Copperhill Books

One of the ingenious features of CAN, the bit monitoring, contributes not only to the enormous reliability of Controller Area Network, but it is ironically also responsible for its biggest drawback, the limited physical distance of the CAN bus. Restricted physical distance is definitely an issue especially for applications such as elevators, security systems and other building technologies. The one and only way to extend the usable bus length without compromising the reliability is to lower the baud rate. In cases where reliability is not the main focus, e.g. diagnostics and low priority messaging, a CAN bus can be extended as far as the reach of the Internet.

Bus Length according to ISO-11898

The usable physical distance in a CAN network depends, first of all, on the applied baud rate as shown in the table below.

Baud Rate [kbits/sec] - Distance [m] – Distance [ft]
1000                                30                        150
500                                 100                      300
250                                 250                      750
125                                 500                      1500
62.5                               1000                    3000

While a distance of roughly 150 ft. at 1 Mbit/sec seems to be restrictive, especially for building technologies, a baud rate of 1 Mbit/sec can nevertheless be considered an overkill for such applications. Experience has shown that the vast majority of automation applications can work sufficiently with baud rates of 500 kbit/sec or even 250 kbit/sec and that includes demanding motion control applications.

In all consequence, CAN (including CANopen, a higher layer protocol on top of CAN) provides communication means between intelligent nodes. The keyword is “Distributed Intelligence”, which results in increased system performance. “Intelligence” means that the nodes are responsible for a major part of their control tasks. The more these nodes process internally, the less they depend on the communication means. The advantage of CAN lies primarily in its vast reliability.

Naturally, there are applications that require high speed and CAN provides that as well, however, with the drawback of limited physical distance.

CAN Repeaters

The bus length extension per CAN repeaters is a myth that is unfortunately being maintained by some manufacturers and vendors of these devices. CAN repeaters provide primarily electrical isolation and signal conditioning.

While CAN can be operated with a simple twisted pair of wires, the quality of the CAN bus cable can be a major factor in terms of maximum bus length. Poor cable quality will quickly knock a signal strength down below a receiver’s threshold. The result will be signal errors and consequently increased bus traffic due to error frames and repeated messages.

CAN repeaters can be used to boost the signal strength and maintain standard bus lengths, but not extend them. The only extension is the one from a poor quality network to a properly functioning network. Ironically, CAN repeaters, due to their internal delay times in the range of milliseconds, will actually shorten the usable bus length in the range of several meters.

CAN-to-CAN Gateways

Some vendors in the CAN business offer a variety of interfaces that support the communication between two separate CAN networks (e.g. the CAN-CBM-Bridge by esd electronics). Such devices allow to extend a CAN network by a factor of two, but they, too, have latency times due to reception, processing and re-transmission of CAN frames. Another disadvantage is that, for instance, both CAN networks cannot exchange error frames.

Such gateways, however, also post some advantages such as message filtering – to lower the bus traffic between the networks – and the use of different baud rates in the networks.

Ethernet Gateways

The ultimate way to extend the reach of a CAN application is the use of Ethernet gateways, which consequently even allow the connection to the Internet.

For instance, the EtherCAN device by esd electronics provides operation modes to either connect two separate CAN networks per Ethernet or allow the monitoring of network activities through the Internet.

In the first mode, you can connect two separate CAN networks per Ethernet to maintain almost unlimited physical distances. Each network will need its own gateway, which in turn will contribute to higher latency times. The EtherCAN also supports message filtering and thus can decrease the number of messages between the networks.

The second mode, operation through the Internet, is supported by the EtherCAN’s internal web server. Imagine, having your application run in Australia, but monitoring and analyzing the bus traffic in your office in Chicago. The EtherCAN comes with an extensive PC software package with CAN analyzing and monitoring features.

Summary

An extension of your CAN network is definitely possible, but it is important that you are aware of the drawbacks. The knowledge of the drawbacks, may they be minor or major, is the first important step to select the right solution.

Per definition, SAE J1939 provides serial data communications between microprocessor systems (also called Electronic Control Units – ECU) in any kind of heavy duty vehicles. The messages exchanged between these units can be data such as vehicle road speed, torque control message from the transmission to the engine, oil temperature, and many more.

