Motherboard expansion slots come in many different shapes, sizes, and speeds. In this video, you’ll learn about the characteristics and speeds of popular motherboard expansion slots.
<< Previous Video: An Overview of Motherboard TypesNext: Motherboard RAM Slots >>
If you’ve ever looked at a motherboard in detail, it looks like a city that you would see if you were flying over in an airplane from 30,000 feet. There’s a whole set of pathways and routes, just like you would have streets going through. There’s quite a lot of interactions between all of these different components on the motherboard.
You will also see that there’s a lot of places where the traffic will flow between adapter cards and maybe flowing up to a Northbridge or a Southbridge. And all of these separate, independent pathways are there. It’s difficult to see in this picture. But if you look very close at a motherboard, you’ll see all these little traces on the motherboard where all of that data is flowing.
You also have the option on these motherboards to be able to expand the capabilities. There are expansion slots built right into the motherboard. So if you wanted to add additional interfaces, some additional video options, you can do that right with the existing hardware that you own in that motherboard.
If you look at that motherboard carefully, you can see the bus that’s on the motherboard itself. Sometimes these are a lot of tiny little roads, these tiny little traces. Sometimes it’s many traces that are on the motherboard itself. And as we see in older motherboards, the wider the bus gave us more bandwidth. Well, that’s not necessarily the case any longer and we’ll talk about that in this video.
We’ll often refer to the clock speed of the bus. That refers to how much data is passing by every second or every time we have a clock cycle inside of that computer. And you can see that the expansion bus itself, it has its own clock. It’s not required to run at the same speed of the CPU.
If we were to look at the specifications of a computer, we normally will see a clock speed associated with the speed of that PC. And that’s usually referring to the speed of the CPU that’s on the motherboard. But there are a lot of different speeds that are occurring on that motherboard. The expansion bus itself has its own clock. It doesn’t run at the same speed as the CPU.
You’ll notice that the numbers of these speeds will be called MHz or GHz. That’s standing for megahertz and gigahertz. If it is 1 megahertz, that is one million cycles per second. A hertz being one cycle per second, therefore a megahertz is a million cycles per second. And obviously, a gigahertz is 1,000 megahertz. So as we look at these numbers and try to decipher just how fast something is running, we need to understand how fast it’s running in Hertz, so that we can make comparisons to other systems or other computers.
Another important consideration is that the speed of the bus does not necessarily correlate back to exactly how much data we’re putting across that connection. In many cases, we can put a lot of data over a connection over one single clock cycle. If you look at something like memory, it’s a good example of that. Something like DDR3 transfers data at 64 times the rate of the memory clock. That’s because it’s able to transfer data in different ways for a single clock across that link. That provides a lot of efficiencies.
So when we’re looking at the bus itself and the speed of the clock of the bus, don’t think that the speed itself limits the amount of data. In some cases, we can put a lot of data across that bus in one single clock cycle.
If we were to look at the architecture of a computer, our adapter slots would be out here connected to what we call the Southbridge. This is usually where we are plugging in these expansion connections. The main processing of the computer takes place at the CPU of course. And it’s the Northbridge that connects all of this higher speed communication, for instance between the memory and the CPU.
The other systems that we’re plugging in for our interfaces or perhaps other connections are almost always happening down here on these PCI adapter slots. Some differences to this might be something like a graphics adapter slot. If you have a PCI Express or an AGP slot, it may connect directly to the Northbridge. So there is this separation inside of the computer where most of the time our slower devices are connected to our Southbridge and the high-speed devices are connected to the Northbridge.
It’s pretty obvious when you look at a motherboard where the expansion slots are. There’s these long slots that we built so that we’re able to add different capabilities. It makes it very simple. If we wanted to plug in an adapter card, we find one that matches. This happens to be a 64-bit adapter card, into this PCI slot.
We line it up with the connectors on the slot. And we push it right into the slot so that it goes all the way in, so that almost none of those copper connectors are visible. When you have it in there, you’ll see that it goes in almost completely into that connection.
So it’s very simple. You power off your computer. You plug in whatever expansion you would like. You power it back on. And now, you have additional capabilities available, all by adding different connectors into those slots.
PCI stands for Peripheral Component Interconnect. That’s a mouthful. That’s why we almost always call it a PCI slot.
It was created in 1994. It’s been around a long time. And it’s so standardized that even the newest motherboards tend to put one or two PCI slots on there just so it’s backward compatible with your older adapter cards.
