Balanced / Unbalanced

September 28, 2013 in AV Design Tips, The Basics by Sam Davisson

UnbalancedWhen I did the first grounding post people seemed to think I was setting the groundworkk for talking about audio. I was actually setting the foundation for all AV. Although its true that any audio problems are related to grounding issues. Balanced and unbalanced interconnect systems, while mostly related to audio, are prevalent in all AV cabling.

RS-232 is an unbalanced interconnect system. RS-422 or RS-485 are balanced. Analog video and SDI are unbalanced. Analog audio, the pro format is balanced. The consumer format, unbalanced. Digital audio employs both types of signal interconnects as well.

Definition: Impedance –
A measure of the opposition to the flow of alternating current through a circuit. Impedance is measured in ohms. The resistance of a circuit to direct current (also measured in ohms) is generally not the same as its impedance, due to the effects of capacitance and induction in and among the components of the circuit.

Balance is defined in terms of the impedance of the two signal conductors with respect to a reference, which is usually ground. If these impedances are equal and non-zero, the system is balanced. If the impedances are unequal the system is unbalanced. A signal conductor with a grounded return conductor is, therefore, an unbalanced.

A small, common-mode, 60 Hz noise, voltage can exist between the chassis of two AC powered devices regardless of whether they are safety grounded (use a three-wire plug) or not. An isolated ground system is designed to minimize this.

Unbalanced Interconnections

An unbalanced interface uses only two conductors to carry the signal from one device to another, one conductor carries the signal and the other is the grounded return. In consumer audio systems this usually consists of a cable with a center conductor and a shield terminated in an RCA Phono plug.

The “RCA Phono” plug was developed by RCA many years ago to be used for short interconnects between a turntable and amplifier inside a phonograph. This unbalanced
interconnect system is simple and inexpensive, but as with many other connector systems has been adopted for uses other than originally intended. It has become the de-facto connector for consumer audio/video equipment. Note, however, that the design was original intended only for short cable runs within the same piece of equipment. Being an unbalanced system it is susceptible to common-mode noise voltages.

A problem occurs when there is a ground voltage (common-mode voltage) between the two interconnected devices. Because of this voltage, a small current will flow down the cable shield between the devices (sometimes referred to as common-mode current, or most commony as a ground loop current). If the cable shield were ideal (zero impedance) this current would not cause a problem. However, since the shield has a finite resistance, a small noise voltage will appear across the length of the cable shield. The magnitude of this voltage will equal the common-mode current times the shield resistance. This voltage is in series with the signal voltage and will add directly to it at the receiver. In other words, an unbalanced interconnect system has no ability to reject common-mode noise voltages.

This coupling can be referred to as common-impedance coupling, and is the result of the fact
that in an unbalanced two-wire system the shield is performing two functions. It is a shield carrying the common-mode noise current, but it is also one of the signal conductors carrying the return signal current

Consider a typical case of the interface between two grounded (3-prong AC plug) pieces of audio equipment. This example is typical some cases would be better and some worse. Also, note that calculating signal to noise is a future discussion.

The shield resistance of a fifteen-foot cable typically will be 0.25 ohms, depending on cable type used. If the 60-Hertz shield current is 250 uA, the voltage developed across the shield will be 62.5 uV. For consumer audio products the reference signal level is 300 mV (-10 dBV). The signal to noise ratio will therefore be 74 dB. In a high end consumer audio sysem, we would most likely be able to hear some 60-Hertz hum in quiet passages of the program material.

You could conclude, at this point, that ungrounded equipment, those using a 2-prong AC plug, might solve this problem by eliminating the ground connections. This sometimes helps, but it does not necessarily eliminate the problem.

For ungrounded equipment, the common-mode ground current can still flow through the inter-winding capacitance of the power transformer. The impedance of this capacitance will normally reduce the current to 100 uA, or less. However audible noise may still be present.

The impedance of the inter-winding capacitance is frequency dependent, more current will flow at higher frequencies (harmonics of 60 Hz) than at the fundamental frequency (60 Hertz). Therefore, the interference will consist of a high frequency buzz which is more audible by the human ear than the 60 Hz hum.

