Hilbert,

I've built EMI/RFI shielded solutions and measured their effectiveness.  Further, I have done this area where MOD, GCHQ, MI5, MI6 and even the CIA are known to be highly active.  So, I've seen a **lot** of classified technologies over the years and had first hand opportunity to put my shielding solutions up against them.

Your knowledge of EMI/RFI shielded enclosures is **extremely basic** and certainly not drawn from practical experience in constructing such structures.  It is so bad, it is almost child-like.  In theory, your description is correct, in practice it does not quite work out that way.  You reply is littered with so many basic errors, I don't even know where to begin.

I have observed jammers penetrate TEMPEST grade enclosures in HF, VHF and UHF.  In fact, I got to see a local variant of this last September and even behind 80dB of double-wall shielding, it wiped out my SDR.  On the web, there are scientific papers on 2.4GHz transmitters that pass straight through double-walled Faraday cages and disable electronics.

As for cables, any cabling penetrating a cage or placed close to an aperture will become an antenna.  Try putting a small copper wire through an aperture of Faraday cage and watch on a spectrum analyser all signals within the bandwidth of the cable jump to full strength.  It does not matter if this cable carries DC, USB data, etc.

You are correct in Independence, but modern jammers are designed to perform impedance matching using arrays.  Impedance mismatches are great for things like EMP, but useless against any modern tech.

Anyway, thanks for your input but I would appreciate it if you just did not bother.  I don't see what you adding to this discussion other than partially correct information.  Have you considered a career with the government?  disinfo agent or something similar??  I think you would be great. 

It has already been decided that I will design an enclosure from scratch, so the discussion is pretty much finished anyway.

Regards,

Mark McCarron

Date: Mon, 20 May 2013 18:09:25 -0400

Subject: Re: [USRP-users] USRP EMI Solution
From: hilbert3141-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org
To: mark.mccarron <at> live.co.uk

CC: usrp-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org

Oh, Marky Marky, Marky.   Your fuzzy-headed lunacy is sooooooo very cute.  Can I have your babies?

A grounded steel enclosure, even one with "holes in it" *does* offer quite a bit of protection from stray RF fields
  entering the box at points you'd rather it didn't.  Any "holes" in the box will leak proportional to the wavelengths

  of the impinging RF fields.  The box basically acts as a wrong-impedance reflector.  Certainly for the kinds of

  isolation levels you're talking about (120dB?  really, you think you need 120dB?) you need to pay attention to

  seams, and the like.  But calling a random metal box "transparent" to RF indicates a profound misunderstanding

  of how RF works, and impedance match, and, well, everything, really.

Consider an example of a "sheet of metal with holes in it" that many of us RF geeks will find quite regularly in our

  perambulations about the relevant technical territories.  Consider, for example, a parabolic dish reflector that isn't

  solid--a "mesh", if you will.   If RF fields acted the way you claim, then the ones heading towards a satellite dish

  must, at the last femtosecond, decide whether they're engaging with a satellite dish, and do the reflection thing, or

  engaging with a "tissue paper" metal enclosure, such as you've previously described.  They are, I guess vastly

  smarter than we have previously given them credit for.  I am humbled.

OK, now on to things other than the box.   DC power conditioning is certainly something to pay attention to. But
  most RF gear already has a fair emount of EM decoupling in the power-supply input circuits, and unless I'm

  mistaken, the Ettus power cable already has one or two ferrites on it, which act as common-mode chokes.

  If you don't know what a common-mode choke is, look it up.   But augmenting that certainly wouldn't be
  a crazy thing to do, if you suspect that your [put your tinfoil hat on, everyone] *JAMMERS* are getting in through
  your power cord.   Adding more decoupling can't really hurt here.  I think the existing decoupling that comes
  with typical RF gear (not just Ettus gear) in general, is reasonably good.

Now, hmmm, on to your USB and RF ports.  I'm not sure how you're going to isolate your RF ports from incoming

  radio waves that you don't like, because, well, they're kinda designed to, you know, pass radio waves (actually,

  the electric-field component of radio waves) across your enclosure envelope.  I guess if you were only interested
  in a very narrow band of frequencies, you could put a geezly stiff filter in front of your RX port, and on your TX
  port, again, geezly stiff and narrowband.  Check with Sinclair and Andrew. They'll sell you geezly-narrow/stiff

  filters for pretty much any frequency you want, as long as you have $$$$ to spare.