You’ll find those 32-bit and 64-bit versions of the PCI slot. And depending on what version of PCI happens to be running on your motherboard, you’ll have different kinds of throughput. For instance, a 133 megabytes per second is one that you will see on a 32-bit PCI slot that’s running at 33 megahertz. The highest will be 533 megabytes per second. And you’re only going to get that if you’re running with the 64-bit PCI card at the higher clock rate of 66 megahertz.
A 32-bit PCI bus is one where we have expansion slots. We would plug in our adapter cards. And they’re connecting over this 32-bit bus to the Southbridge. If this was a 64-bit PCI connection, it has twice as many connections going between a much larger expansion slot and also connecting back to the Southbridge.
Here’s a picture of the PCI slot. You can see the 32-bit PCI slot is much shorter than the 64-bit PCI slot that you’ll find. It has all these little keys on it as well. Notice that each slot is very different looking. And that’s because there are these little sections on the card, these notches that are on the adapter card itself, that both designate the size of the card, along with what voltages are required to use that particular adapter card.
You can see this card has a notch at the 3.3V and the 5V position. And if either of those voltages are available on your motherboard, this particular card will take advantage of them. The 64-bit expansion card, much longer you can see. There’s a notch in it that designates where that’s 64-bit happens to be.
This is another adapter card that will work in both 3.3V and 5V. But if your motherboard doesn’t support one of those voltages, that key space won’t be available and you won’t be able to push that card into the motherboard.
As we became more mobile, we needed a way to take the same functionality as PCI and shrink it down for our mobile devices. So a new standard was created called mini PCI. This had the same type of signaling, the same type of capabilities as PCI, but it put it into a very, very small form factor that worked perfectly for laptops. So we could add a Wi-Fi card, a mobile broadband card, and put it right into the laptop itself, but using a same standard as something that we were very comfortable with, with the PCI format.
Unfortunately, when you plug these into the laptops, they are inside under the covers of the laptop. So unlike a where you could plug into an interface on a PCI adapter card, these particular adapter cards are inside of your laptop. You really don’t have access to it once you install it into the system.
Here’s an example of what I mean by that. This is a mini PCI card. This is a Wi-Fi card that we put inside of a laptop. The cover has been removed. And it’s one that gives you access. You can snap it into the slot and then put the cover right on top of it.
You see there’s a modem connection here. There’s a slot for some additional memory if you wanted to put it into this laptop. Once you install the card, and this one happens to need an interface for an antenna, so we’ll plug in the antenna connection to this. And now, we’ve got this entire system ready to go for this Intel wireless card. We put the cover back on. And now when we start up our operating system, we have a brand new wireless card available for us.
One of the things we found with PCI is as the years went on, it became much harder to get higher throughputs through those particular legacy interfaces. So we created an update to the PCI called the PCI-X. That stands for PCI eXtended.
We really created this for servers that needed a lot of throughput. And they really got a lot of bandwidth out of this update. They were able to get four times the clock speed. It had a lot of abilities to plug in gigabyte ethernet cards or high-speed storage. And you can see that the capacity, the throughput of these devices, went up to 1,064 megabytes per second.
Now, keep in mind this PCI-X does not stand for PCI Express. That’s a completely different format using a completely different method of communication. PCI-X stands for PCI eXtended, which is simply a newer version, an updated version of the PCI standard.
On newer motherboards these days, you’re going to find PCI Express slots. You’ll see that abbreviated as PCIe, with a lowercase “e.” It’s important to delineate the difference between that PCI-X format and the newer and certainly much more common, PCI Express. If you look at a new motherboard these days, they almost always are going to be primarily a PCI Express bus that you’ll see on that device.
You will also see this abbreviated as PCIe with that lowercase e. This is not the PCI-X for the abbreviation. PCI Express is PCIe.
PCI Express also is not that traditional 32 or 64-bit wide bus. With PCI Express, we’re using a serial communication to send traffic in both directions. We call these lanes. What’s nice about this is that we can plug in a very slow PCI Express card and we can plug in a very high speed PCI Express card and they’re using their own lanes to communicate.
In the older PCI style, if you plugged in a slower card, the entire bus had to slow down to that speed. We don’t have that problem with PCI Express.
You’ll see the size of these adapter connections for PCI Express as 1, 2, 4, 8, 16, or 32 full duplex lanes. And you’ll see them written as x1, x2, x4. And the x here is pronounced “by.” so you’ll see a “by 1,” a “by 4,” a “by 16,” and a lot of different interfaces. And they all have different sizes as well. Sometimes it’s very easy to pick out the difference between a by 1 interface and a by 16 interface.
