An integral part of any satellite receiving system is the low-noise block downconverter (LNB), sometimes incorrectly called a “low-noise block”. As satellites use high frequencies of ~3-5Ghz for C-band and ~10-13Ghz for Ku-band transmissions and are very limited in transmission, sending these signals through a coaxial cable at the receiving end is impractical because of the high losses incurred in the coax at these frequencies.
As a result, the LNB was invented to overcome these issues. The LNB that sits at your dish is powered via the coax cable, and is actually a combination of several components:
- A waveguide (scalar ring) which “focuses” the receiving elements on the dish. This is separate in the case of C-band LNBs.
- A set of probes, which are basically the antennas which receives the signals focused and reflected by the dish, along with a blocking “bar” to separate the horizontal and vertical polarity signals (at least, for linear polarity LNBs).
- A set of amplifiers which takes the weak signals and boosts them up.
- A stripline filter which performs a band-pass function to select a band of frequencies to be converted.
- A dielectric resonator oscillator (DRO) to produce the local oscillator frequency.
- Another amplifier and mixer with the DRO which downconverts a section of the received RF into an intermediate frequency (IF) often in the L-band (950Mhz – 2050Mhz) which is easier to send by coaxial cable.
- A voltage regulator and integrated circuit to interpret the voltage and tone to select the appropriate local-oscillator and polarity probe.
- A “multi-switch” (or similar) in case of multiple-output LNBs.
LNBs are now considered rather “commodity” items, easily available for about AU$10-20, and in my experience, most of them function satisfactorily. The LNBs on the market generally have identical (or very similar) specifications for frequency stability, phase noise, gain, and noise figure. At times, these are sometimes “unrealistic” due to the attempts to “grab” your attention when choosing them. In my experience, most of the market LNBs will perform within about 10% of each other, and only extreme DX-ers will care about the differences between them.
Since I have a bit of an excess of 11300Mhz Ku-high band LNBs which are not that useful to me (since I prefer 10700Mhz “wideband” models, and might soon prefer dual-LO universal models in the future), I decided to teardown a pair of single output units.
ST Gold AP8-T2J 11300Mhz Single LNB
This particular LNB I have only used for a short period. This one was acquired for free from some second-hand satellite gear I collected, and it seems that ST Gold is a local importer-brand. In reality, much of the commodity satellite gear is made by only a handful of companies and is just rebranded. Sometimes this is obvious just by the external styling.
It seems like ST Gold didn’t do much of a job rebranding the LNB, keeping the original model name of AP8-T2J. This works out to be an LNB OEMed by MTI (Microelectronics Technology Inc). They have been in the market over 25 years designing and producing LNBs under many names. This one is probably one of their better performing, but older models, which you will notice due to its physical size.
Some satellite DXers have made a habit of “decapping” the cap from the front of the LNB to increase the sensitivity marginally. The front cap itself is very securely fitted, and is almost impossible to remove without cutting through the side and destroying the seal.
The cap itself features a greased silicone gasket, and two concentric channels which mate with the scalar ring itself, ensuring a waterproof seal. This is required since the circuitry is vulnerable to moisture and can creep in via the feedhorn.
The feedhorn can be seen as a “stepped” series of concentric rings, leading to a “pipe” segment. This is a “conical scalar ring”, and sets the focal length of the feed horn. Because this LNB features a standard 40mm diameter mount, it is designed for use with offset fed dishes with an f/d ratio of 0.6. This makes it very suboptimal if mounted on prime focus dishes which have an f/d ratio near 0.3, as it will not see most of the dish area!
Inside the pipe segment, we can see one of the probes for one of the polarizations. This is effectively the “antenna” that receives one polarity. Behind that is a printed circuit board segment in a T shape – with metal on the side facing forward which acts as a reflector/shield for the first polarity, and the other “arm” of the T at 90 degrees is the probe for the orthogonal polarity (not visible from this side).
Without the outer shell, we can see that the main body of the LNB is casted and ground down to size. The flat shell segment at the back houses the necessary electronics.
From the top, we can see some quality inspection markings made on the body, and a casting code of GPTYS-02A.
The electronics chamber is secured by several Torx screws. Certain ports are covered by silicone of different colours – the two near the F connector are normally used to trim the local oscillator frequency. The one in the middle is an additional screw that secures the case. The silicone helps waterproof the unit, as well as keep the adjustment screws in alignment.
Because of the sensitivity of RF to dielectric materials, the PCB used is made of a special material, and many of the areas are not covered by solder resist. From this side, the other probe is visible, as a printed trace. The bottom right quarter of the PCB are the inputs from each probe for each polarity and the first stage amplifier. The bottom middle houses the second stage amplifier and what appears to be a stripline band-pass filter. The left third of the PCB features spaces for two DROs, for universal LNBs, but since this features only a single LO, only the bottom DRO is populated. The middle of the PCB features the mixer and some switches, controlled by the IC in the top right. The other IC in the top left is the voltage regulator.
