[SC.LUG] RE: [SC.LUG]Radio Telescopes = was ISPs

[Martyn Kinder]

See below...

----------------------------------------
Martyn Kinder G0CZD

mailto:martyn@czd.org.uk

http://www.czd.org.uk


-----Original Message-----
From: sc-admin@mailman.lug.org.uk [mailto:sc-admin@mailman.lug.org.uk]On
Behalf Of Ian Molton
Sent: 11 August 2003 22:27
To: sc@mailman.lug.org.uk
Subject: Re: [SC.LUG] RE: [SC.LUG]Radio Telescopes = was ISPs


On Mon, 11 Aug 2003 19:27:31 -0000
"Martyn Kinder" <martyn@czd.org.uk> wrote:

> Use a common local oscillator at a lowish frequency - say 1GHz and fed
> to each receiver via identical length cables. So take 50ohms out of
> the LO to a wilkinson hybrid and you will get two outputs still at 50
> ohm.

:My knowledge of RF is shaky at best. slow down a tad ;-) whats a
:wilkinson hybrid?

To get a proper match and reliable and consistent power transfer within RF
blackbox units, it is essential that the impedance between units is
matched.Most RF systems (TV's and Video systems excluded) use 50ohm as the
base feed impedance.

A single LO will have a 50ohm output and needs to match to a 50 ohm mixer on
the receiver. However, if you split it, you end up with 2 25ohm outputs
(ohms law). So a wilkinson divider does an impeddance transfer to 100ohms
first before the split, so maintaining a 50ohm system.

:a 1GHz LO wont be high enough to get the signal within my receivers
:capability. I need ~9GHz in order to bring the freq. below 1.3GHz (and
:its a cheap receiver so I'd prefer to stay well within the range).

Yes it will. All LO's have elements of multiplication. The normal sat  LNB
use a DRO to generate the LO in the LNB. These run at about 9.0GHZ, mixed
with the 10.5GHz received from the Satellite gives two outputs of 1.5ghz and
19.5ghz. The upper output is filtered out, leaving the lower 1.5ghz
amplified and fed to the sat rx. This is OK for a single system, but DRO's
are not stable and suffer a lot of phase noise. OK if you are resolving a
wide signal, but no use at all if you need to closely compare two signals to
see if they are in phase - ie in interferometer.

Now by starting with a lowish LO (at about 1GHZ) makes components and
measurements (remember the feed to the receivers has to be identical) far
less critical. So now you simply multiply and filter at the receiver (LNB)
the common LO and mix with the received very weak signal.

> It may need a gain block to recover any lost power, depending on
> how much you started with. Feed the multiplied LO (multiplied by 22 in
> this case) to the receiver mixer. Now down convert to the same
> Intermediate frequency - in my case 70MHz. I will now have a pair of
> receivers tuned to the same source, providing an exact in-phase
> composite pair with a 22GHZ input and 70MHz output.

:So I was on the right lines for that implementation. I think initially
:I'll use my plan, just because its simpler.


> the best
> designed feeds will have a main lobe pointing in the direction you
> want ie straight ahead plus any number of scondary lobes. by careful
> positioning of the two dishes, you can cancel many of these lobes out,
> thus providing a much narrower beamwidth.

:Can you expound on this some? diagrams would help. I dont understand how
:'secondary' lobes come about.
See http://www.qsl.net/n1bwt/chap1.pdf

Paul Wade is an acqaintance who has produced the best amateur on-line
microwave antenna book.

> Bear in mind that most astro
> radio sources are point sources, it is important to keep the beamwidth
> as narrow as possible to get the highest resolution.

Thats why Im choosing a high frequency (10GHz). Not as high as yours
though :)

In the sceme of things it is arbritary. Your Sat dish has a 3dB beamwidth of
about 3 degrees. the sun subtends an angle of 1 degree. The Taurus radio
source is about .1 of a degree......

> Now to go VLBI, you have the
> same problems. assuming that you will be using drift scanning, ie
> letting the source pass in front of your aerial, you ideally need two
> aerials a long way apart.

:Correct.


> The further apart gives you higher
> resolution, but ONLY if you can guarantee that the two downloaded
> signals are exactly in phase, otherwise you will not have a clue where
> the resultant beamwidth is pointing.

:Yep. I got that far.

Good

> A few years ago I was in
> discussion with a radio Ham in Paris to build a VLBI on 136MHz. At the
> time, there was no way we could synchronise signals well enough, but
> with GPS it is now (slightly) easier. So we can get Local Oscillators
> locked by using GPS as the source, but what we can't do is measure the
> distance between our two locations to 20cms which we thought would be
> appropriate for this experiment.

:Hm. at 10GHz I'll need to know within ~2cm.

Nope - 3mm

The thoguht occurs that one aerial could be mounted on a skid 2m long.

a *GOOD* GPS can get just sub metre accuracy these days, so we could
mount one antenna fixed, and move the other over 2 metres (one either
side of the GPS's idea of the location) until the strongest lock is
achieved...

GPS is not reliable enough for positioning. the americans build in a certain
amount of uncertaincy which varies from day to day. But what it does give
you is a very accurate time source that will be the same for you as well as
for me. It also provides a common reference system for your local
oscillator....

If we managed that, I would be WELL impressed - since Jodrell bank only
managed 10 times that precision I believe!

Ian Morrison (a JB astronomer) reckoned that one of the most useful things
they achieved was to measure the distance between two dishes to an accuracy
of a few mm.

> knowing this exact logitudinal
> position would be essential or you would be simply adding two signals
> together without understanding anything about the phase relationship.
> worst case, you could be pointing at a huge Radio Source and it would
> be totally cancelled out..... :-(

yeah, that would just suck...
*Serious understatement* and just as likely as getting the correct additive
pair.

> I am looking for a pair of 1.5 metre dishes. This will give about
> 45dbi gain plus another 100db system gain. For amateur use, absolute
> gain is probably less important than low system noise figure and
> reproducibility.

Can you make that two pairs? and centre-fed, not offset, preferably?

1.5m dishes are like rocking horse droppings as they are just manageable. 3m
dishes pop up more often but my wife has said "no" to 2 of those already -
bugger.

Im keen to make this work, so I'd be happy to put some cash in, eg. to
buy a GPS rx for clock and position work. I'm also prepared to build a
dedicated PC for the measurement taking, equipped with a UPS (I already
have the UPS).

Lets see how you get on and if you can clearly identify Taurus A first. If
you can mange that (consider that as crawling) then we will have a go at
walking...

see http://www.ras.ucalgary.ca/radiotel/calibration.html for an excellent
paper describing side or secondary lobes

KM
----------------------------------------
Martyn Kinder G0CZD

mailto:martyn@czd.org.uk

http://www.czd.org.uk