Help Design an unbalanced consumer Line level to Balanced interface

[richiyobs]

i tested the U pad on guitar
compared it with Hi-z input on soundcard
great Fat sound and better transient and as you said too much attenuation, it needs more than 40dB gain and no noise
first played altering the shunt resistor to get the proper level for mic preamp, tought it was the right way to match impedance but still not satisfied
instead varying the series resistor, if i don t go wrong can change a bit the freq response
so i m very courius about the right solution for right Hi-z mod
i like that design! simple and doable ... searched on line for pads topologies but nothign found on the subject, only suggested and calculated values according to attenuation you need

best

 

[Bo Deadly]

Your text says shield connected at both ends but drawing show NC shield at TS end.

When I said "What you are advocating is", I was referring to MisterCMRRs drawing in post #44 which shows the shield connected at both ends.

But I think post #57 is the answer I was looking for which is that (with some liberal interpretation) for gear with a proper protected earth connection, leaving the shield disconnected at one end is probably a good idea (at least Rane seems to think it is). However, for gear without a proper protected earth power cable, the device is essentially "floating" and the common mode noise can completely overwhelm any common mode rejection. In this case, the shields should be connected at both ends to "wrangle" the grounds together.

However, note that this is just masking the real problem which is the device that does not have a proper protected earth. The proper solution, as indicated by MisterCMRR, is to simply fix that gear and give it a proper earth.

Also I think it is important to recognize that most modern gear sold within at least the last 10 years almost ALWAYS has an isolated SMPS power supply. It seems that even though most of this gear does not have a proper protected earth (frequently it's just a plastic enclosure), common mode noise does not seem to be a problem in these devices. This is the scenario where I might wonder if connecting shields is a good idea or not.

My guess would be that for isolated SMPS without protected earth, the shields should be connected at both ends because SMPS are so good now that leakage currents are just not a problem. As long as there's a fat cap between 0V and the other side of the SMPS transformer, switching noise passing through the SMPS transformer onto 0V should not be a problem. But for old crappy SMPS, I could easily see it being worth it to leave the shield unconnected at one end. And for anything that has a proper protected earth, I would also guess "no", only connect at one end.

Of course there are exceptions. One might be a tower PC with a bunch of fans, whirling LEDs and other gobbledygook, ... In this case, there's probably no hope for any USB audio interface connected to it.

Unfortunately it's simply not clear what the best option is in every case. There really should be a little battery operated device with a cable gland and a bunch of different cables to connect to whatever device and cable that analyzes the lines with and without the shield connected and if the output is cross coupled and what the impedance of each line end is and so on. That would be a handy device.

 

[JohnRoberts]

Grounding shields is a well discussed topic. In an ideal world both ends should be grounded. For those of us operating in the real world results vary.

JR

 

[MisterCMRR]

My only defense of the 26 dB figure, is that 22 dB is the attenuation of the Jensen JT-DB-E "direct box" transformer, which is at the heart of what are widely considered the world's finest high-performance direct box products (spec sheet at https://www.jensen-transformers.com/wp-content/uploads/2014/08/jt-db-e.pdf). Since, generally, synths are capable of substantially larger signals than guitars, so a bit more attenuation seems appropriate. There is also a mathematical tradeoff that links input vs output impedances and the loss of a passive attenuator. The design I showed compromises impedances. It's 10.5 kΩ input impedance should technically be at least 22 kΩ to qualify as a "consumer" load to the synth. Its 476 Ω output impedance should technically be the industry norm of 150 Ω or 200 Ω. But very few real microphones hit that target. The Shure SM57, for example, is closer to 350 Ω over most of its bandwidth (see its equivalent circuit in Jensen AN005 at https://www.jensen-transformers.com/wp-content/uploads/2014/08/jt-db-e.pdf). Making an attenuator pad with less attenuation forces these input and output impedances to be farther compromised. So the choice of 26 dB is just what I consider a reasonable compromise.

 

[MisterCMRR]

Leakage currents (those that unintentionally flow from "live" power conductors to the gear's internal "ground" system play a role in these noise problems - as well as the safety standard organizations that put limits on leakage current. Equipment that is agency-approved and sold with a 2-prong power connection is quite safe - and "proper" in that sense.