The lid itself seems to contain “ribs” to shield each section from stray RF. The adjustment screws can be seen to poke through the cover, and their effect on the oscillator is to “trim” their frequency. Some thermal pads can be seen, which takes away heat and conducts it to the casing. Due to the use of a rubber gasket around the flat portion of the lid, there is no need to silicone the outer screws.
So there we are – a tour of an older LNB. So how does this compare to a newer LNB?
BAS Satellite 11300Mhz Single Output LNB
This was the first Ku LNB I owned. I bought this unit from my “local” satellite shop, which talked me out of buying a 10700Mhz LNB on the account I was looking at some FTA satellites. I suppose it wasn’t exactly bad advice, but as a feed/satellite hunter, it soon fell out of favour with me and has been sitting unused after a year. It was good value though, being about AU$15, maybe because they were “getting rid” of stock.
The unit itself actually has no performance issues – after having bought many different units including a dual-output 11300Mhz unit (which I won’t tear down), I found the performance to be very good. Looking around, it seems this might be another MTI unit, as it shares some visual similarities with the present AP8-series LNBs despite being about 5 years old.
Seeing as I wasn’t likely going to use this much again, as I still have other spares, and I have a preference for at least dual outputs, I thought I might as well tear this one down as well.
Again, we start by butchering the front cap and removing it. A very similar waterproofing design is used, although the plastic material is different and shiny, rather than matte.
This particular unit has a slightly different scalar ring, with the inner flange “curved” and one less rib. The probes themselves are “discrete” wire probes, with a blocking “bar” across the middle. This is how the last “broken” universal LNB looked like (which I didn’t actually take pictures of).
The inside of the unit is particularly strange, as the connector itself is angled at a strange angle. The plastic shell actually serves to hold the body at a strange angle. There is silicone where the blocking bar was inserted, and the casting has the letters GP on this side.
Again, some form of quality control marking and grinding of the cast body is seen. This particular unit uses a lot of silicone for weatherproofing.
The cap of the casting has the letters TYS-4GP, which bears some similarities to the MTI LNB above, making me somewhat confident that this is an MTI unit. The two adjustment screws are seen under white silicone, with orange silicone for the cap cover and seam.
The protrusion of the adjustment cap can be seen into the body, with the cap having a “shaped” design to separate and shield internal circuitry sections.
The F connector output is on the middle left. Spaces for the two DROs can be seen in the left-top half, with only one DRO fitted due to the single-LO nature of this LNB. The probe entries and first stage amps can be seen in the right half of the PCB, with the stripline filter middle-bottom. The voltage regulator is much smaller, in the top right, and a different IC is fitted.
A bit of desoldering of the F-connector pin allowed the PCB to be freed and the rear to be exposed. Only a few traces are present, as well as the orthogonal probes. The rest is all shielding/ground plane.
Trimming Adjustment Limits
I resoldered the F-connector pin and reassembled the LNB, but with a difference – I cut off the back-shell so as to make the screws accessible while the LNB was in operation.
A bit of masking tape was used to make sure the cover still held tight, and the 40mm “throat” could be well secured by the LNB clamp of the offset dish. The lack of the front cap meant that my experiments must be undertaken in “dry” conditions only, if I’m not to kill the unit.
One thing I wanted to know is just how far can you move the local oscillator with the screw? I pointed it towards Optus D2 and took some spectras.
Left screw, removed
Left screw, screwed all the way in
The movement wasn’t that much. Using the beacons, we can quantify the difference between the two states.
Screwed all the way in
The range of adjustment is approximately 16.75Mhz.
LNB Local Oscillator Drift
It is often said that DRO LNBs are relatively unstable by narrowband transmission standards, and many of them specify a stability of +/- 1 Mhz at room temperature and +/- 2Mhz throughout the operating temperature range, with an absolute accuracy of about 1Mhz.
I spent some time adjusting the screw until it was “on target”, and then I monitored the beacon carrier frequency over the course of a few hours at a 1Mhz span (i.e. +/- 500khz). During the day, the LNB easily drifted more than 500khz, in a semi-unpredictable manner because of the interactions of sun, wind and rain.
The LNB performs a critical function in every satellite reception system. It seems that, over time, the march of technology with more quiet amplifier ICs and refinements in build have miniaturized the LNB circuitry and reduced the cost. For the price, it is a surprisingly sophisticated, well-engineered piece of RF gear despite the conceptual simplicity. Dealing with 10Ghz is not something easy!
The stability of the LNB, while suitable for wide-carrier usage, is not sufficient for narrow carriers. Despite manual adjustment, carrier drift is experienced due to changes in temperature, which exceeded 500khz either way during the course of a normal day. As a result, making accurate frequency measurements with DRO based LNBs is not possible unless a reference frequency is simultaneously measured.
I wonder how much better “low-cost” PLL-based LNBs would be – probably not much better, but there are of course much better, stable units at much higher costs intended for use with narrow satellite data carriers and professional usage.