Despite its shortcomings, this unbalanced system works well. Especially in cases with short cable runs, and the equipment within the same rack so there is very little, or no, 60 Hz voltage between the chassis of the interconnected devices.

Trouble shooting and eliminating audio issues will come in future posts but understanding the issue at hand should help you develop techniques of eliminating problems before they occur making troubleshooting unnecessary.

Advantages of Twisted Pair Wire

Twisted pair wiring, even when unshielded, is very effective in reducing magnetic field coupling to and from the wire pair. There are only two conditions necessary for this to be true. First, the signal must flow equally, and in opposite directions, on the two conductors.

Secondly, the length of the twist must be less than one twentieth of a wavelength at the frequencies of concern. (One twist per inch will be effective up to about 500 MHz).

The above is true whether the terminations are balanced or not. In addition, if the terminations are balanced, twisted pair wiring will also be effective in reducing electric field coupling to and from the wire pair. Even though field coupling is not the primary noise coupling mechanism in audio systems, it is still a good practice to always twist the signal and return conductors in a cable. (Twisting is especially important in the case of microphone cables.)

Balanced Interconnections

For a balanced interconnect system both of the signal conductors have an equal, and non-zero, impedance to ground. Therefore, three conductors are required, signal+, signal-, and ground or shield.

Professional audio installations require long cable runs of 500 feet or more and have to be able to operate at microphone signal levels of 3 mV (-50 dBV) as opposed to a line level of 300 mV for consumer audio. As a result he maority of professional audio equipment is designed using a three-conductor balanced interface (two signal conductors,signal and signal return, and a grounded shield), using XLR connectors, or in some cases 1/4 inch phone plugs.

This three conductor balanced interconnect system avoids the problem of the shield having to serve two purposes. The signal is now carried on the two internal conductors (usually twisted together) and the shield only acts as a shield and not also as a signal return line. The penalty for this improvement in performance is a more complicated and hence more expensive system.

A 60Hz shield current flowing between two interconnected devices will still produces a voltage drop in the shield, but this noise voltage is not in series with the signal. Rather it will be coupled equally (as a common-mode noise voltage) into both signal conductors.

Since the receiver looks at the difference between the two signal conductors (not the voltage between one of them and ground), the common-mode noise voltage cancels out and is not seen by the receiver.

A balanced interface theoretically would be completely immune to noise and interference. But, in the real world, nothing is perfect. Even if you attempt to make the impedance of the two signal conductors to ground the same there will be some difference, if only a fraction of a percent, and this will limit the degree of common-mode voltage rejection and therefore the maximum noise suppression possible.

For this example I’ll assume that the impedance balance is such that the circuit can provide
60 dB of common-mode noise rejection, a very conservative assumption, and that the other
parameters are the same as in the unbalanced system above. A well designed balanced interface can have as much as 80 to 100 dB of common-mode noise rejection.

Conclusions

When using unbalanced interconnects between audio equipment the primary noise-coupling mechanism is due to common-impedance coupling. The cable shield resistance
is the common-impedance, and any small 60 Hz AC potential between the equipment chassis is the noise source producing a common-mode noise current on the cable shield (ground loop
current).

This problem can be minimized, or eliminated, by any one of the following approaches:

  • Minimize the AC voltage between the two pieces of equipment.
  • Minimize the cable shield resistance.
  • Block the common-mode noise current by using a signal isolation transformer.
  • Use a balanced interconnection system.

Rack Elevations and Math

September 21, 2013 in Tech Talk, The Basics by Sam Davisson

Math_0b935f_1351529

I got sidetracked a little bit with the last post on The Decibel. But now, since we’ve covered Power and Grounding, lets go back to the equipment rack. Before we get into the calculations you need to make and share with the electrical and mechanical contractors lets discuss effective equipment rack layout techniques.

What’s the best way to elevate an equipment rack?

