Backscatter Radar

[VK3OE]

Peter VK5PJ asked: "would it be possible to make this function from a single site"?

This is a rather bigger question, as operating from one site changes the system from a bistatic radar into a monostatic radar. This is because at a single site the receiver will be overloaded by the transmitter so the receiver must be blanked off during transmit. The effect of this is that the minimum range is determined by the length of the transmit pulse plus the receiver recovery time. For a 20ms transmit chirp and a 10ms recovery time for the receiver the minimum range is about 4500km. If the receiver listening time is 170ms (assuming a 200ms transmit-receive cycle) then the maximum range is 25500km. This means that most VHF tropo, Es would not be seen as it is within the minimum range. However it could be very effective for looking at long distance propagation on 6m and HF.

The processing gain would be reduced to 16dB or 19dB for a complex receiver (from the 33 or 36dB for the one second bistatic system) but that is probably not too much loss. The transmit power is also reduced by 17dB because of the short transmit pulse when compared to the one second transmit pulse. So, the overall loss of gain when comparing the systems is 36dB. This starting to get serious and may limit the capability of the system on 6m but it will most probably still be OK on HF.
The range resolution remains the same as the one second system at 75km provided that a 2000Hz wide chirp is used.

The big trick is to get the receiver to recover quickly enough so that the minimum range is not compromised, this will require some diode RF switching for the transmit receive function as well as an input clamp for the receiver to minimise the overload from the transmitter as the receiver will need to be running all the time. The receiver AGC will also need to be turned off. Spectrum Lab can certainly generate the required chirp but getting a transceiver to run the required transmit-receive cycle is a big ask. A separate receiver is most probably the way to go with a custom T/R switch and receiver input limiter.



Wayne VK4WDM asked; distance and that confirms it is California, Has this old guy got it right?

You certainly have it right Wayne, the system gives the distance and you see the echo therefore it is open to California!!!

1. How narrow is the beam? Can you tell if it is the coast of Mexico, California or Alaska?
This is a bit more problematic as the beam-width is set by the combined beam-widths of the transmitter and receiver antennas. You would be lucky to have a beam-width much smaller than 20 degrees so this limits the guess as to where the signal is returning from. Rotating the antennas automatically and plotting on a "A" scope plot would also help... The future...

2. How big does the target have to be? What about Hawaii of some of the smaller Pacific Islands?
The target does not have to be land as the sea surface will also provide a good backscatter signal but is somewhat dependent of the state of the sea surface. Land appeared to provide a better signal in the tests that we did.
3. I presume that the signal gets up so high that intervening land masses don't get in the way so you could, say, check a path from VK4 to ZS?
This is correct for Es and F, in fact the test we did showed a long path return from ZS. For tropo the land intervention will still be an issue, you will see the tropo terminate abruptly and a big signal from the blocking land mass.

4.Will it work on 2m? I have often thought that VK to KH6 on 2m is possible. This might tell us when it might be possible. Also other paths such as: VK6 to Reunion or even Africa, or VK7 to Vk0.
This is also correct. the system has the ability to look in any direction and make an idea tool for checking paths to anywhere. It is not limited by frequency, it can be used in any amateur band to check the propagation in any direction so that the reliance on beacons or TV transmissions etc. can be finally broken.

The system will also directly measure the backscatter as well as the surface reflections, this will be a developing art to determine which effect is being seen. For instance, for tropo there should be a continuous return as the elevated duct will leak along its whole length. Rain scatter on VHF and microwave bands will be interesting to try out. Aurora on 6 and 2 can also be checked assuming that the Doppler shift does not cause too much problem for the chirp receiver!!

The idea of this new system is to break the bounds that amateurs have been previously limited to by always needing to have a signal at the other end. This is no longer necessary. The possibilities are substantial and it will be easy for anyone to have a go. I am looking forward to having the first release software so that testing can begin in earnest and everyone can try it out. The results will generate much discussion and provide a great tool for finding new paths. There is also the possibility of placing internet connected remote stations at good locations but that is another forum post.

Thanks for your questions, 73
Andrew VK3OE
PS. Keep checking the latest release of Spectrum for further developments.

[VK3OE]

I have just received an email from Wolf DL4YHF advising that a new version of Spectrum Lab is available for testing the chirp Backscatter Radar, see below for the links.
Good luck in testing, I have not yet had time to try it out but please let me know how you get on and we can pass comment on to Wolf.

