Ferrite Beads on Inputs

[JohnRoberts]

I don't know if these were standard parts or custom connectors made for us by Neutrik (we used millions of connectors at Peavey), but I recall plastic jacks with serrated metal contacts that would pierce through paint/power coat layers to insure a ground bond to the chassis.

JR

 

[Bo Deadly]

I don't know if these were standard parts or custom connectors made for us by Neutrik (we used millions of connectors at Peavey), but I recall plastic jacks with serrated metal contacts that would pierce through paint/power coat layers to insure a ground bond to the chassis.

Actually I just changed my mind.

I'm going to use Neutrik NC3FBV1 and NC3MBV1 which are plastic vertical PCB mount but they have metal plates on the front that completely encircle the hole as well as a contact on one of the screw holes that is supposed to make a good bond with the panel AND the plate and screw contact are internally connected to pin 1.

Although I'm going to keep the caps and chip inductors on pins 2 and 3 which will now be very close to pin 1. So no separate chassis wire.

So the 48V ground return will be through the chassis and the line input will have pin 1 hard-wired to chassis. No biggie.

Actually if it's an annodized aluminum panel I suppose it might not make contact. Maybe I'll have to scuff up the aluminum just behind the hole so that the plate makes contact. Or possibly I could put a serrated washer on the one screw.

The only issue is that because the power supply is external, the chassis ground will still have to run though the CPC connector bolt inside the enclosure to get to the connector pin for chassis ground and then down the power cable. Unfortunately the only way to fix that would be to use an entirely different connector that is metal and has a pin for it. But the length of the bolt / PCB trace / pin is probably no more than a few cm which according to Google shouldn't be an issue for anything less than ~1 GHz.

 

[Andy Peters]

I don't know if these were standard parts or custom connectors made for us by Neutrik (we used millions of connectors at Peavey), but I recall plastic jacks with serrated metal contacts that would pierce through paint/power coat layers to insure a ground bond to the chassis.

They're a standard item in the Neutrik catalog now.

 

[ricardo]

I'm going to use Neutrik NC3FBV1 and NC3MBV1 which are plastic vertical PCB mount but they have metal plates on the front that completely encircle the hole as well as a contact on one of the screw holes that is supposed to make a good bond with the panel AND the plate and screw contact are internally connected to pin 1.

These are the Neutrik's I'm referring to.

They also make a special plug to deal with gear with a 'pin 1 problem' .. ie gear that hasn't taken RFI seriously/properly.

Actually if it's an anodized aluminum panel I suppose it might not make contact. Maybe I'll have to scuff up the aluminum just behind the hole so that the plate makes contact. Or possibly I could put a serrated washer on the one screw.

Yes.

The only issue is that because the power supply is external, the chassis ground will still have to run though the CPC connector bolt inside the enclosure to get to the connector pin for chassis ground and then down the power cable. Unfortunately the only way to fix that would be to use an entirely different connector that is metal and has a pin for it. But the length of the bolt / PCB trace / pin is probably no more than a few cm which according to Google shouldn't be an issue for anything less than ~1 GHz.

You can use a shielded cable for the power cable.  It's not hard to get your inside decoupling leads to less than 1 cm.  These could be the leads of a 100n ceramic connecting the shield to the nearest chassis screw.

 

[abbey road d enfer]

And I'm not sure it's necessary anyway because if the board is configured as a mic input with 48V power, pin 1 must be connected to a non-chassis ground for the 48V return current 

I believe you did not grasp fully the implications. With phantom power, Pin 1 needs to have a galvanic connection to the phantom supply reference (0V), but it does not need to be a "stiff" connection; having even a few hundred ohms would still provide perfect phantom operation. But "stiff" RF connection betwxeen Pin  I and the power ground and chassis is paramount. Stiff means the impedance of that connection must be very low at RF, so a capacitor with short leads is as good as a short piece of wire. But the impedance at 50/60 Hz must be reasonably low in order to provide good hum rejection, and sufficiently high as to minimize loop currents. Differential ground voltages are usually <1Vrms, so a 100 ohms limits current at about 10 mA.

If it's a line input, ideally pin 1 should just be left unconnected entirely in which case there can be no shield currents.

I beg to differ with the word "ideally". Although leaving Pin 1 unconnected for line level connection may work in "ideal" environments, the need for a proper connection may arise in many circumstances.

I know some people insist it should be connected to chassis ground

Rightly so

so it can be jumpered that way or with the Jensen resistor cap shunt.