SAE J1939 and its companion documents have quickly become the accepted industry standard of choice for off-highway machines. It was all too natural that organizations and manufacturers in the agricultural, military and marine industries, rather than re-inventing the wheel, adopted the proven combination of physical layer, Controller Area Network (CAN), and J1939 as the higher layer protocol for vehicles. However, it is in the specific nature of agricultural and military as well as marine applications that slight modifications, including a name change, were necessary.

These “new” protocols are:

• ISO 11783 (a.k.a. ISOBUS) – Agricultural Industry
• MilCAN – Military Applications
• NMEA 2000 – Marine Applications

ISO 11783 (ISOBUS)

ISO 11783, a.k.a. ISOBUS, is a CAN (Controller Area Network) Higher Layer Protocol based on the SAE J1939 standard for forestry and agricultural vehicles. ISO 11783 was a joint development by manufacturers in the agricultural and forestry industry to address the increasing needs for electronic control in the machinery/vehicles they produce.

ISO 11783 consists of the following parts, under the general title Tractors and machinery for agriculture and forestry – Serial control and communications data network:

• Part 1: General standard for mobile data communication
• Part 2: Physical layer
• Part 3: Data link layer
• Part 4: Network layer
• Part 5: Network management
• Part 6: Virtual terminal
• Part 7: Implement messages applications layer
• Part 8: Power train messages
• Part 9: Tractor ECU
• Part 10: Task controller and management information system data interchange
• Part 11: Mobile data element dictionary
• Part 12: Diagnostic
• Part 13: File Server

The ISO 11783 standards can be purchased through the International Organization for Standardization, ISO. Note: The price tags for each document are extraordinary.

The standard is managed by the ISOBUS group in VDMA (http://www.isobus.net). Note: The web site lacks the substance to be taken seriously. The bulk of the little information that exists is based on marketing material and most documents are in German. Technical information is virtually non-existent.

MilCAN

According to the official web site (http://www.milcan.org) : “MilCAN has been defined by a group of interested companies and government bodies associated with the specification, manufacture and test of military vehicles. The MilCAN working group was formed in 1999 as a sub-group of the International High Speed Data Bus – Users Group (IHSDB-UG) when a need was recognised to standardise the implementation of CANbus within the military vehicles community. The mission statement of this newly formed group was ‘To develop, for various application classes in all military vehicles, a common interface implementation specification based on CANbus’.”

Describing the MilCAN standard is not an easy task and the only reason it found its way onto this web site is due to the fact that it is partly based on J1939. It seems that the creators of the protocol tried to satisfy the protective demands of every European member (in this case especially the Germans and Brits) on one side and American companies on the other. One can only appreciate that the circle of members was not extended any further. MilCAN is an inconsistent mixture of CUP, a protocol developed by the German Army (Bundeswehr), SAE J1939, representing the American side, and CANopen, representing the European side.

As a resullt, there are two variants of MilCAN, MilCAN A and MilCAN B. MilCAN A is based on the 29-bit CAN identifier according to SAE J1939, the major difference being that MilCAN A supports deterministic data transfer and accommodates both, synchronous and asynchronous,  data. MilCAN B, on the other hand, is based on the 11-bit CAN identifier and can, at least officially, make use of devices that have been designed for CANopen. Also officially, it should be possible to mix J1939 devices with MilCAN devices on the same bus.

NMEA 2000

Of all the SAE J1939 derivatives, NMEA 2000 seems to be the only consequent and straight-forward adaptation of J1939.  While taking advantage of a proven and ingeniously designed protocol, NMEA 2000 defines only its own messages.

NMEA 2000 is used for marine data networks providing communication between marine specific electronic devices such as depth finders, chartplotters, navigation instruments, engines, tank level sensors, and GPS receivers.

It has been defined and is controlled by the US based National Marine Electronics Association (NMEA). Information on their official web site (http://www.nmea.org) is somewhat sparse. Another web site, http://www.jackrabbitmarine.com, however, provides in-depth information.

