Damping transformer ringing (Zobel network question)

[orphanaudio]

QE solid state gear isn’t 600 ohm. I don’t want to make it so it is.

Depends on what model QEE module you are working with.
{EDIT}:
Ok dug back through the thread and found one mention of a QEE EQ-312.... (could have saved a LOT of Q&A if mentioned at the top of the post).

QEE 300 & 400 series inductor eq,s are looking for a 9db loss at the input and optimized for 600 ohms to the eq nets. This was typically provided by our QEE-3436 bridging line input, strapped 5k : 600.
The AM-3/AM-4 type discrete amplifiers in these will be ok with direct unbalanced drive out if the signal will not be exposed to short circuits for any length of time (partially inserted patch cables for instance), but a 1:1 output (QEE-3041) is usually recommended for the most authentic tonality and certainly to maintain ability to handle output levels approaching +28dbm.
The QEE-3436 is built using High-Ni laminations that offer extremely low distortion, and is capable of dealing with incoming levels in excess of +30db, even more if strapped for 20k : 600 (appx 15db loss).

(I'll leave the rest of this post in place since it still applies.)

With respect to original console builds from the late 60's to the mid 70's (and quite a few custom desks that were built well into the 80's),.. all our discrete transistor, inductor eqs are expecting a 600 ohms entering the eq nets (the bridging input xfrm may present a 5k impedance to the incoming source, but it exits as 600) and are typically linked to a 1:1 600 ohm output xfrm wherever that signal exits the console. You can of course modify the circuitry if you have the need, but you also have to account for the depressed levels that QEE inductor eq,s require to present unity gain at the output.

2k2 is a very typical secondary load for driving 600 ohms, especially if the transformer exhibits excessive ringing/distortion, but thats not a hard/fast rule as many contributors to this thread have illustrated. Selecting the right transformer for the application is of course the first step, but there should typically be less than a few 10th,s of a db loss if the terminating load is correctly sized and the driving amplifier is capable of supplying adequate current.

 

[Gold]

That’s good info Ken. I’m not going for historical authenticity. I used a pair of MM312S with UTC LS30’s as input transformers and an impedance balanced output for over 20 years in my mastering rig.

I had a pair of QEE 3041. I’m not a fan. I thought they were a definite downgrade from an impedance balanced out.

 

[orphanaudio]

That’s good info Ken. I’m not going for historical authenticity. I used a pair of MM312S with UTC LS30’s as input transformers and an impedance balanced output for over 20 years in my mastering rig.

I had a pair of QEE 3041. I’m not a fan. I thought they were a definite downgrade from an impedance balanced out.

Can you define how your impedance balanced output is implemented?

 

[Gold]

Can you define how your impedance balanced output is implemented?

The buildout resistor is 27R. On the output xlr connector I have a 27R from audio common on pin 3 to pin 1. If I wanted to be more correct I would include the 100uF cap but I didn't.

 

[ruffrecords]

The buildout resistor is 27R. On the output xlr connector I have a 27R from audio common on pin 3 to pin 1. If I wanted to be more correct I would include the 100uF cap but I didn't.

Pin 1 should always be connected directly to chassis and nowhere else. For an impedance balanced output you would connect your 27R between pin 3 and analog 0V and nowhere else.

Cheers

Ian

 

[Gold]

Pin 1 should always be connected directly to chassis and nowhere else. For an impedance balanced output you would connect your 27R between pin 3 and analog 0V and nowhere else.

Cheers

Ian

Pin 1 is directly connected to chassis. On the module Audio Common and Chassis are separated.

If signal + goes to pin 2 what pin should audio common go to? Certainly not pin 1. It has to go to pin 3. The 27R is to float audio common WRT chassis. Isn’t that what an impedance balanced output is?

 

[ccaudle]

The 27R is to float audio common WRT chassis. Isn’t that what an impedance balanced output is?

No, impedance balanced output is where the impedance of pin 2 (signal hot) to circuit reference/gnd and the impedance of pin 3 (cold/signal return) referred to signal reference/gnd are identical, but pin 3 is not driven by an opposite polarity signal. The alternate of impedance balanced is symmetrical drive where both pins 2 and 3 are driven by opposite polarity signals (which presumably will also have identical impedance from pin 2 to gnd and pin 3 to gnd if the driving circuitry is identical for both legs).

what pin should audio common go to?

