July 2008 Edition

inspection solutions

Quieting the EMI racket

Wireless mesh networks can ensure data arrives uncorrupted

By Jeff Wilkinson

Electronics as a group are some of the most mechanically and performance-fragile systems that live in industrial environments. Within those environments are frequent examples of noise and interference sources that disturb or otherwise impair the reliable functioning of electronics.

While heavy steel packaging and seal rings keep harmful, volatile, and corrosive liquids and particles out of sensitive electronics, they cannot protect the emissions of radio waves in wireless networks. Wireless sensors and metrology instruments frequently are being deployed in production areas.

The wireless systems deliver their quality or production data to critical operations downstream and are crucial components of the modern production reality. With this said, wireless data transmission of data is critically needed and simultaneously vulnerable to shop-induced interference.

Radio frequency (RF) waves are the carrier for data in a wireless data acquisition system. These waves are simply energy propagated through free space. When free space is cluttered with other energy forms, intentional radio waves are compromised.

RF is highly susceptible to corruption and alteration via a variety of Electromagnetic Interference (EMI). EMI has been defined as the "degradation of the performance of a piece of equipment, transmission channel, or system caused by an electromagnetic disturbance" (ANSI). EMI can occur throughout the EM spectrum from
0 Hz to 20 GHz or higher frequencies.

There is no such thing as a 100 percent noise-immune radio system. Systems designers must develop robust wireless data collection networks and sensors to be less susceptible to EMI.

However, EMI problems are most prevalent in the RF spectrum. Since in many applications the RF carries the data, then good RF handling must be encouraged to keep the data intact.

If there is EMI, RF-based systems must manage their performance relative to interference if they are to be useful. There is no such thing as a 100 percent noise immune radio system. So with that truth, systems designers must develop robust wireless data collection networks and sensors to be less susceptible to EMI.

There are many techniques that designers can use to offset the impact of EMI-generated noise in a wireless network. The mesh network is an
option that has some distinctive features. First, it has a single and central "gateway" function where all system wide commands and network management can occur. Data from the network also returns here.

Secondly, the sensor/measurement endpoint radios can be active components of the network. Thirdly, numerous routers or repeaters are present and can be added to enable multiple paths for OTA (Over the Air) transmissions.

The mesh is inherently robust to interference by the nature of the system configuration. For example, observe what happens to the OTA flight of the RF. Figure 1a shows how as the endpoint radio acquires data from its measurement tool or sensor and transmits it to the gateway, a plume of EMI from an induction hardener cancels the RF in the immediate vicinity.

Adjacent to the first router, other routers (Figure 1b) have also received the data transmission. Once the blocked router has found no data was received, it cannot pass along any data to the gateway. Simultaneously, the other routers have the good data and attempt to send that data to the gateway.

In a mesh network, the data may make a series of router hops until it reaches the gateway. While this is happening, other copies of the data are en route. When the gateway sees an exact copy of already received data, the gateway discards the additional copies. Endpoint radios on the network have a unique "address" which allows only the intended data to reach the gateway without duplication.

In addition, when EMI is not present and optimal operating conditions exist, the mesh network speeds OTA transmission by constructing a routing table in each network element. The router table creates a predetermined path for data, which allows the other routers in the network to be either idle or available for other endpoints to be received.

Some of the key advantages of wireless data collection are freedom from the bench, host PC’s, wires, and the limitations of bringing a part for inspection to the tool.

However, with wireless systems, data being recorded may not be immediately or readily visible.

Advanced systems for wireless data collection have indicators/enunciators on the data transmission device itself, to show the operator the status of the system and the successful transmission of data. In many systems, the end node is "unaware" that the network or host PC loses power, and therefore, data will not be recorded.

But, if the endpoint has the ability to show the user that data transmission has failed, a few good things can result. The operator can immediately see that there is a problem with the network and correct it. Conversely, without any system status indication, the operator can unknowingly corrupt his own data collection.

With a feedback method of system status oriented to the user, data can have more integrity. However, when the user ignores or misses the feedback, data can still be corrupted. As an additional safeguard to lost data, the endpoint radio can incorporate a means to store data recorded when the main systems is down.

Modern wireless data collection system designs incorporate a storage feature that collects any data "taken" at the endpoint and holds it until the system becomes available again. All the while data is being stored on the endpoint, the endpoint radio alerts the user with the feedback system described above and waits for the system to come back online. Once the system is back on line, the endpoint will dump its data to the gateway/PC host system. No data is lost, even though the main system has failed.

With DataSure, there is a system feature that communicates between endpoint and gateway and seeks to ensure that uncorrupted data arrives at the gateway. If the system tags the data sent with a CRC or parity error, the gateway informs the endpoint to resend the data again from the endpoint’s temporary memory.

Data is then sent again, up to 10 times, insuring good data ultimately gets to the data set. If data is received as good, then the gateway informs the endpoint to discard its temporary memory of that data, so that it won’t get sent again. The temporary memory is only held until the gateway validates successful receipt of the data.

In one test, successful transmissions have been recorded with zero failures in 3.5 million measurements. This is an example how data collected in the field show remarkable performance figures.

L.S. Starrett Co.

Jeff Wilkinson is general manager of the Starrett Advanced Technology Division.

What do you think?
Will the information in this article increase efficiency or save time, money, or effort? Let us know by e-mail from our website at www.ToolingandProduction.com or e-mail the editor at dseeds@nelsonpub.com.