Now, there's the data ports.  USB, in your case, but I'll generalize to 1GiGe as well.    With USB, the cable could,
  clearly, bring in unwanted EMI on the DC path -- again, the power-supply stuff we talked about previously would

  be equally-well applied here.  And again, I think the Ettus gear includes one or two ferrites to act as common-mode
  chokes.  But, well, the data bits.  You can't really filter them, because if you did, they'd stop being bits, and

  be more like greasy smears left on the road by a horrific motorcycle accident.  USB-2.0 uses, as far as I know,

  baseband, differential, signalling -- so there's a strong component at 480Mhz and every odd harmonic of it.

  If you put, let's say, series inductors and shunt capacitors across your D+/D- signals in the USB circuit, it will

  stop being useful for USB entirely.  Now, there *are* USB EMI filters out there, but if you look at the specs, they

  can't be used for USB-2.0 rates (480Mbit/sec) because they just smear the bits all to holy flying heck.

Now, let's say you were using a 1GiGe port for your data bits that need to cross the "event horizon" of your

  oh-so-cool 120dB isolated enclosure.  Here, the story is a *little* better, but not much.  1GiGe uses a symbol

  rate of 125Msymbol/second, with 5-level modulation.  The bandwidth goes from DC up to about 250MHz or so.

  So, theoretically, you could put some 250Mhz low-pass filters in the ethernet signal lines, and then you'd

  only have to worry about your megawatt jammers below 250Mhz -- since you must necessarily pass signals
  across the ethernet connection between DC and 250Mhz (approximately).

Or, you could, you know, turn your radio into a kind of Shroedingers Radio, inside a shiny copper enclosure with

  only a power cable going into it.   You don't get to find out if it's doing anything useful until you open the box.

  But boy, oh, boy, you'd have some really great isolation going on there, buddy.

On Mon, May 20, 2013 at 5:38 PM, Mark McCarron <mark.mccarron-SrPyuzuu+o61Qrn1Bg8BZw@public.gmane.org> wrote:

I don't know what you mean by that statement.  Care to clarify? 

Digital domain filters, such as those in GNURadio, do not protect circuitry.  They manipulate a representation of a signal, not the signal itself.  These are two quite different things.

The design should take a couple of hours if I get the right components.  I have seen filters for the antenna and I'm sure a filter for the DC cable is pretty standard too.  Although, I will need to check the ratings on those filters, they need to be sharp and in excess of 100dB.  I'm still sourcing a USB filter.

Regards,

Mark McCarron

Subject: Re: [USRP-users] USRP EMI Solution
From: patrik-ujVHncHi1eiDV3RqQuXYcQ@public.gmane.org
To: mark.mccarron-SrPyuzuu+o61Qrn1Bg8BZw@public.gmane.org
CC: usrp-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org

Date: Tue, 21 May 2013 00:29:27 +0300

Aren't the purpose of a jammer really designed to to jam?

>In my scenario, I just need to wipe out EMI/RFI interference.  So, a sealed box with filtered inputs is required.  I have surge protection on the cables already, as well as RFI
>conditioning/filters on the mains.  I was hoping someone had designed one already, but it looks like I will need to design it from scratch.  It should not be too difficult, or even
>expensive.
Good *luck* (there is no such thing as luck in RF, you know...)
Did you know that in GRC there is High-/Low-pass filters?

>I'll look into designing this over the next few weeks.
I'm sure that this will keep you busy several weeks...

Patrik

-----Original Message-----
From: Mark McCarron <mark.mccarron-SrPyuzuu+o61Qrn1Bg8BZw@public.gmane.org>
To: usrp-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org <usrp-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org>
Subject: Re: [USRP-users] USRP EMI Solution
Date: Mon, 20 May 2013 21:46:59 +0100

Perhaps some background is required.  I'm in Northern Ireland and military jammers exist everywhere here.  If you know anything about the history, you would know that we have had over 30 years of terrorism/war.  Despite the peace process, certain groups still exist, bombings and shootings still occur regularly.  But not on the same scale as 20-30 years ago.  The jammers are used to prevent the remote triggering of bombs, although most bombs are timed or use command wires to compensate.  In the past, British military foot patrols would have portable jammers that would wipe out TV and Radio.  Basically, HF, VHF and UHF.  At times, this still occurs and tests of jammers around police stations are common place.  Notification of the latter is usually provided.  To be anyway useful, the peak power on the jammers must be quite strong and given their portable nature, they can at times be quite close.

