​​​​Protocol Specification

System Development and Integration Group

10^th December 2010

1 Overview

1.1 References

1.2 Scope

This document provides the publicly available means for controlling HiQnet products.  Included are the following:

• HiQnet product architecture
• All formatting of messages
• Network specific information for implemented transports
• Open-loop and closed loop control methodologies
• Examples utilizing System Architect to facilitate message formation
  
  Although examples are given using specific Devices, detailed documentation is not provided for every HiQnet Device.  Instead, methods of using System Architect to glean that information are provided. It is assumed that readers are familiar with System Architect, the basics of Ethernet networking, and RS232.  Note that it is not intended that third-party control Devices implement all of the methods detailed in this document.  It is assumed that in most cases only a subset of these messages will be implemented.  Some of the information presented is for explanatory reasons only, such that a person desiring to control a HiQnet Device may understand the underlying behavior. Lastly, control of USB products is beyond the scope of this document. 
  
               

2 HiQnet Product Architecture

It is important to have a basic knowledge of how HiQnet products are developed.  We envisage HiQnet as more than just a networking protocol.  Our goal is to develop a system solution for all Harman Pro networked products.  To facilitate that end, we have designed a common model or architecture for all products that will be developed under HiQnet.  This common product architecture is then reflected as a messaging system that enables communication between products.  Finally, HiQnet also consists of reference designs on common physical networks.  The HiQnet messaging protocol was designed to be transport independent, it requires only certain network services. 

The HiQnet product is modeled with a multi-tiered approach.  The top level that usually represents the product itself is called a Device.  The Device must contain at least one Virtual Device, the first of which acts as a Device Manager.  Each Virtual Device can contain Objects and/or Parameters. Objects themselves can contain other Objects or Parameters.  A Parameter contains the state of a single controllable variable.  Below we will define each of these terms in detail.

At every level in the hierarchy there are also attributes.  Attributes are member variables that contain useful data about the containing Virtual Device, Object, or Parameter.  For instance, one Object attribute is the Object Name.  In the case of parameters, attributes are used to hold the parameter's max and min values.

Attributes can either be STATIC, Instance, or Instance+Dynamic.  STATIC attributes are the same value across all Devices that are the same manufacturer/model.  Attributes that are denoted Instance indicates that the Device upon powerup sets the value of the attribute.  Attributes that are denoted Instance+Dynamic are those that are set on Instantiation and can change during the life of the item.

Virtual Devices, Objects, and Parameters all have a unique Class ID and Class Name. Either the Class Name or ID can be used to uniquely identify the HiQnet item and allow the designer to know key information about the item.  For example, from the Class ID of an Object, the designer knows the number, type, and order of Parameters in the object.  From the Parameter Class ID, the designer knows the data type and max/min.

It is important to note that there is no distinction in HiQnet between elements used for signal processing such as a Parametric EQ, control elements such as a mechanical fader, or sensor elements such as an output meter.  Even global items such as passwords and MIDI channels can and should be put inside the basic HiQnet model.  By viewing everything as a parameter, Object, or Virtual Device, the same mechanisms for subscriptions and control can be universally applied across the product.

2.1 Device

Device designates the Device or product itself. Devices are comprised of Virtual Devices.

2.2 Virtual Device

The Virtual Device is a collection of Objects, parameters and attributes that make up a useful unit.  They offer the designers a convenient method of segmenting the product.  As an example, if you examine the structure of the dbx 4800 in the System Explorer you will see that they have separated the product into two sections, one for all the processing objects that can change with presets changes and the other for the global utility section.  At a glance it is very easy to distinguish which parameters will be affected by a preset change. 

2.3 Virtual Device Attributes

All Virtual Devices have the following Attributes:

2.4 Device Manager Virtual Device

Every product contains at least one Virtual Device, the Device Manager. Some products such as the Crown UPS3 have been architected with only the Minimum Device Manager Virtual Device.

2.4.1.1 Device Manager Attributes

All Device managers Virtual Devices have the following Attributes:

2.5 Object

A HiQnet Object is a collection of parameters grouped together for convenience.  An example would be an EQ object or compressor object.  Objects can contain other objects so for example a channel object could be comprised of a gain and an EQ object. 

2.6 Parameter

Within the HiQnet model, the smallest modifiable parameter in a product is held within the parameter.  Examples of parameters include the variables of an audio object, like frequency and the position of a fader on a control surface.  Simple products like a wall controller may contain only several parameters, while others such as a mixing console may contain hundreds of thousands. Typical operations on parameters include 'set' a variable and 'get' a variable; these could translate to setting the frequency of an EQ and getting a delay time for display.

The HiQnet protocol supports several different data types including Unsigned Byte, Float, String, etc.  It is important when you are sending a message to a HiQnet Device that you use the appropriate data format.

2.6.1 Parameter Attributes

2.6.1.1 Minimum Value

Minimum Value is specified in the Parameter's Data Type.  See section 2.7.6 for an explanation on how to retrieve the minimum value of a parameter.

2.6.1.2 Maximum Value

Maximum Value is specified in the Parameter's Data Type except for the BLOCK and STRING types, which will use a UWORD for its maximum.  In the case of a block, the maximum value specifies the maximum size that the variable length block can be in bytes.  In the case of a string, the maximum value also specifies storage, which is twice the number of characters including the NULL because strings are encoded with Unicode.  See section 2.7.6for an explanation on how to retrieve the maximum value of a parameter.

2.6.1.3 Control Law

The processing object uses a control law to recommend how it would like to be controlled.  For example, a Parameter for frequency may want to be logarithmic, a gain SV may want to be logarithmic with extra resolution around 0 dB.

If you have a Parameter that can take on any floating-point value between the Minimum and Maximum, you still want to specify the control law to give a good look and feel to the user.  For example, in the case of a frequency variable, it is often desirable that when the user turns an encoder or pushes the <up> control on a spinner that the next value be logarithmically spaced from the previous value.

The control law may also be used to specify the granularity that a Parameter can accept.  For example, a gain Parameter may have a maximum resolution of .1 dB.

This control law is not needed in the case of an enumerated Parameter, as all steps are known.

2.6.1.4 Flags

Bits 0, 2, and 3 are reserved. Bit 1 is the Sensor Attribute.

1. = Non-Sensor
2. = Sensor
   
   This attribute is used in subscriptions to automatically set the type of subscription to periodic or on change.  Examples of sensor Parameters include output meters, threshold meters, or power amp temperature.  Non-sensor Parameters are things like frequency or MIDI channel.

2.7 HiQnet Addressing

Addressing in HiQnet is split up into three main sections, a 16 bit HiQnet Device address, a 32 bit field that designates the Virtual Device and Object and finally a parameter Address that designates the correct parameter within the Object.  The System Explorer in System Architect always shows any of these addresses with a trailing number enclosed in '[ ]'. 

2.7.1 HiQnet Device Address

The HiQnet Device address is often referred to as the HiQnet address in System Architect.  In the below example, the three Devices are addressed 1, 2 and 3.  These addresses are shown in the System Explorer and the Venue View.

2.7.2 Virtual Device Address

The Virtual Device and Object Addresses comprise a 32 bit number, segmented into an eight bit Virtual Device and a 24 bit Object address.  Using the dbx 4800 as an example, you see that the default configuration shows three Virtual Devices, the Device Manager(Shown with the Device Name “DriveRack 4800"), the Processing Object Virtual Device(shown with the preset name of “Wide Open") and the Utilities Virtual Device.  These have the Virtual Device Addresses of 0, 1, and 2 respectively. 

2.7.3 Object Address

The Object Address in the System explorer is broken into its three eight-bit sections.  For example, the first Input Mixer Object in the DriveRack 4800 is addressed [3.2.0].  The second mixer is addressed [3.2.1].  The Object Address will be unique within that Virtual Device.

The fully qualified 48 bit address of the first mixer then is

[3(Device).0(VD).3.2.0(Object)] 

2.7.4 Source & Destination Addresses in Messages

Some messages are specific as to the kind of HiQnet 'item' they may originate from or can be sent to.  An example is clearer – the SetAttribute message may be sent by a Device Manager Virtual Device, a Virtual Device or an Object.  This asymmetry in permissible Source and Destination addresses is documented using the following convention:

                        'DEVICE'             is a placeholder for any Device Address.

                        'VD'                      is a placeholder for any Virtual Device Address.

                        'OBJECT'             is a placeholder for any Object Address. 

2.7.5 Parameter Index

The Parameter Index uniquely identifies the parameter within an Object.  Continuing with the example 4800, you will see that the Input Mixer has four parameters: Source Select[0], Input Gain[1], Pink Noise On/Off[2], and Pink Noise Gain[3]. 