NMEA 2000 is a modernized version and replacement of an older standard, NMEA 0183. It has a significantly higher data rate (250k bits/second vs. 4.8k bits/second for NMEA 0183). It also uses a binary message format as opposed to the ASCII serial communications protocol used by NMEA 0183. Another distinction between the two protocols is that NMEA 2000 is a multiple-talker, multiple-listener data network whereas NMEA 0183 is a single-talker, multiple-listener serial communications protocol.

J1939 is a higher-layer protocol based on Controller Area Network (CAN). It provides serial data communications between microprocessor systems (also called Electronic Control Units – ECU) in any kind of heavy duty vehicles.

The main advantages of using CAN as a field-bus technology are reduced wiring (CAN requires only two wires between nodes), extremely reliable communication, easy implementation and improved maintenance and service capabilities, which consequently not only produce better vehicle performance, but also help to reduce production costs.

• J1939-based protocols are used in:
• Diesel power-train applications
• In-Vehicle networks for trucks and buses
• Agriculture and forestry machinery (ISO 11783)
• Truck-Trailer connections
• Military vehicles (MiLCAN)
• Fleet management systems
• Recreational vehicles
• Marine navigation systems (NMEA2000)

The protocol features of J1939 are based on two older SAE (Society of Automotive Engineers) specifications:

1. SAE J1708
SAE J1708 specifies on the physical layer of the communication link. It uses RS485 as an electrical layer operating at 9600 baud. (Note: Unlike RS232/485 there are no message collisions under CAN). Messages under J1708 start with a Message Identification Character, followed by the data information and a checksum. The message length is 21 characters (or less) and each data character is 10 bits long. Each character starts with a start bit of low polarity.

2. SAE J1587
SAE J1587 is a joint SAE/TMS “Recommended Practices for Electronic Data Exchange Between Microcomputer Systems in Heavy-Duty Vehicle Applications”. It regulates the communication and standardized data exchange between ECUs based on J1708 networks.

Note: The situation regarding documents/literature on J1708 and J1587 is as dire as with J1939.

The J1939 specification is described by a number of SAE documents, the SAE J1939 Standards Collection:

J1939
Recommended Practice for a Serial Control and Communications Vehicle Network*

J1939-01
Recommended Practice for Control And Communications Network for On-Highway Equipment

J1939-02
Agricultural and Forestry Off-Road Machinery Control and Communication Network**

J1939-11
Physical Layer – 250k bits/s, Twisted Shielded Pair

J1939-13
Off-Board Diagnostics Connector

J1939-15
Reduced Physical Layer, 250k bits/sec, Un-Shielded Twisted Pair (UTP)

J1939-21
Data Link Layer

J1939-31
Network Layer

J1939-71
Vehicle Application Layer

J1939-73
Application Layer – Diagnostics

J1939-74
Application – Configurable Messaging

J1939-75
Application Layer – Generator Sets and Industrial

J1939-81
Network Management

The motor data needed to select a motor are rated speed, rated torque, intermittent torque, and rotor inertia. However, the best servo motor selection criteria is to use the motor’s performance curve (torque over speed) and to verify it with the application requirements. Not all motor data sheets do provide such detailed information, since some manufacturers prefer to define the rated/intermittent torque and the rated speed of their motors in a more conservatively manner. Under certain conditions it is, however, possible to operate motors beyond their rated data.

Again, the following motor data are essential for the selection process:

• Rated Speed
• Rated Torque
• Max. (Peak) Torque
• Rotor Inertia

The following criteria have to be fulfilled:

• The motor’s rated speed is equal to or higher than the calculated (required) speed.
• The motor’s rated torque is equal to or higher than the calculated (System) RMS torque.
• The motor’s maximum torque is higher than the calculated (System) peak torque.
• The ratio of load to motor inertia does not exceed the user-defined safety factor.

In case of servo motors the ratio of load to motor inertia should not exceed 10:1. Otherwise the motor could start jerking.

Matters become a bit more complex when the motor’s performance profile, i.e. torque over speed performance, is used to determine whether or not the requirements are met.