I really wish there was a correct answer to that question. The truth is it’s a very fine balancing act. There are a few rules of thumb but all of them are somewhat situational. But I’ll list as many as I know:

  • Install heavy equipment toward the bottom of the rack.
  • Separate equipment by signal type and/or function
  • Evenly distribute the power/heat load
  • Consider cable lengths
  • Ease if install, in other words keep it simple for the the installers to cable
  • Equipment cooling

Taking these one at a time, I’ll try to explain the logic (or at least my logic) behind them:

Install heavy equipment toward the bottom of the rack
A best practice is to build and test systems in your shop prior to delivering on site. This will help decrease the overall install time. It’s much easier to troubleshoot problems off-site and out of the view of the client. But that means you have to transport the racks. Having the heavy equipment at the top of the rack makes the racks top heavy and increases the possibility of an accident during transport. If your lucky it’s only equipment that gets damaged if something happens and not an employee.

Separate equipment by signal type and/or function
There are a number of things you can do to eliminate the possibility of cross talk and signal noise. This is one of those things. Your much less likely to have problems if you bundle cables by signal type. I practice and suggest bundling your cables in this manner. I also suggest that bundles of different types are kept separate by a minimum of 1/2" except for speaker level cables which should be treated like AC power cables and kept a minimum of 2" from any other signal type. Here are my suggestions for cable type bundles:

  1. Analog, line level, balanced audio
  2. Analog, line level, unbalanced audio
  3. Analog, microphone level audio
  4. Digital audio
  5. Analog Video
  6. Digital Video (HD/SDI)
  7. Digital Video (HDMI)
  8. Digital Video (HDBaseT)
  9. Control Cabling and DC voltages
  10. Network Cabling
  11. Speaker level audio
  12. Other – There always seems to be something that fits in this category. Such as how would you classify a cable with Cobranet information. While it is a CAT5e cable it is neither network or control.

As far as separation by function, I like to keep my video processing equipment separate from my audio processing equipment and source equipment. But here is where exceptions tend to creep in. In order to keep the noise level to a minimum you need to keep any unbalanced audio cables as short as possible, preferably under 15′. This will often dictate that some if not all of the audio processing equipment will need to be in the same rack as the sources.

Evenly distribute the power/heat load
The most important aspect to this is rack cooling. By keeping the loads evenly distributed you keep the heat evenly distributed between the racks. This will simplify cooling requirements and make it less likely that equipment will suffer from heat.

Consider cable lengths
I already touched on one aspect of this. Unbalanced audio cabling needs to be kept as short as possible to maintain a low noise threshold. When we get deeper into things like signal and noise ratios (SNR or S/N) I’ll explain in greater depth as to why this is important.

But there are other cable lengths to consider. The majority of HDMI equipment has cable equalization built into it. That equalization is set for an average HDMI cable of 6′.

HDBaseT suffers when cables are bundled depending on the type of CAT cabling being used and the number of cables in the bundle. With CAT5e UTP cable, a bundle of 7 cables will begin to see catastrophic failures at about 35′. CAT6 UTP bundled in the same manner makes it about 50′. When bundling CAT cable carrying HDBaseT information the best method and the only one that allows for the full 330′ extension is CAT6A.

I know, what about STP cabling. If everything is done correctly, this will work. But a broken shield or a noisy ground floor can actually cause your problems to be greater.

Ease if install, in other words keep it simple for the the installers to cable
This one is pretty much self explanatory, I think. Racking equipment of similar signal types makes building the rack simpler because the installer isn’t jumping from cable type to cable type. It also makes for a cleaner installation.

Equipment cooling
Keeping the equipment cool prevents failure. More importantly it prevents intermittent errors that can be difficult to track down. Your overall rack plan should include looking at where the rack is being installed, how it is being cooled and the method of cooling.

Elevating the Racks
The first thing I do is download all of the equipment specifications for the equipment contained on the Bill of Material (BOM) or estimate. I usually separate them into two folders, field equipment and rack equipment. Then I’ll create a spreadsheet listing equipment, height (in RU’s (rack units = 1.75"), badged current draw, efficiency current, idle current and any other pertinant information I can think of at the time. Badged power is the maximum current draw when a piece of equipment is energized. Efficiency power is the amount of current draw after the equipment is energized and has settled. Idle current is used for some displays and amplifiers and indicates the amount of current drawn when the equipment is not in use.