2.7.6 Parameter Range

The range of a parameter can be found in the SetSV message string that is copied to the clipboard.  In the HiQnet String Settings in the options, “Use Placeholder for Parameter Value" should be enabled (see section 5).  Rightclick on a control on a panel and choose “Copy HiQnet Parameter String".  Paste in to a program such as notepad to find the SetSV message string and which includes the range of the parameter.

02,19,00,00,00,22,00,33,00,00,00,00,00,01,11,06,11,00,01,

00,00,20,05,00,00,00,01,00,01,06,[Float (20 - 20000)]:XX,XX,XX,XX

The value where the XX's are are where the value should be inserted.

2.8 Alternate methods of finding a HiQnet address.

Sometimes it can be difficult to find exactly the correct parameter you need within the System Explorer.  There are two other good ways to find the addresses you need. 

2.8.1 Copy HiQnet Information

The simplest method to find the HiQnet address of an object is to locate the object on a panel, right-click the object, and choose “Copy HiQnet Information".  The figure below shows the right-click context menu when a PEQ is right-clicked on an dbx SC 32.

You can then paste the HiQnet Information in to a program like Notepad which will give you

Name(type of object): 6-band PEQ

Node: (Hex):0x01, (Decimal):1

VD: (Hex):0x11, (Decimal):17

ObjectID: (Hex):0x061100, (Decimal):6.17.0

This provides you with the node address, virtual device address, and object address.  Combined these give you the overall HiQnet address of the object.

2.8.2 Finding and Address using the Custom Panels and the Properties Window

The simplest way to find an address when you don't know it is to find the parameter you want to control on a factory panel and then drag it to a custom panel.  In this example, to get the address for a gain in the input module of the Crown 4200. 

1. Start a new custom panel by going to the Custom Panel tab and selecting “Create"
         
   
    
2. Now open the Channel 1 Source selector factory panel.
   
    
3. Next, holding down the <Control> key, drag the “Routed" fader to the new custom panel
   
    
4. Right Click on the Control to access the Parameter Address Editor
   
    
5. The full address is now visible in the Parameter Address Window.  In this example, this is HiQnet Address 3, Virtual Device 0, Object Address [5.40.1] and Parameter Index 0.
   
    

3 HiQnet Message Format

The following section lists the detailed message formats for the common HiQnet messages.  See section 3.4.1 for an explanation of the datatypes and how they are stored.

3.1 Header

This is the common header for HiQnet messages.  The first field is for HiQnet message version.  The current HiQnet version is 0x02, please use this as the default. 

Optional 'Error' header (FLAGS=0x0008):

Optional 'Multi-part' header (FLAGS=0x0040):

Optional 'Session Number' header (FLAGS=0x0100):

When using multiple header extensions in a single packet they must be added to the end of the header in the same order as they are listed above.

A Device will send an error header back as a reply to a received message with a header extension it does not understand.  Older Devices do not support sessions, for example.  Some newer Devices require sessions always.  Other Devices will support sessions, but allow session-less communication.  So always start a session with a Hello(Query) with a session number, and if the Device replies with a Hello(Error) header, then proceed with session-less communication with that Device.

• If an error message is returned in response to a Hello message, a MultiParamGet(NumParams=0) message will be used for backward compatibility in order to start Keep Alives.
• See the sessions section.
   
  Messages may originate from anywhere in the hierarchy – 0xDEVICEVDOBJECT.

3.1.1 Version

The Version Number indicates the revision number of the entire protocol; it is not used for differentiating between revisions of individual messages.  HiQnet is currently at revision 2.  Devices that communicate with HiQnet version 1.0 include the dbx ZonePro family.  All others use version 2.0.

3.1.2 Header Length

The Header Length is the size in bytes of the entire message header, including any additional headers such as 'Error' or 'Multi-part'.

3.1.3 Message Length

The Message Length is the size in bytes of the entire message - from the 'Version' field through to the last byte of the payload.

3.1.4 Source Address

The Source Address specifies the HiQnet address where the message has come from; this is often used by the recipient for sending back reply messages.

3.1.5 Destination Address

The Destination Address specifies where the message is to be delivered to.

3.1.6 Message ID

The Message ID is a unique identifier that indicates the method that the destination Device must perform.  If there is a payload, it is usually specific to the type of method indicated by the Message ID.  Product-specific IDs may also exist and will be documented appropriately.

3.1.7 Flags

The Flags denote what kinds of options are active when set to '1' and are allocated in the following manner:

Bit 5 must be set for any applications using TCP/IP only on the network interface.  This will ensure that any messages are sent guaranteed (TCP rather than UDP).

3.1.8 Hop Count

The Hop Count denotes the number of network hops that a message has traversed and is used to stop broadcast loops.  This field should generally be defaulted to 0x05.

3.1.9 Sequence Number

The Sequence number is used to uniquely identify each HiQnet message leaving a Device.  This is primarily used for diagnostic purposes.  The sequence number starts at 0 on power-up and increments for each successive message the Routing Layer sends to the Packet Layer.  The Sequence Number rolls over at the top of its range.

3.2 Types of Messages

3.2.1 Request Acknowledgement Message

The ReqAck flag is used to provide a message level service that can be used by the sender to know when the recipient of the message has carried out the specified action.  When the message sender sets the ReqAck flag, the message recipient, upon performing the specified action, will send back the same message to the sender with the ReqAck flag cleared and the Ack flag set.  This provides a valuable mechanism because the Ack is not sent upon receipt of the message which would mean “I have received your request;" instead by sending the Ack upon performing the action this literally means “I have done it."  If the original message had any data in its payload, that data is not sent back with the acknowledge message.  This mechanism can be used to key actions such as the sending of the next message only once the recipient has serviced the first message.

3.2.2 Acknowledge Message Flag

As specified above, a message with the Ack bit set means I have performed the requested action.

3.2.3 Information Message Flag

The Information flag is normally used to denote a response to a request.  Instead of defining a new message ID for the reply to each request, we have decided to encode the reply ID using the Information flag.  For example, the response to a Get message is also a Get with the Information flag set and the corresponding data appended in the payload.  Note that the Information and Ack flags may be used together.  If you receive a Get message with the ReqAck flag set, then your response would be a Get with the Ack and Information flags sent and any corresponding payload appended to the end. 

The Info flag is also used to indicate unsolicited informational messages (a message that is neither a request nor a response to a request).  Again, the Information flag simply means the message is not a request!

3.2.4 Error Message Flag

In the case of an error in a received message, the 'error' flag must be set and an error code and error string appended to the end of the message header.  The original message is then returned to the sender. 

3.2.5 Guaranteed Message Flag

When set, the 'guaranteed' flag indicates the message must be transmitted on a network service with guaranteed delivery.  A cleared flag denotes a preference for the message to be sent via an unreliable, datagram service.  In this latter case, the Routing Layer may in some circumstances (such as proxy) choose to send the message via the guaranteed service instead.

1 – Guaranteed 

0 – Unreliable datagram

3.2.6 Multi-part Message Flag

When set, this flag indicates the message is part of a multi-part message sequence and the message header is extended by the addition of the 'Multi-part' header. 

1 – Multi-part message

0 – Single-part message

Many Devices will not be able to send a single message large enough to contain all the data they wish to transmit (such as the data set required for a preset change).  For this reason, we provide a means of sending multi-part messages where the payload is spread over a number of messages, which together form all the data required for a single method. 

The algorithm for multi-part messages is as follows:

1. Preparing the first multi-part message header
• Set the 'Multi-part' flag.
• Copy the 'Sequence Number' to the 'Start Sequence No.'
• Set 'Bytes Remaining' to be the data outstanding, including this data.  This is the total size of the payload, not including header.
• Transmit the first message.
2. Preparing the remaining multi-part message headers

• Set the 'Multi-part' flag.

• Set the 'Start Sequence No.' to that used in the first multi-part message.

• Set 'Bytes Remaining'to be the data outstanding, including this data.  Does not include the previous messages' payload size.

• Transmit the current message.  This is the last message when 'Bytes Remaining' is equal to this message's payload.

Destination knows when the last message is being received because the 'Bytes Remaining' in the last message's multi-part header is the same as the size of the payload in the last message.

           
 

3.3 Device Level Methods

                           METHOD NAME            MESSAGE ID            PURPOSE

3.3.1 Get Attributes

INFORMATION (response to message):

3.3.2 GetVDList

INFORMATION (response):

3.3.3 Store

The Store method saves various types of performance data into non-volatile local storage such as FLASH.

UBYTE  ubyStoreAction

UWORD  uwStoreNumber

STRING strWorkgroup

UBYTE  ubyScope

The 'Store Action' determines the type of data affected:

The uwStoreNumber parameter identifies a local storage space and undergoes no translation or mapping to another value. 