Now, I've seen quite a few on the list make the mistake that a metal box is some form of EMI/RFI solution.  It is not.  Even a bare metal box has highly limited shielding capacity.  The engineering tolerances, even in laser cut metal, are just not high enough to provide much in the way of shielding.  Its effectively transparent.  The issue is edges, corners and seams and cables entering the box.  Any external input must be filtered.  DC, USB and antennas need either low pass or bandpass filters and this needs to be properly integrated into the box with surge protection.  The box itself requires high pressure gaskets suitable for screwed applications.  In a secure environment, this box would reside within a type of DMZ and only the data cable would pass through to the inside of a secure room, through yet more filters.

In my scenario, I just need to wipe out EMI/RFI interference.  So, a sealed box with filtered inputs is required.  I have surge protection on the cables already, as well as RFI conditioning/filters on the mains.  I was hoping someone had designed one already, but it looks like I will need to design it from scratch.  It should not be too difficult, or even expensive.

I'll look into designing this over the next few weeks.


Regards,

Mark McCarron


Subject: Re: [USRP-users] USRP EMI Solution
From: patrik <at> poes-weather.com
To: mark.mccarron-SrPyuzuu+o61Qrn1Bg8BZw@public.gmane.org
CC: usrp-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org
Date: Mon, 20 May 2013 22:19:24 +0300

>The pulse-like nature of the jammer could cause physical damage.
It could, but I suspect that no govt allow that strong signal

>I'm looking to build an EMI/RFI seal box with about 100-120dB attenuation from 10KHz-40GHz
Usually ~1/4 wavelengt is enough to stop/catch a signal (think of a mesh dish at Ku-band). Ain't your Ettus box enough?

>I just need a good solutions for the antenna, power and USB.
Don't forget that your coax to the antenna can also *catch* signals
Have you tried using a 6V motorcycle battery for the receiver and laptop (unplugged from all AC (charger))?

Adding ferrite beads (choke) on your wires is the most cheapest/easiest solution (can be recycled/found on most electronics).
The fact is, if someone is jamming you're band it is for a purpose, it is for a govt use (at war) or RF band already in use (bought) like TV, cellphone, etc.
If someone is jamming you, you should report it or call them why....

Patrik

-----Original Message-----
From: Mark McCarron <mark.mccarron-SrPyuzuu+o61Qrn1Bg8BZw@public.gmane.org>
To: usrp-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org <usrp-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org>
Subject: Re: [USRP-users] USRP EMI Solution
Date: Mon, 20 May 2013 18:37:04 +0100

Although much less frequent than it used to be, the area I am in can experience occasional jamming.  Whilst this may have no real effect on a wideband system, it could be picked up by the internal circuitry.  The pulse-like nature of the jammer could cause physical damage.

I'm looking to build an EMI/RFI seal box with about 100-120dB attenuation from 10KHz-40GHz.  So, some sort of galvanised steel with EMI gaskets should work.  I just need a good solutions for the antenna, power and USB.

Regards,

Mark McCarron



From: columbo_the_legend-PkbjNfxxIARBDgjK7y7TUQ@public.gmane.org
To: usrp-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org
Date: Mon, 20 May 2013 12:33:11 +0000
Subject: Re: [USRP-users] USRP EMI Solution

What exactly are you looking for in terms of interference rejection/suppression? The boxing as it stands provides pretty much what most people care about in terms of EMI in that it reduces the probability of spurious emission/coupling. If you're looking for double figure suppression on some specific band that is leaking somewhere other than the antenna port, it likely isn't an issue generic enough that someone has already developed an isolation solution. 


From: mark.mccarron-SrPyuzuu+o61Qrn1Bg8BZw@public.gmane.org
To: usrp-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org
Date: Sun, 19 May 2013 01:11:45 +0100
Subject: Re: [USRP-users] USRP EMI Solution

The requirements are different, but how much EMI/RFI shielding do you think the box provides?

I can tell you now it would be around 1dB, if even that.  It really is as effective as tissue paper.  It may couple some near-field effects but not much.

Regards,

Mark McCarron



Date: Sat, 18 May 2013 19:24:04 -0400
Subject: Re: [USRP-users] USRP EMI Solution
From: hilbert3141 <at> gmail.com
To: mark.mccarron-SrPyuzuu+o61Qrn1Bg8BZw@public.gmane.org

The requirements of TEMPEST and the requirements of ordinary commercial EMC are very, very different.


Because a box doesn't meet TEMPEST says essentially *zero* about its suitability for ordinary commercial
  applications in the EMC realm.


I've been inside the photon-locked shielded room at NRAO, and other similar rooms.  Ordinary commercial gear

  *does not* need to meet those requirements.  At all.  Ever.


