The next project

Argh. I bought several different cheap 2.4Ghz Yagi's and found all of them to be junk. One of them had more gain in the reverse direction than forward. However, I haven't tested that particular model yagi.

I've also analyzed the MFJ yagi and found that is badly mismatched: There are 2.4Ghz Yagi's that work, but you don't find them for $10 on eBay. I suggest you test your new antenna to see if it has any gain before blundering forward. If not, I have some recommendations for a replacement.

Good luck.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

Somewhere in storage I have about 5 yagi-type antennas with built in LNA's. A group of us at HP built the antennas to intercept the 2.5GHz(?) signal from Home Box Office, before they started scrambling the signal to their subscribers.

The construction used flat aluminum sheeting strips, judiciously sized, and judiciously placed along the rod. The finished 'tubular' design looked like a skeleton of a paper towel roll. Don't know the gain, but remember the LNA had a NF in the range of 1.2dB [I think, some people actually got their antennas down to 1dB] Would that style of antenna work for this project?

Footnote on the history of those 'pirate' antennas: Homebox Office sent out trucks looking for people who had these types of antennas mounted on their roofs, and scanned for the mixer frequency to find them. They then mailed an invitation to gain amnesty, and start paying HB for the signals. I had heard some people did not and were taken to court over the matter - don't know. Shortly after HB did all this, they started scrambling the signal anyway so nobody could 'pirate' again. But, they did a real simple scramble technique [suppressed Horiz sync signal] which was easy to bypass with a simple little add-on.

I had mounted mine in a tree so wasn't noticeable *and* couldn't tell if it was mine or neighbor's. One day I accidently cut the power supply/control coax with a hedge trimmer into a couple of pieces. This was just before the truck searching, and threatening court campaign. But, HB's programming was so awful, I didn't even bother to fix, replace capble, or even take down the antenna. With the antenna not being electrically active, or visible, I missed out on all the bruhaha.

But I remember the thing was easy to construct and worked quite well. Much better than that rod type yagi.

Just used a square rod, metal shears to cut the 'loops', and a few rivets and hardware cloth screen on the back side.

Ummmm... like this perhaps? The antenna is about 900mm long and has 33 elements. I don't recall the exact frequency but I think it was something in the 1.6GHz range.

None of my clueless customers wanted anything that was big and obvious. Dish, dipole array, and panel antennas were therefore out of the question. A wire grid dish might have sold, but it was too expensive (at the time) to build. So, what was left was a yagi. Someone else was selling disk Yagis so we just copied and improved the design. At 60 wavelengths long, the antenna should have a theoretical gain of about 25dBi. Gain measurements were tricky with the junk we had for equipment, but my guess is we got 20dBi.

The LNA was bipolar because it was cheap and rugged. 3-4dB NF was the best we could do with bipolar. We tried GaAs but I kept blowing them up.

That's called a "loop yagi". It's circular version of a quad antenna. There were some of those around, but for lower frequencies. They work much the same as a disk yagi, except are more difficult to build.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

Inflatable GaAs? Who knew! ;-)

-- Anyone wanting to run for any political office in the US should have to have a DD214, and a honorable discharge.

"Tracking works great" implies this already works. Granted you are ESL, but I suspect this doesn't work so great. Top this off with the satellite band is not the 2.4GISM band. You haven't convinced me you know what your are doing.

If you are going to use a dvb-t dongle as the front end, forget it. That thing is too noisy.

Your antenna should be a biquad or cantenna plus a dish.

I think the service was Showtime, not HBO. It was a MDS feed. I still have the coffee can antenna and PCB out of Texas.

The service was called Premiere Channel. At the time, there was so much theft from Chinese receivers that Premiere hired EE students during the summer to drive around Silicon Valley mostly and spot the coffee can antenna and Chinese receivers.

The other decoder back in the day was when channel 48 (I think) turned their OTA signal to a pay service. How the hell the FCC ever approved that scheme is a good question. To descramble the signal, you had to recover the horizontal refresh, which was placed on the audio. The big problem was reinsterting the sync, since back then TVs didn't have a video input. You could glitch the RF with the recovered horizontal sync and that seemed to work. A National stereo FM demod chip was used to recover the horizontal sync, basically the 15khz signal was close enough to 19khz for a lock.

A lot of the MDS to analog TV block converters have been showing up on ebay.

That would have been MMDS in the US at 2596-2686. It was MDS-TV in other countries on other frequencies. MMDS was DOCSIS 1.0 upconverted to 2.5GHz circa about 2002. I still have a pile of MMDS hardware I converted to wi-fi frequencies.

