820 238 -0.9 [-1.3]
2 26-Jun 1235-1241 501.24821 238 -1.1 [-1.3]
2 27-Jun 1322-1340 501.24821 242 -1.3 [-1.4]
2 28-Jun 1328-1344 501.24823 245 -1.3 [-1.4]
3 30-May 1437-1444 483.25103 242 -1.1 [-1.2]
3 31-May 1533-1540 483.25105 245 -1.3 [-1.2]
4 11-Jun 0201-0212 495.25122 297 -1.5 [-1.3]
4 12-Jun 0247-0305 495.25117 299 -1.3 [-1.3]
5 12-Jun 0236-0240 621.24960 299 -1.5 [-1.5]
5 25-Jun 1153-1158 621.24968 239 -1.6 [-1.7]
5 26-Jun 1344-1348 621.24969 243 -1.4 [-1.7]
6 15-Jun 0517-0524 639.25953 296 -1.7 [-1.7]
6 16-Jun 0602-0608 639.25952 290 -1.6 [-1.8]
7 15-Jun 0604-0614 693.24943 293 -2.2 [-2.2]
7 16-Jun 0642-0649 693.24942 288 -2.8 [-2.6]
8 15-Jun 0650-0653 501.24013 295 -1.4 [-2.0]
9 15-Jun 0657-0707 471.26025 294 -1.8 [-1.8]
9 16-Jun 0739-0744 471.26026 289 -1.8 [-2.0]
10 16-Jun 0447-0451 687.24971 287 -1.3 [-1.1]
11 16-Jun 0533-0540 633.25960 287 -1.8 [-1.7]
12 18-Jun 1424-1427 576.24955 275 -2.5 [-2.1]
12 19-Jun 1531-1555 576.24962 267 -1.6 [-1.5]
12 20-Jun 1703-1706 576.24966 259 -1.1 [-1.2]
13 22-Jun 0813-0819 501.25885 247 -0.7 [-0.7]
13 23-Jun 0855-0902 501.25888 243 -0.8 [-0.7]
14 22-Jun 0818-0823 501.25979 243 -0.8 [-0.7]
14 23-Jun 0859-0905 501.25975 243 -0.8 [-0.7]
15 22-Jun 0830-0838 489.24989 245 -0.8 [-1.0]
15 23-Jun 0909-0918 489.24990 240 -1.0 [-1.0]
16 22-Jun 0904-0917 489.24890 245 -1.3 [-1.4]
16 23-Jun 0949-0953 489.24890 240 -1.3 [-1.4]
17 23-Jun 0935-0946 477.24996 239 -1.1 [-1.2]
17 25-Jun 1116-1127 477.24995 235 -1.0 [-1.1]
17 26-Jun 1218-1229 477.24995 235 -1.2 [-1.3]
18 23-Jun 1018-1020 603.24956 244 -1.3 [-1.8]
18 28-Jun 1501-1502 603.24956 247 -2 [-1.8]
19 23-Jun 1107-1108 519.25983 235 -2.2 [-2.4]
19 24-Jun 1146-1152 519.25982 231 -1.9 [-2.2]
19 25-Jun 1237-1241 519.25982 231 -2.0 [-2.2]
20 26-Jun 1204-1210 507.25995 241 -0.7 [-0.7]

* = bearing in degrees from the tx of the setting moon

** discrepancies between observed drift rate and Doppler are largely due to my spectrum analyser, which takes 40 seconds to complete a scan. If the signal is not present for most of the scan or appears for only a few minutes the error can be as large as 0.5 Hz/min.

Equipment: 24 element Band 4/5 Yagi at 2.5m above ground, horizontally polarized, Elevated (10-30 degrees) and tracked manually on the moon;
2dB NF GaAsFET preamp (JIM,M-75) mounted behind Yagi reflector;
Icom R7000, PC(Mac)-based audio spectrum analyser (bandwidth ~2Hz);
5 MHz frequency standard and 10 kHz comb generator;
moon prediction software.

April, 2003 update from Tony Mann

Here in Perth I've logged some new txs since January, 2003, notably from the Middle East, Sutton Coldfield, UK and northern Sweden. The details are given below.

The one megawatt txs from Sutton Coldfield and northern Sweden were detected with up to several dB signal-to-noise ratio on the spectrum analyser.

Dubai's 2MW tx on ch33 was found, but not as yet the tx on ch41. KTV Kuwait City gives a strong and reliable signal for about 20 minutes on ch24. It's there every time I look, unlike some European and North American txs.

