Choice of Processor

The most important parameter that will govern the choice of the processor will depend upon its ability to satisfy the needs of the application. A typical application may require 4000 MIPS processing power, 2 USB, 1 Ethernet, 3 Serial ports, support for 1 Gbyte NAND Flash, Bluetooth, 802.11b, Audio,  XVGA LCD, touchscreen. Once basic processor and peripheral requirements are satisfied, we may narrow down our choice to the pricing, speed of processor for future enhancement and previous experience with the processor line.

It will also depend upon the software support offered by the processor as also the experience of the design team members with a chosen series of the processor.

Absolute comparison of processors is a difficult task, especially in embedded arena. Unlike desktop arena, there is a lack of the benchmark data comparing processing powers of the embedded processors.

Software Support / Operating System

The choice of Operating System will be decided by the expertise of the existing team and the support for the available processor. Some of the choices that exist for the choice of the operating systems are VxWorks, Windows CE, Linux ( comes in various flavors), Embedded NT, DSP/BIOS, pSOS, eCOS, ATI Nucleius, LynxOS, ThreadX.

It is likely that half of the time the team working on a project would like to keep the operating system same when trying to switch the processor for a different system. This makes sense in many cases, as they feel comfortable with the existing system and are confident that most of their code will work without modification in the new system. However, when there is a change in the underlying hardware, it may make more sense to switch the operating system that is better suited to the new chosen processor or the system.

Availability of Reference Design

It is always preferable to use a proven reference design to proceed. This will minimize the number of errors and design re-spins. The designer is not expected to prove the design. His focus should be to deliver the product with minimal possible additional design efforts. On hardware design perspective, this will mean availability of the schematics design files as well as layout design files. We should not rely upon the pdf files of the schematics of simply the gerber files of the layout. The design team should make it clear to the processor sales team  that they will have a better chance of adaptability only if they provide the reference design files. If the reference design uses a CPLD or a FPGA, the design team must make sure that they have the source code and they understand the tools required to build the Verilog or VHDL source code.

The lower level software also forms the part of the reference design. This must be evaluated by the firmware engineer. If a right choice is made based depending upon the availability of the reference design, it is possible to get the embedded board up and running in first power up.

Use of CPLD / FPGA

Use of CPLDs and FPGAs ( Field Programmable Gate Arrays) provide hardware design flexibility, cost reduction, space saving and many other exotic  features. However, if the designer chooses to use CPLD or FPGA in the design, he must take into account the time required to develop verify and maintain the code for the CPLD/FPGA. It will also depend upon the available manpower and expertise available.

Battery Power Requirements

A key challenge in the design of portable medical device companies is to achieve two goals which have contradictory requirements – the goal of low power consumption and that of high performance.

The medical devices usually will fall in one of the two categories. The first category is that of  small microcontroller based systems that will  be required to operate many months before requiring recharge and or battery replacement.

The second category is of the full fledged processor based system with potentially LCD displays that have battery life in hours or in days at the most.

In both cases designer will have to work with power system design engineers, software team to implement the lower possible current consumption – most importantly in idle state.

EMI/EMC Compliance, EMC 60601-1, EMC 60601-2 requirements

The Safety requirements must be understood before the system is taken to the compliance tests. This is a very broad area and the EMI test , ESD test, safety test requirements must be clearly understood right at the beginning of the design.

EMI test will require that the PCB Layout must follow the schemes that minimizes radiation from the board. This will require power plane planning, proper terminations for all high speed signals. The designer must be prepared for the case when the design may fail in EMI test even after taking care of all possible design rules. The designer should make a provision for the Spread spectrum oscillator and a faraday cage, in case, the design marginally fails in EMI test. This will avoid costly re-spin.

Medical standard require that the fire enclosure for any unit with power rating 15W or greater should be 94V0 or better. The mechanical designer must work with the electrical engineers to find if the power consumption of the board is expected to stay within 15 Watts.

Electrostatic Discharge (ESD) is the abrupt release of charge from one object) to another. Such a discharge can permanently damage or otherwise upset the function of sensitive electronic circuits. ESD requirement can usually be fulfilled if we make provisions for the ESD diodes at all connector pins. ESD tests can be destructive if provisions for ESD safety are not provided with the board.

If power supply modules are used, we must ensure that they have follow required safety standards. All power system should be designed such that the short and open circuit conditions do not create hazard. The power supply regulators should have current limiting features. Reverse battery protection must be provided.

FDA Approval

Once we determines that the device falls into the category of the Medical Device, we will  have to get FDA approval to be able to sell the device in US market.

FDA classifies the Medical devices in three categories Classification identifies the level of regulatory control that is necessary to assure the safety and effectiveness of a medical device. We must understand the processes involved in getting the FDA approval. Different categories of the devices have different requirements. When taking the product requirement we must know the quality norms and requirements that must be adhered to. FDA recommends you validate your device under actual or simulated use conditions to ensure the device conforms to defined user needs and intended uses. As  part of the verification test, the devices verification test must be performed and recorded.

A detailed discussion about the approval process is beyond the scope of this paper. FDA approval process in neither difficult nor technically intensive as far as it does not involve drug and invasion to body parts. However, a prior experience to the FDA approval process greatly helps.

Signal Integrity Requirements

Signal integrity  is a measure of the quality of a electrical signal ( usually high speed). A degradation is the quality of the high speed signals can lead to system failures, data loss.

Most embedded processor based systems now a days have high speed data bus including DDR, PCI-X, USB 2.0, SATA and others. Any failure not to observe signal integrity requirement will lead to unstable system performance and will be catastrophic for the Medical data logging based system. All systems must be analyzed by an expert in the Signal Integrity. Besides that the system must be tested for frequency margin  and voltage margin to see if it fails due to any signal integrity issue.

