>Building Reliability Into Your Digital Telemetry System

From the June 2002 Issue

Building Reliability Into Your
Digital Telemetry System

By Steve Baker, a senior engineer at Welch Allyn Protocol Inc., Beaverton, OR. Contact him at sbaker@monitoring.welchallyn.com.

System capacity. Cost efficiency. Flexibility. Expandability. Ease of implementation. Sure, these features top most lists when hospital CIOs select wireless telemetry technology for patient monitoring. But beware. For real-time, mission critical applications, they are all secondary to one other objective: reliability. Digital telemetry systems must be nearly as reliable as the phone system; otherwise patient safety will be unnecessarily compromised.

Reliability is partly dependent on which band the telemetry system uses, but more dependent on paying attention to fundamentals in communications architecture when implementing the system.

Operating Band Selection

The operating band choices for medical telemetry systems were changed in mid-2000, following an incident at Baylor University Medical Center where a DTV broadcast interrupted medical telemetry in part of the hospital for a short time. The incident led to a June 2000 Federal Communications Commission (FCC) ruling that rendered most medical telemetry systems technically obsolete.

Although the ruling does not require hospitals to replace existing systems, the Food and Drug Administration (FDA) will not give market clearance to any new devices that transmit in the traditional telemetry bands. Further, the FDA encourages medical device suppliers to replace systems in the installed base. The FCC ruling leaves hospitals with the choice between two frequency bands on which to operate patient monitors: the 2.4 GHz Industrial, Scientific, and Medical (ISM) band and the Wireless Medical Telemetry Service (WMTS).

The need to avoid signal interference for patient monitoring applications, however, makes the ISM band the only effective choice. All wireless communication systems experience noise and interference. The ISM band reduces interference better than the WMTS band, thus providing uninterrupted data flow and the reliability medical facilities need.

The best implementation of the ISM band for medical telemetry follows communication rules ( 802.11 Standards) developed by an international standards organization, the Institute of Electrical and Electronics Engineers (IEEE). The specifications provide uniformity on how devices can share the radio spectrum with one another without data loss. These band-sharing standards include bi-directional communication and use of spread-spectrum modulation that minimizes interference. This functionality allows for acknowledgement of packets and automatic retransmission of unsuccessful packets.

In part because of the standard-based specs and the acknowledgement protocols used by 802.11, ECRI last year recommended that medical facilities choose IEEE 802.11-compliant systems for wireless communications applications within hospitals. ECRI reported that the 802.11 protocols “reduce the likelihood that a signal cannot be transmitted due to interference, resulting in a better aggregate signal-to-noise ratio.”

In contrast, the WMTS band has no standards. This situation encouraged vendors to develop proprietary, conflicting infrastructures and protocols that interfere one with another. In a medical setting, interference on the WMTS band can result in a loss of patient data, unless various systems are carefully analyzed for co-existence and frequency use is prudently allocated. The interference problem can be worse in hospitals that mix old and certain new WMTS technologies.

For example, a spread-spectrum device that uses the fairly small 608-614 MHz band can disrupt the transmission of narrow-band technologies operating in the same band. One-directional telemetry in the WMTS band requires a dedicated band because it cannot function without enormous dropouts unless the band is noise-free. While the spread-spectrum system can be “throttled back” to avoid interference with older implementations, it then becomes functionally a narrow-band system.

On the plus side for the WMTS band, it offers legal protections that ISM does not. Unfortunately, this legal protection only holds for intentional transmissions. Interference due to harmonics of motors or computers can be experienced in any given RF band, including the protected WMTS band. The FCC specifically stated in their June 2000 ruling that there is no protection from adjacent channel interference, where milliwatt telemetry transmitters compete against 1 Megawatt digital TV stations.

Recent experience with emergency services illustrates the problem. Last August, The Oregonian newspaper reported that emergency service communications in 28 states were interrupted due to interference from cellular towers on adjacent frequencies. The FCC indicated that it is not responsible for fixing the problem because no one is breaking the law. The Oregonian mentioned that modern equipment would go a long way toward reducing the interference problem. That is what the ISM band offers: a new technology for medical telemetry that resists interference. Guard bands protect the ISM against such adjacent channel interference.