I conjecture this is because the Kuwait City tx has an unobstructed view of the horizon at moonrise over the Persian Gulf. There's also a KTV tx of similar power on ch39, received simultaneously.

KBWT-19 Muskogee, OK is the strongest (and longest duration) signal I've found from the USA. This may be an artefact of my longitude and my location. My south-eastern horizon is partially blocked and that seems to affect my ability to receive many US east coast txs, which peak when the moon is only several degrees above the horizon. (I'm too far west.)

KBWT-19 gives a peak signal-to-noise ratio here of about 18 dB using a 1 dB noise figure LNA and 14dBd gain narrowband Yagi. This is sufficient to just be able to hear it in bfo mode on the R7000 rx. On 18 Jan 2003 I made an audio recording of KBWT-19. The last 5 minutes, displayed as a spectrogram, can be viewed at http://physics.uwa.edu.au/~agm/eme2.jpg It's the broken white line at 1.5 kHz. The signal fades out completely at times. You can just hear it, on peaks above the noise, in a bandwidth ~300Hz (the brown region on the spectrogram), at http://physics.uwa.edu.au/~agm/audio/501.mp3

carrier freq. tx qtf date(s) (MHz) (deg)
Europe:

London? UK 487.22402 116 26/1/03
Sutton Coldfield UK 647.22404 122 23/2/03
Sutton Coldfield UK 703.22404 120-2 23/2/03,29/3/03
Taivalkoski Finland 487.23943 146 27/1/03
Wroclaw Poland 503.23970 124 27/1/03

Storuman Sweden 567.23975-8 34-66 15,16,17/3/03
Gallivare Sweden 567.25008 65 17/3/03
Solleftea Sweden 695.25084 147 23/3/03

Middle East:

Dubai UAE 567.25002-10 72-90 19/2/03;16,18/3/03
Kuwait City Kuwait 495.25010-26 61-120 3/1/3-17/3/3 *
Kuwait City Kuwait 615.24942-51 72-8 16,17/3/03

* 495.25019 120 3/1/03 495.25025-6 72-8 16,17/3/03
495.25018 119 26/3/03

USA:

Mobile, AL WPMI 477.26033-5 273-91 6,8,9/3/03
Mobile, AL WMPV 513.25980 286 8/3/03
Burlington, NC WGPX 483.25003-7 287-91 8,9/3/03
South Bend, IN WNDU 483.25011 274 6/3/03
Goldsboro, NC WNCN 489.23992 274 6/3/03
Decatur, IL WAND 489.24993 267 5/3/03
Milwaukee, WI WVTV 495.23988 274 6/3/03
Muskogee, OK KWBT 501.24740-3 245-92 6,9,21,22/3/03
Iowa City, IA KWKB 507.23929-30 267-74 5,6/3/03
New Orleans, LA WHNO 507.23958-60 268-73 5,6/3/03
Wrens, GA WCES 507.23977 274-87 6,8/3/03
Lexington, NC WTWB 507.24915 273 6/3/03
Crossville, TN WBXX 507.25991 274 6/3/03
Ft Myers, FL WBBH 507.26007 274 6/3/03
Tollaston, AZ KPPX 693.24996 245 22/3/03

Equipment:

Below 520 MHz I use a 14dBd gain narrowband Yagi whose polarisation can be easily adjusted. Above 520 MHz I use a Jaycar (91 el) Band 4/5 Yagi whose gain is nominally 15-19 dB(i?) and whose polarisation is restricted to either vertical or horizontal. Normally I use horizontal antenna polarisation, which matches North America all of the time and Europe most of the time, except when there's Faraday rotation. However for locations 5000-6000 miles away, such as the Middle East (and north Asia), I always have to use vertical polarisation. The preamp is one of 3 having noise figures of 1 dB or better: a Winegard PA-4975 (gain 24 dB) ex USA and now 20 years old, a homemade 1 stage GaAsFET 480-510 MHz (gain 14 dB) and a Research Communications model 9247 (gain 20 dB). The Research Communications unit with a nominal noise figure of 0.5 dB seems to give results not noticeably better than the others, perhaps because my antennas do not have a high enough front-to-back ratio (~30 dB) to reject ground radiation. Also, it cannot be used on the 500 MHz narrowband Yagi as it is unstable with that particular input - very strange.

Rx: R7000 in usb mode with audio into a PC-based audio spectrum analyzer;
0.3-3.3kHz span, resolution ~1 Hz and ~16-32 averages.
High stability 5MHz quartz crystal frequency reference, with markers every 5 kHz.