The severity of "ground loop" issues with an unbalanced interface between two pieces of grounded (3-prong) gear is primarily a function of how physically distant their AC power connections are. The length of the intervening safety-ground (PE for Europeans) conductor in the building wiring (and subtleties of its construction) determines the magnitude of the ground voltage difference between 3rd prongs at the two outlets. Generally, noise will be even worse if they're on different branch circuits. This magnetically-induced voltage difference can be quite high 100 mV or more, making unbalanced interfaces completely unusable. Using 2-prong power connections neatly avoids this problem and, I'm convinced, is why manufacturers are willing to pay the additional cost of safety-rated components, double insulation, thermal switches, etc. in order to use a 2-prong AC power connection.

Although 2-prong AC connections avoid the worst noise issues, the remaining "leakage" currents are not insignificant either. In the US, this current can be as high as 0.75 mA (UL compliant) at 60 Hz - enough to feel a tingle if it flows through you! This current will flow in the shield of an unbalanced interconnect to get to ground (actually via safety-ground/PE to get to Neutral). As it does, Ohm's law dictates that a voltage drop will occur. Consider that the shield resistance of 25 feet of unbalanced cable could easily be 1 Ω. If leakage current is only 0.3 mA, the drop is 300 µV and that noise is added directly to the signal. Assuming a -10 dBV standard consumer reference signal level, that makes signal-to-noise ratio only 60 dB ... not exactly "studio quality."

For higher-power gear, over say 300 W rated consumption, the cost of 2-prong operation becomes high enough that most gear opts for a 3-prong grounding AC connection to comply with regulations. "Cheating" the safety-ground on a 3-prong cord would allow leakage currents as high as 3.5 mA (you'll definitely want to let go if it travels through you!!) - and in some cases 5 mA.

 

[Whoops]

Making an attenuator pad with less attenuation forces these input and output impedances to be farther compromised. So the choice of 26 dB is just what I consider a reasonable compromise.

Thank you so much Bill, that's what I though the problem might be, in altering the input and output impedances. Thanks for explaining it so well.
So -26dBs it is, for sure the Mic Preamp will not clip and in the situations the level is too low, just bring the Gain up.

 

[Bo Deadly]

The severity of "ground loop" issues with an unbalanced interface between two pieces of grounded (3-prong) gear is primarily a function of how physically distant their AC power connections are. ... This magnetically-induced voltage difference can be quite high 100 mV or more, making unbalanced interfaces completely unusable. Using 2-prong power connections neatly avoids this problem and, I'm convinced, is why manufacturers are willing to pay the additional cost of safety-rated components, double insulation, thermal switches, etc. in order to use a 2-prong AC power connection.

Thinking about this more, I have to wonder if new gear should just use an ACDC SMPS and go entirely 2-prong power.

SMPS have been getting a lot of traction and discussion here. Very low noise (mains free) supplies can be made with off-the-shelf SMPS which are already certified for safety. It is easy to use an ACDC SMPS with no mains exposed inside - just feed the power cable through a cable-gland, add spades and screw it down onto the terminal block of the SMPS. There's really no safety reason for the chassis to be earth protected.

Now, as you just explained, earth ground is just a liability. It's just a long wire into which currents can be induced by the magnetic fields which are ever-present and varied depending on nearby power cables in walls, AC compressors, stage lights, etc. An isolated ACDC SMPS will almost completely eliminate that noise at low frequencies and the high frequencies can be shunted or filtered easily with inductors, caps and, my favorite, capacitance multipliers.

As long as all I/O is balanced (or at least impedance balanced would be enough for the more cost effective scenarios), there's simply no need for the safety ground / protected earth in this case. And this is supported by the bulk of new gear using just a DC barrel jack. I thought it was just because they were being cheap or because the enclosure was plastic but maybe it's really because the third prong is not only not needed, it's a liability.

Although one exception would be guitar gear where the guitar needs to be connected to safety earth / protected earth. Guitarists are a special case as usual.

 

[Newmarket]

Thinking about this more, I have to wonder if new gear should just use an ACDC SMPS and go entirely 2-prong power.

...

Although one exception would be guitar gear where the guitar needs to be connected to safety earth / protected earth. Guitarists are a special case as usual.

What's the reasoning for the guitar related exception ?

 

[JohnRoberts]

Thinking about this more, I have to wonder if new gear should just use an ACDC SMPS and go entirely 2-prong power.

many already do, and make that universal mains voltage input for international sales.

