RFI protection concepts

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
User avatar
JR.
Posts: 2418
Joined: Sat May 24, 2008 7:21 pm

Re: RFI protection concepts

Post by JR. » Thu May 29, 2014 9:39 am

Gertjan wrote:
I was looking at the equivalent schematic in the THAT datasheet.
Here is the input redrawn with this equivalent schematic:
Image
First thought is that the Ferrite Bead impedances are small in relation to the THAT input resisters.
But on second thoughts there is off course a lot more in there. Input protection diodes etc.
Maybe enough parasitic capacitance for RF to play with....
I’ll test with input beads in place too.


Regards, Gertjan.
While I have nothing much to add to the RFI discussion, I have opinions about audio references. First I will observe that the pin on the 1246 is labelled "Ref" not GND. I have seen it called 0V and I find that preferable to "audio ground".

The internal 0V reference generally needs an AC and DC path to PS common, which ideally has an AC and DC path to chassis, but I question the value of hard wire or near hard wire (with FB) connection between 0V and chassis ground. In fact that hard wire connection defeats the benefit of having the differential connection that is working to subtract out ground potential errors.

For large structures I have no problem with having multiple different internal 0V reference nodes, with differential connections to maintain signal integrity between the sundry local references.

The main purpose of providing a path between chassis and PS common, it to keep input common mode signals within the expected PS rail voltages to prevent circuitry saturation. Sometimes when rolling your own output differential circuitry it is tempting to use low value resistors for the reference leg of the differential to provide a modest impedance path too.

Inside the chassis you have audio hot and audio low or audio references. Mixing grounds with signals can be problematic.

JR

User avatar
Gertjan
Posts: 12
Joined: Mon Mar 11, 2013 9:12 am
Location: Utrecht, The Netherlands
Contact:

Re: RFI protection concepts

Post by Gertjan » Fri May 30, 2014 3:16 am

JR –

Of course you are completely right!
And my schematic is certainly not in accordance with the AES48-2005 directions.

Ricardo already touched this weak spot:
This is a good scheme. The only caveat might be you have Chassis 'directly' connected at both i/p & o/p. This might give a hum loop but it depends on loadsa other stuff ... like what's connected. Probably OK if its an all portable system.

I like direct connection to Chassis at the outputs (mainly for when you connect to unbalanced gear) while the p1 brigade (Whitlock et al) prefer direct connection at input.

If you don't connect GND to Chassis at either input & output, you still need to get the Shield/GND connected to Chassis at RF via a 100n ceramic with short leads
Ouch! You are touching sensitive spot.
After half my life honing my Star Grounding skills, Pin 1 came along.
After that I have been balancing between good HF practice (multiple grounding) and good audio practices (Star grounding)....

I am revisiting my grounding & pin 1 library.
Already dug the original Muncy article out of the big stack of AES Journals
I am in the middle of rereading papers and rethinking my grounding strategy.
The directions of AES48-2005 make me think that I can have best of both worlds.
The Star Grounding in the Audio Domain, While for RFI thinking in terms of a Faraday Cage, decoupling everything for RF with small capacitors to Chassis (and have only one connection between Chassis and 0V / Audio Ground / Star point).

To put things in context:
The last time I completely wired up a studio is two decades ago.
These days my (Pro) designs are intended to be used for portable use in a bag for film sound recording purposes.

In a film sound bag there is usually one Battery , feeding all gear.
In the past one of the problems was that, if not careful, Battery return current would flow through audio grounds.
These days, it is more common to use a DC-DC converter in the Power Supply, and have galvanic isolation between the Battery and the audio circuit.

So, there are no Power Ground connections (like in a Studio).
Sill I think that AES48-2005 can also work in my situation, and is a better solution all around.

