Home Theater Forum and Systems banner

1 - 20 of 26 Posts

·
Premium Member
Joined
·
15,054 Posts
Discussion Starter · #1 ·
Its been discussed many times in different posts particularly on the Home Theater Shack and it seems there is some confusion on the subject. I will post it here as well just in case some newcomers here also dont understand the difference.

This post will help make this much clearer and hopefully with the input of others clear up the does and don'ts of proper connection to pro audio gear.
I do not want to make this sound confusing so I will spell this out in plain English without all the mumbo-jumbo.

Home consumer audio gear uses unbalanced analog audio connections that is designed for short distances and low cost. These connections are usually "rca" and have a shield and a single positive for the signal.
In the graph below you will see the signal (a sine wave in this case) and the shield connected to ground which means it has no signal on it.


A line level audio signal can be looked at as a low voltage signal and unbalanced audio signals are usually between 200mv and 1volt depending on weather it is a fixed or adjustable level.
Given this "lower" voltage signal, long distant runs are susceptible to interference and thus is only good for short runs of usually less than 20ft (50ft can be done if the cable is of better quality).

The problem with using home audio gear with professional audio gear is that the pro gear uses balanced input and output signals. Amps in particular either use a 1/4" Tip Ring Sleeve (TRS) connector
or an XLR connector

The audio signal used in these connectors is at a higher voltage ussually 1.5v and as seen in the graph below has twice the signal as an unbalanced signal.


There are two alternating signals -1.5v & +1.5v and the shield. The advantage of this connection is that when used for long distances the noise (interference) is not only canceled out by the alternating voltages but because its a higher signal voltage it can travel much longer distances without interference (as far as 1000ft) without a booster. The primary factor that allows longer connections with balanced lines has more to do with impedance than voltage. A balanced line has a low impedance (200 to 600 Ohms) while an unbalanced line typically has a 10k impedance or so.

Now here is the problem, when we try to use pro audio gear with balanced inputs connecting to the home audio equipment with unbalanced outputs you usually get incompatibility problems. These problems consisting of noise (usually 60Hz hum) or the signal coming into the pro gear being to low causing the noise floor to be raised when you boost the signal too much making the background hiss or hum to be amplified to the point that it becomes unusable sometimes referred to as signal to noise ratio.
On some amps or other pro gear you may have the ability to "short" the two connectors on the 1/4" TRS to make it a TS unbalanced input. The misconception of this option is that the voltage signal is also boosted but this is not normally the case and is still lower than what a balanced signal should have meaning that the amp at MAX on the level controls will not output the full amount of power it can. If your running this configuration and your amp levels are at MAX to achieve the levels you want I suspect that you have this problem as pro amps should not need to be run more than 75% of its max level control.

The fix to this issue is to use a line balancing transformer also called a DI box like the Samson S-Convert costing around $50


There are two types of DI bokes Active and Passive, Active generally is a cleaner and more quality way to go and costs more. It also requires phantom power or a 9v battery to operate.
Passive are simply a transformer with no extra power required.

DI's take the unbalanced signal and convert it to a balanced signal and raises the voltage. Adapters can be bought or made to go from XLR to 1/4" TRS as well as seen on the S-Convert there is only XLR outputs. There are many different models to chose from some having only TRS ins and outs others giving you two or more choices.

There are also several companies that make consumer grade high quality external amps that have proper unbalanced inputs so looking at that rout if your planning on getting an external amp can be a good option. The big plus to using some of the pro gear is that the price seems to be much lower per watts and can be found for under $300 new for two channel amps offering well over 200watts X2 like the Berhinger EP lineup.


Please feel free to comment in this thread and add more info.
 

·
Registered
Joined
·
55 Posts
Great contribution. Balanced and Unbalanced signals still getting some minds real confused. But after this thread things will be easier.

I do face this problem with some pro installations where they use domestic consumer receivers and want that signal to go to a pro mix board, or even straight to a pro amp, using the receiver as a surround decoder only, much cheaper than pro decoders...

The client can't understand that the signal is not the same...

For this situation, something like the Behringer DI800 works ok:

 

·
Registered
Joined
·
28 Posts
Time out - in a unbalanced system the shield or the shield ground combination has exactly the same amount of signal as the hot or center conductor! It would be a grouse violation of electrical principals if it didn't. Only the measurement technique of using the chassis/ground as your common reference point makes it appear as if the shield has no signal.

