Attenuating above 100MHz on squarewave (SPDIF digital audio)

Um, won't a few ferrite beads cure everything?

--
Many thanks,

Don Lancaster
Synergetics   3860 West First Street  Box 809  Thatcher, AZ 85552
voice: (928)428-4073 email: don@tinaja.com   

Please visit my GURU's LAIR web site at http://www.tinaja.com

That is the usual case except when cable other than coax or twisted pair is used.

That would depend on where the common-mode signal arose. It is easy for slight deviations from your 'tank' and 'pipe' concept (with which I basically agree) to convert high frequency normal mode signals into common mode signals. An LPF on the source will bring those way down.

--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.

Hi,

We're producing a consumer item that outputs a SPDIF (digital audio) signal. The squarewave produced is 0.5V pk-pk with a data rate of 6.144Mbits/sec. Our previous design failed EMC (Europe) radiated emissions testing on the SPDIF output line, at frequencies around 100MHz and higher. To pass, the testing house had to clip a ferrite around the lead.

For the new design I want to put a basic filter in the circuitry before the SPDIF output socket in order to attenuate those 100MHz and higher signals enough so that it passes the radiated emissions tests. However, we don't want to reduce the slew-rate too much on the square-wave. I think as long as the filter does not attenuate frequencies of 50MHz or lower that would be fine for us.

Could someone please suggest what filtering I could use. I'm thinking of an inline ferrite or an RC or LR low pass filter maybe? I guess this kind of thing is done all the time for SPDIF, so is there a standard recommended filter circuit?

We'll take it along to the EMC testing house and be able to swap values there if needed, but I'd rather get it right, or nearly right, before we get there TBH.

Many thanks, James

I read in sci.electronics.design that James wrote (in ) about 'Attenuating above 100MHz on squarewave (SPDIF digital audio)', on Tue, 28 Dec 2004:

So the emission was **common-mode**: the ferrite (if it's the right sort) absorbs the energy travelling on the cable shield.

Filtering the signal in unbalanced or differential mode won't, or probably won't, affect the common-mode emission.

A ferrite ring or a common-mode choke next to the SPDIF output connector, possibly in a shielding box, would be my first try.

>

--
Regards, John Woodgate, OOO - Own Opinions Only. 
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

....

Yes. (Or even an LC filter as I suggested.) The advice about getting the cable (pipe) and enclosure (tank) well connected is good, but the mechanical aspects of doing that well can be expensive. To reduce the impact of not doing that so well, knocking the input spectrum down will help and, it done with printed inductor(s), cheaper than an external ferrite device.

--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.

signal.

Thanks John - that is key to the solution, I guess.

the

as

an

Our enclouse is aluminium, so maybe that is sufficient without a shielding box? Many thanks for the suggestions.

the

signals

long as

be

an

That sounds a very good approach. Thank you for the ideas. I'm afraid if it is common mode noise then I've asked the wrong question so this solution may not be suitable for my case.

The only issue with this in our case is that the sockets are surface mount and so fitting a ferrite berad would mean using a diferent connector and using fliying leadds, adding to the manufacturing cost. An SMD common mode choke might be the answer though. We are designing the PCB now, so redesign is not a problem...yet ;)

-- James

signal.

the

as

an

get

signals

don't

long

be

I forgot to mention that the RCA phono connector for the SPDIF output has its outer conductor grounded to both the PCB ground and also to the aluminium chassis via a screw, thus there is no track as such going to the outer conductor. The phono connector's inner conductor (for the signal) solders to a track on the PCB. I'm wondering how I might use a common mode choke in this case? (FWIW, I originally chose a phono connector with the screw-mounting, partly with EMC in mind.) Sorry if this is a basic question.

Many thanks, James

can

Thanks Gerhard for the reference. They look interesting. I've also come across some Murata common mode chokes in 0805 form that could be useful. Just wondering how best they should be connected in our case, where the outer conductor of the connector is directly grounded to the chassis and the board's ground?

James

mount

mode

before

would

kind

values

I read in sci.electronics.design that James wrote (in ) about 'Attenuating above 100MHz on squarewave (SPDIF digital audio)', on Tue, 28 Dec 2004:

Without changing that physical arrangement, you can't introduce common- mode suppression. You can only put a ferrite on the cable.

--
Regards, John Woodgate, OOO - Own Opinions Only. 
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

James,

Ferroxcube has nice SMD common mode coils in 2x1 turn or 4x1 turn (which can be connected as 2x2 turn). Maybe you could combine common mode and a differential mode choke

Best regards,

Gerhard

"James" schreef in bericht news:I%eAd.80$ snipped-for-privacy@newsfe6-win.ntli.net...

redesign

6.144Mbits/sec.

the

the

signals

don't

long

be

of

of

recommended

we

Thanks John. Since we are at the PCB design stage, I'm thinking of going back to basics somewhat to solve this properly. Referring to my copy of "EMC for Product Designers" by Tim Williams, he recommends a separate chassis ground and signal ground. We would connect the outer conductor of the phono connectors to the aluminium chassis and chassis ground. These will be bonded by a spacer. The signal ground used over the rest of the board would not connect to the chassis at all, and only to the chassis ground via a single track.

I did try to do that on the original unit, but practicalities meant the ground planes were not kept as separated as they should have been.

Is this scheme likely to work better with regard to passing conducted emissions testing? The unit is driven by an external DC adaptor with 2 wires; should I connect it's low-side to the chassis ground or signal ground?

Thanks, James

I read in sci.electronics.design that James wrote (in ) about 'Attenuating above 100MHz on squarewave (SPDIF digital audio)', on Tue, 28 Dec 2004:

That is a sound principle. Think of the outer contact and the cable shield as a pipe connecting the 'tank' (conducting enclosure/chassis) of your product to the product it is supplying signal to. You don't want any 'leaks' in the pipe.

There is not much room for compromise in this EMC/grounding business.

To chassis ground.

There is a LOT more information on this and similar issues in the June

1995 issue of the Journal of the Audio Engineering Society. It's all practically-oriented, not mathematical. Go get a copy from

--
Regards, John Woodgate, OOO - Own Opinions Only. 
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

signal.

6.144Mbits/sec.

the

the

signals

long as

be

Hi Larry. To clarify this in my mind, are you saying an inline ferrite or RC/LR low pass filter circuit on the signal line, just before the output socket is a good idea (in addition to sound grounding)?

Thanks, James

Sorry - I meant LC rather than LR low pass filter. I also reread your previous reply, Larry, about using printed inductors as a low pass filter - I think you have already answered my question there, thanks.

James

message

or

Many thanks for the advice, Larry.

-- James

You could even use your common-mode choke here to connect the chassis ground to signal ground.

Like this (use courier or other constant-width font for this ascii art schematic):

CM-CHOKE SIG_IN>---~~~~~~~-----------> connector signal pin ----- ----- +---~~~~~~~------+---->connector ground | | SIG_GND CHAS_GND

The above is supposed to show a common-mode choke (a transformer, but sideways). I hope you get the idea. I don't really know how to draw a CM choke in ascii art. Sorry.

Just my $0.02. I don't have only a little experience with EMC, and none with SPDIF. But it seems like this should give you a pretty good shot at getting it right.

--Mac