SMPS have been getting a lot of traction and discussion here. Very low noise (mains free) supplies can be made with off-the-shelf SMPS which are already certified for safety. It is easy to use an ACDC SMPS with no mains exposed inside - just feed the power cable through a cable-gland, add spades and screw it down onto the terminal block of the SMPS. There's really no safety reason for the chassis to be earth protected.

I have killed a lot of brain cells thinking about this over the decades. While working at Peavey we lost one guitar player to electrocution when a mis-wired (RPBG) outlet energized the chassis of his guitar amp. His amp was plugged into the rouge RPBG outlet, while his buddy's amp was plugged into the properly grounded outlet of the electric stove. While trading guitars with his friend he managed to hold both guitars one in each hand, imposing full mains voltage through his core, killing him.

Now, as you just explained, earth ground is just a liability. It's just a long wire into which currents can be induced by the magnetic fields which are ever-present and varied depending on nearby power cables in walls, AC compressors, stage lights, etc. An isolated ACDC SMPS will almost completely eliminate that noise at low frequencies and the high frequencies can be shunted or filtered easily with inductors, caps and, my favorite, capacitance multipliers.

I already shared one example where the safety ground became a local liability, there are others. Many musicians get stung by the ground path coming through the microphone fed from a console plugged into another branch circuit, when energized by a leaky old guitar amp. I am glad to say that modern gear is much safer than it was, but be careful when messing with old legacy gear.

As long as all I/O is balanced (or at least impedance balanced would be enough for the more cost effective scenarios), there's simply no need for the safety ground / protected earth in this case. And this is supported by the bulk of new gear using just a DC barrel jack. I thought it was just because they were being cheap or because the enclosure was plastic but maybe it's really because the third prong is not only not needed, it's a liability.

good luck getting UL to abandon chassis safety grounds. 2-wire SKU require double insulation for all mains conductors.

Although one exception would be guitar gear where the guitar needs to be connected to safety earth / protected earth. Guitarists are a special case as usual.

Guitarists are special. A couple more thoughts on that.

As I already shared we (Peavey) lost at least one guitarist to a rouge mis-wired outlet with a properly grounded, UL approved, guitar amp plugged in (UL sat with us in court and defended our design). In engineering there are always multiple ways to remedy any issue (raise the bridge or lower the water). One strategy that I have supported for years/decades is capacitor coupling the exposed metal parts on guitars to the jack ground, that is usually earth bonded through the guitar amp input jack. Such a coupling capacitor can be sized large enough to suppress hum noise, while small enough to limit current from mains voltage exposure to tolerable amounts. GFCI outlet trip at 6mA because currents less than that not considered a human safety risk.

I strongly advocate for use of GFCI outlets**** wherever guitar amps get plugged in... It wouldn't hurt for FOH consoles to use them too.

now for TMI- Years ago I developed a "muso saver" a bit of kit that musicians could carry with them for personal safety. I abandoned this SKU also for a couple reasons. In my judgement I would never get musicians to invest in safety gear, and secondly I would never get UL to approve my Gordian solution to break the EGC or safety ground path, when significant ground path leakage was detected. I used a standard GFCI outlet and a 3 pole relay to interrupt Line, neutral, and safety ground. This would NOT remedy the energized hazard, but would open up the safety ground path that could lead to death.

Another experiment I did was to hack into a GFCI outlet strip (like image below), and incorporate my cap coupled safety ground inside the strip. This would never pass UL because the safety ground path would not take out any fuse/breakers. If the stray current was coming from the device plugged into this GFCI that would trip removing the mains voltage, if the shock hazard was coming from some external source (like a FOH console) the cap coupled ground would protect the human by holding current to sub lethal levels. I sent a modified unit to respected guitar amp designer I used to work with and he confirmed that it was adequately quiet in use.

Be safe out there... FWIW I'm cheap but I have GFCI outlets installed all around my old 2 wire house.

JR

**** we can buy low cost GFCI protected outlet strips

 

[user 37518]

many already do, and make that universal mains voltage input for international sales.

I have killed a lot of brain cells thinking about this over the decades. While working at Peavey we lost one guitar player to electrocution when a mis-wired (RPBG) outlet energized the chassis of his guitar amp. His amp was plugged into the rouge RPBG outlet, while his buddy's amp was plugged into the properly grounded outlet of the electric stove. While trading guitars with his friend he managed to hold both guitars one in each hand, imposing full mains voltage through his core, killing him.

