A single hardware issue can cause signficant problems on a production network. In this video, you’ll learn about PoE standards, transceiver mismatches, and how to calculate transceiver signal strength.
If you’ve recently installed a telephone on a desktop, an access point, or a camera on your network, then you’ve probably connected it using Power over Ethernet, or PoE. This provides power over the same wire that we’re using to run data, which greatly simplifies the installation of these devices because we don’t have to run a separate cable for the power.
The power that’s put on this Power over Ethernet connection can come from one of two different sources. One is built-in power at the switch itself. You may have a PoE switch. That switch may be providing power. And we refer to that as an endspan.
Sometimes you’ll have a switch that doesn’t support Power over Ethernet, and you’ll need to have a separate device in the middle that injects power onto those wires. That injector is referred to as a midspan because it sits in the middle between the switch and the PoE device.
You’ll often find one of three different Power over Ethernet standards that are being used. The original PoE was a 15.4 watt of DC power output, and it has a 350-milliamp maximum current. This would obviously be for very small devices. Perhaps it’s a simple telephone or a very small access point.
The next standard would be PoE+. With PoE+, we can put additional power on this connection because it supports 25.5 watts of DC power with a maximum current of 600 milliamps. PoE+ is often used for larger telephones, cameras, and other devices.
And the largest of these three standards is the PoE++. As the name implies, this provides more power than the previous standards. You can have up to 51 watts at 600 milliamps maximum current– we refer to that as a type 3 connection– or a type 4 connection that can support 71.3 watts with a 960-milliamp max current.
This larger standard supports much larger devices. There are some laptops that support PoE++. And if you have a large camera that supports pan, tilt, and zoom, that additional power for that camera comes from PoE++. This newer PoE standard provides support for additional Ethernet standards, such as 2-and-1/2-gigabit, 5-gigabit, and 10-gigabit connections.
If you’re installing a PoE switch, and your goal is to power a PoE device, you need to make sure that the device and the switch are compatible. For example, if you have a PoE+ switch, that switch will not be able to power a PoE++ device. You can often look at the switch itself, and it will tell you what interfaces support Power over Ethernet.
It may be a switch where all of the interfaces support Power over Ethernet. This is a 10-port gigabit Power over Ethernet Plus-managed switch, and you can plug into any of these interfaces and use Power over Ethernet. Some switches may have some interfaces that have no PoE support. Some might support PoE+. And other interfaces may support PoE++.
You also might want to look at the switch specifications to see what the maximum amount of PoE power that can be supported using that switch. For example, the switch may support 200 watts of PoE, or it may support 720 watts of PoE. You will need to add up all of the devices that you’re connecting to that switch, determine what their maximum amount of power draw might be, and make sure that that is under the capacity of that particular Ethernet switch.
You might also run into hardware issues when you’re working with transceivers. Transceivers provide a modular connection for our Ethernet devices, but we have to be sure that we’re using the right transceiver for the right connection. For example, if you’re plugging in a fiber transceiver, you have to make sure that the fiber transceiver matches the type of fiber that you’re connecting. There’s usually a wavelength mark on the transceiver, so it might show that it is a 850-nanometer transceiver or a 1310-nanometer transceiver.
All of these need to match throughout the entire link, and it’s usually based upon the fiber that you’re plugging into these connections. So you’ll want to check your fiber specifications and the specifications of the transceivers and make sure all of those match. If they don’t match, you will undoubtedly have signal loss, which will result in error counters increasing, a loss of signal, or slowdown in overall network efficiency.
If you were to just grab a transceiver out of your drawer and plug it into a switch, you might be using the wrong one. These are two transceivers, but they are two very different transceivers. At first glance, they look almost identical. But if you look very closely at these transceivers, you’ll see the wavelengths are marked in very small numbers. And it’s very easy to miss that when you’re plugging into a switch.
That’s also complicated by the fact that once you plug it into the switch, you can no longer read the markings that are on the side of the transceiver. And you may either need to look at the specifications that are shown on the switch, or you may need to physically remove the transceiver just to be able to look at the label that’s on the side.
One of the challenges when connecting these transceivers and fiber connections together is we need to make sure that we get enough signal from one end of the connection to the other. This is especially important if you have a very long run with many connections in the middle, and you may be concerned that not enough signal is getting to the other side. Each of these devices has what’s called a sensitivity level, which will tell you how much signal it’s able to receive and still interpret everything properly. Usually, this is part of the specifications of the device, and it should be well documented.
This is why it’s important to calculate the exact power budget you might be from one end of the connection to the other. We’ll first need to determine how much power is being transmitted by the transmitting device. This is usually measured in decibels per milliwatt. This gives you a total number of decibels that are being transmitted for 1 milliwatt of signal.
We then might want to look at the media that we’re using and determine how long that connection is and how much signal we would expect to lose over that distance. We also want to take into account connectors or splices because each one of those takes away from a little bit of the signal.
Once we determine the total signal loss based on that distance and the number of connections in between, we can subtract that from the total amount of transmitted power. That final value is the amount of power we would expect to receive at the end of that connection. We would then compare that total received power to the sensitivity value of that particular interface, and hopefully the amount of power that we’re receiving over that connection is stronger than the sensitivity value.
Our decibels per milliwatt are decreasing as they move through this media, so our sensitivity values will be a negative value. In this particular example, this is a transceiver that is an SFP transceiver at 850 nanometers, and you can see its receiver sensitivity is negative 17 decibels per milliwatt. So if we perform that calculation of our power budget, and we come up with a received power value of negative 17 decibels per milliwatt or higher, then we have a good signal. If we calculate that power budget, and we end up receiving negative 20 dbm, then we know that we don’t have enough power to be able to receive it properly on this transceiver.