I also have a couple rules of thumb I use when it comes to "head room" when elevating a rack and calculating the amount of current I’ll need to power the rack. I like to keep my "head room" at 20 – 25% of capacity. So, on a 44RU equipment rack, I try to maintain 9-10 blank spaces for future growth. With my power circuit, I’ll try to keep power at 15 – 16 amp draw on a 20 amp circuit.

Discounting amplifiers, electronic equipment uses power pretty efficiently. I calculate that efficiency at 2.5. So to calculate efficiency current, divide the badged current by 2.5. In other words a piece of equipment that has a 10a fuse will draw 4a after the initial power surge. This is an estimate and not an exact measurement. That’s where your headroom comes into play. It’s your safety factor. Using a power sequencer (a device that delays the initial power rush for equipment) you can effectively exceed the current draw of the equipment over the badged listings to about 20 amps and allow you to maximize the amount of equipment on a single circuit.

Rack Math

I guess it’s time to talk about that math I promised. Everyone who creates rack elevations should be familiar with Ohms Law and how to calculate power usage. Additionally they should be able to use these tools to calculate the heat load in BTU’s for a specific equipment rack. BTU stands for British Thermal unit. Our good friend Wikipedia can explain why it’s named what it’s named. I’m happy to not explain it

Ohms Law – V(voltage)=R(resistance)*I(current)
Power Calculation – W(power expressed in Watts)=V(voltage)*I(current)


Heat Load Calculation =
3.41214 x W(power expressed in Watts) BTU/hr

Once you’ve elevated your racks based on the criteria and exception already noted your ready to calculate power requirements required by the electrical engineer and the heat load needed by the mechanical engineer.

I wish I could give you some hard and fast rules as to when to use a power sequencer and when not. I try to use them whenever there are going to be extended times when the equipment is not going to be in use. This allows you to "power down" the rack until it is needed, saving energy. If you are using a power sequencer then for every 20 amps of power you will need one 20a circuit. If not, then for every 15 amps of power you’ll need one 20a circuit. In either case, with equipment energized and stable your rack should be drawing roughly 15 amps.

Calculate the power
Power Calculation – W(power expressed in Watts)=V(voltage)*I(current) or Power (w) = 110v (US) x 15a. Maximum power draw is 1650 watts.

Calculate the Heat Load.
One would think, looking at my my formula’s above that the heat load calculation would be 3.41214 x 1650w = 5630 BTU/hr. But that doesn’t take into account efficiency. Remember the equipment will only be drawing 1650w for a very short time. So divide the 5630 BTU/hr by equipment efficiency of 2.5 and your estimated heat load should be 2252 BTU/hr.

Please note amplifiers play by different rules and will be covered in their own discussion when we can discuss things like different amplifier classes.

The Decibel

September 14, 2013 in Tech Talk, The Basics by Sam Davisson

Singal to Noise Ratio
Honestly, I wasn’t sure what direction this was going to go when I started writing and I’ve spent most of the last week trying to decide what should be next. I was thinking linearly, like a book. But in reviewing and responding to comments on the Power and Grounding post, I realized that perhaps the best approach would be to tackle misconceptions as they arose and then precede from there.

In order to do that, I found that I needed to set the definition of what is probably the most confusing or perhaps better expressed as confused term in the industry.

The Decibel

The decibel (dB) is a logarithmic unit used to express the ratio between two values of a physical quantity (usually measured in units of power or intensity). In AV the decibel is commonly used as a measure of gain or attenuation, the ratio of input and output powers of a system, or of individual factors that contribute to such ratios. The number of decibels is ten times the logarithm to base 10 of the ratio of the two power quantities.

The decibel is used for a wide variety of measurements in science and engineering. In AV we use it for the gains of amplifiers, attenuation of signals, signal-to-noise ratios and sound pressure levels. The decibel offers a number of advantages, such as the ability to conveniently represent very large or small numbers, and the ability to carry out multiplication of ratios by simple addition and subtraction. On the other hand the decibel confusing and cumbersome.