PCI Express doesn’t have all of those separate wires and a parallel communication. It’s simply has a single lane. So a Express by 1 slot is going to have a connection up to the Northbridge. Because we’re using a much faster communication here, we generally connect to the Northbridge. And there’s another lane on the way back. So this is a full duplex communication, going back and forth to that Northbridge.
If we need more throughput, let’s say we’re putting in an adapter card or a device that needs more speed, we can simply plug it into a slot that has got more connections, more lanes of communication. So a by 4 slot is going to have four of those full duplex lanes between the interface card and the Northbridge.
Through the years, there have been different versions of PCI Express that we’ve seen. And you’ll need to look at your motherboard to determine what version of PCI Express it happens to be using. If we wanted to compare the different speeds between those different versions, you can see that version 1.x had 250 megabyte per second lanes in each direction. When we got to PCI Express version 2.x, it doubled the speed up to 500 megabytes. per second.
Although you don’t see it very often, there is a standard version 3.0 and it has the capability of sending 1 gigabyte per second in either direction. And there is currently under way work for version 4.0. And it is said to increase the speed up to 2 gigabytes per second in both direction, when you’re working in a PCI Express environment.
When you look at a motherboard that has PCI Express slots, you’ll see different sizes of the slots depending on the number of lanes they’re using. This motherboard also has some of the legacy PCI connectors. And here’s a PCI Express by 16 slot. Here’s another by 16 slot. You can see they’re very long, especially when you compare them to the PCI by 1 slot.
Those are relatively short, very small connections. You don’t need a lot of real estate. There’s only two lanes there that you’re going to use, really one lane, bidirectional. You’ve got that full duplex connection in that by 1 slot.
If you were to look at the adapter cards for that, they match this. For instance, here’s an adapter card that is a PCI Express by 1 adapter card. And you can see it doesn’t take up a lot of room. That serial communication really allows you to very efficiently transfer information between the adapter card and the motherboard.
Another legacy interface you might run into is the CNR. That stands for Communications and Networking Riser. This was created just so we could add additional network cards or additional modems into our computers. This is a very, very small interface. It’s the small yellow interface here at the bottom. And it was designed just for this purpose.
The idea being on most systems, you really didn’t need very much. There’s so much already integrated onto the motherboard, that we just need an additional slot that was able to perform these particular functions. Well these days of course, motherboards have almost everything included on them. So we don’t often see a CNR slot on our modern motherboards.
Another legacy slot you might run into is the AGP slot. This stands for Accelerated Graphics Port. We almost always call it AGP. And you’ll see it written on the motherboard with the letters AGP next to the slot itself.
Before we had PCI Express, we needed some way to increase the throughput for our video cards. There was a lot of gaming, a lot of high-end graphics that was required. And AGP was built just for that. So we were always able to plug in and get some additional video out of these AGP slots. But now that PCI Express is around, we have much more bandwidth available. Almost all of the new video cards that we see are using PCI Express as its slot.
If you compared a PCI card with AGP, you’d see there’s a lot more connectors on that AGP card. There’s a lot more throughput that we needed. Therefore, we needed more connectors to the motherboard.
Many of these cards also have these notches on them that allow you to lock that particular card into a section of the motherboard. That way if the workstation is jostled or moved, you can be assured that that card isn’t accidentally going to pop out of the slot.
As AGP matured, we also got higher throughputs out of AGP. The initial versions, version 1.0, which ran at 3.3 V, gave us what we call an AGP 1x and 2x speeds that give the speeds up to 522 megabytes per second.
The next version of AGP, version 2.0, was a 1.5V version. But you can see, it was a 4x, which means we were able to get 1.07 gigabytes per second out of that standard. And you can see, we doubled it again with AGP 8x to 2.1 gigabytes per second. We were able to get a lot of throughput through that AGP 8x.
You can also see in that 3.0 version that we had an 0.8V as the standard on that bus. As time went on, we got very efficient with using the power on AGP. And you can see as we went through the different versions, the bus requirements itself dropped from a 3.3V all the way down to 0.8V.
Just like all of the other expansion slots, AGP is very unique. You can’t plug PCI into AGP and vice versa because of the format of the slot itself. And most of the time, you’ll see on the motherboard that it specifically says that’s an AGP connection, so you can be very sure what your plugging into and how you’re connecting to that motherboard.
You can see there have been a lot of different interface types through the years. And all of these buses allowed us to improve the capabilities of our computers. And if you look at your desktop, you’ve probably got one or two of these available so that you can add on additional capabilities to your computer.