To start with the basics, we need to see how the rated speed and rated/intermittent( (peak) torque of a motor is being defined. Basically it is at each manufacturer’s discretion how exactly they define the data. The following picture shows a sample of a motor performance profile:

In this example the rated speed (Vn) is defined as the max. possible speed where the motor still supports the continuous torque (Tcont). The max. speed of the motor is actually higher, but the torque will eventually go down to zero. The intermittent torque (Tmax) may be supported even at higher velocity than the rated speed, however, manufacturers tend to provide data that are on the safe side of the motor operation.

The previous profile may also be a simplified derivation from the actual motor performance. The following picture shows a more complex case:

One way to define the motor data would be to set rated speed, rated torque and intermittent (peak) torque at a safe point in the performance profile as shown in the next picture:

The red lines indicate an example to set the “official” motor data. Besides the point that this is a very coarse definition of the motor’s capabilities, this example also shows that the motor would be rejected, since it would (theoretically) not provide the required peak torque.

The following example shows yet another case where the rated motor data would not meet the application requirements, however, the performance profile does support the requirements.

The rated speed of this sample motor has been defined at 2000 rpm, but the torque requirements are low enough that the motor can support the torque even higher than at 2000 rpm.

The motion control engineer must take all of these described circumstances into consideration:

1. If the motor data sheet does not provide the performance profile (torque over speed or vice versa) of a particular motor, the engineer must use the rated data and make sure the following criteria are met:

• The motor’s rated speed is equal to or higher than the calculated (required) speed.
• The motor’s rated torque is equal to or higher than the calculated (System) RMS torque.
• The motor’s maximum torque is higher than the calculated (System) peak torque.
• The ratio of load to motor inertia does not exceed the user-defined safety factor.

2. If the motor data sheet does provide the performance profile (torque over speed or vice versa) of a particular motor, the engineer should still use the rated motor data. The actual selection, however, is based on a comparison of the performance profile with the application requirements, i.e. the engineer verifies whether or not the performance profile supports each torque at the corresponding speed as defined in the duty cycle.

Excerpt from: A Comprehensible Guide To Servo Motor Sizing by Wilfred Voss
Published by Copperhill Media Corpoation

1.    Overview

The vast majority of automated manufacturing systems involve the use of sophisticated motion control systems that, besides mechanical components, incorporate electrical components such as servo motors, amplifiers and controllers.

The straightforward task for the motion system design engineer is to specify the smallest motor and drive  combination that can provide the torque, speed and acceleration as required by the mechanical set up. However, all too often engineers are familiar with the electrical details, but do lack the knowledge how to calculate the torque requirements of the driven mechanical components. This will lead in many cases to improperly sized motion control applications. The impact, economically as well as technically, will be one of the topics in the following chapters.

Modern motor sizing software packages, such as VisualSizer-Professional, provide the convenience of computing all necessary equations and selecting the optimum motor/drive combination within minutes; they are, however, mainly used to circumvent the timely process of selecting a motor manually. While motor sizing programs can have an educational value to some degree, most of them do not provide any reference on how the equations were derived.

Some basic knowledge of inertia and torque calculations can have a profound impact on the motion system performance. Simple details, like when to use a gearbox in a motion system, may not only help to reduce costs, but will most certainly improve the system performance.

2.    The Importance of Servo Motor Sizing

The importance of servo motor sizing should not be underestimated. Proper motor sizing will not only result in significant cost savings by saving energy, reducing purchasing and operating costs, reducing downtime, etc.; it also helps the engineer to design better motion control systems.

2.1    Why Motor Sizing?

The servo motor represents the most influential cost factor in the motion control system design, not only during the purchasing process, but especially during operation. A high-torque motor will require a stronger and thus more expensive amplifier than smaller motors. The combination of higher torque motor plus amplifier results not only in higher initial expenses, but will also lead to higher operational costs, in particular increased energy consumption. It is estimated, that the purchase price represents only about 2% of the total life cycle cost; about 96% is electricity.

Proper servo motor sizing will not only assure best system performance; it also provides considerable cost savings.