A change in power by a factor of 10 represents a 10dB change in level. Whereas a change in power by a factor of 2 represents a 3dB change (remember the relationship is logrithmic and not linear). To keep things even more confusing a change in voltage by 10 is equivalent to a change in power by a factor of 100 or a 20dB change. Therefore a change in voltage by a fator of 2 represents a 6dB change.

Confused yet? The thing to remember is that the decibel is usually qualified with a suffix to indicate which reference quantity or frequency weighting function has been used. Common suffixes in the AV world are:

Volages

  • dBV – RMS voltage relative to 1 volt, regardless of impedance (the apparent ac resistance of a circuit containing capacitance and/or inductance in addition to pure resistance)
  • dBu or dBv – Originally this was always expressed dBv but changed to dBu to avoid confusion with dBV. The v comes from volt whereas the u comes from unloaded. dBu can be used regardless of impedance, but is derived from a 600 Ω load dissipating 1mW (0dBm)
  • dBmV – voltage relative to 1 millivolt across 75 Ω. Obviously this is typically used in analog video systems
  • dBuV – voltage relative to 1 microvolt. Widely used in television and aerial amplifier specifications. 60 dBμV = 0 dBmV.

Accoustics

  • dB (SPL) – sound pressure level referenced to the nominal threshold of human hearing (0dB SPL)
  • dB(A), dB(B) and dB(C) – These symbols are often used to indicate the use of different weighting filters, used to approximate the human ear’s response to sound. The measurement is still in dB(SPL). These measurements usually refer to noise and the effects on humans and animals. They are in widespread use with regard to noise control issues, regulations and environmental standards. According to ANSI standards, the preferred usage is to write LA = x dB. Nevertheless, the units dBA and dB(A) are still commonly used as a shorthand for A-weighted measurements.

Audio Electroncis

  • dBm – power relative to 1 milliwatt. In audio, dBm is typically referenced relative to a 600 ohm impedance. This measurement has pretty much been retired in line level audio as the typical impedance of a an input on a piece of equipment is greater than 10 kΩ aka Hi Z.
  • dBFS – the amplitude of a signal compared with the maximum which a device can handle before clipping occurs. Full-scale may be defined as the power level of a full-scale sinusoid or alternatively a full-scale square wave. A signal measured with reference to a full-scale sine-wave will appear 3dB weaker when referenced to a full-scale square wave, thus: 0 dBFS(ref=fullscale sine wave) = -3 dBFS(ref=fullscale square wave).

The point behind this is that simply stating that something is 74dB is pointless without the suffix referencing what that measurement is relative to. 74dB is not the same as 74dB SPL or 74dBV.

Power and Grounding

September 7, 2013 in AV Design Tips, The Basics by Sam Davisson

Isolation, the key to a good grounding structure

This is the first of an upcoming series on “The Basics” of designing the AV systems that high end clients expect and the average client will respect.

Quality AV systems don’t happen by accident. They happen by design. Just like building a house, you’ll never reach that true quality threshold without a proper and strong foundation. With all electronic systems, especially AV, that foundation is the grounding structure.

Most designers and installers of AV systems think of grounding as black magic. How often have you or someone you know said that a cable is “picking up” noise or that the solution is “better” shielding or floating the shielding? (Note: In very high end audio systems floating the shield can be essential but in the typical AV installation, if grounding is done properly, should not be required.) Equipment manufacturers are typically of no help as they don’t have a clue as to what’s going on because they design and test their equipment in a defined, never changing environment.

The basic rules of physics are overlooked, ignored, and / or forgotten. Electrical engineering courses rarely even mention practical issues of grounding. As a result, myth and misinformation have become the normal

The preferred grounding technique for AV equipment, which contributes to this sense of mystery, is isolated grounding. The reason for this is noise interference and particularly what is known as common mode noise. In AC power systems, the difference in potential between neutral and ground is one form of common mode noise, since any change in neutral potential relative to ground also affects all of the other power circuit conductor potentials to ground. A more troublesome form of common mode noise is the differences in ground potentials throughout an electrical system. When AV devices are interconnected byway of audio, video or control cables, any difference in ground potentials between the interconnected pieces of equipment is common mode noise to the audio, video or control circuits.

