Mic input transformer grounding questions

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

AusTex64

Well-known member
Joined
Jun 3, 2013
Messages
525
I've noticed on many of their mic input transformers Jensen and Cinemag have separate wires for the Faraday shield between primary and secondary windings, and a wire for the case/can.  Jensen and Cinemag say to tie the Faraday shield to signal ground and the can wire to chassis ground. What's the point of this? Pin 1 adherence? What are the advantages? Inquiring minds want to know.

I also note that the ancients usually tied the Faraday shield to chassis ground (RCA using UTC transformers, for example), and their stuff sounded pretty good and measured well. If the transformer only has a wire for Faraday shield (no can wire) is it best to tie that to signal ground or chassis ground? The can usually mounts with screws or a clamp to the metal chassis, so it manages to get grounded without a wire.

Thank you in advance for wisdom.
 
If you connect the external shield to the internal shield, any noise shunted by it would also be potentially injected inside. So case goes to chassis and internal shield should go to the local quiet audio 0V reference.
 
On the LL1538 Lundahl data sheet "pin8" is connected to ground. I believe that's the signal ground?

http://www.lundahl.se/wp-content/uploads/datasheets/1538_8xl.pdf
 
Actually for a mic transformer that is already inside a grounded enclosure, I'm not sure if there is a huge advantage to connect the mu-metal case to anything since it's really protecting from electromagnetic interference and not so much electrostatic. Also there are a number of Jensen mic transformer datasheets that show dual internal screens in the schematic with the primary screen connected to chassis ground and the secondary screen connected to local ground. This is supposed to yield the best possible CMRR.
 
AusTex64 said:
I've noticed on many of their mic input transformers Jensen and Cinemag have separate wires for the Faraday shield between primary and secondary windings, and a wire for the case/can.  Jensen and Cinemag say to tie the Faraday shield to signal ground and the can wire to chassis ground. What's the point of this? Pin 1 adherence? What are the advantages? Inquiring minds want to know.
You need to analyse how HF parasitics are dealt with. Let's say you have a common-mode spike at the primary. This is capacitively coupled to the case, to the electrostatic shield, to the secondary and to the magnetic core.
Connecting the electrostatic shield to a "ground" isolates the secondary. Now the spike is dumped where? You certainly don't want it to pollute the audio ground. Dumping it to "earth" (Pin 1 gnd/chassis) is the most logical thing to do. But the "earth" conduction path is far from perfect, so there is still a fraction of the spike that is coupled to the secondary. In order to minimize the effect on the active circuitry, it needs to be dumped to the audio "ground".
So there are two different paths with two different coupling capacitances.
I would guess Jensen/Cinemag figured out the capacitance from primary to case/core is higher than that from primary to electrostatic shield, so they chose to "earth" the former and "ground" the latter.
I must say I don't completely agree with their recommandation on a global basis.

I also note that the ancients usually tied the Faraday shield to chassis ground (RCA using UTC transformers, for example),
  That's what I would instinctively do.

  and their stuff sounded pretty good and measured well.
  Well I would say in normal conditions (but what's "normal" today with all the EMI and RF pollution?) it shouldn't make a difference in terms of response.

  If the transformer only has a wire for Faraday shield (no can wire) is it best to tie that to signal ground or chassis ground? The can usually mounts with screws or a clamp to the metal chassis, so it manages to get grounded without a wire.
Without any further analysis of the transformer's specificities, I don't know; measuring the leakage capacitance between pri and shield, and between pri and case would give an indication. The highest one should go to "earth".
 
Thanks all for your replies.

I decided to email Jensen with this question: 

<I’ve read this document carefully: https://jensen-transformers.com/wp-content/uploads/2014/09/Audio-Transformers-Chapter.pdf . I’m still trying to understand the advantage of wiring the case/can of a mic input transformer to “chassis ground”, and connecting the Faraday/electrostatic shield between primary and secondary windings to “signal ground”. Why not just wire both to chassis ground?>

This is the reply from Jensen, written by Bill Whitlock no less:

<Thanks for writing Jensen. Although not explicitly explained in the book chapter, the reasoning behind the different grounding strategies is fairly straight-forward. The Faraday shield has a substantial capacitance to each winding of the transformer. Therefore, any ground-voltage-difference (GVD) between the shield and the zero-volt reference ground for the amplifier it feeds will be coupled through this capacitance ... appearing as noise in the amplifier's signal output. Since even the tiniest voltage differences around this reference point can cause noise issues, it's also critically important that the grounded end of the transformer winding be connected directly at this same point. Remember that even an inch of wire or PCB trace can have noise across it (until we have room-temperature-superconducting copper, at least).  On the other hand, the housing of the transformer has very little capacitance to any of the windings. Therefore, it matters much, much less where it is grounded ... and, for convenience when a metal chassis is involved, grounding to the chassis (which automatically happens in normal mounting) is both convenient and effective at blocking possible nearby electric fields. Now, if the part is a PCB mount type, the can and Faraday shield pin may conveniently be simply tied together to reference ground ... but there is a slight risk that something metallic and connected to the chassis or other ground point may touch the can, and contaminate the reference ground by connecting it to another ground.

Of course, the bigger point here is that ground is not ground is not ground ... a "complication" that most engineers would rather not think about. But in low-level, low-noise circuitry, each point in the ground system (both inside and outside in a interconnected system) must be considered in terms of current flows and the tiny voltage differences that occur in wiring. Although the resistances (often only milli-ohms) and currents (often only a milli-amp) are small, the voltages can easily be tens of micro-volts ... enough to spoil an otherwise near-theoretical system noise floor.>

 
AusTex64 said:
"the reasoning behind the different grounding strategies is fairly straight-forward. The Faraday shield has a substantial capacitance to each winding of the transformer. Therefore, any ground-voltage-difference (GVD) between the shield and the zero-volt reference ground for the amplifier it feeds will be coupled through this capacitance... On the other hand, the housing of the transformer has very little capacitance to any of the windings."
So, your earlier statement "Jensen and Cinemag say to tie the Faraday shield to signal ground and the can wire to chassis ground. "was erroneous. It's the contrary, which makes sense to me. I guess it comes from the way they have drawn the ground and earth symbols, that apparently are the exact opposite of common usage.
 
I'm reading Jensen's response and AusTex's original statement about the guidance from Cinemag/Jensen as the same: Faraday shield to signal ground, can to chassis ground.

Now, his observation that "the ancients" connected the Faraday shield to chassis ground runs opposed to this guidance...

Either way, it's all great info. Thanks for sharing the note from Jensen.
 
rackmonkey said:
I'm reading Jensen's response and AusTex's original statement about the guidance from Cinemag/Jensen as the same: Faraday shield to signal ground, can to chassis ground.
I must admit jensen's response is quite puzzling. They admittedly recommend dumping EMI/RFI into the audio "ground", which is in utter disagreement with the general consensus about the so-called Pin 1 issue.

Now, his observation that "the ancients" connected the Faraday shield to chassis ground runs opposed to this guidance...
And I believe they were right. It's somewhat upsetting being in open contradiction with such a respected figure as Bill Whitlock. Further analysis is needed. On one hand, dumping EMI/RFI into the "earth/chassis" is a brute force answer, that works as much as the earth is solid, on the other hand, making the electrostatic shield equipotential to the audio "ground" is a viable differential approach. Both have their merits, that depend on how strong are the various "grounds" and how susceptible the circuits are. I would think tubes would vouch a more radical answer than solid-state.
 
Without the Faraday shield, capacitance between primary and secondary will couple common mode noise across the load. With the Faraday shield, yes it will shunt it into the audio ground, but it won't be across the load. And as Whitlock said, because of the capacitance between the shield and secondary it must be connected to the amplifier ground and not the chassis.

At least that's my interpretation of everything discussed so far.
 
go back and read Whitlock's response more carefully.

what is the purpose of the "Faraday" shield in the input transformer?  It is not about noise. It is about lowering capacitance between primary and secondary. The can takes care of the noise. The E Shield takes care of pri/sec capacitance.

The E shield between pri/sec is like a highway to the inside of the transformer, can or no can.  So Whitlock explains that any stray noise tied to the e shield will be injected into the transformer windings, thus the concern that they want to ground the E shield at a low noise point.

you are erroneously associating the screen with  noise reduction instead of pri/sec capacitance.

now with power transformers, the E shield is a whole new ball game, now we are talkin noise suppression.

somebody who travels coast to coast to do lectures at AES conventions on the subject of ground noise probably has there stuff together, you would think?