The strWorkgroup parameter is not used and should be set to 0.

The ubyScope parameter is reserved for future definition. 

Devices that are unable to perform the store operation will return an error.

The Store(info) message allows a Device to indicate to a subscribed Device that a storage location has been modified.  The source of the data stored into non-volatile storage is not inferred.  The payload indicates the storage location that has been modified.

Store(info) allows synchronization between multiple System Architects subscribed to a Device whenever a change in the configuration state occurs.

INFORMATION:

3.3.4 Recall

Activates various kinds of performance data. 

3.3.5 Recall Action Determines Type of Data Affected

Recall Action

1. – Parameters    (parameters only)
2. – Subscriptions            (Subscriptions only)
3. – Scenes                      (1 to N PARAM + Subscriptions)
4. – Snapshots     (All PARAMs + Subscriptions)
5. – Presets                      (Config + Snapshot)
6. – Venue            
   
   For actions 0 to 4, RecallNumber identifies a local storage space and undergoes no translation or mapping to another value. 
   
   For action 5, RecallNumber identifies a 'venue recall number', which each Device translates into a 'local recall' and 'local action' via the 'Venue Table'.
   
   The Venue Table for Devices can be examined and modified through System Architect.  See the Tools/Venue Recall button on the ribbon.
   
   Some Devices are not required to do anything for a specific 'Recall Number'; these may enter the enumerated value 'No Action' in their Venue Table.
   
   The Workgroup Path indicates which Devices are to respond to a 'venue recall'.  Devices that are outside of the specified workgroup will take no action.  For all other recall actions this parameter is not used and should be set to 0.
   
   The Scope parameter is reserved for future definition. 
   
   Devices that are unable to perform the requested recall will return an error.  See the Event Log sectionfor the format of the Event Log Subscription Information message that is sent from Devices when an errors occur.

3.3.6 Locate

The 'locate' method requests that the receiver makes itself 'visible' to the customer by flashing its 'Locate LEDs'.  If available, these are typically located on the hardware panel of the product. 

The locate method is compulsory for Device Manager Virtual Devices. Virtual Devices and Objects may optionally choose to support it.

The method has a single parameter:

                            UWORD uwTime             - Locate time in milliseconds

0x0000 – Turn off locate LEDs

0xFFFF - Turn on locate LEDs.

Time periods between 0x0001 and 0xFFFE indicate a period of time during which the locate LEDs must flash.  After the time period is completed the LEDs must be turned off.

The locate method will flash the LEDs at a rate of 2Hz.  This allows the 'locate' signal to be differentiated from product-specific flashes which may be active on the same LED (some products only have one LED).

3.3.7 Locate Message

3.4 Event Log

Each HiQnet Device has an Event Log.  Items reported into the Event Log such as protocol errors or product-specific errors can be transmitted onto the network.  If you subscribe to Device foo's Event Log and Device bar sends foo a malformed packet, because you are subscribed to foo's Event Log, foo will send you an event log message telling you it has received a bad message from Device bar.

3.4.1 Event Log Data

3.4.1.1 Category

Category identifies a sub-system within the product into which associated Event IDs are grouped.  There may be no more than 32 event categories, those already declared as are follows:

1. – Unassigned
2. – Application
3. – Configuration
4. – Audio Network
5. – Control Network
6. – Vendor Network
7. – Startup
8. – DSP
9. – Miscellaneous
10. – Control Logic
11. – Foreign Protocol
12. – Digital I/O
13. – Unassigned
14. – Unassigned
15. – Control Surface
16. – 31 Unassigned
    
     
    
    The Category is represented in some messages by an enumerated UWORD and in others as a ULONG bit-field.
    
     

3.4.1.2 Event ID

The Event ID identifies the actual event that triggered a log entry.  These are held within a zero based enumeration with the range of a UWORD.  The Event ID may be 'overloaded' across event categories; that is to say, an Event ID such as zero may mean 'Device Started' within one Category and 'Preset Recalled' in another. 

The Event ID range is divided into two sections:

0x0000 – 0x7FFF Æ Global Event IDs common across all products

0x8000 – 0xFFFF Æ Custom Event IDs specific to a Device Manager Class ID

3.4.1.3 Event ID Definitions

The Global Event IDs for each category are given below:

Control Network Event IDs

3.4.2 Requesting Event Log Client Subscriptions

A client requests an event log subscription by sending a SubscribeEventLog(I) message to the Device Manager Virtual Device of the Device it is interested in receiving events from.  The 'Category Filter' indicates which categories of event the client wishes to hear about.

The category is a field of type ULONG, a '1' indicates a subscription; '0' means no action (it does not mean unsubscribe).  Each category is represented by a bit in the ULONG.  The same client may subscribe multiple times to the same Device.  The server receiving the SubscribeEventLog message will perform an OR operation on the current filter settings and the ones in the message payload; the result will be the actual categories subscribed to:

The 'Max Data Size' allows the client a chance to limit the size of any 'additional data' sent to it as part of an event log entry.  It is the server's responsibility to ensure this data is not larger than the figure specified by the client.

3.4.2.1 Sending Subscribed Events

Once an event log subscription is activated, the server will send the client a RequestLogInfo(I) message.  Typically, there will be one message per event, but the server is able to package multiple events per message and so the client must be able to handle a message containing multiple events.

3.4.2.2 Cancelling Client Subscriptions

A client may cancel a category of event log subscription by sending the server an "Unsubscribe Event Log" message.  The 'Category' in the payload specifies the category of events the client is unsubscribing from; this may be a single category or multiple categories.

Event Log subscriptions, along with all other types of subscriptions, are automatically cancelled if you send the Device a Goodbye message.

A '1' bit represents 'unsubscribe', a '0' represents 'do nothing'. 

For example:

3.4.2.3 Protocol Errors

A Protocol error is a special kind of event which occurs when a Device receives a HiQnet message it cannot service.  This may be for any number of reasons but usually because the message is incorrectly formatted, addressed to the wrong destination, or contains out of range parameter data.  All HiQnet Devices must be able to generate and handle error messages.

The error must be reported back to the message sender by returning a HiQnet 'error' message.  The error message contains both an error code and a networkbyte-ordered Unicodestring representation of the error which may be used by a technician for diagnosing the fault.  The protocol error codes which may be used in the message are the same as the Event IDs enumerated in the 'Control Network' category of the Event Logging section. 

Unlike event logging, when a protocol error occurs, the error message is always returned to the sender, regardless of any event log settings. 

3.4.2.4 Subscribe Event Log Messages

3.4.2.5 Unsubscribe Event Log

3.4.3 Request Event Log

3.4.4 Request Event Log INFORMATION (response):

3.4.4.1 Priority

The Priority field allows events in the Event Log to be assigned one of three levels of importance, these are enumerated as follows:

1. - Fault
2. - Warning
3. - Information
   
   The Priority is represented by a UBYTE.

3.4.4.2 Sequence Number

The Sequence Number denotes the order that events occurred in.  This field is important for products that do not have a real time clock with which to generate the time and date fields of the event log.  The sequence number starts at 0 the first time the unit is powered on and continues to increment by one for each generated event.  The sequence number must be preserved in non-volatile storage so that it persists across power cycles and firmware upgrades.

The Sequence Number is a ULONG.

3.4.4.3 Time

Time represents when the event was generated and logged.  The format is 24 hour clock, with two digits for hours, minutes & seconds separated by a colon For example:

17:25:47

The Time is transported via the STRING data-type.

3.4.4.4 Date

Date represents the day the event was generated and logged.  The format is year, month & day separated by hyphens. For example: 2004-12-13

The Date is transported via the STRING data-type.

3.4.4.5 Information

This is a string giving any additional information about the event.  The data-type is STRING.

3.4.4.6 Additional Data

Additional Data is a BLOCK field for events which carry event-specific data.  This could for example, be a copy of a HiQnet message which when processed triggered an internal error within a Device.  The format and purpose of the data may also be event-specific, it is not required that the recipient should necessarily be able to understand or want to use the extra data. 

The maximum size of this data is a decision left open to the product designer.  However, the product must be able to truncate the data to the maximum size requested by a client that subscribed to the event log.

If an event does not carry additional data then the length of the BLOCK must be set to 0.

3.5 Introduction to Parameters

3.5.1 Data Type Definition

A small set of data types are used in HiQnet to represent parameter values.  All parameters must use one of the following data types:

                               4

The format for BLOCK is a variable length of memory that may be used for any kind of data.  The first two bytes are a UWORD that contains the size of the block in bytes not including the UWORD itself.  The maximum size of the BLOCK is constrained to 65536 bytes.  Because blocks can be used to represent any kind of structured data, it is assumed that the sending and receiving sides know how to format the data. 

