From IoT factory to operating room: how to design a better communication system

“The foundation of Industry 4.0 is a reliable communication infrastructure. Decision makers use infrastructure to extract data from machines, field devices, and factories. To ensure the reliability of robots and human-machine interfaces, we must first have a deep understanding of the underlying technology options.

“

The foundation of Industry 4.0 is a reliable communication infrastructure. Decision makers use infrastructure to extract data from machines, field devices, and factories. To ensure the reliability of robots and human-machine interfaces, we must first have a deep understanding of the underlying technology options.

Although the factory floor and the operating room are completely different, the equipment used must operate reliably and accurately, which is critical to the tasks performed. As devices require smarter systems, more data, and higher fidelity, their demand for bandwidth continues to increase. At the same time, faster communication interfaces must provide the same reliability and safety while resisting environmental hazards and electromagnetic compatibility (EMC). EMC means that the system can play its intended role in its operating environment, without generating electrical noise, nor being excessively affected by electrical noise.

Robot and machine vision

Vision-guided robots can provide higher flexibility and higher production reliability in high-value manufacturing environments. Without visual guidance, the robot can only perform the same task repeatedly until it is reprogrammed. With machine vision, robots can perform more intelligent tasks. For example, in the production line, they can scan defective products on a conveyor belt, and the adjusted robot can pick up defective products, as shown in Figure 1. In hazardous EMC environments (such as factory automation), the reliability and effectiveness of the vision/robot interface is determined by the selected wired transmission technology. There are many ways to implement machine vision camera interfaces, including USB 2.0, USB 3.0, Camera Link, or Gigabit Ethernet.

Figure 1. Camera machine vision and robot-Ethernet, USB or Camera Link interface

Table 1 compares several key indicators of the USB, Ethernet, and Camera Link standards. Industrial Ethernet has many advantages. The maximum length of cables using 2 pairs of 100BASE-TX and 4 pairs of 1000BASE-T1 standards can be up to 100 meters, and the maximum length of single twisted pair cables using the newly launched 10BASE-T1L standard can be up to 1 km. , And the EMC performance is high. The cable using USB 2.0 or USB 3.0 should not exceed 5 meters, unless a dedicated active USB cable is used, and protection diodes and filters are required to improve EMC performance. However, as industrial controllers generally use USB ports with bandwidths up to 5 Gbps, this provides designers with some advantages.

Camera Link requires industrial controllers to be equipped with dedicated frame grabber hardware. USB or Ethernet does not require an additional frame grabber card for the industrial controller. The Camera Link standard first appeared at the end of 2000 and is the most commonly used interface for machine vision systems. Nowadays, USB and Ethernet-based machine vision cameras are more widely used, but applications that require pre-processing of multiple cameras are still using Camera Link and frame grabbers to reduce the main CPU load. Compared with Gigabit Ethernet, even at the basic speed, the amount of data output by the Camera Link standard is as much as twice that, and the output distance is shorter. The Camera Link physical layer is based on low-voltage differential mode signals (LVDS). Since the common mode noise coupled to each line is effectively eliminated at the receiver, it is inherently EMC robust. The EMC robustness of the LVDS physical layer can be improved by electromagnetic isolation.

Through the use of Ethernet on the link between the camera and the robot, and the industrial controller using the IEEE 802.1 Time Sensitive Network (TSN) switch, the synchronization of the industrial camera and the robot operation can be achieved to the greatest extent. TSN defines the first IEEE standard for time-controlled data routing in switched Ethernet networks. ADI provides a full set of Ethernet technology, including physical layer transceivers and TSN switches, as well as system-level solutions, software and security features.

Human Machine Interface (HMI)

Human Machine Interface (HMI) is often used to display data from a programmable logic controller (PLC) through a human-readable visual representation. Standard HMI can be used to track production time, while monitoring key performance indicators (KPI) and machine output. Operators can use the HMI to perform multiple tasks, including turning the switch on or off, and increasing or decreasing the pressure or speed in the process. HMIs are usually equipped with an integrated display; however, an HMI with an external display option has many advantages. The HMI device using the external high-definition multimedia interface (HDMI®) port is smaller and easier to install in a console using a standard DIN power rail, and can also be used to monitor PLCs.

When using HDMI, the cable length can be up to 15 meters, which is convenient for routing to the touch screen and control room, as shown in Figure 2. In an industrial environment, extending HDMI over longer cables is challenging because EMC hazards can affect wiring. When the motor and pump are connected to a DIN rail PLC, indirect transient overvoltage may also appear on the HMI.

Figure 2. Human Machine Interface (HMI) with Ethernet and RS-485 input, and HDMI output.