If you have a battery powered player and a battery powered amp/speaker and only connected the center conductor of the interconnect between them, nothing good would happen.

People often loose sight of the concept that whatever signal goes out the hot lead must return on the shield and/or whatever other (sometimes noisy) paths that it can find.
 

·
Premium Member
Joined
·
15,054 Posts
Discussion Starter · #4 ·
in a unbalanced system the shield or the shield ground combination has exactly the same amount of signal as the hot or center conductor! It would be a grouse violation of electrical principals if it didn't. Only the measurement technique of using the chassis/ground as your common reference point makes it appear as if the shield has no signal.
Very true and thanks for the clarification, however to make it understandable to someone who has no understanding of the whole process needed to explain this properly its better to just say it simply the way it was worded.
 

·
Registered
Joined
·
28 Posts
Sorry, didn't mean to let it get too complex too fast. I just get upset when people loose track of how important the signal return system is!
I also get upset when people think that a ground rod (or a new ground rod) is a place that noise can just be dumped into and like magic the noise disappears.
 

·
Registered
Joined
·
9 Posts
Great addition TonyVdb.

This is one of the topics I have the hardest time explaining. If you don't mind I may use it to teach some of the volunteers I work with.:clap:
 

·
Registered
Joined
·
63 Posts
That sine wave sure looks like a distorted triangle wave to me :)

In many cases to connect an unbalanced signal to a balanced input, you simply have to short the cold and GND together at the balanced input connector. No need for transformers. You need DI units when you have impedance mismatches that need to be accounted for, such as guitar pickups (which are high impedance).
 

·
Premium Member
Joined
·
15,054 Posts
Discussion Starter · #9 ·

·
Banned
Joined
·
66 Posts
Balanced and unbalanced devices can often be connected directly to each other, with proper cables and wiring. I second Syd’s recommendation of the Rane paper, with a simple to understand chart with pictures and all.

Here’s the link: http://rane.com/note110.html

There are two reasons you’d need a transformer or “bump box” between gear with differing input styles:

1. Level differences. Consumer stuff can mostly be considered to operate around -10dBu, pro gear around +4dBu. That’s an uncomfortable 14dB difference. If inputs and outputs are adjustable (mixer input gain, for example), you’re good to go, if not, you need a “gadget” to do it. Strictly, a transformer isn’t the best way to pick up 14dB. In fact, that’s a lot to do with a transformer, and usually only happens with mic input transformers which will saturate at line levels, and tend to have poor transient response (Jensen transformers excluded, of course, but they cost real money). Losing 14dB can be done with a resistive pad. Active bump boxes take care of this in both directions. In the original post, the term “DI Box” was used for a device to interface bal/unbal devices. A DI Box is specifically for connecting an instrument to a mixer. Instrument levels are different than consumer audio gear, and a DI can include special conditioning filters and has a pass through so a player can plug his axe into his amp, and it’s output is usually at mic level, all in all, not the best box for the job. The device in the picture on the first post is the correct item, though.

2. Noise immunity. Sometimes you need the common-mode noise rejection of a balanced input to kill ground loop noise, or noise picked up on a long cable. Sometimes. If the run is short, and you get hum, there are other things to do, mostly cost-free, that relate to grounding (another thread?). Frankly, after building and maintaining a bunch-o-studios, I can count on a few fingers the times in the last 35 years I absolutely had to have a device to fix this within a single studio (between rooms is different). These situations are solvable by proper grounding and connections. Again, see the Rane paper.

One note, sort of an aside. All cables designed for audio, balanced or unbalanced, include a shield to help reduce noise pickup. This shield works for capacitive coupling, or electromagnetic pickup (RF), but is essentially transparent to magnetically induced noise, line hum fields from power transformers. This fact is not commonly realized, but it’s physics. There are no practical magnetically shielded cables. It’s important to drive longer cables with low source impedance outputs to reduce the effects of all three types of noise coupling, and take advantage of a balanced input to cancel noise coupled in from the outside. Steel conduit does provide some magnetic shielding.

All balanced inputs are NOT created equal. Some have higher noise rejection ability than others, but nobody specs this (the spec is Common Mode Rejection Ratio, CMRR). Odd, because its the only reason to use a balanced connection, and the only thing not specified. The reason is, good CMRR is hard to design for, and it's not a single figure spec, but rather a dB ratio over a specified bandwidth, like 70dB from 0 - 100KHz for example. Manufacturers don't want to spend the money do to that, so they don't spec CMRR.