"Hello Andrew,

I have just uploaded a new SL version (V2.75 b05) which has the Chirp
Radar configuration built inside the "Quick Settings" menu (under 'Other
amateur radio modes' .. 'Chirp Radar Experiment'). The second
programmable button (which shows 'Radar: Transmitting' or 'Radar:
Receiving') can be used to switch between RX and TX. The RX display
still uses the old time-domain scope, but the trigger now uses a
sample-counter based time (it doesn't need the received signal for
triggering). Averaging is off by default, but it can easily be turned on
in the TD-scope window.

Some experiments with additional windowing are still pending, but I
thought you (or other users) would like to test this early beta version.

The installer can be downloaded from here:

http://www.qsl.net/dl4yhf/spectra1.html

And the preliminary documentation of the Chirp Radar config is here:

http://dl4yhf.ssl7.com/speclab/chirp_fi ... hirp_radar

(Note: the old pages at freenet.de were discarded when freenet suddenly
decided to charge money for hosting a few MByte of webspace)

Cheers,
Wolf . "

73,
Andrew Martin
VK3OE

[VK4ABW]

Hello Andrew. I have been following this topic and after your last post realised that i may be able to contribute.

My current setup consists of 6 totally separate radios and antenna's on 6 mtrs. Off this, 4 towers each have a 13el long boom yagi (60ft) with full rotation and are fed from my study via 5-50 or 4-50 LDF hardline. These 4 towers are separated roughly 80mtrs from each other which allows me to easily have several receivers parked on 50.110mhz whilst transmitting on the band. There is some de-sense between my northern tower and southern towers but it is very minimal.
The point i'm getting at here is that i have no problems transmitting 1,500watts on my southern tower whilst simultaneously listening on my northern tower. I will have to do some checking to see what the minimal recovery times are, but i am pretty sure that if i turn off the agc on the Icoms, it would be extremely quick.

Do you think it is worth while for me to install and test the latest version utilising my qro permit for testing purposes

Gary
VK4ABW

[VK3OE]

Hello Garry,

Absolutely, give it a go!!! The critical issue is that you may not be able to see "in close" (see answers to questions above). However if you can get the receivers to recover quickly then the arrangement should enable you to see from about 4500km out to around 27000km, from your own station if the total transmit and receiver recovery time are about 30ms. It may also help your receiver if you put some limiter diodes (low threshold reversed in parallel RF diodes with a 200ohm resistor also) across the RF input to the receiver so as to ensure that the receivers have the best chance of recovery. If your antennas have enough decoupling to give an S5 to S7 level then the close in will also work but given the distance that is probably not achievable.

Good luck, let me know how you get on.

73
Andrew
VK3OE

[VK3OE]

Hello,

I have recently modified the latest version of Spectrum Lab to work as a Chirp Radar using one transceiver, monostatic mode, so that a short chirp is sent and when the receiver recovers return signals can be seen. I used a 756PIII with VOX, and AGC set to fast. The receivers recovers fast enough so that returns past 5700km are seen. An example of a return on 21 MHZ is attached, together with a no return trace. I used 64 x averaging to stabilise the trace. The direct signal was obtained through the monitor function, but if this is not available, then a small amount of the sound card output should be coupled back to the input to get a direct signal reference. The distance is calculated by multiplying the time from the direct signal to the return by 150,000 km to get that actual distance. If your receiver can recover more quickly then the first signals will be received from closer in. There is about 33dB of gain lost running in this mode but given that I was using just 50 watts and a dipole on 21 MHZ then that does not seem to be a problem. This should also be viable on 6m but there has not been any propagation to try it out. Of course if you want close in reflections then you will have to run in bistatic mode, with a distant receiver so as not to overload the receiver.

Monostatic Amateur Radar, 21 MHZ, Dipole, 50 Watts IC756PIII

If anyone wishes to get the Spectrum Lab config file for the maonstatic radar application please email me VK3OE@bigpond.com

73

Andrew
VK3OE

[VK4GHZ]

Andrew, this may be a dumb question (?), but why is there a discrepancy with the direct signal time in the two graphs above?

With the same test equipment/procedure, shouldn't the direct signal back time in ms be constant?

[VK3OE]

Hello Adam,

Yes, not a dumb question at all, in fact I have been pondering this for some time and it appears that there are two issues.

Firstly there is drift is due to differences in the sound card input and output clocks. This can be compensated for In Spectrum Lab and shows as a slowly moving (in time ) pulse in the time domain scope when the system is running.

The second issue is that the chirp receiver is not yet synchronised to the chirp generator when the system is started so that the location of the direct signal pulse is arbitary at the moment. There is a big issue of how to time everything so as to get the correct location of the impulse from the chirp matched filter receiver. Ongoing development !!