The disk yagi in the photo was from a different era, probably the

1970's. MMDS came much later, starting in about 2000. Some time around 2001, Sprint used their MMDS licenses to deliver wireless internet on the same frequencies. Sorry, but my memory doesn't have better dates.

Measuring my disk yagi, the director is 4.7 cm in diameter. The conventional design requires a diameter = 0.3 wavelengths. Grinding the numbers, that's 15.7 cm = 1 wavelength, or about 1.9 GHz. I vaguely recall that was the frequency used for Smog Angeles system of the early 1970's. (The disk diameter for a Wi-Fi disk yagi is 3.7 cm). I should probably fire it up to see if it still works, but I don't recall the power supply voltage.

Drivel: This was the shiny new MCI MMDS "survey" van photographed somewhere in Tennessee(?) in about 2001. The diamond shaped antenna is a California Amplifier antenna, which includes a Wi-Fi band reject filter to help with interference problems. They were trying to figure out why their spectrum analyzer couldn't see anything on the Wi-Fi frequencies. Try to visualize what would happen if the approximately 80 ft (extended) tower were deployed on not very flat ground. Note the sagging rear springs on the van. This abomination disappeared about a week later.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

I tried the old loop yagi with fender washers back in the day. Nothing ever beat the coffee can.

What I got off of ebay a few months ago was California Amplifier 31732. They go for about $10 including shipping, well if you buy at least two. Lots of mods on the internet. I don't see any on ebay at the moment.

If you look down the list of frequency ranges on the MiniCircuits xformer catalog page, you'll see the problem: Most xformers crap out at about 1GHz MHz. The problem is that they're trying to be broadband. It's much easier and less lossy to use a resonant transformer, such as the 1/2 wave balun. It also solves the problem of going from a balanced load (folded dipole) to an unbalanced feed (coax cable).

Before you suggest it, a gamma or T-match can be made to work. The problem is that the location of the tap is somewhat critical. The T-match will require a hairpin stub, which is also a problem to tune. The gamma match might create a skewed antenna pattern. With a sweep generator and either directional coupler or return loss bridge, it can be tuned, but the 1/2 wave balun is much easier.

VSWR is not as important in receive as in transmit. Still, it would be nice if the antenna were reasonably well matched. The problem with the RCB is that it will only display the vector sum magnitude of the antenna impedance. It will not show whether it is capacitive or inductive. Therefore, there's no hint which direction to tune or cut (longer or shorter). One thing it's really good for is demonstrating how easily effected the antenna tuning can be by nearby objects.

Or a panel antenna. However, for tracking satellites, it will probably need to be a dish. A 24dBi barbeque grill dish isn't really that big but it's also not very close to a true parabola. A solid dish would be better.

However, you'll need more than a radio control servo to move it around.

Well, you can take the 3dB polarization loss going from RHCP to linear, or just build a proper CP feed. The feed beamwidth has to be matched to the width of the dish, so you can't hang any old CP antenna in place of the stock linear feed. A CP patch or helical feed antenna will probably work.

Note that the dish has the same alleged 15dBi gain as the yagi with a much narrower pattern and far fewer side lobes.

It's a common affliction. I have it also: Quite comfortable. I keep wanting to cover it with aluminum foil and see if will play like an antenna, but haven't bothered.

Note that they're using a 24.3dBi gain antenna in the link budget calcs. 0.8dB NF at 35dB gain is not going to be a cheap or easy to build LNA.

Good luck.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

On a sunny day (Wed, 31 Jul 2013 23:36:38 -0700) it happened Jeff Liebermann wrote in :

Yep, that is limited.

I will have to study those solutions, no experience with that.

That is a nice paper, I was just reading it, and already gave me some ideas. I have a big movable satellite dish mounted to the wall south facing, but it only moves in a pattern following earth curve for geostationary sats at the equator (40,000 km up).

An other large construction like that (and this is noisy too :-) ) is not something I would prefer. The small yagi would be nice, I would plant it in the garden for a start,

I have a lot of reading to do obviously.

But what is it used for? Plants?

Hey we are talking about 280 Euro, more than 300 $...

I bought a few of these on ebay:

This is the PCB:

This site has the circuit diagram, bottom page, seems exactly the same, following where the PCB tracks go:

2 x RF super low noise preamps 1 RF amp 1 Mixer / cable driver 2 oscillators with GHz resonators (fun stuff! 9.75 GHz and 10.6 GHz), 1 FET control chip, voltage regulator too :-) so far I spend 5 $.

The idea is now to use these parts to make that 300+ $ down mixer... The SMA connectors have already been ordered (and that is the next project after this one). I have 2 of these LNBs, so if I kill one of them super low noise FETs, I can have an other go (and there are 2 per LNB, once for H polarization and one for V). A tip: Those FETS are max 1.5 or 2 V or so , and do NOT use a meter with a 9 V battery, I modified an analog meter I have from ebay for only a few $ so it works on resistance range with a 1.5 V battery. :-) And still use a 10 k resistor in series. The control IC for the FET bias (negative voltage generation) comes with the 5 $ LNBs.

