The bus of a motherboard allows communication between different components, and standardized expansion slots provide a way to easily upgrade our computers. In this video, you’ll learn about the bus architectures and how to use expansion slots to customize our PCs.
If we look at a motherboard here from above, you can see it’s like a little city. There’s all these little different components on the motherboard. And all of these components need to be able to communicate with each other.
That’s why internally, on the motherboard, there are a number of what we call buses. These are communication paths that go between all of these different components. There are also these expansion slots on the motherboard. And the expansion slots also give us a way to communicate back to CPU and memory and other components on the motherboard. And in this video we’ll focus on these expansion slots and the bus that is used to communicate information from these expansion slots.
When you hear someone talk about how much information you can put through one of these expansion buses, you may hear them talk about the width of the expansion bus. That’s a little bit of an old way to describe it, because these days we’re using different methods to communicate information in and out of that expansion bus. And we’ll learn more about that in just a moment.
But the width effectively means how much traffic we can push through a connection. And you’ll see that we have bandwidth ratings these days that describe how much data we’re able to move in and out of that expansion slot.
There’s another important characteristic of the expansion bus, and that’s how fast the bus is going. We refer to this as the clock speed of the bus. Every bus has its own clock. It’s able to send information at a certain rate of speed.
We measure the speed as the number of cycles that we can perform in a single second. The cycle is called a hertz. If we can perform one million cycles in a second, we call that a megahertz, or one MHz, if we wanted to abbreviate it.
If you want to do 1,000 of those megahertz, it would be a gigahertz. Generally the faster we can have this clock rate go, the more information we’re going to be able to transfer in and out of that bus. You would think then that one clock cycle would equal one transfer of information.
But there are some techniques that we can use to transfer more than one piece of information on a single clock cycle. Take for example DDR3 SDRAM. This is the memory that we would have inside of our computer. It can transfer 64 times the amount of the memory clock speed. So although the clock on the memory is going at a certain rate, we can increase the amount of data transfer, through a number of different techniques, to go 64 times what that clock speed might actually be.
Here’s a block diagram of a traditional PC architecture. And although a number of these components are now collapsed into single chips, and we may have different components plugging into different parts of the motherboard, this is a very good overview that shows you what all of the different components might be. Some of these we might recognize, like the CPU and the memory. And others like the northbridge or southbridge are topics that we’ll discuss in a future video.
You can see that this is a motherboard that supports PCI adapter slots. But of course, if it’s PCI Express, these slots are still going to be in this motherboard. But they may be connected to other parts of the motherboard chipsets.
These adapter slots are the ones that we’re going to use to connect all of these different components, to expand the capabilities of our motherboard. If we were to look at the motherboard itself, you can see here are some of those PCI expansion slots. They are different sizes and different types of connections you would use. If you’re going to be purchasing a component to go in a motherboard, you have to make sure you’re purchasing one that’s going to fit the type of slot you have for that motherboard. Fortunately, there are some very standard sizes for PCI expansion slots, and as long as you’re getting the right size, it should work fine in the motherboard that you have.
Once we find the component that we’d like to add, like this network interface card, you can see that it slides right into this particular connection. And all we have to do is push down on this card very gently. And it will very easily slide into the slot. If you’re ever working with your components on the motherboard, you never want to force anything into a slot. You want to be sure everything moves very, very easily and that you don’t bend any of the components on that motherboard.
Let’s look at more details of these legacy conventional PCI connections. They stand for Peripheral Component Interconnect. You’ll never hear anybody refer to it that way. They will almost always call it a simple PCI connection or a PCI slot.
It’s a very common interface. It’s been around since 1994. It has a 32-bit and a 64-bit bus length. It depends on how much data now we can transfer over that expansion bus. And it communicates in a parallel mode. We’ll look at what that means in just a moment.
The amount of throughput will vary depending on what bus we happen to be using. And a conventional PCI connection can go anywhere from 133 megabytes per second all the way up to 533 megabytes per second if you’re running a 64-bit bus at 66 megahertz.
When we talk about a PCI bus as being parallel, that means that for a single transfer, all of the bits are transferred at one single time. To be able to do that, we need some way to communicate across many different connections simultaneously. On a 32-bit PCI bus, we’ve got 32 connections between, in this case, the southbridge and the PCI expansion slots themselves. If it’s a 64-bit bus, then we need twice as many. We need 64 lines of communication on that bus going all the way across.
This is a challenge as we begin going faster and faster and needing wider and wider buses. We begin to run out of space on the motherboard. And it becomes a challenge, making sure we’re able to get data from one side to the other all at once across that entire bus.
If we look closer at the 32-bit PCI slots and the 64-bit PCI slots, you can see there is a little bit of a difference between them. It’s because they are designed to fit a very specific size card into these slots. And these little notches that are built into the slots are designed to fit cards that will support a certain type of power requirement.
If we look at the 32-bit expansion card, you can see there are these notches at the bottom that designate how much power this particular card can use. This is an adapter card for a modem. And this modem can support both a 3.3 volt type connection or a five volt type connection. So it depends on the type of motherboards you have and what type of power it supports. In this particular case, this card will support either one of those voltages.