The signal integrity engineer must be involved at the very beginning of the design cycle, when the team is making critical architectural decisions and selecting component technology. Signal integrity engineer should be able to tell the amount of the efforts that will be required to simulate, if any, the high speed signals and the amount of the time and effort it may take. The Signal integrity and the testing team must develop a comprehensive simulation and measurement strategy that applies the appropriate level of analysis to each bus in the system.

Continued availability of components

As soon as the design starts all the components must be procured, at least in smaller quantity. This exercise will make any long lead component known ahead of time. The PCB designer must get samples of unverified components and ensure that the Library is correct not only by checking the design, but, also by physically inspecting the component.

Medical devices usually have much longer life than the consumer goods. This demands that  the component selection requires additional thought for assurance of continued availability.

The designer should make a checklist of the components that may go out of life in a specified amount of time. A possible alternate source must be listed. If the component does not have any alternate source it must be pointed out if in the check list. All such components should be grouped together and, if possible the electrical nets should go to an add on card connector for a possibly different source alternative.

Analyze the Potential Risk Areas

No matter how good the design is, there are always some risk areas which can derail the deadline. All possible risk areas must be analyzed and documented. This will make everyone aware of the potential pitfalls. Attention must be paid to the high risk areas. It often happens the most of the meeting times are devoted to discuss petty things, like GUI interface and look and feel while more important discussion regarding core system design and the driver software development, potential hardware/ software integration issues get un discussed as these are difficult to follow.

The Engineering Manager’s only way to be able to analyze the risk will depend upon his past experience and his ability to keep himself updated of the technologies followed behind the design. This will be in addition to the feedback and analysis of his team members.

Mechanical Enclosure design

Medical devices have been historically designed with best exotic appearance. The design team must have a team member who not only can apply design techniques, but can also be creative in design. The creativity aspects can only be known from his demonstrated past experience. The team members should ensure the requirements are delivered and rest should be left to the imagination of the enclosure designer.

Vikas Shukla is currently working as Senior Design Engineer at BL Healthcare. He has degree in Computer Science and Engineering from IT-BHU, Varanasi, India. Mr. Shukla has over 15 years of experience in design of microprocessor-based systems. His expertise includes signal integrity, architecture and design of remote patient monitoring systems. The views expressed are his own.

An object may become negatively charged by gaining electrons on its surface from another object. The object that transfer electrons become positively charged. In practical medical application a charged human body or object can come in contact with the medical appliance. The extra charge will discharge in the appliance and may or may not cause failure. The designer must ensure that the medical appliance is able to withstand static electricity.

Failures caused by Electrostatic Discharge can happen in a number of different ways. The electrostatic discharge can propagate to the ICs fusing together a junction. It may permanently damage the medical device. In medical applications it can cause the device to malfunction in the middle of the process of data acquisition or transmission causing potentially serious medical hazard condition. The handheld medical devices are more susceptible to human induced ESDs. While permanent failure of the ICs are less likely, though not impossible, the button malfunctioning can induce functional errors. Designers must resolve and test these issues.

Methods of Preventing ESD and ESD induced errors

A number of different strategies exist to prevent and ESD induced error. The designer should implement these to comply with the ESD safety standards as well as to avoid any failure and malfunction.

Shunting ESD Energy

Zener Diodes, Metal Oxide Varistors, transient voltage suppression (TVS) diodes, and regular  (CMOS) or bipolar clamp diode can be used to shunt the ESD energy thereby providing protection to the ICs. These are designed to absorb electrostatic discharge (ESD) energy that is introduced from I/O ports and travels through the connector onto the system board. ESD-protection diodes thus provide  protection against ESD-induced system malfunction and/or damage to ICs. The designer may add an ESD diode at all the connectors pins to prevent any ESD from entering into the ICs.

For signals at  higher data rates ( for example USB 2.0, IEEE 1394, DVI, PCI Express), the parasitic impedance of these protection devices can create signal integrity issues.

Figure 1 : Shunting ESD Energy using protection diodes.

PCB Layout for  ESD

PCB spark gaps can supplement the Electrostatic Discharge protection.  The Spark gap is created between the Signal Pins and the connector etch pad. The etch structure provides a small gap between points. This encourages a breakdown from  high voltage across the clearance. The clearance area must be free from solder mask because the solder mask has the tendency to increase the voltage breakdown of the gap. The breakdown voltage of a small spark gap is given by V = (3000pd+1350) where p is the pressure in atmospheres(1) and d is the distance in mm. A 20 mil (0.508 mm) gap will provide a voltage breakdown of 2850V. This will help designs pass the required ESD levels.

Software Key Debouncing

In handheld medical devices using membrane buttons,  ESD can momentarily induce pulses that can induce a false button press signal.  In certain conditions it is equivalent to creating hazard condition. For example, assume a medical appliance in which patient has to answer questionnaire about if he had taken a medicine. A false button press due to ESD may generate erroneous alarm and incorrect conclusions.

The designer may implement a software debounce technique in addition to the ESD diode protection to overcome this type of failure. The ESD is a very short duration phenomenon ( usually in nano seconds) and therefore a software debounce technique where the it checks for key press after an interval of 20 to 100 milliseconds provides sufficient safegaurd against the ESD induced malfunctions.

Vikas Shukla

Vikas Shukla is currently working as Senior Design Engineer at BL Healthcare. He has degree in Computer Science and Engineering from IT-BHU, Varanasi, India. Mr. Shukla has over 15 years of experience in design of microprocessor-based systems. His expertise includes signal integrity, architecture and design of remote patient monitoring systems. The views expressed are his own.