To ensure the robustness of the system, the interface technology needs to be carefully selected. With the rapid development of industrial Ethernet, fieldbus technology (such as CAN or RS-485) has become popular. According to industry sources, more than 61 million RS-485 (PROFIBUS®) nodes have been installed worldwide, and PROFIBUS process automation (PA) equipment has increased by 7% year-on-year. The installed base of PROFINET (Industrial Ethernet Implementation) is 26 million nodes, and the number of devices installed in 2018 alone reached 5.1 million. 1 As mentioned earlier, the use of Ethernet-based technology can achieve high EMC performance. This is because electromagnetic is written into the IEEE 802.3 Ethernet standard and must be used at every node. RS-485 devices can include electromagnetic isolation to improve noise immunity; protection diodes can be integrated on-chip or placed on the communication PCB to improve resistance to electrostatic discharge and transient overvoltage.

HMI usually needs to resist electrostatic discharge and use ESD protection diodes to improve signal robustness. For industrial HMIs, integrated reinforced isolation can protect operators from electrical hazards. Although reasonable isolation solutions for Ethernet and RS-485 are currently provided, today, video transmission is mainly isolated by expensive optical fibers, which support gigabit transmission speeds. The latest developments of ADI on electromagnetic isolation technology (such as the ADN4654/ADN4655/ADN4656 series, whose data rate can exceed 1 Gbps) provide designers with a competitive and lower-cost alternative solution.

Endoscope

Surgical imaging, including endoscopes, is a unique application that must provide high-fidelity images while ensuring patient safety. The previous generation of endoscopy equipment was called a video endoscope, which used a series of glass lenses and a light pipe to transmit the image from the imaging head to the charge coupled device (CCD) sensor. Using visible light as a medium to transmit images from the patient to the endoscope, this method can isolate harmful currents, but it does not perform well in terms of manufacturing cost and image quality.

Recent surgical imaging equipment overcomes these challenges by turning to digitization, and from CCD to CMOS sensor, which is easy to expand in size and can be embedded in the camera head. After using a CMOS camera, there is no need to connect multiple lenses in series, and the overall image quality can be improved. The reduced production cost makes the use of disposable surgical endoscopes possible, so there is no need to worry about disinfection. The camera is further reduced, making minimally invasive surgery possible.

After turning to the digital endoscope, a high-speed Electronic interface must be provided between the CMOS image sensor (contacting the patient) and the camera controller (CCU). LVDS and Scalable Low-Voltage Signaling (SLVS) layers have gradually become commonly used physical layers for this interconnection, providing high bandwidth and relatively low power. 4 This interface is different from the interface in the video endoscope. It is currently electronic and may be able to transmit dangerous currents. Because it does not have the isolation of the optical medium, the system must be designed to isolate the patient from potentially harmful currents.

Figure 3. The electronic interface of a digital endoscope with a CMOS image sensor.

For any medical system connected to mains power, patient safety is paramount. The IEC 60601 medical electrical equipment standard places strict requirements on components that protect the patient (MOPP) from harmful voltages. To use high-bandwidth solutions to transmit image data while meeting these stringent security requirements presents a major challenge for system designers. The electronic video transmission from the CMOS image sensor to the endoscope CCU is such an example, and a high-speed connection that meets safety requirements needs to be established between the two. ADI’s unique solution performs high-bandwidth transmission within a credible security barrier to meet the requirements of the IEC 60601-1 standard.

Medical display

Other medical equipment, such as ventilators and electrocardiogram (ECG), are directly connected to the patient for respiratory assistance and monitoring. Information about the patient will be displayed on the graphic display that comes with the medical device for easy viewing by the operator. According to the IEC 60101 standard, the display in the medical device is known, credible and certified, and can be used as a medical device. For any off-the-shelf external TVs and monitors, this cannot be guaranteed. In order to ensure the safety of patients, isolation should be added to the external connections between medical equipment and peripheral equipment to protect patients. For traditional low-speed interfaces (such as RS-232, RS-485 and CAN), this isolation may not be important and can be achieved using standard digital isolators.

On the other hand, the isolation of the video port from the external display poses unique challenges. The bandwidth requirement of the standardized interface of the display far exceeds the bandwidth that can be achieved using an appropriate number of optocouplers or standard digital isolators. Trying to isolate the entire signal chain of the video interface will further increase the complexity. For example, the HDMI 1.3a protocol includes not only the conversion minimized differential mode signal (TMDS) used to transmit video data, but also the exchange of video/format information, power circuits, and detection and display (receiver) connections between devices. Disconnected bidirectional control signal. 5 All of these factors must be considered when adding electrical isolation that the system designer sees as an obstacle. In many cases, it may not be possible to add a safety barrier to these display ports using the previous method, so the medical system does not include an external display port. ADI provides a reference design for electrical isolation of commonly used video protocols (such as HDMI 1.3a), so that when patients need to be protected, additional security protection can be directly added.

