RFI protection concepts

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
User avatar
Gertjan
Posts: 12
Joined: Mon Mar 11, 2013 9:12 am
Location: Utrecht, The Netherlands
Contact:

RFI protection concepts

Post by Gertjan » Tue May 27, 2014 7:21 am

I would like to discus RFI protection concepts.
I am using the concept below. XLR pin 1 to ground, and ferrite bead / ceramic cap filters on signal leads.

Example from a recent design:
Image

However, on the outputs my configuration is different than several examples from very well respected people, see below.

From the RHAT1646 datasheet:
Image

El Cheapo from Ricardo:
Image

It seems this is a different idea, of just providing a low impedance to ground, and a high impedance into the circuit, thus steering HF energy to ground instead of into the circuit.
My idea is just HF filtering the output leads, just as any other “antenna “ connecting to my box.

Any ideas & opinions over which concept works better and why?

I am a film sound recordist. After education in Electronics I have been designing part of my gear (Hi-Fi and Pro audio) for about 30 years.
The biggest RFI problems in film sound recording are nearby Cellular Phones (cast & crew, 900MHz-2,4GHz) and our own Wireless Microphones (usually around 600-800MHz)

Regards, Gertjan Miedema.

User avatar
JR.
Posts: 2410
Joined: Sat May 24, 2008 7:21 pm

Re: RFI protection concepts

Post by JR. » Tue May 27, 2014 9:20 am

The best answer is probably to test alternate ways with typical problem RFI sources. I suspect all 3 ways will work, one may be slightly better than the others.

Your schematic appears like a mirror image on the output as input, but they are different. The output side has a relatively low source impedance.

The THAT schematic looks like they are using the ferrite beads as dual function to isolate output capacitance like the typical output resistors do.

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 » Tue May 27, 2014 10:38 am

Thanks for your insight!

When filtering, you need an impedance to drop the unwanted signal over (the L’s in my circuit).
Otherwise you are depending on the source impedance for how effective the filtering will be.

HF signals might not adhere to this logic....

Indeed, in my schematic, the low output impedance from the 1646 is in parallel with the filter C, degrading the filtering.
The filter is designed to start working around 10MHz, and 1646 Small Signal Bandwidth is also 10MHz

Although I am happy to have a well-equipped Bench, It’s all for Audio frequencies....
Above 20MHz I am blind. No Calrec CB-Radio for testing either.

Normally I would just build different strategies, and test what works best for me.
This is more difficult with RFI filtering.
Also, as there are many sources of RFI problems, it might take years before you encounter a condition which proves to be a problem.

That’s why I like to hear from people who are experienced in this field.

User avatar
mediatechnology
Posts: 3260
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: RFI protection concepts

Post by mediatechnology » Tue May 27, 2014 9:07 pm

Although I am happy to have a well-equipped Bench, It’s all for Audio frequencies....
Above 20MHz I am blind. No Calrec CB-Radio for testing either.
Thank you for joining us and posting here!

I have to admit to having little experience in the RFI area other than the usual studio tech work and broadcast (FM mostly) transmitter experience.
Both have occasionally been in high RFI environments.
In the studio area it was usually AM stations in the near field of the studio. I (luckily) rarely had problems.
In broadcast it was usually FM transmitters (and an occasional AM) miles from the studio.
I feel for the AM guys who have transmitters co-located with studios.
They not only have the RFI issues but lightning to deal with.

What I do have is a Motorola R2600 400 kHz to 1 Ghz Generator/Analyzer.
If the BBC or someone has defined a test procedure or standard I'd entertain using the R2600 to test some of these outputs.

I like the added build-out resistors in your first example.
The added resistors would seem to limit rectification of signals picked up in the output.

ricardo is the resident expert having dealt with this in theatre applications.