Transmitter Placement and Interference Sources. Unlike traditional medical telemetry (including WMTS), bi-directional ISM technologies use electronic access points (APs), or wireless hubs, instead of “dumb” antennas. Many devices can connect or “associate” with an AP, and the AP allows one device at a time to communicate with another device or with the hard-wired local area network. Some APs, such as IT devices, represent a potential source of interference, but good wireless system design—including a site survey and appropriate placement of APs—will minimize interference and virtually eliminate dropout.

A typical 2.4 GHz ISM-band data transmitter has an indoor range of about 30 meters and is only strong enough to cause substantial packet re-transmission when within about two meters of another telemetry monitor’s intended receiver. The relatively short range means that the hospital controls the only sources of interference, e.g., APs installed for IT purposes. While a leaky microwave could cause more packet re-sends than a data device, it will not adversely affect ECG data dropout rates if it is at least two meters from telemetry equipment. The question to ask is, “How many interference sources of significantly higher power than the patient monitor will be significantly closer to an AP than the monitor?” The answer is almost always zero.

Spread Spectrum. Another reason ISM band technology manages interference better than WMTS is its use of spread spectrum technology. Spread spectrum increases the effective signal-to-noise ratio of a communication system. Processing gain, reported in decibels (dB), is used to report the effectiveness of a spread spectrum system. A system with a higher processing gain can more effectively extract the signal in the presence of noise. In other words, adding processing gain is akin to decreasing the noise in the communication channel. Since dB are on a logarithmic scale like the Richter scale, adding 10 dB of processing gain means the system has 1/10th the noise of a system with no processing gain (0 dB). Adding 20 dB of processing gain equates to 1/100th the noise of the system. In general, the more spreading that occurs, the higher the processing gain. With 79 MHz of bandwidth, the ISM band can spread data more than is possible in the WMTS band, which may have as little as 1.5 MHz available.

Frequency Hopping versus Direct Sequencing. Frequency Hopping Spread Spectrum (FHSS) has 79 non-overlapping 1-MHz wide channels and a processing gain of 19 dB while Direct Sequence Spread Spectrum (DSSS) has 11 overlapping 22-MHz-wide channels with a processing gain of only 10.4 dB. This means FHSS has more than seven times as much processing gain as DSSS, and is the most robust use of spread spectrum currently available for medical monitoring.

DSSS, upgraded to 11 Mbps in 802.11b, offers higher data transmission rates than 802.11 FHSS, but higher data rates have more dropout, making it even less reliable for real-time, mission critical applications. When the data must get through, as with vital signs monitoring or the New York Stock Exchange, 802.11 FHSS is the solution of choice. Another reason DSSS is not suitable for these applications is that redundant coverage is usually impossible to achieve because only three of the 11 DSSS channels don’t overlap. FHSS allows for the installation of many APs in small or large areas, so one needs only to install more APs to support the required patient monitor load.

A single, narrowband source of interference that merely decreases the data throughput for FHSS can cripple a DSSS system. It is well known that Bluetooth degrades DSSS data rates more than it does FHSS data rates. The irony is that most hospitals that install DSSS because of its purported higher data transmission rate do not understand that the range for 11 Mbps transmission is typically 4.5 to 9 meters (less if there are walls). APs are usually installed so that for most of the hospital, the range allows only 1 or 2 Mbps data rates.

It is clear to see why FHSS is being chosen for mission critical applications. It is also no coincidence that the FDA has given market clearance for medical monitors in the ISM band from the major manufacturers, all of which selected FHSS: Agilent (M3 and M4), GE Marquette (Dash 2000 and Dash 3000), Welch Allyn (Micropaq and Wireless Propaq CS), Spacelabs (Ranger) and Criticare (MPT2.4). These manufacturers chose FHSS in large part because it is the spread spectrum technology with the highest interference immunity available for patient monitoring.

Conclusion

While one option for hospitals is to re-tune legacy telemetry systems to use the WMTS band, this is merely an investment in old technology that does not increase system performance (or decrease the total amount of dropout). An ISM-band based system that uses FHSS is a better option for hospitals that prefer a more robust communication system that will help them reduce patient risks.