On a sunny day (Thu, 01 Aug 2013 07:43:40 GMT) it happened Jan Panteltje wrote in :

I found this yagi calculator:

It runs in Linux in wine (win 32)

For a RG58 coax the balun loop is only 4.1 cm, also double checked against some other program.

So I have backup if things do not work...

Via this link:

Oh-oh.

Before you embark on such a construction, I suggest you borrow a 24 dBi 2.4 GHz commercial dish and try a little experiment. Try to aim the dish at a known distant 2.4 GHz source, such as wireless access point. The key word is "try" as you are likely encounter difficulties. The problem is that the -3dB beamwidth of the 24 dBi dish is only about 8 degree wide: My measurements show 5 (vertical) degrees beamwidth but my accuracy is probably lacking. Trying to manually aim the dish, in 2 axis, is not easy. You can get a rough idea of what it's like by drawing an 8 degree angle on a piece of cardboard and trying to optically align it with a distant object. When I align high gain wireless link antennas, I either boresight the antenna (by removing the dish feed and sighting down the feed mounting tube), or use a laser. Extra credit for trying to track a moving RF source with the 24 dBi dish.

The fundamentals of a dish are fairly simple. The key part is the illumination angle of the dish feed. In transmit:

- Too wide, and you lose power in transmit off the edge of the dish. - Too narrow and you lose transmit gain by illuminating only part of the dish. In receive:

- Too wide an illumination angle means the feed will pickup terrestrial interference off the edge of the dish.

- Too narrow and you lose receive gain because the feed does not "see" all parts of the dish.

50% efficiency is considered typical. Also see f/D (focal length to diameter) ratio to see how it affects the required illumination angle. Note that most of the W1GHZ book is on feeds, with only passing mention of the parabolic reflector. That's it. Everything else is computation and construction details.

It's a chair or sofa. Add cushions and it's quite comfortable. Unfortunately, it's also rather difficult to get in and out.

$370 US dollars.

If you dig into the specs for RF front ends, the MMIC integrated devices will do about 1.5 dB NF at 2.4 GHz. If you switch to various discrete FET designs, you might see 1.0 dB NF. To get 0.8 dB, someone spent some time selecting the best devices, individually tweaking the parts for best NF, and probably has a huge pile of RF test equipment. I would have charged more than $370 for all that.

Incidentally, the same principal applies to antenna design. Anyone can build a 2.4 GHz wi-fi antenna with up to maybe 10 dBi gain. The antenna bandwidth is rather wide and construction tolerances are not very critical. However, as the gain increases towards 24dBi, the bandwidth of the antenna becomes more narrow and the cut tolerances of the elements become more critical. Miss an important dimension and you may have plenty of gain, but on the wrong frequency, or the reverse. Low gain antennas can probably be built without much testing and tweaking. High gain antennas need tweaking, tuning, and a large pile of test equipment.

Nice paper weight. That's a Ku band LNB (low noise block downconverter). Moving the LNB about 30% from 13 GHz to 10 GHz is probably not possible. You'll need new larger X-band WR-90 waveguide. You'll also need to "stretch" the interdigital filters on the PCB.

You might be able to remove the components and use them on a different PCB for building a 2.4 GHz downconverter.

The various AO-40 satellite receiver stations are a good start. Something like this, based on an MMDS front end: Note the size of the dish and the helical dish feed. I have the same dish:

The journey of a lifetime starts with a single SMA connector.

I have boxes of Ku band LNB's. I live in a forest. Getting satellite TV reception through the trees is not a trivial exercise. Replacing the junk LNB's supplied by Dish and DirecTV with lower NF versions is usually required. The result is a rather large collection of noisy LNB's, from which I steal microwave components.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

This looks like a good 2.4 GHz feed and balun construction, except that the exposed parts of the coax cable are too long. Also, exposed copper should be plated with something as copper oxide is a lousy RF conductor.

I use it and it works. Note that it assumes that you're going to need a 4:1 balun for the yagi.

I like this one a little better: I have several others. The fun thing to do is run the same the design through several Yagi design programs and compare results.

Careful. The length is easy. It's 1/2 wavelength at 2401 MHz or 6.25 cm. Multiply that by the velocity factor of the coax. That's 0.66 for low density polyethylene dielectric such as RG-58c/u, 0.697 for RG-316/u (common coax used for 2.4 GHz), or 0.78 to 0.83 for various flavors of semi-rigid coax. I've measured the velocity factor on some of the coaxes I've used, and found that it varies, mostly dependent on the quality of the coax.