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

Re: RFI protection concepts

Post by ricardo » Tue May 27, 2014 9:24 pm

The caps MUST be AT the socket and connect to the Chassis by the shortest possible leads. Routing via ferrites or inductors will significantly increase the path taken by RFI to Chassis ... and sensitivity to clicks, pops & mobile phones. 1" is definitely too much if you are serious about RFI proofing.

My arrangement, also developed independently by Schoeps and the THAT circuit will be hugely better than yours. But its difficult to stress how important the physical aspects are.

The ideal is to get any RFI to the Faraday Cage (Chassis) without ever entering the box. On ALL other sockets, the 'shield' must still go to Chassis at RF .. even (especially) if you can't make a direct DC connection.

http://www.neutrik.com/fl/en/audio/204_ ... tlist.aspx

Their circumferential capacitor acts like my 100n ceramic from p1 to Chassis where I can't make a DC connection .. but with zero length leads. Their ferrite bead is a superior alternative to the 10R on pin1 recommended by Bill Whitlock in some of his papers.

But there is ALOT of very expensive 'pro' gear which will tell everyone you have received an SMS message which is why some studios ban mobile phones.

In the absence of a Calrec CB radio, its likely your mobile phone is the easiest test source. If your gear doesn't burp if your phone receives a SMS while lying on top ot it, its probably OK and better than most. I can tell you that the biggest task is to make sure the interference isn't being picked up by the other stuff connected to it.

So please test the difference and let us know how you get on.

You'll notice I don't even talk about filters on the input/output etc cos they all pale into insignificance compared to the above. Just think how the RFI gets to the Chassis and remember 1" is an excellent aerial.
__________________

I'm really out of touch with EMI/RFI standards having been nearly 2 decades in the bush. But I can tell you that in the early 80's, the EU EMI standards & testing including the BBC's were completely inadequate to ensure no clicks, pops & buzz from lighting in TV studios. That's why we built the Calrec CB radio.

The major mike makers had transmitters sorted out cos the broadcast organisations were major customers from da early days.

That's not to say, the big makers always get it right. There's a very famous German mike that has serious p1 problem and a pic of its internals appears in one of the classic papers describing the problem (Whitlock?). Dunno if they've sorted it out today.

User avatar
mediatechnology
Posts: 3260
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: RFI protection concepts

Post by mediatechnology » Tue May 27, 2014 9:58 pm

its likely your mobile phone is the easiest test source.
My vote is a phone in GSM mode with its 100 Hz (or so) RF burst rate.
At one time I had WiMax modem (Clearwire, pre-4G), a Samsung GSM phone on the Verizon network, and a Verizon LTE phone (mostly 700 MHz band).
Only the GSM phone produced rectification and ingress into telephone handset cords.
The 4G LTE phone and WiMax are still with me and I haven't heard a peep out of either.

Speaking of Faraday cages: A microwave oven makes an excellent Faraday cage for testing.
I put my smartphone in one the other day and watched 14 or so GPS and Glasnost satellites, WiFi, 3G and 4G drop instantly off the map.

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

Re: RFI protection concepts

Post by ricardo » Tue May 27, 2014 10:17 pm

mediatechnology wrote:My vote is a phone in GSM mode with its 100 Hz (or so) RF burst rate.
At one time I had WiMax modem (Clearwire, pre-4G), a Samsung GSM phone on the Verizon network, and a Verizon LTE phone (mostly 700 MHz band).
Only the GSM phone produced rectification and ingress into telephone handset cords.
The 4G LTE phone and WiMax are still with me and I haven't heard a peep out of either.
Didn't realise there were big differences. We're supposed to have 4G in Oz but its still GSM in Cooktown.
A microwave oven makes an excellent Faraday cage
Wayne sets the latest must-have fashion craze for Golden Pinnae gear :D

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 » Wed May 28, 2014 4:19 am

Gentleman, I really feel honoured with your detailed insights, and appreciate your effort, especially Ricardo’s

Main point from Ricardo is that One should think in terms of shorting RF to Chassis / Case, shunting the RF around the circuit. And not think in terms of filtering Inputs / Outputs.