The conventional method of servo motor sizing is based on calculations of the system load, which determines the required size of a motor. Standard praxis demands to add a safety factor to the torque requirements in order to cover for additional friction forces that might occur due to the aging of mechanical components. However, the determination of the system load and the selection of the right servo motor can be extremely time consuming. Each motor has its individual rotor inertia, which contributes to the system load torque, since Torque equals Inertia times Acceleration. The calculation of the system torque must be repeated for each motor that is being considered for the application.

As a result, it is not an easy task to select the optimum motor for the application considering the vast amount of available servo motors in the marketplace. Many motors, that are currently in action, have been chosen mostly due to the fact that they are larger than required and were available short-term (e.g. from inventory). The U.S. Department of Energy estimates that about 80% of all motors in the United States are oversized.

The main reasons to oversize a motor are:

• Uncertain load requirements
• Allowance for load increase (e.g. due to aging mechanical components)
• Availability (e.g. inventory)

Not only is the power consumption higher than it should be; there are also some serious technical considerations.

2.2    Technical Aspects

Oversizing a motor is naturally more common than undersizing. An undersized motor will consequently not be able to move the load adequately (or not at all) and, in extreme cases, may overheat and burn out, especially when it cant dissipate waste heat fast enough. Larger motors will stay cool, but if they are too large they will waste energy during inefficient operation. After all, the motor sizing process can also be seen as an energy balancing act.

AC motors tend to run hot when they are loaded too heavily or too lightly. Servo motors, either undersized or oversized, will inevitably start to vibrate or encounter stalling problems.

One of the major misconceptions during the motion design process is that selecting a larger motor than required is only a small price to pay for the capability to handle the required load, especially since the load may increase during the lifetime of the application due to increased mechanical wear. However, as demonstrated in the picture below, the motor efficiency deteriorates quickly when the motor operates below the designed load.

Picture 2.2.1: Example Efficiency vs. Load

Picture 2.2.1 shows an example of two motors, 10 HP and 100 HP. In both cases there is a sharp decline of the motors efficiency at around 30% of the rated load.

However, the curves as shown in the picture, will vary substantially from motor to motor and it is difficult to say when exactly a motor is oversized. As a general rule of thumb, when a motor operates at 40% or less of its rated load, it is a good candidate for downsizing, especially in cases where the load does not vary very much. Servo motor applications usually require short-term operation at higher loads, especially during acceleration and deceleration, which makes it necessary to look at the average (RMS) torque and the peak torque of an application.

There are, however, advantages to oversizing:

• Mechanical components (e.g. couplings, ball bearings, etc.) may, depending on the environment and quality of service, encounter wear and as a result may produce higher friction forces. Friction forces contribute to the constant torque of a mechanical set up.
• Oversizing may provide additional capacity for future expansions and may eliminate the need to replace the motor.
• Oversized motors can accommodate unanticipated high loads.
• Oversized motors are more likely to start and operate in undervoltage conditions.

In general, a modest oversizing of up to 20% is absolutely acceptable.

High efficiency motors, compared to standard motors, will maintain their efficiency level over a broader range of loads (see picture 2.2.2) and are more suitable for oversizing.

Picture 2.2.2: Example High/Low Efficiency Motors

2.3    The Objective of Motor Sizing

The main objective of motor sizing is based on the good old American sense for business: Get the best performance for the lowest price. As we have learned from a previous chapter the price contains the following components:

Purchasing Costs 2%
Repair, Service, Maintenance, etc. 2%
Operating Costs (Electricity) 96%

In order to get the best performance for the best price it is mandatory to find the smallest motor that fulfills the requirements, i.e. the motor that matches the required torque as close as possible. The basic assumption (which is true for the majority of cases) is that small torque is in direct proportion to smaller size, lower costs and lower power consumption. Smaller power consumptions also result in smaller drive/amplifier size and price.

From a technical standpoint it is also desirable to find a motor whose rotor inertia matches the inertia of the mechanical setup as close as possible, i.e. the optimum ratio between load to rotor inertia of 1 : 1. The inertia match will provide the best performance. However, for servo motors a ratio of up to 6 : 1 still provides a reasonable performance. Any higher ratios will result in instabilities of the system and will eventually lead to total malfunction.

