Where wireless meets intrinsic safety

How to deal with the constraints of putting access points in hazardous areas.

By Ian Verhappen

In an ideal world, it wouldn’t be necessary to worry about explosive atmospheres. Unfortunately, plant environments are far from ideal, with most process plants being about 80% Class I, Div. 2 (Zone 2) and another 10% Class I Div. 1 (Zone 1), so some protection is the norm. This is one constraint faced by network designers.

If we could install all the required wireless network access points in unclassified areas, things would continue to be easy. Unfortunately, this isn’t the case, and we need to place access points anywhere they’re required—unless, of course, we wish to spend significant effort with corresponding impact to system complexity by forcing the design to fit the area classifications. This is constraint two.

Consistent with our bias toward running one cable to every device with Power over Ethernet (PoE), defined in the 802.3af/at specifications, we’d prefer to go that route if feasible. The most common PoE today is based on IEEE 802.3af and 802.3at, which is backwards compatible with 802.3af. IEEE 802.2af provides power up to 15 watts per port, and 802.3at provides up to 25 watts per port to support higher-powered devices. The standards divide PoE devices into five classes based on the maximum average power available at powered device: Class 0 is 13.0 W, Class 1 is 3.84 W, Class 2 is 6.49 W, Class 3 is 13.0 W and Class 4 is 25.5 W.

Download: 2016 State of Technology Report: The rise of configurable I/O

Another reason for using PoE is to assure reliable power to the access point: PoE can be sourced from the interface room on the same UPS as the control system, rather than field-powered at a local panel. Power sourced from field panels is more susceptible to interruption and quality issues such as surges. Partially in response to this and similar demands, UPS units are now available for Class I, Div. 2 (Zone 2). These units add complexity to design and maintenance, and because they’re field-mounted, are likely to have lower reliability than equipment in a clean interface room environment. Constraint three.

Galvanic isolators designed for PoE manage the power on the CAT 5/6 cable, so in the event of an arc or spark, the result will not cause an explosion.

Normally, at three strikes you’re out. In this case, we still have options. Because explosion-proof enclosures are metal, the “box” also works well as a Faraday cage, keeping radio signals out. Manufacturers have developed antenna arrangements that transition from the safe environment inside the enclosure to the outside to capture signals, so explosion-proof remains one choice we can use.

This configuration of safe area inside the enclosure with the other side exposed to the explosive atmosphere is similar to galvanic isolation, where one side of the isolation barrier is intrinsically safe, while the other side is unclassified. Galvanic isolators are often used in pairs, with equipment between the isolators considered equivalent to intrinsically safe (IS).

Some manufacturers make galvanic isolators designed for PoE installations that manage power on the CAT 5/6 cable, so if something happens on the cable to cause an arc or spark, the result won’t cause an explosion.

Ideally, the full system would be suitable for deployment and maintenance in any classified area, which means an IS system. The definition of  IS is: amount of energy that can be released in the hazardous area in the event of fault is voltage/power-limited to less than 30 V and less than 4 Watts.

At least one manufacturer has developed a fully IS Ethernet solution with a range of products, and also offers an IS interface to third parties for incorporation into their equipment with a maximum 12-VDC output. The company’s offerings are 200 mA at 10.9 VDC-Ex ia IIB (Class I, Div. 1, Groups C, D) output and 400 mA at 11.8 VDC-Ex ib IIB (Group B) output, and include IS PoE network devices.

Unfortunately, the above IS solutions are manufacturer-specific, which means you must purchase all the elements of the solution from the same supplier. This may not be a problem since most organizations tend to limit the number of suppliers used in their facilities.

Show Comments
Hide Comments

Join the discussion

We welcome your thoughtful comments.
All comments will display your user name.

Want to participate in the discussion?

Register for free

Log in for complete access.


  • You CAN actually put the wireless network access points in unclassified areas (or zone 2 which does not require intrinsic safety) so wireless sensor networks with intrinsically safe TRANSMITTERS can be easy – provided you pick the right wireless topology. Only the wireless sensors need to be in the hazardous area requiring intrinsic safety. However, this requires a FULL mesh topology. There are many wireless networks using IEEE 802.15.4 radio, but they are not the same. By full mesh topology I mean capable of 7 “hops” or more from the sensor to the gateway allowing the data to be routed around the metal obstructions in the plant. WirelessHART already supports 7 hops or more. 2-4 hops currently supported by other networks is not sufficient. By using a full mesh topology, you only need the gateway at the edge of the plant unit. There is no need for powered backbone routers in the middle of the plant units associated with other network technologies. If there is any weak spot in the WirelessHART network, simply drop in an additional device as a router – a battery powered device, intrinsically safe, needing no power wiring, therefore very simple to deploy. I agree that intrinsically safe PoE is an issue at the moment. Apart from the Ethernet switches, are there any field devices actually supporting it? Another problem is that there are no standard IS parameters (I, U, P, L, and C) so lots of engineering to demonstrate safety. When you replace any part with another model you have to redo it again. So, avoid this problem by placing the gateway at the edge of the plant unit; unclassified or zone 2.


RSS feed for comments on this page | RSS feed for all comments