Audio common doesn't go to any pin, it is the reference node for the circuit, which becomes the common mode voltage source as seen by the downstream receiver. An XLR connector has signal hot, signal cold or signal return (both terms have been used), and shield. Note that there is no pin defined as audio common.

On the module Audio Common and Chassis are separated.

You typically need one point where you tie the reference node to chassis so that you won't develop any appreciable common mode voltage between chassis and circuit gnd that could capacitively couple between them. Usually the output of the power supply is the appropriate connection point.

The buildout resistor is 27R...If I wanted to be more correct I would include the 100uF cap but I didn't.

The impedance of a 100uF cap at 60Hz is right around 27 Ohms, so by not including the cap your balance is off by 100%. Or 50%, I guess depending on which way you want to look at it.

 

[Gold]

No, impedance balanced output is where the impedance of pin 2 (signal hot) to circuit reference/gnd and the impedance of pin 3 (cold/signal return) referred to signal reference/gnd are identical, but pin 3 is not driven by an opposite polarity signal. The alternate of impedance balanced is symmetrical drive where both pins 2 and 3 are driven by opposite polarity signals (which presumably will also have identical impedance from pin 2 to gnd and

That's what I have. The same impedance from pin 2 and pin 3 to ground (ignoring the capacitor).

Audio common doesn't go to any pin, it is the reference node for the circuit, which becomes the common mode voltage source as seen by the downstream receiver. An XLR connector has signal hot, signal cold or signal return (both terms have been used), and shield. Note that there is no pin defined as audio common.

Huh? What is signal + supposed to reference to if it doesn't go to a pin? If it was a straight unbalanced out Audio Common would have to go to an output pin. It wouldn't work otherwise.

 

[ccaudle]

What is signal + supposed to reference to if it doesn't go to a pin?

I don't quite follow the terms in your question. I assume by "signal +" you mean the in polarity output, that connects to pin 2, which is variously labeled signal +, signal positive polarity, signal hot. Did I miss any?
As far as what it "references," in circuit engineering the reference node refers to the circuit node that is used to measure voltages against, usually colloquially called "ground" but that term is somewhat ambiguous.

That's what I have. The same impedance from pin 2 and pin 3 to ground

But you said it connected to the shield pin. The resistor should specifically connect to the reference node of the output circuit close to the output driver. You want the 0-signal voltage on the circuit side of the output isolation resistor and the voltage on the "inside" of the pin 3 resistor to be literally the same. If you put a differential 'scope probe between pin 1 of the output connector and the circuit reference (aka "ground") node with a downstream device connected, the voltage will not be 0 because of shield currents flowing through pin 1 to the chassis. That voltage is indistinguishable from signal voltage to the downstream differential receiver, so if that pin 1 noise is what is connected to the resistor to pin 3, even with a perfect diff input you would still have noise that could not be rejected by the diff input because it violates the design assumption that any interfering noise will be common mode on both the hot and cold/return signals.

If it was a straight unbalanced out Audio Common would have to go to an output pin.

Why would you put an unbalanced output on an XLR when you could make an impedance balanced output? That would just be a waste of a connector.

 

[Gold]

I don't quite follow the terms in your question.

I think there is some confusion. I don’t have a question. You and Ian said the impedance balanced output I made is wrong. I’ve been explaining what I did.

I am strictly talking about making the unbalanced output into an impedance balanced output.

But you said it connected to the shield pin. The resistor should specifically connect to the reference node of the output circuit close to the output driver.

I said audio common connected to chassis pin 1 via a 27R.

Why would you put an unbalanced output on an XLR when you could make an impedance balanced output? That would just be a waste of a connector.

Now I’m really confused. This whole thing is about an impedance balanced output.

You said audio common shouldn’t go to an output pin. Let’s pretend the output connector is a TS that is isolated from the chassis. What goes to S of the TS?