Properly Conceived Isolated Grounding System

Properly Conceived Isolated Grounding System Through a 3 Phase Transformer

There is much confusion over what an “isolated ground” (IG) is, how an isolated ground technique is implemented, and why it is used. Isolated ground is allowed in the U.S. by the National Electrical Code (NEC) and in Canada by the Canadian Electrical Code (CEC). But it should be noted that in both cases, isolated ground is an exception to the standard grounding requirements. NEC 250-74 and 250-75 allow IG wiring only where required for the reduction of electrical noise on the grounding circuit. In practical terms, what this means is that not all electrical contractors or electricians are versed in building isolated ground systems.

Common Myths About Earth Ground and Wires

Unfortunately, as electronic equipment developed the term “ground” became sort of a generic term. In AC power systems ground refers to a common connection point, typically earth ground. In a DC power system, such as what is powering my laptop, ground also refers to a common connection point for power return. Thus, the very meaning of the term ground has become vague, ambiguous, and often quite fanciful but make no mistake these grounds should never have thier common connection point in common.

Some electricians have a strong urge to reduce unwanted AC voltage differences by “shorting them out” with massive conductors, the results are most often disappointing. Other electricians think that system noise can be
improved experimentally by simply finding a “better” or “quieter” earth ground. Many indulge in wishful thinking that noise currents can somehow be skillfully directed to an earth ground, where they will disappear forever!

Here are some common myths about grounding:

  • Earth grounds are zero volts – presumably with respect to some “mystical absolute” reference point. This leads to fanciful ideas about lots of ground rods making system noise disappear. The fact is soil resistance between ground rods is much higher, often significantly than a wire between them.
  • Wires have zero impedance (is the apparent ac resistance of a circuit containing capacitance and/or inductance in addition to pure resistance) – therefore they can extend a zero-voltage reference to many locations in a system, eliminating voltage differences. In fact, wires are quite limited:
    • The DC resistance of a wire applies only at very low frequencies and is directly proportional to its length.
    • The inductance of a wire is nearly independent of its diameter (gauge) but is directly proportional to its length and increases at bends or loops.
    • A wire resonates (becomes an antenna) when its physical length is a quarter wavelength

Something to consider, are earth grounds even necessary for low noise system operation?

Equipment Rack Treatment for Isolated Ground SystemUpdate 9/11
To answer the above question: no earth grounds are required for safety, not for low noise system operation!

But what I really wanted to update here today, something I left out of my post, was the equipment rack treatment an AV integrator needs to do in order to maintain a nice clean isolated ground system.

Unless your are installing the equipment rack on a wood floor, the rack must be isolated from the floor. There are a number of ways to successfully accomplish this. While not called out, the drawing above used wooden 2×4’s underneath the rack but many equipment rack manufactures also sell ground isolation kits made from polycarbonate strips that work perfectly fine.

If your bolting the rack down you’ll need to use a nylon washer to keep the bolt itself from making the electrical path. If your simply setting the rack on the floor with no requirement to bolt it down be sure to include leveling feet with rubber covers to provide your isolation. Finally, if the rack is portable, the rubber wheels should do a good job of providing the required isolation.

Install a copper bus bar in each rack. This bus bar is connected to the isolated ground system (at the isolated ground sub panel) and then connected to the equipment rack via a 12 gauge stranded wire. The bus bar must be isolated from the equipment rack at all other points. Additionally most AV equipment provides a grounding lug on the back of the equipment. Again use a 12 gauge stranded wire to connect equipment chassis to the copper bus bar. Note: some older equipment may strap signal ground to chassis ground. If this is the case be sure and remove the strapping mechanism.

All AV conduits coming into the rack must be isolated from the rack. Most electrical supply outlets provide isolation bushings for this. In addition your equipment rack vendor may be able to provide isolation knockout panels. Finally, be sure and use an approved IG power strip.

Following these guidelines will ensure that your IG system remains so.

As I discuss future AV basic design topics grounds and dealing with grounding issues will be a constant theme. As I said, we are simply setting the foundation at this point, not only of your AV system but for future posts as well.

404's powered by true Google Search API