Jim
 

·
Registered
Joined
·
28 Posts
One note, sort of an aside. All cables designed for audio, balanced or unbalanced, include a shield to help reduce noise pickup. This shield works for capacitive coupling, or electromagnetic pickup (RF), but is essentially transparent to magnetically induced noise, line hum fields from power transformers. This fact is not commonly realized, but it’s physics. There are no practical magnetically shielded cables. It’s important to drive longer cables with low source impedance outputs to reduce the effects of all three types of noise coupling, and take advantage of a balanced input to cancel noise coupled in from the outside. Steel conduit does provide some magnetic shielding.
Jim
This is not always true! There are some cases when a co-ax or other un-balanced cable circuit can attenuate the induced field.
I have an old copy of Henry W. Ott's book "Noise Reduction Techniques in Electronic Systems".
He has examples of co-ax and/or unbalanced circuits that attenuate 0, 13, 27, 55 even 80 dB's. It really depends on how the circuit deals with this common mode interference. Of course the magnetic field goes right through the shield, but the same interference signal voltage is on the shield and the center conductor, some circuits can deal with this.
 

·
Banned
Joined
·
66 Posts
Well, interesting post, but the statement of mine you highlighted is true, which you confirmed in your post. The shield does nothing to prevent mag induction to the shielded wires inside.

From what you say about the Ott book, it all happens in the balanced input, which is essentially what I said. But the shield still has no effect on magnetic induction. Again, see the Rane paper, which outlines how unbalanced to balanced connections can take advantage of common mode rejection.

The problem with using unbalanced cable in a balanced circuit would be getting unequal noise signals on both shield and center conductor through any form of coupling. In this case shield design may change how much of the noise signal is a perfectly common mode. Double shielding or braided shields have more cross-section area than the inner conductor, and therefor lower series resistance, which upsets the balance. The combination of a foil wrap and drain wire may be closer to the resistance of the inner wire, but still out of balance. Also, since the shield acts to block (and pick up) capacitive and electromagnetic signals, those will signals be out of balance and not be canceled at the input.

The impedance of the driving amp has a big effect too. A very low Z output makes is hard for coupling to occur because the coupling mechanisms are usually of fairly high effective source Z, with (magnetic induction being an exception, often having a lower source Z). Using a low source Z also swamps out capacitive HF loss for the first 1000ft or so. Low driving source Z has a huge effect on preventing coupled noise.

By the way, it's possible and effective to use unshielded twisted pair wire with good balanced drivers and inputs to achieve low noise long line connections. Telephone companies did this all the time, and though most areas of the US there are digital facilities now, only a couple of decades ago high quality broadcast audio circuits with response to at least 15KHz were built with unshielded pairs miles long. My personal experience includes equalizing leased dry pairs (actually an unshielded alarm circuits) flat to 20KHz with a dynamic range well in excess of 16 bits. The loop was probably three miles long at least, though there was no way to tell. I've also wired one studio with single pair unshielded Cat 5. The key is in the driver and receiver. It's not at all that shielding is unnecessary, though. Most balanced inputs don't have high CMRR over full bandwidth, so shielding helps with some noise pickup.

Perhaps we could agree to say that shielding on wire attenuates capacitive and electromagnetic coupling, and does nothing to stop magnetic induction. Shielding stops some noise from getting in, a good balanced circuit cancels noise that does get in. And, short wireds, low source impedance, and proper grounding may make connecting dissimilar input styles possible without additional interface boxes.

What say you?

Jim
 

·
Registered
Joined
·
28 Posts
Well, I disagree Jim. Henry W. Ott is the authority on the subject! Bill Whitlock and Jim Brown use him as the main reference. Co-ax is never a balanced circuit interconnect. It's all about common mode interference rejection. If the noise on the shield and on the center conductor are the same, then it's up to the circuits common mode rejection. Ott shows up to 80 dB rejection. Try this, take a long freebie RCA phone interconnect, solder a resistor across one end (say 100 Ohms more or less) lop the interconnect through some magnetic fields with the other end plugged into some input circuits. The circuits with good common mode rejection will have little hum. Go down to the library (it may take a collage library) and check-out one of his books (they cost almost $100, way to much to buy).
Not much of his writing is available on the web. But a little is at his site:

http://www.hottconsultants.com/
 

·
Registered
Joined
·
28 Posts
Now before this gets everyone totally confused. For the above noise rejection to work, you need some atypical conditions. The "noise voltage" & current in the shield needs to be identical to the "noise voltage" & current in the center conductor. But in most circuits the shield "noise voltage" has many other sneak paths (like other interconnects, chassis and power cord safety grounds). If the shield and center conductor "noise voltage" are not identical then the common mode noise rejection goes out the window.
 