I already shared one example where the safety ground became a local liability, there are others. Many musicians get stung by the ground path coming through the microphone fed from a console plugged into another branch circuit, when energized by a leaky old guitar amp. I am glad to say that modern gear is much safer than it was, but be careful when messing with old legacy gear.

good luck getting UL to abandon chassis safety grounds. 2-wire SKU require double insulation for all mains conductors.

Guitarists are special. A couple more thoughts on that.

As I already shared we (Peavey) lost at least one guitarist to a rouge mis-wired outlet with a properly grounded, UL approved, guitar amp plugged in (UL sat with us in court and defended our design). In engineering there are always multiple ways to remedy any issue (raise the bridge or lower the water). One strategy that I have supported for years/decades is capacitor coupling the exposed metal parts on guitars to the jack ground, that is usually earth bonded through the guitar amp input jack. Such a coupling capacitor can be sized large enough to suppress hum noise, while small enough to limit current from mains voltage exposure to tolerable amounts. GFCI outlet trip at 6mA because currents less than that not considered a human safety risk.

I strongly advocate for use of GFCI outlets**** wherever guitar amps get plugged in... It wouldn't hurt for FOH consoles to use them too.

now for TMI- Years ago I developed a "muso saver" a bit of kit that musicians could carry with them for personal safety. I abandoned this SKU also for a couple reasons. In my judgement I would never get musicians to invest in safety gear, and secondly I would never get UL to approve my Gordian solution to break the EGC or safety ground path, when significant ground path leakage was detected. I used a standard GFCI outlet and a 3 pole relay to interrupt Line, neutral, and safety ground. This would NOT remedy the energized hazard, but would open up the safety ground path that could lead to death.

Another experiment I did was to hack into a GFCI outlet strip (like image below), and incorporate my cap coupled safety ground inside the strip. This would never pass UL because the safety ground path would not take out any fuse/breakers. If the stray current was coming from the device plugged into this GFCI that would trip removing the mains voltage, if the shock hazard was coming from some external source (like a FOH console) the cap coupled ground would protect the human by holding current to sub lethal levels. I sent a modified unit to respected guitar amp designer I used to work with and he confirmed that it was adequately quiet in use.

Be safe out there... FWIW I'm cheap but I have GFCI outlets installed all around my old 2 wire house.

JR

**** we can buy low cost GFCI protected outlet strips

The amount of performers/musicians killed by holding a guitar or holding a microphone is staggering.

This happened just a couple of months ago here in Mexico. I feel bad for her, gladly she didn't die.

 

[JohnRoberts]

The amount of performers/musicians killed by holding a guitar or holding a microphone is staggering.

I have been paying attention to this for decades and not a "huge" number of musicians reported actually killed.

Many have been shocked but short of a lethal dose. Many of those sub lethal shock events are not widely reported (and when, often misreported as electrocution, like the video posted).

In connection with my old day job I also paid close attention to clergy shocked in baptismal pools, etc.

Mains voltage can still be lethal so always be careful.

This happened just a couple of months ago here in Mexico. I feel bad for her, gladly she didn't die.

The good (less bad) news for musicians is that public events often have multiple witnesses and support staff who can assist musicians at risk.

JR

 

[user 37518]

I have been paying attention to this for decades and not a "huge" number of musicians reported actually killed.

Many have been shocked but short of a lethal dose. Many of those sub lethal shock events are not widely reported (and when, often misreported as electrocution, like the video posted).

In connection with my old day job I also paid close attention to clergy shocked in baptismal pools, etc.

I agree it is not a major cause of death, but the fact that it happens, and happens more often than one would expect indicates that it is a serious problem. Most are just "shameful" electrocutions like the lady in the video, but it is a serious problem with audio. I guess every singer knows what it feels to sing into a microphone and feel a "tingling" on the lips.

 

[Bo Deadly]

What's the reasoning for the guitar related exception ?

Because the guitar is unbalanced with a high inductance coil (which is usually poorly shielded) exposed to EMI along with all of the metal bits which includes the strings of course. So if 0V didn't have some relation earth, you would get noise from being bombarded with EMI all around us. The metal bits could charge up relative to earth and cause popping noises as you touch and release the strings.

But there is also a safety reason in that there are many opportunities for an uncontrolled voltage to contact that 0V, such as through a guitar amp or any of the other unbalanced guitar gear, in which case it might be good if there was a better path to earth than through your body and into the shield of a mic or some other bit of earthed metal you happen to touch.

