Be careful when hooking up the BSC621 C/Ku LNB!!!

[N9WOS]

When hooking or unhooking or doing maintenance on the coax feeding the LNB, be very careful!

The circuitry in the C band LNB that switches between the C/Ku LNB’s is a bit anemic. It can’t stand reverse polarity/current flow!!!!

Reverse current flow can be applied to it by stray AC and DC voltages that exist between the house wiring and the ground that the dish is placed on.

Stray voltages can be applied to the LNB RCVR port (power input side) when you disconnect and connect the coax to the LNB system. On most coax assemblies, the center conductor usually extends past the barrel of the connector. The center conductor will make contact before the ground conductor will. So standing AC/DC voltages will be applied to the center conductor for a momentary time.

This probably applies to the BSC621-2 DiSEqC 1.2 LNB also, if it uses the same LNB switching system components.

The switch in the LNB uses a combination of signal switching diodes and a couple of transistors to switch power between the two LNB’s It cuts the power to the one it’s not using to prevent bleed through because it doesn’t use a two stage diode switch. So the only way that current gets through to the LNB’s is through active components. The problem is a result of the fact that they don’t have any reverse polarity protection diodes on the power input (RECV side) of the switch. If the input goes negative there is nothing to shunt the current to ground on the power input side of the switch.

Each of the LNB’s themselves have a voltage regulator running off the input to each LNB. It is standard design practice for regulators to have an internal reverse voltage protection diode on the input to ground, output to ground, and between the input and output. As both the regulators in the LNB’s have. So there is reverse voltage protection diodes in the LNB’s themselves (internal to the voltage regulators) So, when the input voltage goes negative, the only current sink is the input to the LNB’s. (diodes in the regulators)

Now you should notice the problem by now. When the input goes negative, it is shunted by the voltage regulators in each of the LNB’s in the assembly. But the regulators are AFTER the switch. The internal switch components are not designed to handle the reverse current flow. It lets the magic smoke out of the transistors that turn on and off the power to each LNB.

Basically. There is a place for the reverse current to go, but it must go through the switch to get there. And the switch can’t take it.

The surface mount transistors can handle a good amount of current in the forward direction. That is when the are switched on and there is zero volts drop across them. But in reverse, the circuitry biases them in the on direction just enough to let enough voltage and current stand across the transistors that it overheats them.

A simple diode connected to the DC takeoff inductor on the power input side (rcvr side) would solve that problem.

The only other danger threatening the LNB is positive over voltage. Technically, the maximum positive input that most regulators are designed to handle is +30V. But normal voltage regulators are relatively tolerant to momentary over voltage, so in the real world it isn’t much of a problem when hooking and unhooking the coax. But still, if they wanted to get fancy, they could put an over voltage crowbar on the input by using a SCR and a 30V zener diode. That would protect it from almost all stray voltage conditions, excluding catastrophic incidents.

But I, myself, would be happy if the head people on this forum would forward it to the LNB manufacturer that it would be a good idea to put a reverse voltage protection diode on the power input side (RCVR side) of the switch.

You may ask how I figured this out. I had bought two C/KU FTA systems for a couple C band dishes that was already in service. One for me, and one for a friend. With each system, I got the LNB of the type listed. One started acting funny after I had been doing some coax work. I could tell it was the internal switch. I could get C band, and KU band when hooked directly from the KU LNB. But I couldn’t get anything through the switch on the KU side. I verified it by moving the LNB to the other system.

Being that it was just after the saudon return period. One week after to be exact. (It always works that way) And I didn’t want to take the time shipping it out to the manufacture. And considering my extensive electronic repair experience. (And my never ending curiosity!!!) I opened it up and fixed it myself.

It took out the transistor feeding power to the KU side in the switch. The transistor feeding the C band side was bloated but still functioning. I replaced the surface mount transistors and added the reverse protection diode into the system to stop it from happening again.

The reason I am posting this now, is because the second LNB I had bought had the same thing happen to it at my friends place. A lose coax connection took out both transistors totally. I just finished repairing it, and doing the diode modification.

The absence of the diode levees this LNB particularly susceptible to stray voltage. So if you have one I would suggest that you have a grounding wire between the receiver and the LNB before you hook it up. That would shunt any stray current and voltages between the house wiring and the dish mount.