Eastern Australia UHF TV received in Perth

For the last 3 nights I've been detecting weak moonbounce signals on Australian ch35 from Mt Ulandra NSW (the same QTH as ABMN0), 576.24966 MHz 34.8S,147.9E 1.6 MW Z H, only 200 miles from Sydney. Conditions were best on the 19th when the signal peaked at 5-6 dB above the noise and was around for over 20 minutes. The distance from Mt Ulandra to Perth is about 2,000 miles.

This tx is one of several UHF txs in Australia that now have 1.6 MegaWatts and are omnidirectional with horozontal polarisation. There are even more that are directional with 2 MW! See for example: http://www.aba.gov.au/broadcasters/pdfrtf/tv_7.pdf

In contrast the strongest 5MW signal from the USA here one night was 12 dB above the noise, but it averaged down to 8 dB.

Clearly it is possible to see txs down to 1.5 MW. There are also a couple of possibilities in Dubai, UAE at 1.7 MW on European channels 33 and 41.

On the 20th I also attempted to look for ABLV-40 Latrobe Valley (1.6 MW) and found something on 611.22381 at 1710 UTC. But moonset is getting very late here (after 1 am), so I won't be doing this again soon.

Options for improving system noise performance

Even though the Jayacar 91 element X UHF antenna has reasonably high forward gain, other antennas are available. For example, the Televes Pro-75 UHF TV antenna has 13-16 dBd gain from ch21-69. This is some 2 dB higher than the Jaycar 91 X. If one was to stack two Televes Pro-75 antennas, the forward gain would range from around 15-18 dB. Televes make a specially designed 5006 combiner for the purpose of stacking two Pro-75 antennas.

The ARRL handbook (1984-1991 editions), feature several GaAsfet preamplifiers for the 432 MHz UHF amateur band. If these designs were modified for the ch21-69 UHF TV band, interesting results could be obtained. For example, with a dual-gate GaAsfet design, the input stage could be tuned to 500 MHz, and having a untuned resistive output, wide bandwidth could be obtained.

Spectrum analyzer programs

Tony Mann uses a Macintosh PC with a homebrew program written by a colleague. Ian Roberts uses SpectraLab, but it's not freeware. There's also Spectran and Spectrum Lab, which are free.

Because the received levels of DXTV video via EME are extremely weak, visual display on an audio spectrum analyser is the optimum way of detection. The bandwidth of audio spectrum analyzers are typically less than 2 Hz, hence detection of extremely weak signals is possible.

Freeware spectrum analyzer programs

DL4YHF's Amateur Radio Software: Audio Spectrum Analyzer ("Spectrum Lab") http://www.qsl.net/dl4yhf/spectra1.html

Weak signal page. Collection of freeware spectrum analyser software. http://users.pandora.be/on4ant/weaksignal_pages.htm

SR5 Spectrum Analyzer. http://www.ar5.ndirect.co.uk/html/sr5.html

IARU Region 2 Monitoring System SPECTRUM ANALYSIS SOFTWARE http://www.echelon.ca/iarumsr2/fftsw.html

SpectraLab http://www.soundtechnology.com/download-center.htm

Spectra software for windows and soundcard by Nino Porcino, IZ8BLY . Accurate spectrum analyzer in the range 0 - 2500 Hz. http://iz8bly.sysonline.it/Spectra/index.htm

USA UHF TV databases

471-521 MHz USA UHF EME targets. Compiled by Doug Smith, W9WI. http://www.w9wi.com/articles/uhflow.html

W9WI TV database compiled by Doug Smith, W9WI. http://personal.bna.bellsouth.net/bna/w/9/w9wi/

EME resources

Collection of EME (Moonbounce) Station URLs compiled by W6/PAOZN. http://www.nitehawk.com/rasmit/ws1_1.html

Weak signal page compiled by W6/PAOZN. Includes FFT spectrum analyzers, DSP, PSK31, and other weak signal articles. http://www.nitehawk.com/rasmit/dsp50.html

Experimental moonbounce UHF TV DX reception. Todd Emslie's recent EME DX.

Using EME to Locate VHF-SHF Transmitters. By Ian Roberts, ZS6BTE.

EME software

Moonsked EME program by David Anderson, GM4JJJ.

The NEW 32 bit Windows9x EME Software by F1EHN.

The new F5SE EME Link Budget Software.

VK3UM's fine Moon Tracking Program provided by Ian, G3SEK.(ver 8.08)

VK3UM's Tracking Board Software.

VK3UM's Tracker software.

VK3UM ver. 8.12

UHF TV antenna

Jaycar LT3182 UHF 91 element TV yagi.