·
Registered
Joined
·
58 Posts

·
Registered
Joined
·
79 Posts
Balanced and unbalanced devices can often be connected directly to each other, with proper cables and wiring. I second Syd’s recommendation of the Rane paper, with a simple to understand chart with pictures and all.

2. Noise immunity. Sometimes you need the common-mode noise rejection of a balanced input to kill ground loop noise, or noise picked up on a long cable. Sometimes. If the run is short, and you get hum, there are other things to do, mostly cost-free, that relate to grounding (another thread?). Frankly, after building and maintaining a bunch-o-studios, I can count on a few fingers the times in the last 35 years I absolutely had to have a device to fix this within a single studio (between rooms is different). These situations are solvable by proper grounding and connections. Again, see the Rane paper.

One note, sort of an aside. All cables designed for audio, balanced or unbalanced, include a shield to help reduce noise pickup. This shield works for capacitive coupling, or electromagnetic pickup (RF), but is essentially transparent to magnetically induced noise, line hum fields from power transformers. This fact is not commonly realized, but it’s physics. There are no practical magnetically shielded cables. It’s important to drive longer cables with low source impedance outputs to reduce the effects of all three types of noise coupling, and take advantage of a balanced input to cancel noise coupled in from the outside. Steel conduit does provide some magnetic shielding.

Manufacturers don't want to spend the money do to that, so they don't spec CMRR.

Jim
Good points here. However, in regard to magnetic shielding, higher quality DI boxes are MU shielded to offset this issue. If a voltage (i.e. noise) is induced in a balanced system by a magnetic field, the voltage is in phase in both sides of the line and therefor would be cancelled by the differential effects of either the transformer or differential amplifier. But it's better to not have to worry about magnetically induced noise in the first place and make sure you're using quality gear that is magnetically shielded.

And finally, reputable manufacturers always specify CMRR. Key word being reputable.
 

·
Banned
Joined
·
66 Posts
Co-ax is never a balanced circuit interconnect.
I never said it was.
It's all about common mode interference rejection. If the noise on the shield and on the center conductor are the same, then it's up to the circuits common mode rejection.
Yes, I agree, and did basically say that too.
Ott shows up to 80 dB rejection.
Yup, that's do-able, and even better.
Try this, take a long freebie RCA phone interconnect, solder a resistor across one end (say 100 Ohms more or less) lop the interconnect through some magnetic fields with the other end plugged into some input circuits. The circuits with good common mode rejection will have little hum.
I don't need to do that experiment. Yes, I agree that would be the case, though this is not a real world condition.
Go down to the library (it may take a collage library) and check-out one of his books (they cost almost $100, way to much to buy).
Not much of his writing is available on the web. But a little is at his site:

http://www.hottconsultants.com/
Thanks, I'll get right on that.

But I must offer my sincerest apology to you for being such a poor communicator. I try my best, but clearly it hasn't worked. Perhaps senility is setting in a decade before it should. I tried to state things clearly, but it seems you think we disagree. Please forgive my ineptitude.

You seem to think we disagree, yet I can find nothing in my posts or yours that we actually disagree on. For example, I said "There are no practical magnetically shielded cables.", which you quoted, highlighted in red, and replied "This is not always true!", then went on to state, quoting from the Ott reference, "Of course the magnetic field goes right through the shield...", which agrees with my statement. So do we agree on this or not? I'm confused.

Then, there's the exchange I've quoted above. Hmm.

Ok, so the only thing I can figure out is that you may not realize that the other two kinds of noise coupling (not magnetic) stop at the shield, and therefore, won't be common mode in an unbalanced cable, and a balanced input won't deal with it. Or perhaps you agree on that too? Whew.

I believe the following statements to be true, and having read the resources posted by Syd from Jim Brown and Bill Whitlock, are in harmony with them too. At least, to my feeble brain they seem so.

What do you think? If you think one is false, can you site a verbatim reference? I could stand to be educated, I've been wrong before, will be again.

