Operation Manual
CTs Multi-Channel Power Amplifiers
page 34
page 35
CTs Multi-Channel Power Amplifiers
Operation Manual
Each component has three identifiers that are used in the HiQnet
system:
• Media Access Control (MAC) Address
• Internet Protocol (IP) Address
• HiQnet Address
Let’s explain each identifier.
Media Access Control (MAC) Address:
This is a physical
address that specifies a specific physical network component.
This address is predefined during manufacture and cannot be
modified. No two network components have the same MAC
address and each component is uniquely identified by its MAC
address. This 48 bit address is located on the outside of every
HiQnet component and is shown in hexadecimal format as 6 two-
byte octets such as:
00-0A-40-00-DC-2F
Internet Protocol (IP) Address:
The IP address of a
compo nent is assigned during the setup of the network and
identifies the IQ component to a larger audience. Because the
MAC address is limited to a single physical LAN,
communications between LANs is impossible solely through the
use of the MAC address. The IP address allows components and
controllers on other networks to communicate with each other
through the use of an IP Router as shown in Figure 8.2.
This allows the network designer to isolate network traffic from each
other. For example, when using wireless devices to control HiQnet
components, the bandwidth limitation of wireless devices will not allow
them to reside on the same network with CobraNet. As a result, the use
of an IP router and a separate network allows the wireless con trollers to
access the IQ components through the router using IP addressing. The
IP address appears as four numbers (0-255) sepa rated by periods,
such as:
192.168.0.1
HiQnet Address:
HiQnet or TCP/IQ components will be addressed
with a unique number from 1 to 64,000 and are assigned within the
software. The Subnet Mask is four numbers (0-255) separated by
peri ods. For example, a subnet mask of 255.255.25.0 and an IP
address of 192.168.0.127 says that all IP addresses in that LAN are in
the IP address range of 192.168.01 to 192.168.0.255.
If a controller is on a different network than the audio components,
communications must occur through an IP router. discovery requires
the use of broadcast Ethernet data packets onto the network. These
broadcast commands cannot be done through an IP router, therefore a
Discovery Proxy is required. The establishment of this proxy is done in
by naming the IP address of a known component on the component
network as the Discovery Proxy. In System Architect it is done in Tools
> Options > Network Settings > Manage Network Connections.
8.2 A Closer Look at CobraNet
Licensed by Cirrus Logic, CobraNet is a protocol, firmware and
hard ware that lets you transmit digital audio over a 100Base-T Fast-
Ether net network.
This section provides an overview of CobraNet. Peak Audio has
pro vided a number of technical papers that describe CobraNet in detail.
These papers are available at www.cobranet.info.
8.2.1 Fast Ethernet
(100Base-T)
Fast Ethernet (100Base-T) runs at 100Mbps (Megabits per second). It
uses a hub or network switch to interconnect all network nodes
(com ponents). This forms a star network (see Figure 8.3). In a star
network, all PCs are connected to a centralized hub or switch.
Messages are sent from the originating PC to the hub, which then
transmits the mes sage either to the whole network or, if a switched hub
(network switch) is used, to the destination device only.
In larger Fast Ethernet networks, additional hubs, concentrators, and
other network hardware are used to form a larger network, as shown in
Figure 8.4.
Today, commonly available networking cards are 10/100Base-T
capa ble, which allows them to be used on either 10BaseT or 100Base-T
networks.
The maximum length of cable for CobraNet is the same as for Ethernet:
100 meters over CAT-5 copper cable, 2 kilometers over multimode
fiber. Proprietary “Fast Ethernet via single mode fiber” networks can
reach even further.
Figure 8.4 A Multiple Star Network
8 Network and CobraNet Basics
(continued)
Figure 8.3 A Star Network
8.2.2 Audio Specs
Sample rate:
CobraNet can handle sample rates of 48 kHz or 96 kHz,
but PIP cards work only at 48 kHz. They cannot receive audio from
devices operating at 96 kHz.
Bit depth:
Set each channel to transmit 16-, 20- or 24-bit audio data
as desired.
