Five top things no one thinks about for PoE — but should!
Power over Ethernet (PoE) is a concept being applied to new and different applications every day. These applications include school public address systems, LED signage in mass transit stations and home thermostats, security cameras, WiFi access points and VoIP phones. But with so many new applications pulling their power from the Ethernet jack, some engineers are overlooking key analog considerations in these designs. This article will discuss critical recommendations to think about for engineers designing a PoE-powered device.
Paul Schreier, Texas Instruments
By Paul Schreier, Texas Instruments
Power over Ethernet (PoE) is a concept being applied to new and different applications every day! School public address systems, LED signage in mass transit stations and home thermostats are all joining the PoE craze – along with the traditional PoE applications like security cameras, WiFi access points and VoIP phones.
But with so many new applications pulling their power from the Ethernet jack, some engineers are overlooking key analog considerations in these designs. After being in the PoE business for a while now, here are some critical recommendations to think about if you’re designing a PoE powered device (PD):
1) Voltage rating matters: Does your PoE PD have a 100V rating? If you can’t tell from the datasheet, probably not. Employing a PoE PD pass FET that can withstand surges up to 100V is important for a robust solution. Even with a 58V-rated transient voltage suppression (TVS) diode in place before the PoE PD, that diode isn’t likely guaranteed to clamp over temperature until up to 93V. In lab tests, we measured spikes up to 98V at the PD! Please see Figure 1 to get an idea. Spikes like these will destroy PDs rated at 60V or 80V – common voltages on lower-end PD solutions.
Figure 1. Even with a TVS diode, 98V can be passed through to the PD IC.
2) Inrush current protection is key: You might think that most PoE-enabled switches (known as power sourcing equipment or PSE) will control inrush current on the switch side, so client-side current-limiting isn’t too important. But what about your customers who use an inexpensive power injector in cases where PoE is not available from their Ethernet switch? Most low end PoE injectors (devices that sit in series between the wall jack and the client) don’t do any type of negotiation or “soft start” while applying power to a client. Make sure your PD can current limit so your capacitors don’t pop off the board during a live insertion, if your customer might use one of these injectors!
3) Design for an auxiliary power supply: If you want your device to have the option to be powered locally by an AC/DC adapter or some other backup power source, many PD products make this difficult – you almost have to “trick” the PD into thinking it’s connected to a PoE switch before it operates. See Figure 2 for a common configuration. Even if you can get your PD to operate, many times you will need to provide a 24V or higher power module as a power supply – most PoE PDs with onboard DC/DC controllers won’t start up with input voltages between ~18V. However, some newer PDs with onboard DC/DC converters are designed to elegantly switch between local and PSE power – and have the capability to start up as low as ~9V. This makes an inexpensive 12V AC/DC adapter a much better solution for a local power source.
Figure 2. Choose a PD front end to accommodate local power supplies.
4) An isolated power supply is a good idea: Though the IEEE 802.3af standard doesn’t specifically call for an isolated DC/DC topology (like flyback or forward), many regulatory bodies and testing houses require it for safety certification if there are exposed conductors on the device like a USB port or any other interface. With no physical boundaries between the high voltage PoE source and the user, it is possible for a voltage spike or short circuit to harm a user. Therefore, many designs end up requiring a 1500V isolation requirement for safety certification – making an isolated flyback design the most preferable design topology. My informal estimate is that over 90% of PoE clients request an isolated topology for these reasons.
5) For a PoE PD front end plus DC/DC controller, the smallest packages aren’t always the best: Because nominal PoE input voltages can range up to 57V, IC vendors need to carefully select a package and pinout to meet “creepage and clearance” requirements from UL and other testing houses. It becomes very difficult to find an appropriate pinout in a leadless package like QFN or SON. Though slightly larger, the older SOIC packages usually provide a more suitable package for PoE PD ICs.
To address these issues, narrow your search to a PoE PD front-end plus DC/DC converter that takes these issues into account. It’s best to minimize the pins and package size required without compromising safety compliance or providing unnecessary features like a non-isolated DC/DC mode. Look for a device that makes designing functionality for an external power supply easy with a specific pin available to prioritize PSE or local power. One example might be the TPS23753.
These tips have been derived from numerous experiences with PoE customers and trying to build better PoE IC solutions. There’s obviously a huge amount of detail involved with each one of the above five items, but just being aware of each of them will go a long way toward getting your 802.3af PoE PD design off on the right foot.
References
– To download a datasheet or request samples of the tps23753, visit: http://focus.ti.com/docs/prod/folders/print/tps23753.html.
– For more information on the above and other PoE solutions, visit: www.ti.com/poe.
About the Author
Paul Schreier is the Product Marketing Manager for the Linear Power Business Unit at Texas Instruments. Paul received his Computer Engineering degree from the University of Nebraska at Lincoln. Send your comments to Paul at: ti_paulschreier@list.ti.com.
Advertisement