Many thanks,
Andrew,
VK3OE

[VK5PJ]

Hello Andrew et all,
while pondering other matters I had one of those eurka moments, I think that any possible frequency error between stations in Bistatic mode can be easily adjusted out of the equation . If the transmitting station runs a fixed tone of, well lets say 1000Hz for this example, then the Rx station can simply measure it in Spectrum Lab and then adjust their Rx freq to make it read 1000Hz, once this simple process is done, they can move on to running sweeps safe in the knowledge they are on the same freq. In this way even an old valve rig with a propper VFO could be used, as long as it is stable over time.

This does not help with drift over time but if this is suspected, then they can periodicially re-calibrate using this simple procedure.

Any thoughts, have I missed a clanger

[VK3OE]

Hello Peter,

Yes, running a 1000Hz tone will work very well as one would assume that in bistatic mode (separate TX and RX) the direct signal is strong enough.

73,
Andrew
VK3OE
PS Wolf is working at improving the Spectrum Lab implementation as it still has a few issues which limit the system capability, but it is basically working.

[VK4ABW]

Hello again Andrew.

Question: is the ms time to distance (x axis) being looked at? as a user set option it would be great.

This would alleviate using a calculator to work out my return distances. I did get a few returns on the w/end around 90ms & 195ms using a single 13el yagi and 200w pointing NE. I'm using a USB interface for all keying plus tx/rx audio, so the system is operating damn quick. Only using the IC7700 barefoot at present as the qro amp will introduce too much delay for monostatic operation.

Another possibility is to hook up one of my TE amps in FM mode (fast switching), which will allow me to chirp 400w.

I am unable to use the monostatic radar during TEP to the NW due to chinese crud occupying the whole band

Gary
VK4ABW

[VK3OE]

Hello Garry,
WOW !!!! Those returns are from 13,500km and about 30,000km, very similar to the returns that David and I got from initial testing!! That would suggest N America and SAfrica on the long path. The distance calculation is time X 150000km. That fast keying is critical as otherwise the distance starts too far away.

I am working with Wolf, the author of Spectrum Lab to improve its capability. We are working to be able to set the direct signal to 0 so that a distance scale works for the monostatic modes. For the bistatic mode timing between stations is a critical issue, but if timing and frequency are stable then we could introduce an offset to set the direct signal to 0 in bistatic mode as well.

We are also working on a moon bounce mode with a 4 second cycle time but having some problems to get it correct. We are not getting the full processing gain and the averaging is not fully operational, it needs a complex average instead of a magnitude average to get the full gain from averaging (It all has to do with noise having a random amplitude and phase and the signal having at least a more or less stable phase during the measurement period).

Pity about the Chinese crud, the improvements that we are planning to the system in Spectrum Lab may be able to partly reduce the effects of that but .....maybe 52MHz!!

Thanks for the report.

73
Andrew
VK3OE

[VK5PJ]

Hello Gary,

I am unable to use the monostatic radar during TEP to the NW due to chinese crud occupying the whole band
Gary
VK4ABW

Could you let us know if you have tried the monostatic RADAR while the TEP is there? The processing of the signal in Spectrum Lab should kill most of the stationary signals, this is one of the beauties of a swept waveform, carriers that are not sweeping at the same rate are factored out in the processing by the FFT's (Fast Fourier Transforms) that are going inside the system. Unless there is a receiver AGC / Overload condition, you may be suprised what comes out at the end. (I have some HF RADAR and sounder background, through time at Jindalee in Alice Springs)

Regards,

[VK4ABW]

Andrew. I use WSJT65 in ping mode for moonbounce on six very successfully. Pretty sure its version 4 as the current version 7 doesn't support it anymore. It is good for testing faraday shifts back to originating station.

regarding the returns signals i saw on the w/end, they were pronounced but not as high as your examples. The distant one was only evident for about 10 minutes but the closest one seemed to be 'ranging' in and out as well as building and fading?? At the time of testing there was a high level of particle noise in that direction. I tested for about an hour on several frequencies to rule out any birdies. i also checked with spectran running in waterfall mode for any rubbish and all was clear. Next time i see something i will screen capture to get the exact ms.

I adjusted the noise level of the bistatic radar down a couple of db (about -55db) so that i could watch the floor level and one thing i noticed was that the noise curve between 0 -100ms seemed to be going up and down very slightly, like an ocean wave. It also moved in time very slightly, so i am not sure what is happening here. This seemed to repeat over and over but was extremely slow, so i assume it may have been affected by the particle noise i was hearing

WRT to TEP, the amount of rubbish, plus the warbler was throwing S9 crap across all of 50mhz.....i didn't venture there. As all my 13el yagi's are peaked for 50.2mhz, i've been testing around that frequency for maximum performance.
(the JA's were there anyway after my test, so i worked a couple)

Stay tuned.