How much of that is Deane Jensen talkin? probably a  sizable chunk,



 
CJ said:
what is the purpose of the "Faraday" shield in the input transformer?  It is not about noise. It is about lowering capacitance between primary and secondary.
Right but I thought that the reason for "lowering" the capacitance was because it's coupling common mode noise. So the capacitance allows high frequency noise to bypass the transformer and just go straight through as common mode noise and thus it ruins CMRR. So the electrostatic / Faraday / E shield shunts that noise to ground. Then it's not across the load and so it's not amplified.
 
squarewave said:
Without the Faraday shield, capacitance between primary and secondary will couple common mode noise across the load.
That is correct and that's what we want to correct.

  With the Faraday shield, yes it will shunt it into the audio ground, but it won't be across the load.
  If it was a perfect shield, it would do two things: it would create an equipotential barrier between pri and sec that would eliminate couplling of CM noise to the sec AND it would dump CM interference ino the ground. But it's not a perfect shield, because it's often constituted with an additional winding, that has imperfect coverage and an intrinicd inductance that makes its grounding questionable.

And as Whitlock said, because of the capacitance between the shield and secondary it must be connected to the amplifier ground and not the chassis.
Considering all the various factors in the equation, I think the answer is not that simple. And in any case it is contrary to the EMI/RFI recommandations, that all interference should be dumped to chassis/earth before it reaches the circuits and pollutes the audio reference (analog ground).
 
So the best solution would be to have two separate shields, one on primary winding  connected to the chassis, and second one on secondary winding connected to the reference ("signal ground") of the secondary winding.  8) . 
 
CJ said:
go back and read Whitlock's response more carefully.

what is the purpose of the "Faraday" shield in the input transformer?  It is not about noise. It is about lowering capacitance between primary and secondary.
  Actually, it doesn't change anything about the pri/sec capacitance. It is a mathematical view of a Y/Delta transform. The reduction in inter-capacitance is effective only as much as the connection of the screen to a strong equipotential node.
So the analysis based on how effective the shield is to dump interference to ground is as much legitimate.

  The can takes care of the noise.
Actually it takes care of electrostatic coupling from other parts of the build AND it's primary role is magnetic shielding.

The E Shield takes care of pri/sec capacitance.
More correctly, it takes care of electrostatic coupling between pri and sec, which makes a difference.

The E shield between pri/sec is like a highway to the inside of the transformer, can or no can.
  Actually, if it was not grounded, it would do nothing in that respect.

  So Whitlock explains that any stray noise tied to the e shield will be injected into the transformer windings, thus the concern that they want to ground the E shield at a low noise point.
  There is a balance to be made between reduction of differential voltage attaining the input stage and correlated pollution of the audio ground. I'm not sure the answer is the same in all circumstances. An individual case study must be done before bringing a judgment.

  you are erroneously associating the screen with  noise reduction instead of pri/sec capacitance.
  I think you are erroneously taking these subjects as separate, when they  are actually two sides of the same coin (electrostatic coupling).

  somebody who travels coast to coast to do lectures at AES conventions on the subject of ground noise probably has there stuff together, you would think?
  Bill Whitlock is like an evangelist; he has to bring simple answers for simple persons. He cannot let doubt install itself in a lecture that is necessarily too short to actually go over all the details.
 
moamps said:
So the best solution would be to have two separate shields, one on primary winding  connected to the chassis, and second one on secondary winding connected to the reference ("signal ground") of the secondary winding.  8) .
You're absolutely right! Most often there is not enough room in mic input xfmrs for an additional screen; also this decreases coupling betwen pri and sec, which is damageable to the frequency response.
 
moamps said:
So the best solution would be to have two separate shields, one on primary winding  connected to the chassis, and second one on secondary winding connected to the reference ("signal ground") of the secondary winding.  8) .

I haven't gone back to look, but I seem to recall the Jensen mic splitter transformers have separate shields per winding with 1:1 or 1:1:1 or 1:1:1:1 connections. 
 
Back
Top