If we look at a 64-bit expansion card, it’s a similar type of connection. There are notches for 3.3 volts and five volts. And you can see there’s a notch in the back designating that this is a 64-bit card. And you have those additional card connections on the back of this particular ethernet card.
Another expansion card interface type you’re going to need to know for the A+ Certification is the PCI-X. This is the PCI eXtended. It was based on the conventional PCI, but it was designed for servers. Servers need a lot of throughput. We’re sending information back and forth over these expansion slots. And usually there’s many people using these servers. And so there’s higher bandwidth. We had four times the clock speed, up to 1,064 megabytes per second of throughput. And just like the conventional PCI, the PCI-X is also a parallel communication.
And you want to be careful because we’re about to look at a type of interface called a PCI Express. And although there is an X in PCI-X, this is not the PCI Express. This is a PCI eXtended. Those are two very different kinds of standards, and you want to make sure you do not confuse the two.
When you’re working with today’s motherboards you’re probably going to use the PCI Express. And you can see the abbreviation for PCI Express is PCIe. This has effectively replaced PCI, PCI-X, AGP and other types of expansion buses, because we’re able to do a lot more with the PCI Express. It’s a lot more flexible in the way that it communicates on the motherboard. And it goes much faster than these older expansion types.
Instead of communicating over parallel connections, PCI Express communicates serially, which means it sends a single bit at a time from one side to the other. You may think that that is a slower way to communicate, but in fact, it’s a much faster way to communicate, as you’ll see by some of the speeds associated with PCI Express.
You’ll see that there are also a number of different sizes of slots for PCI Express. There is the one, two, four, eight, 16, or 32 full-duplex lanes. That means we have communication going back and forth in both directions simultaneously.
You’ll see that they are written with this x1, x2, x4, but it’s actually pronounced, “by one,” “by two,” “by four,” “by eight,” et cetera. And you’ll see, as we look at the motherboard, it’s very easy to pick out the different sizes of the PCI Express because they are physically different lengths on the motherboard.
As you can see in this block diagram of PCI Express, it’s a very different picture than the PCI that had 32 lanes, or 64 lanes, of communication. In the case of the PCI Express x1 slot, you can see there is communication going on that lane one direction, and one lane the other direction. And that’s it. We have a very simplified way of communicating. It’s a much easier way to architect the motherboard. And we can still get very high throughputs.
If we look at a PCI Express x4 slot, you’ll see there are four lanes set up in both directions between the controller hub and the express slot itself. So it’s very easy to scale this up as we need faster and faster bandwidth. And we’ve got a lot more room to work with on the motherboard.
As the capabilities of our systems have improved through the years, we’ve been able to improve the amount of throughput for PCI Express. And there are different flavors and different versions of PCI Express. And you’ll see them written with version numbers so that you can understand exactly what the throughput is for those particular version.
If we look at the per-lane throughput in each direction, the version one of PCI Express allowed us 250 megabytes per second per lane. When we get up to version three, which is what we’re at these days, we’re looking at a gigabyte per second per lane in a PCI Express. So you can see these numbers are much higher than we ever got with a PCI or a PCI eXtended architecture. And soon we’ll be getting even faster throughput. Version four of PCI Express– it’s expected somewhere in 2016– we’ll get us about two gigabits per second per lane.
If we look at our motherboards, we can easily find those PCI Express slots. And you can see that there are different sizes of slots available. On my motherboard, for example, is a PCI-E, for PCI Express, x1, that’s the x1. You can see it’s a little short slot. Here’s another PCI Express x1 slot. Here’s a x8. That’s the x8 slot for a PCI Express. And down here is a x4 slot.
So you will see that the larger slot allows us to have faster and faster communication lanes. That way, if we’re using something like a video card like this one, you can see these interface connections at the bottom for the expansion will go into those larger slots so that you have more throughput for these particular components.
With laptops and mobile devices, we obviously don’t have the real estate that you might have on a desktop computer, so you don’t have a traditional PCI Express slot. Instead you have the same type of communication as PCI Express, but you have a much smaller slot to connect to.
This is the Mini PCI and the PCI Express mini card. This is where we’re really shrinking down the bus so we can use it inside of these mobile devices. So we might want to plug-in a Wi-Fi adapter, maybe a mobile broadband connection, and we want to plug it into a standardized bus type on these mobile connections.
We usually are opening up the laptop and plugging it into the laptop on the inside. There might be a port on the outside of the laptop that gives us just enough access. Or we may have to take off the entire back of the laptop just to install these cards on the inside.
I have a laptop that allows me to take off a small panel on the back. And that gives me access to both the memory slot that you can see here on the right and the mini PCI Express connection. This happens to be for a wireless connection. You can see there is connectors here for the antenna plugging into this main connection. Here’s a modem that’s connected up here. And I can easily move things in and out of this mini PCI connection by simply removing the card and putting something else in its place.