These LNB are used to convert the Ku TV emissions (from 10.7 up to 12.5 GHz) into a more reasonable frequency to be feed through long runs of coaxial cable (950 - 2150 MHz). They are characterized by its very low noise figure and its very high conversion gain (usually over 55 dB).
There are other LNB working at 4 GHz and 2.7 GHz, but they are quite rare and difficult to get, at least in Spain. These LNB have change a lot in these years, from the old ones that only converted one portion of the band with a single polarization, up to the most modern ones, called universal LNB, that works in two segments with both polarizations.
Today, the most easy to find, and the cheapest one, is the so called universal LNB. This is the typical diagram block of one of these LNB:
A universal LNB has a circular waveguide with an illuminator optimized for the parabolic dish. Inside the waveguide, there are two probes (the antennas), one for vertical and the other one for horizontal polarization. These probes are directly connected to a pair of amplifiers. From the output of these amplifiers, one signal is selected (the one from the desired polarization) and its feed again into a second amplifier. The next stage is a RF passband filter, with its output connected to one of the mixer's inputs. The other input is connected to one of the two available local oscillators. The mixer's output is feed into the IF amplifier and then into the coaxial cable, using a cheap F connector.
This F connector is the only electrical connection of the LNB, so the selection of desired polarization, and desired local oscillator must be done using this coaxial connector. In fact, the LNB uses a very simple mechanism to select the polarization/local oscillator. If you supply the LNB with 12 volts, you get one polarization, and if you supply it with 18 volts, you will get the other polarization. I say one and other polarization, because the real polarizations is dependant of the physical mounting position of the LNB.
To select the local oscillator, you must impose a 22 kHz tone to the supply voltage with an amplitude about 1 volt. If the tone is present, the LNB uses the 10.6 GHz oscillator. If not, the 9.75 GHz oscillator is used. You can get this tone with a diode in series with the LNB supply, and sort-circuiting it with a transistor 22000 times in a second.
Now, let's look inside a LNB. This LNB is somewhat older, and have a single local oscillator, so it do not respond to the 22 kHz tone:
We can see the illuminator, at the end of the waveguide. This illuminator has a plastic protective cap, removed here so you can see the the choke rings inside the illuminator. There are also four screws, if we remove them, we can see the internal electronics:
Almost all the LNB is shielded. We can also see a big screw in the shield. This is the fine tuning of the local oscillator. In a universal LNB you will see two big screws, one for each oscillator, and usually marked as L (9.75 GHz) and H (10.6 GHz). The shield usually acts also as heatsink for the voltage regulator (here a 7808). You can remove all screws to remove the shield, so you can see the components:
Let's use the small thumbnail to do a guided tour:
1.- It's a 74HC14, but in older LNB you can find a ICL7660. It is configured as oscillator tension inverter to get the negative voltage to polarize the RF FETs. It is also part of the 22 kHz tone detector in universal LNB.
2 and 3.- They are the soldered probes you can see inside the waveguide. They are DC connected to the FET transistors, so if you touch one probe, the electrostatic electricity in your body will kill the transistors. They are very delicate, so don't touch or manipulate them without adequate protection.
4 and 5.- The FET transistors. They used to be exquisite from the RF point of view, this is, they have a very low noise figure.
6.- Second amplifier, it is also a FET transistor. If look carefully, there is no switch between the first transistors and this one: The LNB supply one or the another transistor to select the desired polarization. I suppose this is a cheap method that works fine. A PIN diode (or similar) will be expensive, and will require more circuitry to work well at these frequencies.
7.- This is the pass band filter. Most ham radio mods acts at this point.
8.- This is the DRO (Dielectric Resonant Oscillator). It is a piece of ceramic with nice properties at microwaves, just like a quartz crystal, but not so stable.
9.- The mixer-oscillator. Many LNB have an IC for this purpose, Others have a pair of FET oscillators and a mixer with one or two diodes.
10.- FI amplifier. A bipolar transistor in this case, but sometimes you can find here an MMIC based amplifier.
The LNB as a 10 GHz receiverThe 10 GHz band covers from 10.0 up to 10.5 GHz. The LNB works from 10.7 up to 12.5 GHz. They are only 200 MHz apart. This small distance and the fact that a LNB is a really simple device will make possible to us to make the LNB to work in the 10 GHz band with little or not modification at all.
There are two points where we can make mods to the LNB to get a better 10 GHz performance: The RF band pass filter and the IF amplifier.
But many times there is no need to modify anything. Most LNB works nicely down to 10.3 or 10.25 GHz, so they can be used around 10.368 with great success. Other LNBs must be modified in some way to get it work in the 10 GHz band.
If we want to modify the filter, usually we will need to expand the filter tracks, not a easy task, so many hams introduce in the filter small quantities of some dielectric material, so the working frequency of the filter moves down. The dielectric material must be selected very carefully: It can be very lossy at 10 GHz so, maybe the modified filter have more attenuation at the 10 GHz band that the unmodified filter.
The other point that can be modified is the intermediate frequency amplifier. If the amplifier uses bipolar transistors, many times we can archive a better response increasing the collector inductance. This will give us more gain, or lower attenuation, but before do this, we must check if we really need more IF gain.
The LNB is designed to have very high conversion gain to overcome the great coaxial run to the satellite receiver, so the signal level at its output is very high.
If we connect a receiver expanded walkie or scanner at the LNB's output it can be easily overloaded by the strong signal coming from the LNB. If this is our case, we do not need more IF gain, otherwise, we need an attenuator.
Symptoms are very easy to detect, because usually an overloaded receiver shows S-meter indication without input signal. It is the same effect observed with walkies or scanners connected to a big antenna in a RF populated areas.
The best way to solve this is to construct a DC injector with incorporated attenuator. The attenuator can be one of the cheap 0-20 dB TV attenuators. Remember to shield the entire DC injector. You will be working around 600 - 700 MHz, in the TV band, and you don't want to hear TV signals in your receiver while operating at 10 GHz.
You can use small ceramic capacitors and a small coil or RF choke. In fact, the DC injector is a diplexer to separate DC current from a UHF signal, so any air wounded multi turn coil will do the job.
Voice can be used as long as it is somewhat wide. In practice Wide Band FM (WFM on walkies and receivers) is the narrowest usable mode with a LNB. Narrow band FM could be used as long as you use a 5 kHz step on the receiver, and look about 200-300 kHz (or more) up and down the center frequency. If you are lucky and find it, you must be able to make human manual automatic frequency control... not very easy, as the signal can jump several kHz in any direction in less than a second. SSB/CW is completely outside the capabilities of most LNB.
So, the combination of a scanner/walkie and LNB is a nice wide band FM receiver, just like the mode used with Gunnplexers, but incredibly more sensitive.
PrecisionBecause LNB are designed to receive several MHz wide signals, their DROs are not very precise adjusted. It is very easy to find LNB with local oscillators 5 or 10 MHz away from the nominal frequency. You can adjust them easily with a known signal (for example, with a frequency marker) and the adjusting screw.
It's very easy. Just tune the DRO's screw to get the known signal exactly at the theoretical output. For example, if you use a 10.368 GHz signal, you need to tune the DRO so you listen it just at 10.3680 - 9.750 = 618.0 MHz. At this point, you can affirm your DRO is working at 9.750 GHz, plus-minus its own stability.