Gary

ohh...by the way....i did turn the beam and the blips disappeared.

[VK3OE]

Hello Garry,

What can I say but a great THANK YOU!!!!! for your detailed testing, that is extremely encouraging. I was not able to do such testing of the system at the time of the original work because of a lack of automation.

The system performance will also be improved once we fix a few residual problems which will greatly increase the gain.

Again many thanks and I look forward to your further testing and some pics !!

73
Andrew
VK3OE

[VK4ABW]

Andrew. I tested the system again lastnight looking straight into my TEP zone (aka 328 deg) but unfortunately no return signals were seen. By the looks of the radar screen, I could see the tailend of dekey and 28ms later my rx is recovering. As the band was reasonably quiet, i set the radar at -80db and this is when the problems started. I lost that pronounced start of the receiver and end of tx points, which resulted in a just a horizontal sound line from 0 - 200ms? I closed the program and started again but this time there was no noise showing on the spectran but the radar screen still had a flat line? So, i shut the pc down and restarted again but the problem is persisting.

My USB device is working fine and so is the radio as i proved both by using spectran in waterfall mode and all was good. I loaded my .usr file and the program was fine but now every time i load the monostatic.usr file the problem comes back. The radar screen is displaying the rx sound (but not correctly) and the keying screen is empty??? yet it is still keying the radio and chirping??.

Any ideas?

[VK3OE]

Hello,

Today we tried out the backscatter radar again when there was some Es around. VK3OE transmitted and VK3AUU received. We used the Spectrum Lab software with one second chirps and the config file used was "ChirpRadar.usr". This sets up Spectrum Lab so that it transmits one second chirps and the time domain scope shows the returns. The scope is initially set to display one second as the actual timing of the transmit pulse is not synchronous to the receiver. Once a a signal is received it can be zoomed in to using the red box zoom. The results for today at 30 degrees is attached.

The direct signal is the first signal at 693 ms ( our beams were side on so that the direct signal was close to the noise) and the return signal is at 727ms at a S/N of over 40dB, a difference of 34ms or 5100km. There may also be weak signal returns at 703ms (1500km) and 710ms (2550km) which would be consistent with Es at the time.

We tried at 330, 0, 30 and 60 degrees but the 30 degree return was the best. Pity there is no one out at T33, V7.

The basics were published in 2/2010 of DUBUS and there will be a full presenation at GippsTech 2010.

[VK2ZRH]

PRESENTATION ON THE BACKSCATTER RADAR - SYDNEY 1ST SEPTEMBER

For anyone in the Sydney area keen on hearing about and seeing results from Andrew VK3OE's backscatter chirp radar, I'll be giving a presentation at the monthly meeting of the St George ARC on Wednesday 1st September. The presentation is an updated version of the one I gave at GippsTech 2010 back in July. Here are the details:

The Practical Propagation Explorer
A Digital Backscatter Chirp Radar for Amateur Use

Venue: Kyle Bay Scout Hall, Donnelly Parade, Connells Point
Time: 7pm for 7:30pm

I'll be giving an outline of the principles of the chirp radar, show the sort of results achieved and an analysis of the propagation revealed. All from some smart software and common amateur equipment.

Take a look at "the cutting edge"

73, Roger Harrison VK2ZRH

[VK1FMSY]

Roger,

Will there be a presentation available via e-mail for those interested, but unable to attend?

I am quite fascinated by this topic.

Kind regards,

Sean
VK1FMSY

[VK4XA]

Roger,

Will there be a presentation available via e-mail for those interested, but unable to attend?

I am quite fascinated by this topic.

Kind regards,

Sean
VK1FMSY

+1 here too Roger
I am interested as well but the trip down from FNQ is a bit expensive nowadays.

[PA5BW]

Hi Andrew,

I stumble over your postings a bit late. Just read a remark on the HPSDR reflector (see http://www.openhpsdr.org) on your experiments with Chirp.
Very very nice work!

For my work I just completed a detailed long haul (300 Km) UHF propagation measurement campaign of 500-days, using the analogue TV-transmitters as a beacon signal, And as they are switched of here now, I did realize it was a "last chance"occasion. Although we still have a lot of questions on propagation issues. So we were already pondering on solutions. Maybe you have already found one!

Is it possible to derive the absolute field strenght from the reflection? By using time gating and then summing up everything that comes back between two predefined delays? That would be great for my purpose.

As on the pictures: I see you display the vertical axis in "volts"??? What is really on the vertical axis?
Do you have graphs for me from a reflection tha has a dB-scale on the vertical axis (need not be calibrated)?

73, Ben PA5BW
pa5bw@xs4all.nl