Strings are Unicode and stored using the String data type.  Like BLOCK, the actual string data is prefixed with a UWORD count indicating the length of the string in bytes.  Strings sent over the network are to include the NULL character at the end of the string.  Because we are using Unicode, the length used in the String format is 2 * (strlen + 1).  For example:  for the string “Hello World," the count will be 24. Note, strings are network-byte-ordered Unicode while on the wire.

3.5.2 Sensor/Non-Sensor

Sensor parameters are those that update periodically such as a meter.  Non-sensors are normal parameters that update only when changed. Examples of non-sensors include Frequency or Gain.

3.6 MultiParamSet

Set 'NumParam' parameters values within an object or Virtual Device. 'Param_ID' specifies the particular parameter to set.  The payload example below shows an array of three Parameters.

3.7 MultiParamGet

Get 'NumParam' parameters values within an object or Virtual Device.

'Param_ID' specifies the particular parameter to get.

3.7.1 INFORMATION:

3.8 MultiParamSubscribe

Subscriptions are used so that the client may be automatically notified when a parameter has been changed.  Because the HiQnet model is a peer-to-peer model, you may specify the receiving destination parameter.  This might be useful when your controller only has a few parameters in it that you want to map across the network.

The sensor rate is the fastest that the client wishes to receive updates for sensor parameters.  Based on workload, the server may choose to send the updates slower.  The sensor rate is ignored for non-sensor parameters.

3.9 MultiParamUnsubscribe

3.10 ParamSetPercent

ParamSetPercent sets the value of a parameter using a Percentage Value.  Unlike ParamSetPercent allows simple control of any parameter without the need to provide control law conversions or format the value in the native format required by the parameter Class ID.  Indeed, when using ParamSetPercent, no prior knowledge of the parameter's attributes are required because the burden of dealing with Data Type formatting, range limiting and control law conversions are off-loaded from the controller to the controlled.

Let's compare controlling a parameter with ParamSet and ParamSetPercent.  Suppose you wanted to perform a ParamSet on a Param with a Class ID of ClassPeqFreq and the following attributes:

                           ParamClass                       ParamClassPeqFreq

                                 Data_Type                  ULONG

                                 Max                             20000

                                 Min                             20

                                 Control Primitive        LOG

The attributes state that the Param is a frequency value in the range 20Hz to 20,000Hz and stored in a ULONG.  The Control Primitive recommends a logarithmic mapping between the surface control and the Param.  This ensures that adequate resolution is given at the bottom end and the control has a good feel to the user when they adjust the frequency. 

ParamSetPercent requires no prior knowledge of the parameter's attributes.  A control surface wishing to set the value just sends down a percentage value between 0 and 100%.  The ParamSetPercent method will take care of everything – even the logarithmic Control Primitive. 

ParamSet, in contrast, requires that the frequency value sent to the parameter be formatted in the manner dictated by the parameter's Class ID – a ULONG in the range 20 to 20,000.  Further to that, the Control Primitive recommends a LOG conversion; the control surface wishing to do this would need to provide a translation from the linear units that an encoder or fader outputs into the log curve requested by the parameter.  Of course prior to this, the control surface must have had prior knowledge of the parameter's Class ID or had issued a query for the attributes.             

The Percentage Value is encoded as a UWORD in 1.15 fixed-point format:

In 1.15 format, bit 15 is used to indicate the sign and the remaining bits 0 to 14 are used to represent the fractional part of the number. 0x8000 represents –1.0 and 0x7fff represents1.0 – 1/32768. The method uses an implicit scale factor of 100%, so 0x8000 = -100% and 0x7FFF = approximately 100%.

3.10.1 ParamSetPercent Message

3.11 ParamSubscribePercent

The ParamSubscribePercent method sets up a percentage-subscription.  It functions pretty much the same way as MultiParamSubscribe, except subscriptions are sent out using the ParamSetPercent message, rather than MultiParamSet.

UWORD      uwPublisherParamID

UBYTE      ubySubscriptionType

HIQNETADDR hiqSubscriberAddress 

UWORD      uwSubscriberParamID

UBYTE      ubyMode

UWORD      uwModeParamID

UWORD      uwSensorRate

uwPublisherParamID –Param being subscribed to

ubySubscriptionType – Set to zero

hiqSubscriberAddress – Object to send the subscription to

uwSubscriberParamID – Param to send subscription to

ubyMode – Sets display mode for Parameter

uwModeParamID – Subscriber Param to receive the 'ubyMode'

uwSensorRate – Period rate in ms for sendingsensor Params

The ParamSubscribePercent method is invoked by a subscriber to request that the publishersend the subscriberthe value of a Param whenever it has been changed by a third party; this is the subscription.  The Param being subscribed to is specified by the PublisherParamIDparameter.  Where the subscription is for a sensor Param, the subscribed Param value will actually be sent out on a periodic basis, whether it has changed value or not.  The subscriber can request a specific rate via the SensorRate parameter.  Whenever the Param value, is transmitted to the subscriber, the message is sent to the location the subscriber requested; the hiqSubscriberAddress and uwSubscriberParamID.

3.11.1 ParamSubscribePercent Message

This message shares the same payload as the MultiParamSubscribe message.

4 HiQnet Network Model

The HiQnet architecture is transport agnostic.  Each network type has a common layering structure as shown above.  Message layer functions have been defined above; we now dive into the routing and transport layers.

4.1 Routing Layer

4.1.1 Routing Layer Introduction

The Routing Layer in HiQnet performs two main functions – the abstraction of different networking architectures, and the routing of messages between them.  The goal of abstraction is to isolate the Application Layer code from the differing characteristics of the underlying protocols and Physical Layers.  The Application Layer code need not know that a particular Device is on TCP/IP or Token Bus, or Ethernet or RS485; it just needs to know that the Device can be communicated with over a standard interface.  The Routing Layer takes care of locating the Device and sending messages to it over the appropriate network.  In this function, the Routing Layer may also be queried by the Application Layer code for network specific information; this can be useful for monitoring purposes or when a user wants direct control over specific settings such as the IP address of a Device. 

The second main function of the Routing Layer is to route messages across different Packet Service Layers.  A product may serve as a bridge onto a different network type or a router between similar network types.  Messages are received in on one Packet Service Layer, and in the Routing Layer passed up to the Message Layer service if the message is for this Device, or onto another Packet Service Layer if the message is actually for a different Device. 

The Routing Layer's routing function is not to be confused with any routing that may occur within a protocol's Network Layer.  The former routes a HiQnet message across a variety of Network Protocols and physical media, whilst the latter typically only routes a datagram within a specific Network Protocol. 

4.1.2 Transmitting Messages

The Routing Layer offers two services for transmitting messages, Datagram and Guaranteed. The service used to convey a message across the network is determined by the 'Guaranteed' flag in the HiQnet Message Header. 

With some messages, in some circumstances, the Routing Layer will choose to override the service specified in the message header and send the message out on a different service. 

4.1.3 Datagram Service

The Datagram service provides a fast connectionless means of transmitting messages to either unicast or broadcast addresses.  The service does not guarantee delivery nor protect against duplicate, missing, or out of sequence messages.  Typically, this service is used for locating other Devices on the network (Device Discovery) and for transmission of sensor/meter data.

4.1.3.1 Guaranteed Service

The Guaranteed service provides a reliable, connection-oriented means of transmitting messages to unicast addresses. The service will guarantee delivery and protect against duplicate, missing, or out of sequence messages. Typically, this service is used for all HiQnet messages except those that are more appropriately sent via the Datagram service.

4.1.3.2 HiQnet Device Addressing

All products need a HiQnet Device Address prior to initiating or receiving messages across a network.  This can be obtained either by 'address negotiation' or by hard coding.  Devices that have negotiated for their initial address must also be ready to accept an address received via a message from the network. 

It is strongly recommended that devices in the HiQnet system have their node addresses fixed.

4.1.3.3 Negotiating a HiQnet Address

The Device as initially shipped (or factory reset) will see that its HiQnet Address is zero, so it will set its HiQnet Address to a number between 1 and 65535 and use broadcast messages to "negotiate" its address with other Devices already on the network.  After that process settles down, the HiQnet Address will be fixed at that chosen number.  Devices never re-negotiate their HiQnet Addresses once they are non-zero.

The SetAddress message is normally used to set the Device to a specific non-zero HiQnet Address so that it makes sense and fits into a user's specific numbering scheme.

4.1.4 DiscoInfo

The DiscoInfo message is central to all Routing Layer activities.  For this reason, we introduce the message early because a basic grasp of its various forms and uses is essential to understanding almost any aspect of the Routing Layer.