Open up a standard commercial-grade or ham-radio grade receiver/transceiver sometime.  You'll find that there's
  no extensive RF gasketting (if any at all).  Just a metal box that is bonded, in several places to the system
  ground, and the RF connector grounds are common with the cabinet.  They aren't "tissue paper", but they aren't
  TEMPEST-qualified either.



I've been involved in radio since 1986.  Installed a crap-ton of RF equipment, done system designs, etc, etc.

  For ordinary commercial and amateur-radio applications.  NONE of it was TEMPEST-qualified, because, well,

  for ordinary commercial purposes, it didn't need to be.







On Sat, May 18, 2013 at 6:38 PM, Mark McCarron <mark.mccarron-SrPyuzuu+o61Qrn1Bg8BZw@public.gmane.org> wrote:

Hilbert,

I've designed EMI/RFI solutions and TEMPEST secure rooms to protect computing, comms and humans.  That box may as well be tissue paper.

Grounded means nothing other than general electrical safety (even then there are caveats to this).  An EMI/RFI shield extends a ground reference around an object.  That's not as easy as it sounds.

Regards,

Mark McCarron



Date: Sat, 18 May 2013 18:24:30 -0400
Subject: Re: [USRP-users] USRP EMI Solution
From: hilbert3141-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org
To: mark.mccarron-SrPyuzuu+o61Qrn1Bg8BZw@public.gmane.org


The metal is *grounded*.  It doesn't matter whether it's painted, or not.  The paint issue is along the seams in

  any type of cabinet.  Ideally, you want all the seams to be one continuous ground, and if you have a cabinet

  that fits-together mechanically, that's where you'll typically find RF gasket material between two non-painted
  surfaces.  But the paint, in and of itself, is irrelevant.


Certainly, if you're going to be operating any kind of RF device like this in a *very high* EMI environment, you might

  want to build yourself a faraday cage with welded seams, and improved RF choking on all non-RF
  ingress/egress points.

The existing cabinetry is more than adequate for most applications.  Hardly "transparent".


For RX, you only have to get the "RF leakage through the cabinet" to significantly below "RF I would have received

  from the antenna in the same EM field".


And the secondary  emissions from the device itself are required to pass various tests, as far as I know,
  like FCC and CE testing.  Again, in order to pass those tests, the cabinet can't possibly be "transparent".

  That's just absurd.





On Sat, May 18, 2013 at 5:58 PM, Mark McCarron <mark.mccarron-SrPyuzuu+o61Qrn1Bg8BZw@public.gmane.org> wrote:

That's not an EMI shield.  The metal is painted and would be effectively transparent to EMI/RFI.  What is required is a continuously conducting surface with proper sealed inputs and EMI gaskets.

Its incredibly hard to properly seal anything from EMI/RFI.

Regards,

Mark McCarron



Date: Sat, 18 May 2013 17:42:09 -0400
From: hilbert3141-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org
To: usrp-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org
Subject: [USRP-users] USRP EMI Solution


Mark writes:

>Has anyone considered constructing an EMI shield box for the USRP?
>
>Regards,
>
>Mark McCarron



You mean, other than the one that comes with it?  The box is metallic, grounded, and most of the perforations in it are a tiny fraction of a wavelength up to a few GHz.  If I were

operating in high RF fields above about 3GHz, I might replace the fan grille with
some steel mesh.



The USB cable Ettus supplies includes a common-mode choke, as does the power cable.


For lower frequencies (below 50Mhz) adding more common-mode chokes to those cables will likely help, and they're available on eBay reasonably cheaply.


But, as far as I know, the existing steel cabinet was engineered to provide EMI isolation, and given its material and construction, it can't help but do so, due to physics...



--
Hilbert (Godamn) Transform

hilbert3141-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

Purveyor of fine Hilbert (Godamn) Transforms since 2013



_______________________________________________ USRP-users mailing list USRP-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

_______________________________________________
USRP-users mailing list
USRP-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

--
Hilbert (Godamn) Transform

hilbert3141 <at> gmail.com

Purveyor of fine Hilbert (Godamn) Transforms since 2013


_______________________________________________
USRP-users mailing list
USRP-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

_______________________________________________ USRP-users mailing list USRP-users <at> lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

_______________________________________________
USRP-users mailing list
USRP-users <at> lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

--

Hilbert (Godamn) Transform

hilbert3141-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

Purveyor of fine Hilbert (Godamn) Transforms since 2013

_______________________________________________
USRP-users mailing list
USRP-users-p6fHTpcPDZaz3Dx2OeFgIA@public.gmane.org
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com