The series arrangement of the "Jensen" RC does not provide the needed "very low impedance at RF"; it should be a parallels RC.

Also, the last 3-4 posts have all been about the shield. Meaning pin 1 and the "pin 1 problem". I understand what you're saying and I appreciate the comments. But my original question was about RF filtering the input pins 2 and 3.

Filtering pins 2 & 3 is almost inseparable from the Pin 1 issue, since common-mode noise is referenced to "ground". If you want to filter out the voltage between either pin 2 or 3 and Pin 1, you have to make sure that Pin 1 is a valid reference.
I must say this can be debated, since it's actually the voltage between pins 2 & 3 and chassis that is the issue, so, yes, pin 1 could be left floating, as long as a proper connection between shield and chassis is provided. Pin "G" may be used as shield connection; as long as all the wiring is done in accordance, why not. This arrangement was advocated years ago by Gotham Audio, who had a specific 3 cond+ shield cable; this was a way of avoiding issues created by the then usual practice of connecting Pin 1 to audio ground. Anyway, whatever name you give it, you must have a connection that carries the shield/chassis/earth equipotential.

It's not clear that that question was every answered

I believe most of them were, but you seem to be too far engaged in your design to seriously reconsider. 

a) the shield is probably much more important wrt filtering than pins 2 and 3

As I wrote, it's a matter of voltage difference, so in that respect, all pins have equal importance; when a-b=c, is a more important than b?

  b) if there is RF on 2 and 3 it's already too late.

Why "too late"? Assume there's always RF on pins 2 & 3.

The question ultimately becomes, how much attenuation is provided by blocking RF on the leads

That's a good question, and there may be several answers, according to the various environments, but today, most of the issues are with cell phones and industrial noise. I would say most of the pollution is between 1 MHz and 1 GHz, far enough from audio frequencies to make filtering  relatively easy.

even though it can still be capacitively coupled to other parts of the circuit.

That's why it is important to keep the parts that cannot be filtered as short and small as possible.

There are plenty of schematics with LC filters in inputs that shunt to ground on a PCB. Can they all be wrong?

Some may be wrong. When the CE regulations came in, many designers did not know exactly how to tackle the issue, so they used parts that sellers told them would solve their problems, but without mentioning that implementation is as important as the filters themselves. But TC electronics, for example, have a perfect implementation of RFI/EMI protection, although built on a PCB.

 

[arnyk]

Similar to the other topic about ferrites on output pins, does it not follow that there should be caps and ferrites on inputs

You might find a number of answers to these questions here:

http://www.rane.com/pdf/old/ms1bsch2.pdf  and similar addresses for other similar products.

This is a PDF in the Rane obsolete product archive which contains detailed  parts layout and schematics for a number of different  products of theirs that had mic inputs.  Their products had/have  a good reputation for EMI resistance, etc.

For example MS1  is a family of mic preamps that they produced over several decades. MS1A and MS1b are part of the sequence of the products in this series.  You can see how their approaches to these problems evolved.

 

[Bo Deadly]

Regarding Neutrik connectors, as others have pointed out, many connectors have a metal spike just inside one of the screw holes. This is what is described in the documentation as the "panel" contact. It makes contact with the shell. The combo series also have two little metal panel contacts. This makes contact with the shell and optionally a ground pin. If there is no ground pin, it is connected to pin 1.

So a part like NC3FAV1 has this spike which connects to the shell and directly to pin 1. That makes the pin 1 to chassis connection about as short as one could hope I think. Although, my understanding at this point is that one would not want to connect anything to pin 1 like a ground plane or the shield of a cable because it will just help radiate whatever RF might be on pin 1. Any such ground planes or shields should be connected to the chassis elsewhere.

 

[Bo Deadly]

This is a PDF in the Rane obsolete product archive which contains detailed  parts layout and schematics for a number of different  products of theirs that had mic inputs.  Their products had/have  a good reputation for EMI resistance, etc.

For example MS1  is a family of mic preamps that they produced over several decades. MS1A and MS1b are part of the sequence of the products in this series.  You can see how their approaches to these problems evolved.

Interesting. So they used ferrite beads and they put them before the caps. Right now I have them after the caps.

 

[ricardo]

http://www.rane.com/pdf/old/ms1bsch2.pdf  and similar addresses for other similar products.
..........
For example MS1  is a family of mic preamps that they produced over several decades. MS1A and MS1b are part of the sequence of the products in this series.  You can see how their approaches to these problems evolved.