The DiscoInfo message is used to:

• Announce the arrival of a Device on the network.
• Search for other Devices on the network.
• Exchange routing information between Devices.
• Keep a network connection Alive 
  
  The DiscoInfo message takes two forms, the Query and Info; these are often referred to as DiscoInfo(Q) & DiscoInfo(I).  The Q indicates that the 'Info' bit in the Message Header flags field is zero, the I indicates that the 'Info' bit is set to one.  The payload is the same for both forms of message, and their usage is now explained in more detail.

4.1.4.1 Query

DiscoInfo(Q) is a Get type message which serves two purposes.  The primary purpose is to find a Device on the network; the secondary purpose is to pass onto the receiving Devices some additional information about the sender.

4.1.4.2 Info

DiscoInfo(I) is a message which may be a reply to a DiscoInfo(Q), or it may be sent as an informational message.  Either way, the primary purpose is to convey routing information about the sender to the recipient. 

4.1.5 NetworkInfo

The Routing Layer has a number of sub-systems that rely on the exchange of Network-specific information.  Within this document, for the purposes of clarity, this information is referred to generically as 'NetworkInfo'.  There is a NetworkInfo structure for each type of Protocol/Physical media combination. The Packet Layer is responsible for providing a NetworkInfo structure representing each of its particular network interfaces.  In implementation terms, this is the means by which the Routing Layer achieves its aim of abstracting different network architectures.  By grouping differences together within a common structure supplied by the Packet Layer we are able to design and implement the Routing Layer in a generic manner.

The NetworkInfo structure forms a part of the DiscoInfo message payload and this is the means by which each Routing Layer communicates network-specific information between Devices.

So that the Routing Layer may identify what kind of NetworkInfo structure is attached to a DiscoInfo payload, each type of NetworkInfo is identified via an enumerated UBYTE.  This is called the NetworkID and is enumerated as follows:

1. – TCP/IP
2. – Reserved
3. – Reserved
4. – RS232
   
    

4.1.6 Device Arrival “Announce"

A Device will announce its arrival by transmitting 5 DiscoInfo(I) messages on to the network at regular 2 second intervals.  Prior to transmitting, the Device must pause for a random period of between 0 and 2 seconds.

HiQnet Address Negotiation will have been completed before the Device Arrival procedure executes.

4.1.7 Device Departure “Goodbye"

A Device which has power-fail detection (or can be manually shutdown as in the case of closing a GUI), must notify other Devices it is communicating with of its departure from the network via a 'Goodbye' message.  The Goodbye allows the remaining Devices to free any resources associated with the departing Device without having to wait for their Keep timeouts to activate the cleanup.  In some situations, such as restarting the Device, this also allows the network to be rediscovered more quickly.

A Goodbye is NOT sent when a Keep Alive timeout occurs.

The closing Device must unicast the Goodbye message to each Device within its Routing Table which has a Keep Alive active.  After this point the Device is assumed to be unavailable and not able to receive or send messages.  The only way the Device may re-enter the network, is to follow the same procedure as for a newly connectedDevice.

4.2 Device Discovery on Demand

Device Discovery is the name given to locating a specific Device on the HiQnet network.  In doing this, the Routing Layer makes use of the lower layers in the HiQnet model services supplied by the underlying layers in the HiQnet model (e.g. TCP/IP on an Ethernet network, Token Network Services on a 485 bus) to determine the presence of HiQnet compatible Devices residing on a given network interface.  The purpose of this level of discovery is to discover the network information necessary to address a Device at a particular HiQnet address.

Each Network Interface must be able to specify the HiQnet Device Address that it can be found on and the Network Address of the Network Interface.  A Network interface may be found on only one Device, but a Device can have more than one Network Interface.  Some types of Network Interface include Ethernet, 485 Token bus, 232 Full Duplex, etc.

The Device Discovery Algorithm is used when a Device wishes to transmit a message to another Device, for which it has no available route.  This 'demand driven' approach minimizes the size of the routing table by only requesting routes for those Devices that need to be contacted.

4.2.1 Searching for a Device

This section illustrates the steps required to search for a specific Device on the network.  The Device to search for is identified by its HiQnet Destination Address; if the Broadcast HiQnet Address is used then the search returns ALL Devices presently on the network – including those Devices which mistakenly have a HiQnet address of zero.

1. Search the Routing Table for a route to the required Device.  If a route already exists (the Device Address searched for matches a Device Address in the table), use it.
2. If no route exists, network-broadcast on the datagram service a DiscoInfo(Q) message specifying the requested Device in the 'HiQnet Device' field.
3. Wait three seconds for a DiscoInfo(I) message(s) to arrive with the HiQnet Device address which matches that requested in step 2.  If the correct message does not arrive, repeat from step 2 for a maximum of three attempts.  If after three attempts the no matching response has been received, exit with a 'not found' error.
4. If a DiscoInfo(I) message is received from the Device being searched for, create a new entry in the Routing Table using the following information extracted from the message:
   
   Device Address
   
   Serial Number
   
   Cost
   
   Max Message Size
   
   Network Address
   
   
   The remaining item, the Interface is expected to be supplied by the Routing Layer itself – it must know which interface the DiscoInfo(I) came in on.
   
   If the search specified the HiQnet broadcast address, many DiscoInfo(I) messages may arrive in return to the DiscoInfo(I).

4.2.2 Keep Alive/Device Departure

Having found a Device and established a means of communication with it, the Routing Layer becomes responsible for tracking the Device's continued participation in the network.  If, for whatever reason the Device should become un-contactable, it is the responsibility of the Routing Layer to notify the Application Layer that the Device has become unavailable.  The 'keep alive' mechanism is the means by which the continued presence of a Device is checked and monitored.

We strongly advise the use of a guaranteed service (where available) for Keep Alive.

4.2.2.1 When to initiate Keep Alive

Keep Alives are only triggered by the Hello/Hello(info) messages

4.2.2.2 Keep Alive Period

The Keep Alive Period (KAP) determines the maximum period of inactivity allowed on a route before it is 'timed-out' and deemed invalid.  A source Device specifies the KAP it requires of the destination by placing it in all the DiscoInfo messages it sends. 

The Keep Alive Period is specified in milliseconds and has a DataType of UWORD. The minimum permissible period is 250ms. Normally the KAP is 10 seconds (10000ms).

4.2.2.3 Guaranteed

For each guaranteed connection (such as TCP), a Device must transmit a DiscoInfo(I) message 'KAP' milliseconds after it last transmitted a DiscoInfo(I) or Application Layer message. The period 'KAP' is the 'Keep Alive Period' that was specified in the latest DiscoInfo message received from the source Device.  The destination Device shall time-out a route when it has received no messages within the timeframe of the Keep Alive Period.

If the connection is terminated to save resources, the keep alive must transfer to the Datagram service.  If at a later time, the guaranteed connection is reestablished the keep alive must move back to guaranteed service.

4.2.2.4 Datagram

For each discovered route that uses the datagram service exclusively (such as UDP), a Device will unicast a DiscoInfo(I) to the destination 'KAP' milliseconds after it last transmitted a DiscoInfo(I) or Application Layer message. 

4.3 Table of Routing Layer Message IDs

4.3.1 DiscoInfo

4.3.2 GetNetworkInfo

INFORMATION MESSAGE:

4.3.3 Request Address

4.3.4 AddressUsed

4.3.5 SetAddress

4.3.6 Goodbye

4.3.7 Hello Query

See the Sessions chapter discussion. A Hello(Query) does not yet have a header extension, but the answering Hello(Info) does.

After bootup, a Device will generate a random session number between 1 and 65535.  Then, for each new session until the next bootup, the Device will increment the session number.  This avoids accidentally using the same session number for two consecutive sessions in case of session breakage.

The Hello and Hello(info) message payloads also contain a mask to indicate the supported header flags.

The minimum support for the flag mask is:  0x01FF (see the Flags chart)

• Bit 0 – Request Acknowledgement
• Bit 1 – Acknowledgement
• Bit 2 – Information
• Bit 3 – Error
• Bit 4 – Reserved
• Bit 5 – Guaranteed
• Bit 6 – Multi-part message
• Bit 7 – Reserved
• Bit 8 – Session Header Extension
  
   
  
             

4.3.8 Hello Info

After a session has been established between two Devices, each Device will put the OTHER Device's session number in all message header extensions before sending the message to the other Device.  Each Device will check all received messages, and ensure the SESSION NUMBER in the Header Extension matches its own local session number that it generated for this session.