Gigabit digital isolation

When video and camera applications require high bandwidth and reliable security, system designers can use the new choice of ADN4654 series LVDS digital isolators. These devices provide dual-channel isolation with data rates of up to 1.1 Gbps per channel, which represents a major leap in speed for digital isolation. They are packaged in a 20-pin SSOP and provide a total throughput of 2.2 Gbps, which is greatly reduced compared to solutions based on traditional digital isolators.

Figure 4. Block diagram of ADN4654 Gigabit LVDS isolator.

Take the video link as an example. 24-bit color can be transmitted at 60 Hz with a resolution of 1920 × 1080 (1080 p). To transmit the required information across the isolation barrier, a total bandwidth of 4.4 Gbps is required. Typical fiber optic solutions have sufficient bandwidth, but to switch from copper media to fiber optics, serializers, deserializers, and electro-optical converters are required. Solutions that use standard digital isolators also require serializers, deserializers, and more than 30 channel isolations, each running at 150 Mbps. When adding isolation to a simple high-bandwidth interface, both solutions will incur overhead for system designers.

Utilizing the Gigabit data rate of ADN4654 can reduce the complexity of the system, and the 4.4 Gbps bandwidth can be achieved using only two devices. Each device has two channels, four in total, and each channel runs at 1.1 Gbps. With the high channel bandwidth, any SERDES modules in the signal chain are no longer needed. In systems that require isolation of more than one interface, both space and complexity improvements can be taken into consideration.

The physical layer interface operating at a rate higher than 1 Gbps has strict jitter and skew requirements to ensure reliable communication. Any components added to the signal chain (such as digital isolators) must have minimal jitter and skew so as not to affect system performance. Excessive jitter and skew may affect the receiver’s sampling margin and increase the overall bit error rate. The ADN4654 can achieve industry-leading skew performance on a given channel, up to 100 ps, ​​and 600 ps between devices, so it is very suitable for isolating these high-bandwidth interfaces. ADN4654 will only bring the least jitter, the maximum random jitter performance is 4.8 ps rms, and the maximum peak-to-peak jitter is determined to be 116 ps, using PRBS-23 (Pseudo Random Binary Sequence) mode. The mode run length is less than 23 bits, which is very common, and in the protocol with a shorter run length of the coding scheme (such as 8B/10B coding), the jitter performance is improved, and the improvement exceeds these values.

The ADN4654/ADN4655/ADN4656 devices utilize internal LDO regulators to provide flexible power supply configurations that can be used in a variety of channel configurations. The ADN4654 is available in a 20-pin wide-body SOIC package or a space-saving 20-pin SSOP package. The SOIC package provides 5 kV rms isolation and 7.8 mm creepage and clearance, making these devices suitable for 1 MOPP (from a 250 V rms power supply) to IEC 60601 standards. By using the package to increase the creepage distance and gap of the device to more than 8 mm, they can be used as a component of a 2 MOPP isolation system.

Figure 5. ADN4654-based systems can easily isolate high-bandwidth interfaces.

Isolate HDMI according to circuit note CN-0422

When adding security isolation to the video interface, the complexity of the video protocol itself will become a major challenge. It must be recommended to isolate each video, control, and power signal, which is a very difficult problem for equipment manufacturers. Plug-and-play design solutions help shorten the system development time required to implement functional design.

Since its introduction at the end of 2002, HDMI has become one of the de facto standards for commercial high-definition televisions and displays. HDMI has been able to succeed because of its feature set and reliable interoperability.

The EVAL-CN0422-EBZ reference design can be used as a plug-and-play solution for users who want to add electrical isolation to the existing HDMI 1.3a video port. Combine iCoupler® technology to transmit the required power and high-speed video and control signals across the isolation barrier.

Figure 6. EVAL-CN0422-EBZ reference design, used to isolate the HDMI 1.3a protocol.

The video data in the HDMI 1.3a protocol is transmitted in four TMSD lines: three data lines and one clock line. Each line must be isolated separately. Traditional digital isolators do not support the high bandwidth or differential characteristics of TMDS, so they are not very suitable. Although TMDS is slightly different from LVDS, it can be compatible with LVDS-compliant devices through simple passive components. These passive components combine two dual-channel Gigabit ADN4654 isolated LVDS transceivers to isolate all four TMDS lines. A pixel clock frequency of up to 110 MHz can be achieved, and a resolution of 720 p is supported when the frame rate is 60 Hz.