While driving, the only accidents I've caused were all while backing up.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

On a sunny day (Thu, 01 Aug 2013 09:23:25 -0700) it happened Jeff Liebermann wrote in :

I have done the tracking linearity and absolute error measurement versus the value requested by the predict program, and it is within about 2 degrees with this servo setup. No way will I point that accurate by hand, and ISS moves rather fast, I have seen it once when the sun got reflected from it, gone in a minute or 2 IIRC. I recognized the shape, went inside, looked up where it was on internet, and sure enough. So with a direct line of sight and a few degrees the yagi should work, especially if a lot above the horizon.

I have considerable experience aligning satellite dishes, and I wrote a software (open source) positioner too,

my 80 cm motorized dish goes all the way from east to west, you CAN align it by hand (beep signal in xdipo), but you would get tired very fast, its heavy... Note that xdipo has a satellite position calculator build in. The big problem with all this is to find the exact north (or south) The advantage of already having this dish is that it is aligned and I can point it exactly south, to within a degree. Here you see xdipo working with a PCI wintvnova sat receiver car, remote via ethernet on the laptop: I added some features such as QPSK constellation display, and other signal relevant stuff.

Yes, that looks like a trap :-) You'd have to roll over to get out...

One of the plastic black things fell of today, it is quite think alu they used for the directors. Alu profile came today, things seem to fit for mounting the servos.

I will of course NOT use that PCB and wave guide, just use the parts! And these transistors do .5 dB with a huge gain at 2.4 GHz!

That is the idea!

Right, cut some trees? Build a tower? I do not know much about US sat TV channels, here we have hundreds (no kidding) of free to air TV and radio channels. PLus Russia, China, Arab world, at one point I could even get Cuba and South America, relayed by an uplink in Spain IIRC, just not NASA TV, but that works on internet, The sat dish is worth every cent of it, had this one for 10 years now, seen some bad storms too, still same LNB, huge signal.

On a sunny day (Thu, 01 Aug 2013 09:41:30 -0700) it happened Jeff Liebermann wrote in :

I have ust downloaded that, wil try int in Linux wine.

I agree :-) That is actually what I did. It seesm to me that at 4.1 cm the round box on my yagi could well house that coax loop, its circumference is 8 cm..

Well, yes, in Germany I was parking after driving from where I live to Berlin, what I did *not* see (very low) was the hook on the car behind where I was parking (for a hanger), and it punched a hole in my car and JUST missed the LPG (liquid gas) tank by a few cm. Well many lives were spared :-) The thing is I have been almost dead so many times, almost is no problem.

These ginormous yagis remind me of people trying to make very high order analog filters. The component tolerances eventually make the additional stages useless. The dish on the other hand can provide a lot of gain and is itself not a tuned circuit. Your problem is then reduced to the feed, which like you stated is the low gain, easier to build antenna. Hence the biquad feeding the primestar wifi hacks.

I've seen military application where the dish feed looks like a log periodic, i.e. they are doing some very wide band RF. Since the dish is frequency independent.

S-band satellite reception is really tough, unless you talking about a blow torch like XM. And there isn't a lot of TVRO gear you can tweak for S-band other than MMDS, which really isn't satellite grade.

C-band, on the other hand, is hackable. It is at a scale where you can roll your own horns. C-band itself is not uniform across the world, so you can get a LNBF from one country that has er um interesting properties in another country. People also grind the oscillator element to change the frequency.

Don't remind me. I did a quadrature SSB modulator using RC filters. Something like this with more RC stages: No hybrid or integrated devices allowed in 1972. Using off the shelf components and tolerances, I couldn't get the carrier suppression down to the -50dBc target. What surprised me was that when I was finally allowed to break the budget and use 1% tolerance and bin selected parts, I still couldn't make it play. I finally built a hand trimmed prototype that met the spec, but wouldn't stay put over more than about 10C temp range.

What, me hack?

The Allen Telescope Array (SETI) uses a log periodic feed. 0.5 to

11.2 Ghz:

Well, the problem with the S-band birds is they tend to have really eccentric orbits. For example, the defunct AO-40 satellite: See "View from above orbital plane". You get really good signal levels at perigee and weak signal at apogee. However, for this project, the ISS is in low earth orbit at about 400 km, which should produce usable signals.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

I have what looks like the same feed off an old MDS antenna. Any secret sauce in hooking this up to a dish? I have some of those oval dishes that I rescued from Craigs before they ended up in a landfill.

The dishes look like this:

One was from satellite internet. I have the BUC as well, though I suspect I can only get into trouble playing with it.