Considering that the sales website has a warning about hooking up external DiSEqC switches with the power on. (or to that effect) If the same manufacture that is making the switches is also the ones making the LNB, then it may be highly likely that it is also the reason for the switches dieing when hooking them up.

So Please ask the manufacture to think about putting a diode on the inputs and output of their switch products (this LNB in particular) to protect against stray reverse voltage..

And on external switches, have diodes between the inputs and outputs. That way, if the LNB is at a higher voltage than the RCVR side, then it don’t’ reverse the power switching element (transistor.) That can happen when you are hooking up an ungrounded LNB’s to the switch.

That is why linear regulator designers put them in all the 78** and 79** series regulators. Just for those reasons.

It is a moderate design oversight in this product, and it needs to be corrected badly!!!!!!!!!!!!

[rosey1]

You could also turn off the power before doing work on your system.

[N9WOS]

The power to the receiver was unplugged.

But the receiver is grounded by other wires to the rest of the house wiring.

There was about a 14V AC between the equipment ground and the distant satellite dish ground. The run between the house and the dish is about 200 Feet.

When it’s dry, only about 10mA flows. But when it’s wet, it can get up to 100+mA between the house and the dish.

The main line that goes by the house is single phase, so there is always a voltage drop on the HV neutral that varies depending on load on the line. That voltage is imposed on all the houses that are grounded to that line. So it is normal for houses around here to have a small AC voltage between the house ground and a pole driven into the ground at a distance.

[Sadoun]

Nice 1st post. Welcome to the forums. Your suggestion will be forwarded to the factory. Thanks

[BYX]

Very nice well written report.
I'm personally delighted with all the information provided, therefore I give you a thumbs up for it.
With all due respect I would like also to point out that in the FTA hobby no one is an "extra class" FTA licensee per say, because there is no such thing.
In most of the cases I dare to say that if to a FTA hobbyist a new LNB fails he or she will return it to the store and if it fails after some time of use you just trash it and buy a new one.
Again I encourage you to continue with the same devotion of HAM, but FTA is a different hobby.
FTA is more of an entry level hobby that requires a more basic set of skills for the hobbyist.
Also with all due respect to other members of this forum, some don't even know that with their FTA systems and dishes on C or Ku bands they are dealing with microwave technology.
Keep up the good work and welcome to the FTA hobby.

[tmostad]

Thanks for the informative report.

I bought a 621-2 and the Ku LNB was dead either on arrival or after installation (I of course don't know which). Can you imagine a case where the switch dying could kill the Ku band LNB?

Your description of the ground loop current to me suggests that a simple grounding block for lightning protection is in some respects a bad idea. It seems there could be a better method for preventing the coax from conducting a lightning strike into the home without a low impedance ground connection.

[N9WOS]

Remember that the 621-2 has a universal KU LNB so it has a 22Khz switch in the KU LNB in addition to the DiSEqC switch that is in the C band section.

If the 22Khz switch uses the same switching system as the switch in the c band side, then it will have the same weaknesses.

I have never had the opportunity to take apart a 621-2 so I am assuming that the universal LNB uses two distinct RF amp paths, with two distinct LO’s and two distinct mixers, with a 22Khz switch that selects between the two outputs.

It may use the 22khz two switch between the two local oscillators feeding the same mixer, and RF path. If that is the case, then I don’t really have any idea of the possible failure modes.

[RFenergy]

N9WOS,

Welcome. Technical stuff is welcome too. I was W(zero)GMT until
I let it expire. This FTA is potentially a technical hobby (it uses a
lot of electronics). The comment from BYX about the various levels
of electronics experience/knowledge that meet here is valid. Maybe
a techie forum would work....I would like that. SMT don't scare me.
And, I mean hardware/firmware....not the unspeakable area, (crypto)

I just bought a 621-2 for my 10' Fiber Glass Channel Master dish.
Could you take a couple of close up photos of the pcb to show the
mod. These designs were done by mortals and thus are not perfect.
the pressure from marketing to get it done yesterday and cheaper
does affect the quality. Ground loops are problems that occur at
all levels of scale, from boards to buildings. In reality ground is never
at ground potential.