The HDMI protocol includes other low-speed signals for control: Display Data Channel (DDC), Consumer Electronics Control (CEC), and Hot Plug Detection (HPD). DDC is used to allow the source to read the display EEID data from the EEPROM and exchange related formatting information. The CEC signal allows functions to be shared among multiple connected source and sink devices. When detecting that the HPD has an additional source (indicating a device connected to it), the HPD is set by the receiving device. These control signals are isolated using two ADuM1250 devices, and when necessary, these signals can be isolated in both directions. Using the ADuM1250 can greatly simplify the design challenges associated with implementing bidirectional isolation channels.

The reference design includes an isolated DC-DC power converter ADuM5020 for powering the display (receiving) side of the isolation device. According to standard requirements, 275 mW will be transmitted to the HDMI cable to support the receiving device. The reference design is used to isolate HDMI source devices, but an isolated power supply circuit can be easily used to isolate HDMI sink devices.

Industrial Ethernet

For machine vision applications, ADI’s multi-protocol Ethernet switches, Ethernet physical layer transceivers, and all platform solution product portfolios ensure seamless connectivity and operational efficiency.

ADI’s fido5100/fido5200 REM switch series includes two 2-port industrial Ethernet embedded switches, which can be connected to any processor, including any Arm® CPU and ADI’s fido1100 communication controller.

By using these industrial Ethernet embedded switches, you can choose the type of processor that suits your application without being forced to use a specific vendor’s protocol stack. REM is connected to the processor’s memory bus and looks the same as any other peripheral on that bus. The storage cycle of REM is reduced to 32 ns (125 Mbps for a 32-bit bus) to support the 12.5 µs cycle of EtherCAT and the 31.25 µs cycle of PROFINET IRT. Data is transmitted back and forth between switches using priority channel queues, so real-time data transmission can interrupt non-real-time data transmission without delay. These queues are managed by the switch driver and interface with the protocol stack to achieve the most efficient data transmission possible. This also means that application software does not have to bother to manage switches, set low-level registers or track complex time management processes.

Another performance advantage of industrial Ethernet embedded switches is their priority channel technology so that they will not be affected by network loading. This advantage ensures that your application can be up and running at all times. REM switches intelligently filter data packets to prevent interference traffic from the processor, manage low-priority traffic according to the processor load, and ensure that high-priority data packets are sent in time, regardless of the total data packet load.

ADI’s ADIN1100, ADIN1200 and ADIN1300 industrial Ethernet physical layer devices (PHY) are designed to achieve robustness in harsh industrial environments. These products have completed extensive EMC and reliability testing and are suitable for applications that require predictable and secure communications. Utilizing industry-leading low-latency and low-power PHY technology, the product portfolio supports data rates of 10 Mbps, 100 Mbps, and 1 Gbps. They are specially developed to maximize data transmission and signal integrity, adopt a small package, and support multiple MAC interfaces at the same time. The industrial Ethernet physical layer kit is suitable for operation in an extended industrial environment temperature range, and can provide a high level of reliability for today and future industrial Ethernet applications. The ADIN1100 10BASE-T1L PHY provides a 10 Mbps Ethernet connection via a single twisted pair cable up to 1 km long and supports hazardous area use cases (intrinsically safe zone 0 applications), which are sometimes referred to as Ethernet-APL. ADIN1100 provides Ethernet connections for intrinsically safe certified devices, such as HMIs, industrial video cameras and thermal cameras operating in hazardous areas.

What first products does ADI offer?

This article describes the application requirements for safe and reliable high-bandwidth video or camera interfaces in industrial and medical applications, and discusses important technical options that can be used while using these interfaces while maintaining key performance. ADI provides innovative solutions, including:

The industry’s first Gigabit digital isolator series, ADN4654/ADN4655/ADN4656, provides new options for isolating high-bandwidth interfaces.

The industry’s first electrically isolated video and camera port, which helps reduce cost and complexity compared to bulky fiber optic solutions.

System solutions that have been tested to comply with regulations have reduced testing and compliance problems. One example is a reference design tested according to the HDMI standard.

A complete set of industrial Ethernet products, including technology, solutions, software and security features, these products are designed to connect the real world to the factory network and then to the cloud.

in conclusion

ADI uses its deep domain expertise and advanced technology to help partners connect future industrial suites and networks. The industry’s first gigabit electrical isolation technology provides an alternative to isolate video and camera interfaces in various medical and industrial applications. ADI’s Ethernet solution utilizes TSN Ethernet switches and low-latency, low-power, long-cable physical layer transceivers to ensure reliable transmission of critical data in harsh industrial applications.