 

[abbey road d enfer]

Huh? What is signal + supposed to reference to if it doesn't go to a pin?

Signal + is referenced to signal -. In other words, signal on Pin 2 is referenced to signal on Pin 3.
Internally, signal + is referenced to "ground"; that's why signal - must also be referenced to "ground". The balancing network (typ. R and C in series) is to be placed between Pin 3 and "ground".

 

[Gold]

The balancing network (typ. R and C in series) is to be placed between Pin 3 and "ground".

It’s an unbalanced output. There is signal + and audio common. Audio common goes to pin 3. I have the 27R between pin3 and Chassis.

 

[ccaudle]

I don’t have a question.

You wrote a statement which was punctuated with a question mark, so I interpreted that as a question.

I said audio common connected to chassis pin 1 via a 27R.

Wait, does audio common connect to pin 3, and pin 3 also connects to pin 1 with a resistor?
That would give you the worst of all worlds, you both have an unbalanced connection, and you have injected shield noise directly into your audio common.

You said audio common shouldn’t go to an output pin.

Yes, in the context of balanced connections on XLR connectors that is correct, none of the XLR pins should connect directly to audio common.

Let’s pretend the output connector is a TS that is isolated from the chassis.

The "S" in a TS connector is "sleeve" or "shield," and a balanced connection has two signal connections inside a shield, so not really applicable to the conversation about balanced connections on XLR connectors.
A TS isolated from chassis would be asking for EMI problems, the shield should be bonded directly to chassis.
A TRS connector would be the equivalent of an XLR connector in phone plug format.

If you wanted an unbalanced connection for some reason, and decided to use a TS connector for that, the S would be screwed directly to the chassis (not isolated), and that point on the chassis would become the common chassis to audio ground connection point.

But the original context was balanced outputs on XLR, so the best plan would be impedance balanced output, which would have a second isolation resistor and capacitor right next to the ones connected to the single ended output driver, with the connection of the second resistor and capacitor being to circuit ground right next to the output driver circuit, and a pair of wires running from XLR connector pins 2 and 3 to those components. The connection is "impedance balanced" because if you compare the impedance from pin 2 to the circuit ground and the impedance from pin 3 to circuit ground they will be the same.

 

[Gold]

Signal + is referenced to signal -. In other words, signal on Pin 2 is referenced to signal on Pin 3.
Internally, signal + is referenced to "ground"; that's why signal - must also be referenced to "ground". The balancing network (typ. R and C in series) is to be placed between Pin 3 and "ground".

Wait, does audio common connect to pin 3, and pin 3 also connects to pin 1 with a resistor?
That would give you the worst of all worlds, you both have an unbalanced connection, and you have injected shield noise directly into your audio common.

It looks like you and Abbey disagree.

 

[ccaudle]

It looks like you and Abbey disagree.

I assure you we do not. I think part of the confusion may be that it is more difficult to describe verbally. Tonight or tomorrow I will draw a picture, that should avoid some of the confusion.

 

[ruffrecords]

It’s an unbalanced output. There is signal + and audio common. Audio common goes to pin 3. I have the 27R between pin3 and Chassis.

There is definitely some confusion. You want to make an impedance balanced output from an unbalanced signal. Audio common should not go to pin 3. It should go to pin 3 via a 27R resistor. Pin 1 should go directly to chassis at the XLR. At one point in your system Audio common will be connected to chassis (usually at or near the power supply).

Edit: Here is a simple picture to illustrate the difference.

Cheers

Ian

 

[Gold]

There is definitely some confusion. You want to make an impedance balanced output from an unbalanced signal. Audio common should not go to pin 3. It should go to pin 3 via a 27R resistor.

That’s the confusion. I’ve seen it done the way I did it but have never seen it done the way you describe.

Pin 1 should go directly to chassis at the XLR. At one point in your system Audio common will be connected to chassis (usually at or near the power supply

It is that way. I’m not sure how the way I did the 27R is functionally different? Either way the receiving end pin 3 sees 27R to chassis.

 

[ccaudle]

I’ve seen it done the way I did it but have never seen it done the way you describe.