The secret ingredient which isn't shown is how their PCB is laid out .. tracking, ground plane, connections to chassis etc.

You can immediately tell if a maker understand low noise design by how they lay PCB tracks.

But look at the Chassis symbols on their schematic.  They are emphasizing that p1 on the XLRs and the caps go DIRECTLY to chassis AT the XLRs .. not via ground planes, extra leads bla bla

 

[abbey road d enfer]

Although, my understanding at this point is that one would not want to connect anything to pin 1 like a ground plane or the shield of a cable because it will just help radiate whatever RF might be on pin 1. Any such ground planes or shields should be connected to the chassis elsewhere.

The whole idea of the "stiff RF connection" is that any RF carried by the shield should be dumped into the chassis. There should be nothing to radiate.
Regarding the ground plane connection, there are two possibilities:
a) the ground plane is a shield: thus it is an extension of the case, and should be connected at pin 1 (but not only, since there may be several Pin 1's and other points worth connecting).  That's what I would do when the unit creates its own RF/EMI.
b) the ground plane is a low resistance path for audio ground; in that case, it does not need to be connected directly to chassis ground. Hierarchical grounding should be applied. That's the solution I would use for equipment that does not generate RF.digital noise.
In mixed designs, the distinction is not 100% clear cut. The ground plane used as a shield may also be used as the "dirty" ground where decoupling caps meet, and the ground plane used as low-res audio ground certainly acts as an electrostatic shield helping with x-talk and unwanted coupling between sensitive tracks.

A schemo does not explicit the intricacies of grounding. All these "AGND" points appear just as one single entity, but in reality they are disseminated all over the PCB estate. How they are managed is crucial.

 

[Bo Deadly]

The whole idea of the "stiff RF connection" is that any RF carried by the shield should be dumped into the chassis. There should be nothing to radiate.

But there is always some distance between pin 1 and the chassis. Albeit small, it's inductance creates a resistance over which RF can be radiated. Yes?

If yes, then connecting anything at pin 1 (as opposed to at the chassis even though it's only maybe 1 cm away) will only serve to distribute that RF further inside the enclosure?

 

[PRR]

> distribute that RF further

A big AM station has a 1,000 foot antenna and throws signal far.

Replace that tower with a 1 inch antenna. Range will be greatly reduced.

Yes, if you connect a 50,000 Watt transmitter to an inch of wire off pin 1, you will have a relatively strong signal IN the box. So don't do that. Mostly we have big transmitters miles away, and teeny transmitters feet away. The RF coming in pin 1 is not large. Getting that energy to chassis ASAP is usually sufficient.

(In the 1930s some audio studios were right UNDER the transmitter tower, as transmitters increased from 5,000W to 50,000W. More obsessive grounding is found in some vintage gear.)

 

[JohnRoberts]

> distribute that RF further

A big AM station has a 1,000 foot antenna and throws signal far.

Replace that tower with a 1 inch antenna. Range will be greatly reduced.

Yes, if you connect a 50,000 Watt transmitter to an inch of wire off pin 1, you will have a relatively strong signal IN the box. So don't do that. Mostly we have big transmitters miles away, and teeny transmitters feet away. The RF coming in pin 1 is not large. Getting that energy to chassis ASAP is usually sufficient.

(In the 1930s some audio studios were right UNDER the transmitter tower, as transmitters increased from 5,000W to 50,000W. More obsessive grounding is found in some vintage gear.)

Back in my pre-Peavey lifetime we once sold a console to a studio in the main beam of a strong AM tower (960 kHz). My console grounds became much more obsessive over-night.

JR

 

[abbey road d enfer]

But there is always some distance between pin 1 and the chassis. Albeit small, it's inductance creates a resistance over which RF can be radiated. Yes?

That's why some insist on having the metal shell jumpered to pin 1 in plugs, as well as on the socket. Also check the Neutrik EMC series. However, this necessary only in critical cases, as others have mentioned.

 

[arnyk]

(In the 1930s some audio studios were right UNDER the transmitter tower, as transmitters increased from 5,000W to 50,000W. More obsessive grounding is found in some vintage gear.)

Much more recently than that. In the mid 60s a friend who had a 1st class commercial license took me along to visit  a client who turned out to be WHFI (Birmingham, MI).  The station had fallen on hard times and relocated their entire operation including business office and origination studios to a building at literally the base of the antenna. So we pull in and there is the live DJ for that time slot spinning LPs and reading commercials while my friend did his routine on-air checks.