4.4 Packet Service Layers

The Packet Service Layer interfaces to the Routing layer above and the Reliability Layer below.  The function of the Packet Service Layer is to provide a means of sending and receiving single HiQnet messages to and from the underlying network architecture.  For example, a Routing Layer may sit upon two Packet Service Layers, one for TCP/IP and another for RS485.

4.5 TCP/IP Packet Service

HiQnet defines a method for packing multiple messages into a single UDP packet.  TCP/IP based Devices making use of UDP must adhere to this standard to ensure interoperability across based products.  Since TCP is stream-based there is no similar requirement for that protocol.

4.5.1 Reliable (TCP) Packet Service

TCP is a stream-based mechanism, accessed usually via a sockets interface.  To minimize overhead, the programmer may wish to aggregate several messages into a single buffer (aka packet) and present them to the TCP socket in one hit; this is an implementation choice for the programmer.

The IANA designation for HiQnet is 'HiQnet-port 3804/tcp Harman HiQnet Port'.

4.5.2 Datagram (UDP) Packet Service

In HiQnet, a sender has the ability to put several messages together into a UDP packet and send them.  The receiver is expected to have the ability to separate out these into distinct messages as long as the maximum buffer size on the receiver has not been exceeded.

The IANA designation for HiQnet is 'HiQnet-port 3804/udp Harman HiQnet Port'

4.5.3 NetworkInfo

This section describes the 'NetworkInfo' structure for an IPv4 based Packet Layer.  The Routing Layer relies on the NetworkInfo to describe each network interface it is bound to.  This structure is appended to every DiscoInfo message the Routing Layer sends out and the information contained within used by each recipient Device to build its Routing Table.  Some GUIs may also display this information to the user. 

In a multi-homed Device, there will be one NetworkInfo structure for each and every network interface.  The NetworkInfo contains the following information:

4.5.3.1 MAC Address

The MAC Address uniquely identifies the network interface associated with the message payload.  On Devices with a single interface, this parameter could be considered redundant since the Serial Number is typically sent within a message and is constructed from the Device's MAC address.  However, on Devices with multiple interfaces, a message may originate from any one of the secondary interfaces and this MAC address allows the recipient to identify each of them. 

4.5.3.2 DHCP Flag

The Routing Layer does not explicitly use the DHCP flag, but this information is exchanged between Devices for informational purposes where there is a requirement to show a map of the network, a GUI for example.  One exception to this is the 'SetAddress' message, which uses the NetworkInfo block to 'set' whether the Device uses DHCP to obtain an IP address.

4.5.3.3 IP Address

The IP Address identifies the network address of the interface a message came from.  In many cases this will be the same as the source address presented by the receiving interface.  However, when a Device is acting as a proxy for others, this address will differ to the source presented by the receiving interface. In terms of implementation, the safest approach is to use this IP Address when it is supplied in preference to the source address. 

4.5.3.4 Subnet Mask

The Routing Layer uses this information to help determine whether it may unicast to another Device.

4.5.4 Gateway

The Routing Layer does not explicitly use the Gateway parameter, this is exchanged between Devices for informational purposes where displaying this information could be useful.

4.5.5 Use Case – Closed loop control of a HiQnet product via TCP/IP – addressing already fixed.

Suppose you desire to control a HiQnet product via TCP/IP.  The Devices all have their IP addresses and HiQnet addresses already fixed.   Assume we have a thirdparty control Device (CD) that wants to control a level in a HiQnet product (HP).

In this case, use the following steps

1. Upon booting of the control Device (CD), it will broadcast on port 3804 a DiscoAnnounce
2. To find the route to the HP, CD will IP broadcast a DiscoQuery with the HiQnet address of CD in the payload.
3. HP replies with a DiscoInfo – giving CD the route information required for HP to open a TCP/IP connection to CD
4. HP opens a TCP connection to CD on port 3804 and then sends a Hello to start Keep Alives
5. CD replies back with a HelloInfo
6. HP sends a ParamSubscribe message to CD
7. CD replies with an ParamSet(I) message, synchronizing the value to HP
8. HP is now ready to send ParamSet messages to CD.  If any other party changes the parameter in question, then HP will automatically receive the change.  If HP would like to get a notification that CD has actually changed the Param, then all ParamSet messages should set the Ack flag. 
9. Periodically HP and CD will exchange Keep Alives via TCP based on the original specified rate. 

4.5.6 Use Case – Open Loop control of a HiQnet product via UDP

Suppose you desire to control a HiQnet product via UDP.  The Devices all have their IP addresses and HiQnet addresses already fixed.  Assume we have a thirdparty control Device (CD) that wants to control a level in a HiQnet product (HP).

In this case, use the following steps

1. Upon booting of the control Device (CD), it will broadcast on port 3804 a DiscoAnnounce
2. To find the route to the HP, CD will IP broadcast a DiscoQuery with the HiQnet address of CD in the payload.
3. HP replies with a DiscoInfo – giving CD the route information required for HP to transmit to CD
4. HP is now ready to send ParamSet messages to CD via UDP
5. Session numbers, Hello(Q/I) and Goobye are not used 
   

5 HiQnet String Settings

System Architect has the ability to copy HiQnet strings to the clipboard to make it easy to use these strings in your control application.  The format of the string that is copied to the clipboard can be customized to meet your needs and these settings are configured from the HiQnet string settings options.

Make sure that “Enable HiQnet String Visibility" is checked. This turns on the feature.

If you are controlling the HiQnet device via IP (Ethernet), then select IP.  If you are using serial then select RS232.  There are some subtle differences in the strings that are created by System Architect between IP and RS232.

Depending on how you need the strings formatted, choose either decimal or hexadecimal, and use the Custom Formatting section to customize further. Pre-defined custom formatting can be selected by choosing AMX or Crestron from the “Format" drop down list.

The Source Address is the address of your device. This is the address that the HiQnet device will reply to.

Enable “Use Placeholder for Parameter Value" if the HiQnet message string should include placeholders for the parameter value instead of the current value for the parameter. See section 2.7.6

6 RS232 Packet Service

This section is outlined as follows:

Introduction

Section 1 – Getting Started/Basic Command Structure

Section 2 – Setting Up and Maintaining a Communication Connection

Section 3 – Recalling preset data

Section 4 – Calculating Checksums

Section 5 – Feedback

Introduction

        The purpose of this section is to show how to connect to a HiQnet Device product, generate and format command strings, calculate a checksum, and receive feedback using the Harman HiQnet Device protocol.  This quick guide is for developers who want to rapidly achieve communications with a HiQnet Device using an open loop approach.  While a closed loop driver would be desired, some developers may choose to implement an open loop driver.  This document will assist developing an open loop driver to control most aspects of a HiQnet Device.

Throughout this section specific reference of Object IDs, Parameter IDs, and Parameter Values are shown.  These specific references originate from a dbx SC.  The same principles are used for any HiQnet device. 

6.1 Getting Started/Basic Command Structure

This section is for RS232, all messages and examples have RS232 formatting included.

IP Connections

      All information contained for RS232 Packet Service Connections is valid for TCP/IP connections except that no FS, FC, or checksum characters are transmitted within a command.  Also, the TCP/IP connection also does not need any PING, ACK/NAK, RESYNC/RESYNC_ACK, or Guaranteed ACK because UDP and TCP have their own link layer.  The IP port is 3804.

6.1.1 Baud Rate

dbx HiQnet Devices are fixed at: 57.6 kbps, 8N1.

6.1.2 Big Endian

Multi Byte data types are sent Big Endian, which means they are send most significant byte first.  If a 16 bit word 0x1234 is sent, it is presented to the transmit register as 0x12 first then 0x34.

6.1.3 Data Types

6.1.4 Resync Request / Resync Acknowledge

The communications protocol uses special characters to synchronize both ends of the serial connection and to keep the connection open.  To synchronize the serial drivers, Resysnc Request and Resync Acknowledge are used.

      Resynch_Request                   0xFF Resysnc_Acknowledge             0xF0

Since these characters are not accepted from the rest of the protocol, the resynch procedure will issue a string of Resync_Request and Resync_Acknowledge bytes to flush the receiving state machine.  To synchronize with a HiQnet Device send 16 Resynch_Request and 261 Resysnc_Acknowledge bytes. 

6.1.5 Ping

To maintain a connection, a PING byte lets the other side know that you are still connected. 

                              Ping                 0xF0 0x8C

If a message has not been sent within the last second, send a PING.  A timeout of 2.5 seconds will result in the HiQnet Device attempting a resync (0xFF's, 0xF0s).  

6.1.6 Resync_Acknowledge Byte

Send this byte at the beginning of every command.  This byte will keep the HiQnet Device from attempting a resync when using open-loop implementation.

                                        Sync                      0xF0

6.1.7 Frame Start Bytes

To indicate the start of a frame, two bytes are used.  The frame bytes will always precede a Message Header.