The way Ian described is the only thing that "impedance balanced" could refer to. The "balanced" in "impedance balanced" refers to the impedance from pin 2 to circuit reference being the same as the impedance from pin 3 to circuit reference.

Either way the receiving end pin 3 sees 27R to chassis.

That is just a short-hand approximation. You don't actually care about the impedance to chassis, you care about the impedance to your output driver reference node (aka "ground"). With the connection you describe pin 3 sees 27 Ohms to the chassis, then whatever resistance the chassis has to the point where the circuit ground connects to chassis, then whatever resistance (and inductance) is in the wire from the to the circuit ground, then whatever resistance is in the circuit ground up to the output driver section.
That extra resistance isn't much, and the added resistance is actually fairly trivial, but the important point is that whatever interference currents are on the cable shields now also "see" that connection back to the circuit ground. Ohms law shows that the current flowing across the chassis will generate a voltage proportional to the current and the impedance (v=i*z). Since that interference current is only going via the pin 3 connection, it shows up as a voltage on pin 3 but not on pin 2, i.e. it is differential, i.e. it is indistinguishable from the signal, so the downstream differential receiver has no way to reject that noise voltage which results from the noise currents flowing across the chassis.

If the resistor is connected to circuit ground right beside the output driver, then that noise current which flows across the chassis may still create a few mV of noise voltage, but if the circuit ground is connected to the chassis at only one point, then the entire circuit "rides" on that noise voltage, so the noise voltage at the downstream receiver will be exactly the same on pin 2 and pin 3, so the differential receiver can reject that noise (since it should amplify only the signal difference between pin 2 and pin 3).

And I see where you were confused, which is partly Abbey's fault as well:

The balancing network (typ. R and C in series) is to be placed between Pin 3 and "ground".

The quotations around "ground" in that statement were to call attention to the fact that the R+C must be between pin 3 and circuit ground. I often use the term reference node for the same thing, because the term "ground" is ambiguous, and misused in a lot of contexts.
Your description of your connection is that the resistor connected between pin 3 and pin 1. Pin 1 is not "ground," pin 1 is the cable shield pin, which should connect to the chassis as overall shield. The circuit reference node should tie to the chassis shield at one point, usually close to the power line protective earth.
Even though they eventually connect together, power line protective earth, shield, and circuit reference node are not the same thing, even though some people refer at various times to all three as "ground." They each have different functions, and they all have resistance and inductance. Though small, those connection details of resistance and inductance are important for audio. Only 1mV of noise is -57dBu. It does not take much current to generate 1mV across steel chassis, and maybe a couple of screws, probably a wire here or there. That difference of where you connect the resistor is the difference between a dead quiet connection, and an annoying hum that makes you want to through your gear out the window.

 

[Gold]

That is just a short-hand approximation. You don't actually care about the impedance to chassis, you care about the impedance to your output driver reference node (aka "ground").

The reference point for audio common is at the power supply output 0V. That 0V reference point (+28vdc-0V- -28vdc)connects to chassis at the mains inlet and nowhere else.

XLR pin 1 is connected to the xlr shell (chassis) and nowhere else. Mains inlet is rear panel left side. XLR’s are on the rear panel right side. I don’t imagine running a wire over to the mains inlet from pin 3 is any lower impedance then bonding pin 1 to the xlr shell.

 

[abbey road d enfer]

The reference point for audio common is at the power supply output 0V. That 0V reference point (+28vdc-0V- -28vdc)connects to chassis at the mains inlet and nowhere else.

XLR pin 1 is connected to the xlr shell (chassis) and nowhere else. Mains inlet is rear panel left side. XLR’s are on the rear panel right side. I don’t imagine running a wire over to the mains inlet from pin 3 is any lower impedance then bonding pin 1 to the xlr shell.I

in this specific case, seen from a DMM's POV, Pin 1, chassis, "ground" and reference point are identical, but remember that reference point and PE are connected together only for safety reasons.
In terms of audio performance, there is no obligation to connect "ground" to "earth".
One can imagine an implementation where the reference point would have no direct galvanic connection to Pin 1.
No battery-powered apparatus actually needs galvanic connection between Pins 2-3 and Pin 1.