.jpg)
The waveguide section of the LNBF - Tuning
The white plastic cover over the front of the LNBF just pops off. Easier said than done since it's a tight fit, but if you keep pulling eventually you'll get it off. Looking into the LNBF you can see the various components shown below:
Tuning the LNBF for 10Ghz EME
The LNBF uses a waveguide probe to receive the signal. In fact it has two, one for vertical and one for horizontal polarization, but I'll deal with that later. The probe is presumably optimized for something like 11.7GHz since most "universal" LNBFs can actually cover from 10.7Ghz to 11.7Ghz and then by some local oscillator tricks also cover from 11.7 to 12.75 Ghz. Whatever it is optimized for, it's almost certainly not 10.368Ghz. To get it best for 10.368Ghz you need to tune the probe by put a screw though the wall of the waveguide opposite the vertical polarization (front) waveguide probe. Just drill a hole, tap it with 4-40 threads, then insert a 4-40 brass screw with locknut through the hole and tweak it for best performance by the method described below.
This image shows tuning screws opposite both the vertical and horizontal probes.
Connect up the receive system. At this point you can be using an unmodified LNBF as it come out of the box. The exact frequency of the LO doesn't matter (+/- 500kHz is OK for this test) and any frequency drift doesn't matter. You're IF receiver must be tuned to the right frequency to receive incoming signals at around 10.368Ghz, so that would be 618Mhz if you haven't changed the 25MHz crystal in the LNBF. If you have changed it then you tune to whatever IF frequency puts you on 10.368GHz. For example if you are using a 25.4 MHz reference frequency, the LO will be on 9906 MHz(25.4 x 390) and at 10368 the IF will be at 462MHz (which many VHF/UHF all-mode transcievers can tube to).
You then need to be able to measure the system audio (AGC OFF) or IF noise power. If you are using an RTL-SDR dongle to look at 618Mhz, the software you are using (e.g. SDRSharp, HDSDR) probably has signal strength indication. With the right interface to your radio or RTL-SDR dongle you can use a program like SpectraVue in continuum mode to measure average noise power in a selected bandwidth. This is the best method since it gives you a "strip chart" display of power vs. time. The measurement you need to make is the difference between the receiver audio or IF noise power with the LNBF pointing straight up into the sky and then down at the ground. You then adjust the penetration of the screw into the waveguide to give you the largest difference between the two readings. With a typical Avenger KSC321S-2 you will probably see something like a 4.5 to 5.5dB difference when the tuning screw is in the optimum position. Making a few reasonable assumptions, a cold sky to ground noise difference of 5dB at 10Ghz corresponds to a noise figure of about 1.4dB
Could any other tuning be done to improve performance?
Possibly. I'd assume the input stage of the LNBF is optimized for something like 11.7GHz because the LNBF is designed of Rx between 10.7 and 12.7GHz. Looking at the circuit, the only input matching appears to be done by the length and width of the microstrip that connects the waveguide probe to the gate of the 1st stage FET (often a NE3503M04). It might be possible to modify this for a better match to give a lower noise figure, but it would not be easy and I have not tried to do it. It would probably be quite easy to make things worse!
Something might be gained by removing the rear (horizontal) probe and replacing post behind the vertical probe with something that blocked the whole waveguide. Then moving that rear "wall" to the optimum distance from the vertical probe. Again I have not tried to do this. It's unlikely it would make much difference.
What about the other (horizontal) probe
As I mentioned there are two probes in the LNBF waveguide, one vertical (at the front) and one horizontal (at the back). You can switch between them by changing the voltage supplied to the LNBF. In the 12-14v range the front (vertical) probe is active. With 16-18v, the LNBF switches to a different first stage which is connected to the rear probe. That probe has to bend through a right angle before it enters the waveguide. It sits behind the vertical probe and also behind a vertical post that sits behind the front probe to act as a rear tuning stop for it. By experiment on several LNBFs, I found that the rear probe give less ground to sky noise than the front probe, even when both have been tuned to 10.368GHz. I assume this is because of the more difficult geometry (it has to bend through a right angle, and the fact that there's a bunch of "stuff" in front of it in the waveguide (even though that "stuff" is for vertical polarization). The bottom line is, for whatever reason, receive performance is better when using the front proble and physically rotating the LNBF through 90 degrees than by changing the voltage and switching probes.