[hpsdr] Mercury Spurious Emissions

[Steve Bunch]

Jeremy,

On Sep 10, 2011, at 11:26 PM, Jeremy McDermond wrote:

> ***** High Performance Software Defined Radio Discussion List *****
> 
> On Sep 10, 2011, at 4:54 PM, Joe Martin K5SO wrote:
> 
>> ***** High Performance Software Defined Radio Discussion List *****
>> 
>> 
>> On Sep 9, 2011, at 5:52 PM, Jeremy McDermond wrote:
>> 
>>> It seems as if Mercury is radiating spurious signals into the antenna connector.  The symptom is showing up on 144.390 MHz.
>> 
>> Using my IC910 with a whip antenna a few feet from my HPSDR enclosure (Antec) I too can see a signal at about S3-S4 strength showing up at 144.390 (or image thereof) when my three Mercs are running.  However, the signal is NOT being radiated from the Mercury antenna connector.  In fact, when I directly connect the input of the IC-910 to the Mercury antenna connector the signal appearing at 144.390 MHz (or it's image, of course) actually goes away.  The emissions you are seeing are coming from the bus or the boards and not being output through the antenna connector, it seems.
> 
> I'm not seeing the same thing on my system.  If I disconnect the antenna connector from the system, I do see the S3-S4 signal, which isn't such a big deal.  If I connect the antenna to the system it becomes at least an S9 signal, if not more.  Without the antenna connector it isn't noticeable on the mobile that's like 20 ft away.  With the antenna connector on, it's S9 on that mobile.  Interestingly if I merely touch the shields of the BNC, I see this behavior.  This seems to indicate that it's coming through the ground side connection.

The braid of your coax is just becoming an antenna for the spurious signals.  As Joe is showing, they aren't flowing through the center conductor.  Put a handful of ferrite beads over the coax at the connector and you'll lower that, perhaps by enough to help.  The root cause is that there are unbalanced RF currents flowing around your boards, power supply, Ethernet cable, etc., and you're just giving them a new way out.  That simply will not go away without boxing up everything and shielding it.

Here's a high-level view of what's happening.  (If some EE wants to jump in and correct or expand on this, please do!)

Working backwards from the spacing of the spurs, let's just accept for the moment the assumption that the 3.072MHz bus clock is the actual source.  So the current flowing out of the bus driver that's driving that 3.072MHz signal flows through conductors on the driving PC board, through the connector, to the Atlas bus, to each of the PCB's and receivers on the bus line as well as the open ends, and back through ground lines on the boards and bus.  Since the bus lines are unterminated and don't match the driver characteristic impedance (bus drivers rarely do except in controlled-impedance situations), the signal also reflects off the ends of the bus, returns to the driver, which typically damps it and re-reflects a smaller version.  So in addition to the 3.072MHz signal and its odd harmonics that you already had flowing over this complex path, you also typically get a spike (a crude approximation of a delta function) and a damped train of reflections -- all those represent harmonic-rich energy flowing through your bus line and ground returns.

The current path for the signals and returning current is not a nice compact path like it would be if the driver signal and return were going over coax or twisted pair.  So the electromagnetic fields around the conductors carrying outbound and inbound current flows aren't equal-and-opposite and therefore don't cancel.  So it radiates.  All the conductors carrying unbalanced currents are radiating, and there's seldom a single thing you can change to fix it.  The originating PC board, the bus, and the receiving boards all contribute to the radiation.  When you plug something like an Ethernet cable or antenna onto this, some of the unbalanced current will try to flow out through it and become part of the radiator if you don't stop it (with chokes, bypassing it to ground, etc.).  You have to either eliminate the unbalanced currents or shield the assembly (a good RF box is basically a great big short-circuit for any loose currents to mostly keep them confined to flowing inside the box, combined with a Faraday cage to keep radiated energy from escaping).  That's pretty much it.

The bottom line is that what you're seeing is not broken, or atypical, it just works in a way that's inconvenient to your usage.  The best fix is to box up the Atlas and all the boards, and follow good RF practices for everything going in and out of the box.  Which is unfortunate for your demo purposes.

Steve, K9SRB

>> That such VHF signals are present in the immediate vicinity of the HPSDR rig is not very surprising of course, no more so than similar type signals emanating from any bus-driven computer system, since the bus has relatively large amplitude square wave signals on it.   I was concerned that something was being radiated via the Mercury antenna connection, as was initially suggested...or at least I that's what I thought that's what the suggestion was.  Well, it isn't, in any case.
> 
> I'm not necessarily suggesting that this is a design issue or a general issue with Mercury in general.  I do know that the behavior of mine differs when I have the antenna ground connected.  I'm just trying to figure out whether I have a problem with my particular Mercury.  Of note is that my portable antenna is only around 10 feet away from the radio and there are no walls between the radio and the antenna.  So, proximity could be an issue, but it certainly seems for me that the antenna makes a difference.
> 
> I made a quick and dirty video showing the issue:
> 
> http://www.youtube.com/watch?v=BxJJFe9PjFc
> 
>> In this case it seems that standard, or perhaps elaborate, rf shielding techniques should be able to reduce the bus or board emissions to an acceptable level in the vicinity of the HPSDR rig, as has been already suggested.  Anyway, no wonder I didn't see much with the spectrum analyzer except for the 122.88MHz clock and it's harmonics when I connected directly to the BNC connector on Mercury earlier.  The problem you see isn't coming from there, clearly.
>> 
>> For your demos, in which folks may have handhelds right up next tot the HPSDR rig you may need to simply have a super shielded setup and not run with the covers off, but that hardly is conducive to showing people what the rig looks like.  At least I think you by now have a good idea what the issue actually is, right?
>> 
>> 73,  Joe K5SO
> 
> --
> Jeremy McDermond (NH6Z)
> Xenotropic Systems
> mcdermj at xenotropic.com
> 
> 
> 
> _______________________________________________
> HPSDR Discussion List
> To post msg: hpsdr at openhpsdr.org
> Subscription help: http://lists.openhpsdr.org/listinfo.cgi/hpsdr-openhpsdr.org
> HPSDR web page: http://openhpsdr.org
> Archives: http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/


 1315747944.0