 

[JohnRoberts]

I agree it is not a major cause of death, but the fact that it happens, and happens more often than one would expect indicates that it is a serious problem. Most are just "shameful" electrocutions like the lady in the video, but it is a serious problem with audio. I guess every singer knows what it feels to sing into a microphone and feel a "tingling" on the lips.

If she didn't die, she was not electrocuted, at least based on the American definition of the word "electrocuted".

1. To kill with electricity: a worker who was electrocuted by a high-tension wire.
2. To execute (a person sentenced to death) by means of electricity.

Because the guitar is unbalanced with a high inductance coil (which is usually poorly shielded) exposed to EMI along with all of the metal bits which includes the strings of course. So if 0V didn't have some relation earth, you would noise from being bombarded with EMI all around us. The metal bits could charge up relative to earth and cause popping noises as you touch and release the strings.

But there is also a safety reason in that there are many opportunities for an uncontrolled voltage to contact that 0V, such as but not limited to exposed metal, in which case it might be good if there was a better path to earth than through your body and into the shield of a mic or some other bit of earthed metal you happen to touch.

repeating myself... a capacitor in series between the guitar's exposed metal parts and the jack ground will provide adequate shielding without risk of lethal shock hazard. If static build up is an issue (I never heard reports of that) a megohm resistor could discharge the capacitor.

==

For another repeat, I advocate use of GFCI/RCD outlets provided for live stage backline, and in studio gear.

JR

 

[Bo Deadly]

repeating myself... a capacitor in series between the guitar's exposed metal parts and the jack ground will provide adequate shielding without risk of lethal shock hazard. If static build up is an issue (I never heard reports of that) a megohm resistor could discharge the capacitor.

Ok. But I don't see people re-writing their guitars like that. And that wouldn't protect the plug shell or jack plate.

 

[user 37518]

If she didn't die, she was not electrocuted, at least based on the American definition of the word "electrocuted".


repeating myself... a capacitor in series between the guitar's exposed metal parts and the jack ground will provide adequate shielding without risk of lethal shock hazard. If static build up is an issue (I never heard reports of that) a megohm resistor could discharge the capacitor.

==

For another repeat, I advocate use of GFCI/RCD outlets provided for live stage backline, and in studio gear.

JR

Ok John, sorry for not being anal enough, she received a non-deadly electric shock, which threw her to the ground and produced a traumatic experience for her. You know what I meant, a bit of sympathy for non-native English speakers wouldn't do any harm.

 

[JohnRoberts]

Ok John, sorry for not being anal enough, she received a non-deadly electric shock, which threw her to the ground and produced a traumatic experience for her. You know what I meant, a bit of sympathy for non-native English speakers wouldn't do any harm.

Sorry I was not trying to criticize you personally, if anything I am more critical of the Video headline that said "electrocuta". I don't speak Spanish but I suspect that means electrocuted.

News media in the US routinely overstates routine shock events as electrocutions. This is not a new complaint from me. Engineering is about precision but some people do find me offensive, sorry.

JR

PS; phase/polarity, buss, etc

 

[JohnRoberts]

Ok. But I don't see people re-writing their guitars like that. And that wouldn't protect the plug shell or jack plate.

People have, I recall a popular thread on Gearslutz years ago with a number of successful anecdotal reports.

The life saving benefit is from the strings not being bonded to the input jack ground.

JR

 

[user 37518]

Sorry I was not trying to criticize you personally, if anything I am more critical of the Video headline that said "electrocuta". I don't speak Spanish but I suspect that means electrocuted.

News media in the US routinely overstates routine shock events as electrocutions. This is not a new complaint from me. Engineering is about precision but some people do find me offensive, sorry.

JR

PS; phase/polarity, buss, etc

No worries, I guess that it was one of those "it's not the same through the internet" communication problems.

 

[Newmarket]

Because the guitar is unbalanced with a high inductance coil (which is usually poorly shielded) exposed to EMI along with all of the metal bits which includes the strings of course. So if 0V didn't have some relation earth, you would get noise from being bombarded with EMI all around us. The metal bits could charge up relative to earth and cause popping noises as you touch and release the strings.






But there is also a safety reason in that there are many opportunities for an uncontrolled voltage to contact that 0V, such as through a guitar amp or any of the other unbalanced guitar gear, in which case it might be good if there was a better path to earth than through your body and into the shield of a mic or some other bit of earthed metal you happen to touch.