I am aware of the importance of short connections.
Caps & Beads were intended to solder directly on the connectors, not on the PCB.
Still good to be reminded.

I am a (happy) user of the Neutrik EMC connectors Ricardo is linking to.
They provide a good solution when having to send Audio to a Broadcast Camera with a Pin 1 problem.
Indeed it’s good to study them close. There is more cleverness in there than noticed at first sight.

Indeed, at the moment my GSM mobile is my main RFI testing tool.
Compounding the difficulty of testing with a GSM is that it transmits proportionally to network field strength.
I live in the Netherlands, where the density of the cellular network is high. So low output of the GSM transmitter.
However, when shooting in a big Hospital, with lots of concrete and shielding, suddenly I hear the pip-pip of transmitting GSM’s.
They just scaled their transmitters up to their 2W max transmitting power in an effort to keep connection.


This is my new RFI proposal, based on the new insights provided by Ricardo:
Image
RF on inputs and outputs is directly shunted to Chassis / Case.
On the output I added ferrite beads to enlarge the HF impedance into the 1646
On the Input adding Beads is useless, as their Impedance is swamped by the internal input resistors of the 1240/1246.
I added an indication what is on the PCB, and what’s not.


Taking the idea of steering RF around the circuit one step further.
What about including beads in the audio path to audio ground, see below:
Image
To clarify the idea the audio ground is redrawn, and the Chassis / Case is also drawn.
Adding the ferrite beads makes the RF path though the PCB much higher impedance than though the Chassis / Case.

Are these beads introducing problems in other areas I am overlooking?

This is for one-off, as good as possible, cost no object Projects.
So mass production, cost calculating arguments are not important in my case.

Regards, Gertjan.

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

Re: RFI protection concepts

Post by ricardo » Wed May 28, 2014 7:50 pm

Gertjan wrote:Indeed, at the moment my GSM mobile is my main RFI testing tool.
Compounding the difficulty of testing with a GSM is that it transmits proportionally to network field strength.
I live in the Netherlands, where the density of the cellular network is high. So low output of the GSM transmitter.
However, when shooting in a big Hospital, with lots of concrete and shielding, suddenly I hear the pip-pip of transmitting GSM’s.
They just scaled their transmitters up to their 2W max transmitting power in an effort to keep connection.
Thanks for the info about GSM phones. This beach bum only got to know about mobile phones circa 2005. When I lived in civilised parts last Millenium, mobile phones were rare, expensive & the size of lunch boxes.
On the Input adding Beads is useless, as their Impedance is swamped by the internal input resistors of the 1240/1246.
...
What about including beads in the audio path to audio ground, see below:
Image

Adding the ferrite beads makes the RF path though the PCB much higher impedance than though the Chassis / Case.
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

I know some of you prefer direct connection of GND at both input & output. Anyone care to comment?

Ferrite beads on the input signal line do help. Its the impedance at RF that's important and that's more to do with the input devices than the resistors. But as this looks like a line level device, you don't have to be as paranoid as with a mike preamp or mike. As JR points out, you only have to reduce the RFI below a certain level .. not eliminate it completely.

Your ferrites on the GND connections are like the Neutrik EMC connectors so good.

IIRC, i couldn't find any specs for the first ferrite beads I used. I might have got them from the local dressmakers shop :mrgreen:

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 » Thu May 29, 2014 5:21 am

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 a sensitive spot of me.
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.


Ferrite beads on the input signal line do help. Its the impedance at RF that's important and that's more to do with the input devices than the resistors.
You are probably right.

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.


IIRC, i couldn't find any specs for the first ferrite beads I used. I might have got them from the local dressmakers shop
These days we have got a wide range of beads, made from different materials, for different frequency ranges.

These are the Beads I am using in this application:

Image

My biggest problem is to keep the different types apart after buying, as they all look alike....

Regards, Gertjan.

Post Reply