I am off, reading, thinking, and testing..... Gertjan.

ricardo
Posts: 262
Joined: Sun Apr 10, 2011 7:24 am

Re: RFI protection concepts

Post by ricardo » Fri May 30, 2014 3:48 am

Gertjan wrote:The directions of AES48-2005 make me think that I can have best of both worlds.
The Star Grounding in the Audio Domain, While for RFI thinking in terms of a Faraday Cage, decoupling everything for RF with small capacitors to Chassis (and have only one connection between Chassis and 0V / Audio Ground / Star point).
Yes. If you think along these lines, you are on the right track.
... In the past one of the problems was that, if not careful, Battery return current would flow through audio grounds. .. So, there are no Power Ground connections (like in a Studio).
Most of the 'Star-Grounding' recommendations are to deal with buzz from rectified PSUs .. and also unbalanced lines. eg that's why I like to make the Chassis connection at the Output of the box. But that makes it more complicated at the Input to deal with Pin 1

Your pro portable gear with only balanced connections to other stuff, in some ways is some ways less of a problem, (though I'm sure modern evil digital introduces new problems

If several items are powered from one battery pack, are there separate leads from the battery to each item of gear?
____________________

PS. I think you are right about the ferrites on you Line i/p being ineffective because of the resistors. For some reason, I thought you had mike inputs. :oops:

User avatar
Gertjan
Posts: 12
Joined: Mon Mar 11, 2013 9:12 am
Location: Utrecht, The Netherlands
Contact:

Re: RFI protection concepts

Post by Gertjan » Fri May 30, 2014 4:26 am

Ricardo –

In a film sound bag all gear (mixer, recorder, radio mike receivers) has it’s own lead to the Battery (or power distribution). That’s why the galvanic isolation in de Power Supplies is helpful.

Indeed “modern evil digital” introduces new problems!
Some Boxes radiate RF up to the frequencies used for the Wireless Microphones.
As Recorder and Receivers are packed close together in a bag, the Radio Mike range can be reduced considerably.

Also digital Recorders and Radio Mike Receivers can have quite a bit of RF on their Audio outputs.
These are main reasons why RFI proofing is important to me.

It has to be said that Digital Recorder Manufacturers are making big improvements in RF Shielding their Cases and filtering their In/Outputs

And we can’t live anymore without the capability of recording 6-8 tracks on location.

Regards, Gertjan

User avatar
JR.
Posts: 2418
Joined: Sat May 24, 2008 7:21 pm

Re: RFI protection concepts

Post by JR. » Fri May 30, 2014 9:41 am

I am not familiar with AES 48-2005 but using small capacitors to provide a low impedance RF path while being open circuit at mains hum frequencies is an old technique called hybrid grounding, Mostly practiced by broadcast industry, but in recent decades RF levels have increased so this is more widely used.

Whatever keep reading, there has been a lot written on the subject, I have a couple books from the '70s about grounding and shielding (Morrison,Ott).

JR

User avatar
Gertjan
Posts: 12
Joined: Mon Mar 11, 2013 9:12 am
Location: Utrecht, The Netherlands
Contact:

Re: RFI protection concepts

Post by Gertjan » Fri May 30, 2014 10:58 am

Indeed I am reading Henry Ott, Neil Muncy, Bill Withlock, Jim Brown.
Most out of my own library. Amazing how much one can forget....

The Hart of the Matter:

Neil A. Muncy, 1995:
Image


AES 48-2005 Standard in a nutshell, 2005:
Image

Gertjan

User avatar
JR.
Posts: 2418
Joined: Sat May 24, 2008 7:21 pm

Re: RFI protection concepts

Post by JR. » Fri May 30, 2014 1:31 pm

Yes, AES 48 2005 is right way... It makes me crazy when people cut off the safety ground pin, or lift shields.

The fact that the aes standard is dated 2005 suggests that the industry may have lagged in best practices for some time.

JR

{edit- perhaps amusing anecdote. I recall attending a dealer seminar in Mexico and was told by a mexican dealer that the way to tell if an amplifier was new or used, was check if it still had the ground pin... Only new never used amps had ground pins. :oops: [/edit]

User avatar
Gertjan
Posts: 12
Joined: Mon Mar 11, 2013 9:12 am
Location: Utrecht, The Netherlands
Contact:

Re: RFI protection concepts

Post by Gertjan » Fri May 30, 2014 4:02 pm

JR,
I really had to laugh when I read your Mexican anecdote!
Gertjan.

ricardo
Posts: 262
Joined: Sun Apr 10, 2011 7:24 am

Re: RFI protection concepts

Post by ricardo » Fri May 30, 2014 5:17 pm

JR. wrote:Yes, AES 48 2005 is right way...
Yes. This is good for RFI. I'll add a caveat about dirty decoupling caps to 'signal reference'.