                               Frame_Start  0x64

                               Frame_Count 0x00, or 0x01-0xFF

For an open loop implementation, use a Frame_Count of 0x00.  This indicates to the receiver that it need not acknowledge the receipt of the message.

6.1.8 Basic Command Structure (Unacknowledged – Open Loop)

6.1.9 Number Parameters

How many Parameters are being changed within one Frame.

6.1.10 Parameter_ID

Each control inside an object is assigned a particular Parameter (state variable) ID.  For example, the fader inside a mixer, a particular source inside a router, or a mute button.

6.1.11 Data_Type

Data types held by the Parameter_Val 

Data_Type = 1 for a single unsigned byte of range 0 to 255 (i.e. a Mute button, or Route) Data_Type = 3 for two unsigned bytes of range 0 to 65535 (i.e. a fader)

6.1.12 Parameter_Val

This variable holds the value of the Parameter_ID in question.  For example, if a Router is the object we want to control, the Parameter_Val will be a single byte of a value 0(NONE), 1 (Source 1), 2 (Source 2), etc.  The Parameter_Val for a fader within a Mixer object consists of 4 bytes and will hold a float value between –inf – 20dB].

6.1.13 CCIT checksum

The CCIT checksum is an 8 bit CRC byte used to validate the Header and Payload.  This byte can be calculated by initializing the checksum byte to 0xFF and passing each byte though the following calculation.  The Network_CCITT_8_Table is included in the Appendix.

UCHAR update_bcc(UCHAR current_bcc, UCHAR new_byte) 

{ return Network_CCITT_8_Table[current_bcc ^ new_byte];

}

The checksum is calculated on all bytes in the Frame, Header, and Payload. See Calculating Checksum section for detailed information concerning checksum calculations.

6.2 Setting Up and Maintaining a Communication Connection

In order to send commands to a Harman HiQnet Device unit from a 3^rd party controller, a connection should be kept alive by the continuous sending and receiving of specific commands.  There are two different levels of “keep-alive" or “heartbeat" commands that should be continuously exchanged.  The first is a Ping designed to keep the transmission layer connection open.  This is done by sending a 0xF0 0x8C every 1 second.  The HiQnet Device will respond with a 0x8C.  If the user desires to only send commands and not receive any feedback from the HiQnet Device, then only this connection must be maintained.  If the 3^rd party controller is not capable of parsing commands, storing independent variables, or receiving commands, then the user may send pre-formed commands with a 0xF0 preceding all commands.  This will open a connection just long enough for the HiQnet Device to accept a command.  Setting the protocol to RS232 in the HiQnet String Settings (see section  will automatically prefix the message string copied to the clipboard with the necessary 0xF0.

The second is a Disco command and it is designed to keep a connection open at the Protocol layer.  If a user wants to receive feedback from the HiQnet Device, a Disco must be sent from the 3^rd party controller at least every 10 seconds to keep this connection open.  To initiate a connection with the HiQnet Device, a Disco Broadcast should be sent.  Any HiQnet Device Devices on this connection will answer back. 

The Disco Broadcast command:

RS232 NetworkInfo is as follows:

Typical values used in the RS232 NetworkInfo are:

                                                      Baud Rate= 57600

                                                      Parity= 0

                                                      Stop Bits= 1

                                                      Data Bits= 8

                                                      Flow Control = 0

When using the serial port it is suggested that the Source Node (3^rd Party Device) address 0x00 0x33. So the actual command sent out the serial port is:

0xF0 0x64 0x00 0x02 0x19 0x00 0x00 0x00 0x3E 0x00 0x33 0x00 0x00 0x00 0x00 0xFF

0xFF 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x20 0x05 0x00 0x00 0x00 0x33 0x01 0x00

0x10 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0xFD 0x01 0x02 0x03

0x04 0x00 0x00 0x27 0x10 0x4E 0x20 0x04 0x00 0x00 0x00 0xE1 0x00 0x00 0x00 0x08

0x00 0x(checksum byte)

The HiQnet Device will respond to this Disco message with its address (node).  Parse this incoming message in order to get the HiQnet Device's address.  Once the address is found you must reply to all incoming Disco messages (send at least every 10 seconds whether a Disco is received or not) or the connection will be shut down.  The reply is the same message as the broadcast command, but with the HiQnet Device's address inserted in place of the 0xFF 0xFF.  The incoming Disco message will be:

It will be assumed that inthis example if the dbx node address is 0x00 0x20 (decimal 32).  

0x64 0x00 0x02 0x19 0x00 0x00 0x00 0x3E 0x00 0x20 0x00 0x00 0x00 0x00 0x00 0x33

0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x24 0x05 0x00 0x00 0x00 0x20 0x64 0x00 0x10

0x59 0x22 0xF0 0xAA 0x2A 0x0C 0x11 0xDE 0xBE 0x67 0x00 0x0F 0xD7 0x00 0xC0

0x1B 0x00 0x00 0x07 0xD0 0x27 0x10 0x04 0x01 0x00 0x00 0xE1 0x00 0x00 0x00 0x08

0x00 0x0B(checksum)

So the 3^rd party controller needs to send the following command for every incoming Disco (send at least every 10 seconds whether a Disco is received or not):

0xF0 0x64 0x00 0x02 0x19 0x00 0x00 0x00 0x3E 0x00 0x33 0x00 0x00 0x00 0x00 0x00

0x20 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x20 0x05 0x00 0x00 0x00 0x33 0x01 0x00

0x10 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0xFD 0x01 0x02 0x03

0x04 0x00 0x00 0x27 0x10 0x4E 0x20 0x04 0x00 0x00 0x00 0xE1 0x00 0x00 0x00 0x08

0x00 0x5F(checksum byte)

Once these two “heartbeats" are being continuously exchanged, the connection is open and the 3^rd party controller can send and receive commands as necessary.

6.2.1 Guaranteed Acknowledgement

      Every command that is sent or received starts with the 0xFS 0xFC start frame bytes.  If the Frame Count (second byte) is anything besides a 0x00, then it is a guaranteed service message and a 0xA5 must be sent to acknowledge this message has been received.  The HiQnet Device will try and initiate a re-sync if it doesn't receive the 0xA5 Ack for every guaranteed message sent. The user may send a 0xA5 after every received message if the user doesn't want to deal with guaranteed service requirements.

6.2.2 Resync

      The HiQnet Device will send at least 261 0xFF's if it believes it is out of sync with the controller.  The HiQnet Device will not accept any serial commands in this state.  This happens when the 0xF0 0x8C heartbeat or the Disco commands, is not being sent at required intervals.  This will also happen if a 0xA5 is not received for each guaranteed packet sent. 

6.3 Recalling Preset Data

Recall           0x0125 (Message ID)

HiQnet devices support recalling preset data.  The most common preset to recall is the scene.  To change a scene on a HiQnet Device, issue a recall scene command (Preset type = 2).  If you are using subscriptions, the changed parameter values will come back to you in multiParameterset messages.

The Workgroup Path and Scope

Here is an example for recalling Device Preset 1.

0xF0 0x64 0x00 0x02 0x19 0x00 0x00 0x00 0x21 0x00 0x33 0x00 0x00 0x00 0x00 0x00

0x20 0x00 0x00 0x00 0x000x01 0x25 0x00 0x20 0x05 0x00 0x00 0x03 0x00 0x01 0x00 0x02 0x00 0x00 0x00 0x01 (checksum byte)

6.4 Calculating Checksums

This document will show two ways of calculating checksums.  The first method is usingcode to perform the checksum operations systematically.  The second method is using a Microsoft Excel Spreadsheet and manually entering commands into the spreadsheet.  The spreadsheet will then calculate the checksum.