I got my first system from Sadoun last summer. The 90cm Ku dish
and rotor is functioning very well. There was a bit of a learning curve but
thats fun. I do a lot of work with Microchip's PIC series of micros and
plan to build the diseqc reference design (which uses a PIC) to break
out the ASCII. I want to know what the Mercury2 is commanding
(that unit is a little buggy I believe). My real blue sky project is a PC
based spectrum analyser for the LNB output band 950-2150Mhz. Before
that a simple agc on the cheap sat finder (another "challenged" design).

Was it just a series blocking diode at the center conductor entry?

RFe (one of the four fundimental forces in nature)

[N9WOS]

Hmm… all right…….

I have pulled the LNB and I will exercise the macro function on my camera.

I will need to drag up a schematic program to make a schematic of the section in question so it will be easier to understand the problem.
Stay tuned……

[N9WOS]

I just decided to use MS paint to illustrate so here is the pics.

LNB is an aerial view of the LNB circuit board.

LNB2 is a little graphical representation of the circuit layout in the switch area.

LNB3 is highlighting the changes I made.

I added the diode and replaced one of the transistor switches. The other transistor switch that powers the C band LNB was still good so I left it in place. The one that switches the KU LNB was bad so I replaced it with a TO92 style transistor that has a little more tolerance to abuse.

The transistor switches just switch handle the power to each LNB. They are isolated from the RF patch via radio frequency chokes. So short lead length is not a high priority. I left the leads a bit long so it would be easy to solder them in place.

The signal switching is handled by two RF switching diodes that are forward biased when it’s particular LNB is powered up.

You want to put the reverse protection diode into the circuit after the RF choke (not directly to the input terminal) so that it is isolated from the RF path, so the diode’s capacitance doesn’t shut the signal to ground.

On the LNB 2 picture I gave a graphical representation in of the reverse protection diode in the 7808 in the c band LNB so that it would be clear where the reverse current is shunted to in when the input swings negative. The KU LNB also has a regulator which has a reverse protection diode too. So there is two protection diodes but both are after the switch.

The reverse protection diode I put in basically mirrors the one in the 7808 voltage regulator. But it’s before the switch so the reverse current doesn’t have to go through the switch to get to it.

[N9WOS]

LNB TR shows a side view of the replacement transistor.

LNB diode shows a side view of where the diode goes.

LNB switch is a schematic of the transistor switch area in particular. It shows why the switch fails.

A transistor usually has a relatively high collector to emitter voltage tolerance. Called the breakdown voltage or (BV ceo) And they usually have a relatively high base to collector breakdown voltage called (BV cbo). That is the to connection which normal have a large voltage across them. But the emitter to base breakdown voltage called (BV ebo) is usually 2 to 6 volts for most transistors.

Under normal conditions the (in black) the emitter to base voltage is around 0.7V But if you have reverse polarity coming in. The voltage across the E to B junction goes up to the incoming voltage that is because the collector is shorted to ground by the reverse protection diode in the voltage regulator of the particular LNB. If the incoming voltage is more than -2 to -6 volts, then the transistor is going to break down and start conducting in reverse. Even with 250 or so milliamp of stray current and the switching transistors will be on their way to toasty-vill. And when the magic smoke comes out, you can’t put it back in.

It only takes 10 to 20volts of stray AC voltage to do the damage. Low enough that you would not feel it if you were holding the lnb/dish in one hand and the coax end in the other. But when you are connecting the coax (or when a lose connection forms during normal service), the transistors will go poof.

The components and regulators can stand large forward spikes (30 to 60V) without any problem. Stray forward polarity tolerance is high enough that running across anything higher in the filed is unlikely. The problem comes when the input goes negative. 10 to 20 volts is well within the normal stray voltages you would find in the field environment. That is why they need to put a diode in the input of the switch to sink the any stray negative current surges, and spikes that comes with lose connection and making or breaking coax connections during assembly of the system.

I can tell that they anticipated that it was a possible problem because they put a diode in the base circuit of the transistor switches. That prevents the bias resistors from conducting reverse current and turning the transistors on when biased in reverse, but they overlooked a critical problem. The emitter to base junction can not stand that high of voltages. So the bias resistors don’t need to conduct base current to cause the transistor to start conducting. It will do it on it’s own when the E to B voltage gets above 2 to 6 V.

That is why I have a big arrow pointing to the diode with the statement ”Diode won’t save it”
So to the design team. Nice try, but no cigar.

I have two LNB’s that I have repaired from the same damage. One took out one transistor (the one show) and the other took out both transistors. After adding the diode in both units I have had no more problems.