As I said, my insistence on direct connection to Chassis at i/p or o/p but not both is from having to do gear that could be connected to unbalanced as well as 'pro' balanced stuff. (Should start a separate thread for this.)

A sorta half way house is P48V. In AES 48 2005, the P48V return is via the Chassis. This may or may not cause problems depending on loadsa stuff.

The best treatment of RFI for unbalanced domestic gear was an article by Jim Sugden (of Sugden Class A amps) in HiFi News 1970's or maybe 60's. It was about phono preamps which didn't click & pop.
It makes me crazy when people cut off the safety ground pion, or lift shields .. I recall attending a dealer seminar in Mexico and was told by a mexican dealer that the way to tell if an amplifier was new or used, was check if it still had the ground pin... Only new never used amps had ground pins. :oops:
For a large part of my working life, this was NOT funny.

In Oz, 'everyone' is responsible for WHS. If someone has unsafe work practices, he can be fined.

In the UK, its slightly different. In every organisation, a director or senior manager, ie moi, was legally responsible and could go to prison if unsafe practices were found. I did a lot of shouting at engineers who lifted safety grounds from test gear in R&D :evil:

But I can't put my hand on my heart & claim never to have done this evil thing ;)

User avatar
Gertjan
Posts: 12
Joined: Mon Mar 11, 2013 9:12 am
Location: Utrecht, The Netherlands
Contact:

Re: RFI protection concepts

Post by Gertjan » Sun Jun 01, 2014 12:12 pm

Here is my new concept for Grounding and RFI protection:
Image

- For grounding I am following AES 48-2005, but Star Grounding as much as possible.

- For RFI protection I treat the Case as a Cage of Faraday, decoupling all incoming leads directly to Chassis.
- When the lead goes into a low impedance I add a Ferrite Bead to make this impedance higher for RF

I don’t like the AES48 recommendation to create the Star point at the 0V of the Power Supply.
It is OK in a perfect (balanced audio) World, but:
- When the signal is not (perfectly) symmetrical (e.g. Unbalanced output, or cheap badly balanced source) the Potential differences between chassis and Star Ground come into play.
- If the plus and minus Power Supply lines are not equally loaded (e.g. single supply signaling, relays, peak leds etc.) or when using 48V Phantom, there will be a current in the 0V wire, creating a voltage to be added to the Star Ground Potential.

Muncy’s reason to use the 0V of the Power supply for the connection to Chassis is to direct the leakage current of the Power Transformer back to Mains Ground/Earth, without sharing 0V or Signal conductors.

With unbalanced outputs in mind, and currents being biggest at the Output, I choose to make the Star-chassis connection at the Output connector. (for this All Line Level application)

When dealing with Mic inputs and Phantom powering, it might be best to make this Star-Chassis connection at the input:
Image

- Cleanest Ground at the most sensitive point in the Circuit
- No Phantom return current through the Chassis

Of course, when all connectors are on the same panel there is not much difference between these two versions.

In my portable applications I don’t have the complication of the Mains, with associated Ground/Earth pins.
But I do have to consider combinations with other equipment where the 0V of the common Battery is directly connected their Ground/Chassis

Unfortunately I was unable to find the "Phono preamps which didn't click & pop" article by Jim Sugden.

I found it hard to test for RFI with iPhone induced RF (while ringing, texting, data transfer)
Testing with Line level Inputs and Outputs it was hard to get any “pip-pip-pip” at all.
Even with the Phone laying on coiled-up Cable. Even without any RFI filtering.

Yet in the field I do encounter these Smartphone pips with some regularity. (especially the big Hospital location comes to mind)
Looks like the root of this problem is more in Pin 1 problems, and Ground loops than RFI filtering...

Regards, Gertjan.

Post Reply