I suspect that unless it has been made illegal, it can happen to this day.

 

[ricardo]

But there is always some distance between pin 1 and the chassis. Albeit small, it's inductance creates a resistance over which RF can be radiated. Yes?

If yes, then connecting anything at pin 1 (as opposed to at the chassis even though it's only maybe 1 cm away) will only serve to distribute that RF further inside the enclosure?

If pin 1 is properly decoupled, you can connect anything you like to it inside the box.  The evil is RF currents flowing along a significant path inside the box.

That's why I (and other serious mike designers like Dip. Ing Wuttke) prefer the caps AT the pins ... and the inductors/beads after that.

Read the Neutrik EMC plug info which explains important points.

In the 60's , most competent mike makers got to grips with MW transmitters.

In the 70's & 80's, the bogey man was thyristor theatre lighting.  The BBC once sent me a small crate of Calrec mikes (the entire spare stock for the N of England) for modification after I sent them a modified Calrec.  It was the only mike which didn't buzz in their new TV studio.

Today, mobile phones are the biggest danger.  There are studios (and hospitals) which insist you leave your phone at reception cos there is VERY expensive gear which will announce to all & sundry that you have received a TXT message.

Doing RFI work on microphones isn't straightforward.  One requirement is a mike preamp with headphones & metering which is 100% immune to RFI with 100m of cable and 200R at the other end. 

But certain mike circuits are also kinder to less than perfect preamps under RFI

 

[arnyk]

The secret ingredient which isn't shown is how their PCB is laid out .. tracking, ground plane, connections to chassis etc.

True. Some of this can be inferred from the parts layout, but this is an area where counter-intuitive designs can abound.

Here's an overview of preamp design from one of the leading mic preamp chip suppliers:

http://www.thatcorp.com/datashts/AES129_Designing_Mic_Preamps.pdf

More detailed documents, some covering the same info:

http://www.thatcorp.com/datashts/dn140.pdf

http://www.thatcorp.com/datashts/More_Analog_Secrets.pdf

In some areas of audio circuitry evaluation boards and reference designs that include details about board layout are generally available, but for mic preamps, not so much it seems.


If the project entails building a number of channels of the preamp, then it may make sense to  pick up an relatively inexpensive classic well-protected mic preamp from eBay or other used gear source in order to test, modify, and inspect its wiring and grounding.

Two preamps that are usually available and have inputs that are well-protected from EMI and RFI are the Symmetrix SX 202  and previously mentioned Rane MS1.



No parts layout or ground plane designs, but there are fairly detailed explanations and design equatioins.



You can immediately tell if a maker understand low noise design by how they lay PCB tracks.

But look at the Chassis symbols on their schematic.  They are emphasizing that p1 on the XLRs and the caps go DIRECTLY to chassis AT the XLRs .. not via ground planes, extra leads bla bla
[/quote]

 

[ricardo]

You can immediately tell if a maker understand low noise design by how they lay PCB tracks.

If the project entails building a number of channels of the preamp, then it may make sense to  pick up an relatively inexpensive classic well-protected mic preamp from eBay or other used gear source in order to test, modify, and inspect its wiring and grounding.

Two preamps that are usually available and have inputs that are well-protected from EMI and RFI are the Symmetrix SX 202  and previously mentioned Rane MS1.

I pontificate on this issue in LNprimer.doc in my Yahoo MicBuilders Files .. mentioning the SX202 and its faults, discussing PCB design but no sample PCBs.  8) You have to join.

It is likely the Rane is good from looking at the arrangement of the components so their PCB is worthy of study.

 

[arnyk]

You can immediately tell if a maker understand low noise design by how they lay PCB tracks.

SX 202 Schematic here:  https://www.gearslutz.com/board/attachments/geekslutz-forum/463216d1429584560t-schematic-resistor-labeling-question-symetrics_sx202_preamp_sch.pdf_1.jpg  or http://www.symetrix.co/kb/SX202_sch.pdf

SX 202 circuit component side picture here:  https://www.gearslutz.com/board/attachments/diy-electronic-build-refurbishment-photo-diaries/52807d1205418057-symetrix-sx202-mod-sx202-board.jpg

SX 202 circuit board land side picture here:

(with some mods)  http://montemcguire.com/images/sx202-gnd.jpg