Latency:
The transmitter and receiver buffering required for reliable
transmission is selectable between 1.33, 2.66 and 5.33 ms (64, 128,
and 256 samples).
8.2.3 Bundles and Audio Channels
You will be assigning all audio channels to groups called Bundles for
transmission over the CobraNet network. Bundles were formerly called
Network Channels.
A bundle can carry up to 8 audio channels. Up to eight bundles can be
sent in a 100 Mbps line, depending on the signal’s bit depth and
sam pling rate.
The usual assignment is 8 channels at 20 bits. You can use fewer
channels per bundle, but maximum size bundles are suggested for the
most efficient use of network bandwidth. If 24-bit data is desired, then
only 7 audio channels can be loaded into a single Bundle.
In System Architect, you create audio connections between sending
devices (transmitters) and receiving devices (receivers). For example, a
mixer could be a transmitter, and a power amp could be a receiver.
You assign each connection a Bundle number. The Bundle number
indicates which devices are communicating with each other.
Also in System Architect, you address each connection as Multicast or
Unicast. A
Multicast
(Broadcast) connection is from one transmitter
to multiple receivers. A
Unicast
connection is from one transmitter to
one receiver. For example, if you send a mixer’s digital audio signal to
several power amps, the signal would be Multicast. Unicast is most
used on “switched” networks requiring more than 64 channels.
You specify a connection as Multicast or Unicast by the delivery
address it you give it in System Architect. 255 and lower is Multicast
(Broadcast); 256 and up is Unicast.
8.2.4 The Conductor
A CobraNet system is coordinated by one audio device in the network
called the Conductor. It regulates the CobraNet digital audio traffic on
the CobraNet network. That is, the Conductor is the master clock for
the rest of the network.
You choose one device to be the network conductor based on a prior ity
scheme. The Conductor indicator will light on the CobraNet device
that is serving as the conductor.
The Conductor handles time-division multiplexing and clock recovery
on the network. The Conductor periodically broadcasts a well-defined
“beat” packet to all components to recover synchronous timing
infor mation. The “beat” also specifies which addresses can transmit at
cer tain time slots within the “beat” period.
Each transmitting device is allowed to transmit in a given time-slot
within the Conductor-controlled isochronous cycle time. The time-slot
is determined by the assigned bundle priority. Higher bundle pri orities
receive lower-numbered positions and are transmitted first.
8.2.5 Repeater Networks
CobraNet can operate on repeater networks or switched networks.
Repeater networks use low-cost Ethernet hubs. Switched networks use
network switches.
Hubs send all data packets to all connected nodes. That is, all devices
in the network receive data — even the Unicast data. Switches send
specific packets to specific devices.
Up to 8 full-loaded Bundles (64 audio channels) are allowed on a
repeater network. More Bundles may be allowed if they are loaded
with less than the full eight audio channels.
Generally, a repeater-based CobraNet network must carry only
CobraNet data to guarantee reliable transmission of audio
pack ets. Use of repeater networks is not recommended.
8.2.6 Switched Networks
A more complex CobraNet network can be built using Ethernet
switches. Switches do not simply broadcast each and every
packet to all nodes. Instead, they check each incoming data
packet to determine its destination and (very quickly) transmit the
data to only that destination port. This allows for more network
data flow, more Bundles and more audio channels.
In effect, each network port in an audio component has 100MB of
bandwidth. The network can be as large as 100MB times the
number of ports on the network.
Another advantage of switched networks is “full-duplex”
connec tions between components. A full-duplex link allows
simulta neous send and receive over the same Ethernet
connection.
In theory, each connection can simultaneously send and receive
up to 64 channels per node, for a total of 128 channels. However,
current CobraNet components can handle only 64 channels (32 in
and 32 out).
Switches and full-duplex connections allow up to 128 channels
per component, and greater than 3000 individual Unicast
chan nels.
Switched networks also allow controllers (such as computers) to
be on the same network as CobraNet devices.
8 Network and CobraNet Basics
(continued)
Summary of Contents for CTs 4200USP/CN
Page 23: ......