In many cases it makes sense to add a gear between motor and the actual load. A gear lowers the inertia that is reflected to the motor in direct proportion of the transmission ratio. This scenario allows to run smaller motors, however, with the price of the gear added to the system. On the other hand the price reduction by using a smaller motor/drive combination may more than just compensate for the gears price.

In review the objective of motor sizing is to:

• Get the best performance for the best price
• Match the motors torque with the load torque as close as possible
• Match the motors inertia with the load inertia as close as possible
• Find a motor that matches or exceeds the required speed

3.    The Motor Sizing and Selection Process

The motor sizing and selection process is based on the calculation of torque and inertia imposed by the mechanical set up plus the speed and acceleration required by the application. The selected motor must be able to safely drive the mechanical set up by providing sufficient torque and velocity.

Once the requirements have been established, it is easy to look either at the torque vs. speed curves or motor specs and choose the right motor.

The sizing process involves the following steps:

1. Establishment of motion objectives

A written outlining of the motion control application will help to establish the necessary parameters needed for the next steps.

• Required positioning accuracy ?
• Required position repeatability ?
• Required velocity accuracy ?
• Linear or rotary application ?
• If linear application: Horizontal or vertical application?
• Thermal considerations Ambient temperature ?
• What motor technologies are best suited for the application ?

2. Selection of mechanical components

The engineer must decide which mechanical components are required for the application. For instance, a linear application may require a leadscrew or a conveyor. For speed transmission a gear or a belt drive may be used.

• Direct Drive ?
• Special application or standard mechanical devices ?
• If linear application: Use of linear motor or leadscrew, conveyor, etc. ?
• Reducer required Gearbox, belt drive, etc. ?
• Check shaft dimensions select couplings
• Check mechanical components for speed and acceleration limitations

3. Definition of a load (duty) cycle

The engineer must define the maximum velocity, maximum acceleration, duty cycle time, acceleration and deceleration ramps, dwell time, etc., specific to the application.

• Define critical move parameters such as velocity, acceleration rate
• Triangular, trapezoidal or other motion profile ?
• If linear application: Make sure the duty cycle does not exceed the travel range of linear motion device.
• Jerk Limitation required ?
• Consideration of thrust load ?
• Does the load change during the duty cycle ?
• Holding brake applied during zero velocity ?

4. Load calculation

The load is defined by the torque that is required to drive the mechanical set up. The amount of torque is determined by the inertia reflected from the mechanical set up to the motor and the acceleration at the motor shaft.

• Calculate inertia of all moving components
• Determine inertia reflected to motor
• Determine velocity, acceleration at motor shaft
• Calculate acceleration torque at motor shaft
• Determine non-inertial forces such as gravity, friction, pre-load forces, etc.
• Calculate constant torque at motor shaft
• Calculate total acceleration and RMS (continuous over duty cycle) torque at motor shaft

5. Motor Selection

The motor must be able to provide the torque required by the mechanical set up plus the torque inflicted by its own rotor. Each motor has its specific rotor inertia, which contributes to the torque of the entire motion system. When selecting a motor the engineer needs to recalculate the load torque for each individual motor.

• Decide the motor technology to use (DC brush, DC brushless, stepper, etc.)
• Select a motor/drive combination
• Does motor support the required maximum velocity ? If no, select next motor/drive.
• Use rotor inertia to calculate system (motor plus mechanical components) acceleration (peak) and RMS torque
• Does motors rated torque support the systems RMS torque? If no, select next motor/drive.
• Does motors intermittent torque support the systems peak torque? If no, select next motor/drive.
• Does the motors performance curve (torque over speed) support the torque and speed requirements? If no, select next motor/drive.
• If the ratio of load over rotor inertia exceeds a certain range (for servo motors 6:1) consider the use of a gearbox or increase the transmission ratio of the existing gearbox. Servo motors should not be operated over a ratio of 10:1.

The following flow chart demonstrates the motor sizing and selection process:

Excerpt from: A Comprehensible Guide To Servo Motor Sizing by Wilfried Voss
Published by Copperhill Media Corpoation