1. Shielding on a cable doesn't block magnetic inductive coupling.

2. Shielding on a cable does block pickup of an electric field and electromagnetic field of an RF source.

3. A noise signal may be magnetically coupled into an unbalanced cable equally on the shield and inner conductor.

4. A noise signal may be capacitively coupled to a cable, but the shield will act as the second plate of the "capacitor" and prevent it from coupling to the inner wire(s). In this case the noise signal will not be equal on the shield and inner conductor(s), not common mode.

5. An RF noise signal may be induced onto the shield of a cable, but the shield will carry most of that signal, hopefully to ground. This noise signal will not be common mode in an unbalanced cable.

6. Any signal presented in the common mode to a balanced input will be canceled to the extent of that input's Common Mode Rejection Ratio. No balanced input is perfect, and none have equal CMRR at all frequencies. CMRR can be reduced by imbalances caused by the driving circuit, and some balanced inputs are more sensitive to this than others.

So, if you don't mind, help me see what it is that we disagree on?

Again, apologies to you for my not writing clearly and causing this misunderstanding.

And more importantly, sincerest apologies to all other readers for making "Understanding Unbalanced vs Balanced Audio Signals" less than understandable. Perhaps we should trash this thread and start over? I promise I'll just keep quiet next time!

Thank you Syd for posting the references. Everyone reading this thread should take the time and at least read the Whitlock paper, and probably forget everything else posted in this thread. He says it well. Jim Brown's slides make some excellent points about RF in audio interfacing, a good read too. Jim's an old friend, he used to bring me hot new gear to evaluate/test/try to break, and I often did evaluate, test, and break it. He's a good teacher.

Jim
 

·
Banned
Joined
·
66 Posts
Good points here. However, in regard to magnetic shielding, higher quality DI boxes are MU shielded to offset this issue.
Of course. Transformers used in a DI should be MU metal shielded. But you can't put MU metal around a cable. Mu metal in it's raw state must first be formed, then heat-treated to become mu metal. Once it is, it is quite rigid and brittle, and expensive, not suitable for a shield around a wire. We've be talking about wire here, and there is no practical magnetically shielded wire.
And finally, reputable manufacturers always specify CMRR. Key word being reputable.
If only that were true. It's sort of half-true. They publish a partial CMRR spec, if at all.

Just to be certain, I checked published specs on two devices from Shure, which I though was "reputable". One, the DFR22, has no published CMRR spec. The other, an SCM-262 mixer, did, at a single frequency. Popped over the Electrovoice. The first power amp they list, the 7100, no CMRR spec. I know neither the Shure nor the EV devices are made for the recording market, but both are "pro" devices with balanced inputs. Not conducting a survey, but if you check, I think you'll find CMRR is mostly not included in the specs, and if it is, it's usually the mic input only, and it's stated as a vague figure ("greater than 70dB"), and at only one frequency. Mackie, for example, on the 1604 VLZ3 specs CMRR only on mic inputs, and then as >70dB @ 1KHz. Again, I haven't done the survey, but I don't recall many manufacturers, reputable or not, publishing a CMRR graph. Even Jensen, Bill Whitlock's company, in their IsoMax audio isolator product, doesn't print a graph, but at least specifies CMRR at two frequencies, 60Hz and 3KHz. Hard to believe a figure that changes wildly over the audio band woudn't require a graph to represent. Nobody publishes a CMRR graph because it wouldn't look good, and most customers wouldn't understand it anyway. Big numbers at 60Hz (117dB, for example) smaller ones at higher frequencies, and as Whitlock's paper details, testing was only recently standardized, and certain balanced inputs CMRR is highly affected by driving devices, which further complicates the spec. Nope, it's not the companies reputation, it's just hard to write a spec people would understand.

Jim
 

·
Registered
Joined
·
79 Posts
Bottom line is that magnetic coupling really isn't an issue (rfi & emi are) unless a voltage is induced in the conductors of a balanced line.. and the state-of-the-art in technology today is that the CMRR for most professional studio equipment is high enough to virtually eliminate any unwanted signal as a result.

Really, a much bigger problem in most studios is the noise associated with single-coil guitar pickups.
 

·
Banned
Joined
·
66 Posts
Add grounding issues to that. Ground loops put huge demands on CMRR capabilities.

Is there some reason hum-bucking pickups are not used more, and aren't more effective? I'm not a guitar player.
 
1 - 20 of 26 Posts
Top