6.4.1 How to calculate a checksum using code for the HarmanHiQnet Device:

//shown in AMX Netlinx syntax

//CCIT copied from “Full Duplex Data Link Specification .PDF"

//$ = hex

CHAR CCIT[] = {$5E,$BC,$E2,$61,$3F,$DD,$83,$C2,$9C,$7E,$20,$A3,$FD,$1F,$41,

$9D,$C3,$21,$7F,$FC,$A2,$40,$1E,$5F,$01,$E3,$BD,$3E,$60,$82,$DC, $23,$7D,$9F,$C1,$42,$1C,$FE,$A0,$E1,$BF,$5D,$03,$80,$DE,$3C,$62, $BE,$E0,$02,$5C,$DF,$81,$63,$3D,$7C,$22,$C0,$9E,$1D,$43,$A1,$FF,

$46,$18,$FA,$A4,$27,$79,$9B,$C5,$84,$DA,$38,$66,$E5,$BB,$59,$07, $DB,$85,$67,$39,$BA,$E4,$06,$58,$19,$47,$A5,$FB,$78,$26,$C4,$9A, $65,$3B,$D9,$87,$04,$5A,$B8,$E6,$A7,$F9,$1B,$45,$C6,$98,$7A,$24, $F8,$A6,$44,$1A,$99,$C7,$25,$7B,$3A,$64,$86,$D8,$5B,$05,$E7,$B9,

$8C,$D2,$30,$6E,$ED,$B3,$51,$0F,$4E,$10,$F2,$AC,$2F,$71,$93,$CD, $11,$4F,$AD,$F3,$70,$2E,$CC,$92,$D3,$8D,$6F,$31,$B2,$EC,$0E,$50, $AF,$F1,$13,$4D,$CE,$90,$72,$2C,$6D,$33,$D1,$8F,$0C,$52,$B0,$EE, $32,$6C,$8E,$D0,$53,$0D,$EF,$B1,$F0,$AE,$4C,$12,$91,$CF,$2D,$73,

$CA,$94,$76,$28,$AB,$F5,$17,$49,$08,$56,$B4,$EA,$69,$37,$D5,$8B, $57,$09,$EB,$B5,$36,$68,$8A,$D4,$95,$CB,$29,$77,$F4,$AA,$48,$16, $E9,$B7,$55,$0B,$88,$D6,$34,$6A,$2B,$75,$97,$C9,$4A,$14,$F6,$A8,

$74,$2A,$C8,$96,$15,$4B,$A9,$F7,$B6,$E8,$0A,$54,$D7,$89,$6B,$35};

//COMMAND COPIED FROM System Architect String Export 

CHAR DBX[] =

{$64,$00,$02,$19,$00,$00,$00,$1F,$00,$33,$01,$01,$01,$00,$00,$20,$01,$01,$01,$00,$01,$00,$02,$20 ,$05,$00,$00,$00,$01,$00,$02,$01,$01};

BUTTON_EVENT[dvTP,65]

{

PUSH:

    {

    LOCAL_VAR CHAR BCC; //CHECKSUM

    LOCAL_VAR INTEGER I; //LOOP COUNTER

    BCC = $FF;  //INITIALIZE CHECKSUM

    FOR (I = 1; I<= LENGTH_ARRAY(DBX);I++)

                                {

                                BCC = CCIT[(BCC BXOR DBX[I])]; //bitwise XOR each element of command

                                }

    }

}

Checksum (BCC) is $08 for the DBX[] command stated in the example;

6.4.2 Serial String Method

The System Architect String Export will calculate and append the proper checksum to every message desired.  This tool is useful for checking the validity of the 3^rd party controller checksum method..            

6.5 Feedback

The HiQnet Device utilizes a command called 'subscribe' to manage all feedback traffic.  Once a subscribe command has been issued, the HiQnet Device will send a string containing the current state of the Parameter (or Parameter's) every time the Parameter changes value.  This will allow a 3^rd party controller to get Parameter updates when a user changes values via the front panel, HiQnet Device GUI, or from another 3^rd party controller.  The HiQnet Device will continue to send feedback until a 'unsubscribe' command is issued, or until the unit loses power.

6.5.1 ParameterSubscribeAll

This message is used to subscribe to every State Variable under an Object or Virtual Device.  If this message is sent to a Mixer Object, then a feedback string will be generated if a user changes any values within that particular Mixer Object.

The ParameterSubscribeAll command: 

Sample String:  Subscribing to Router on a SC32 (source address 0x0033, dest address 0x0020):

0x64 0x00 0x02 0x19 0x00 0x00 0x00 0x24 0x00 0x33 0x01 0x01 0x01 0x00 0x00 0x20

0x01 0x01 0x01 0x00 0x01 0x13 0x02 0x20 0x05 0x00 0x00 0x00 0x33 0x00 0x00 0x00 0x00 0x01 0x00 0x00 0x00 0x01 0xF9

6.5.2 ParameterUnSubscribeAll

This message is used to unsubscribe to every State Variable under an Object or Virtual Device. 

The ParameterUnSubscribeAll command: 

Sample String:  UnSubscribing to Router on a SC32 (source address 0x0033, dest address 0x0020):

0x64 0x00 0x02 0x19 0x00 0x00 0x00 0x200x00 0x33 0x01 0x01 0x01 0x00 0x00 0x20 0x01 0x01 0x01 0x00 0x01 0x140x02 0x20 0x05 0x00 0x00 0x00 0x33 0x00 0x00 0x00 0x00 0x010x49

Note:  If the HiQnet Device issues a resync request that means the connection has been shut down, and all Parameter's will need to be re-subscribed.  All Parameter's will need to be subscribed to whenever the AC power is lost to the HiQnet Device unit.

Appendix

IP Connections

      All information contained above for Section 6 RS232 connections is valid for TCP/IP connections except that no FS, FC, or checksum characters are transmitted within a command.  Also, the TCP/IP connection also does not need any PING, ACK/NAK, RESYNC/RESYNC_ACK, or Guaranteed ACK because UDP and TCP have their own link layer.  The IP port is 3804.

7 Sessions

For third party control sessions are optional.

Communication between HiQnet Devices is predicated on the history of messages having already been sent. Subscriptions need to be established before subscription values can be sent.  After a configuration has been described to another Device, this configuration is expected to persist.  Session IDs ensure that communication with a particular Device ID has not been broken and that the Device has not rebooted and re-presented before Keep Alive Period has elapsed.

It is possible that Devices can reboot and present themselves within a Keep Alive Period.  This raises the possibility that messages intended for a previous boot cycle may be delivered to a Device on its next boot cycle.  This is undesirable.

Putting the session number in the header of a HiQnet message allows the detection of a break in a HiQnet session.  If the session number does not match the session number expected a session break is detected and coherency between the two Devices is assumed to be lost.

7.1 Starting a Session

A session is opened between two HiQnet Devices with an exchange of Hello(Query)/Hello(Info) messages.  When a Hello arrives, the payload contains the session number(A) that sending Device has generated for this session.  The response to a Hello(Query) is a Hello(Info) with a payload containing the session number(B) the answeringDevicehas generated for this session.  In the header extension of the Hello(Info) is the original session number(A).

Once session numbers are exchanged, any message can be guaranteed to be within the session only if session numbers are checked on a frame per frame basis.  Each Device expects to see its own session number in the header extension portion of the header of every message it receives.

7.2 Detecting a Session Break

A session number is created to be associated with the destination Device.  The session numbers are transmitted through Hello/Hello(info) messages, creating a pair of session numbers identifying the session.  Purging of messages for previous sessions is performed when a new session number is presented in either a new Hello or Hello(info) message or a conflicting session number is found in a HiQnet Header. Keep Alive is triggered.

Associating a session number with a HiQnet Device ID allows a remote Device to determine when a new instance of a Device is presented.  Specifying a session number in the HiQnet header allows a session break to be detected.  Incoming messages intended for the broken session are discarded, as are outgoing messages intended for the broken session.

7.3 Characteristics of a Session

The following list represents characteristics of a session:

• Hello/Goodbye messages are bookends for a HiQnet session.
• If a Hello or Hello(info) message arrives while a session is currently open to the same Node ID, the current session is closed, and the new session supersedes.
• Hello and Hello(info) messages will trigger Keep Alives.
• Each live session between two Devices has a unique pair of session numbers for those Devices.  If a session with a remote Device is closed and reopened a new unique session number is used.
• After booting, the first session number shall be random (not the same session number as on the previous boot cycle).
• If a session number is indicated in the HiQnet Header Flags field, the session number in the header extension is checked against the session number that was generated locally when the session was started with the remote Device.  If the session number in the header does not match the session number that the remote Device is echoing back, the incoming message is blocked.  An error message is sent to the Device that sent the invalid session number.
• A goodbye message is sent to a Device that sends an invalid session number in the header extensioncausing it to be destroyed on the remote Device.  No session number is allowed to be in the header of this goodbye, since no session exists for this remote Device ID.
• If a Goodbye is sent on an open session, the session number is included in the header extension.
• On a Keep Alive timeout, the current session is closed. No Goodbye is sent. Nor is any other message.
• If an error message is returned in response to a Hello message, a MultiParamGet(NumParams=0) message will be used for backward compatibility in order to start Keep Alives.
• Hello/Hello(info) must use guaranteed delivery service. The GUAR bit in the Flags field must be set to 1.
  
               

7.4 Sessions Use Cases

In the following use cases, "Node" means "HiQnet Device." "MultSVSet" means "MultiParamSet."

When conflicting session 3457 is presented by node 1, no error is generated.  The new session supersedes the old session.

END OF DOCUMENT