-
We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
are all transistors in a cpu the same?
- Thread starter CTho9305
- Start date
The same as in type? Or as in size?
There are typically two types of transistors on a modern CPU: pMOSFET's and nMOSFET's (p-type and n-type Metal Oxide Semiconductor Field-effect Transistors). These combine to form CMOS (Complimentary Metal Oxide Semiconductor) - the complimentary referring to the fact that both compliment each other. The entire CPU is formed from these two types of transistors in varying sizes. The Pentium used BiCMOS - which added another 3rd type of transistor called a BJT (Bipolar Junction Transistor), but BiCMOS is not very common nowadays.
The two main types of transistors are differentiated primarily by the 'carrier' of charge. nFET's have electrons as the primary charge carrier and pFET's have 'holes' as the primary charge carrier (a hole is an absence of an electron - if this doesn't make sense, then don't worry, plenty of other people can't picture this either). BJT's are differentiated by the fact that work fundamentally differently than MOSFET's.
There are other components on a CPU: diodes, capacitors and resistors are common. The first two are usually made from tying off nMOS FETs in various configurations, the third is usually made with a highly resistive material stretched out for a long distance.
There are typically two types of transistors on a modern CPU: pMOSFET's and nMOSFET's (p-type and n-type Metal Oxide Semiconductor Field-effect Transistors). These combine to form CMOS (Complimentary Metal Oxide Semiconductor) - the complimentary referring to the fact that both compliment each other. The entire CPU is formed from these two types of transistors in varying sizes. The Pentium used BiCMOS - which added another 3rd type of transistor called a BJT (Bipolar Junction Transistor), but BiCMOS is not very common nowadays.
The two main types of transistors are differentiated primarily by the 'carrier' of charge. nFET's have electrons as the primary charge carrier and pFET's have 'holes' as the primary charge carrier (a hole is an absence of an electron - if this doesn't make sense, then don't worry, plenty of other people can't picture this either). BJT's are differentiated by the fact that work fundamentally differently than MOSFET's.
There are other components on a CPU: diodes, capacitors and resistors are common. The first two are usually made from tying off nMOS FETs in various configurations, the third is usually made with a highly resistive material stretched out for a long distance.
You'll get a better answer from one of the CE/EEs that frequent this board (I'm CS myself), but the short answer is no. Various transistors will be optimized for switching speed, or power consumption, etc, depending on what its used for. For example, if a set of logic is part of the critical (speed-limiting) path in the processor, it will probably be created with faster switching speeds. For more specifics on how that's done, someone else needs to take over. 🙂
PM, I have a question for you: I've always been curious as to the exact difference in function between a BJT and a FET.. I know that FETs are voltage-controller versus BJTs are current controlled, but why exactly are FETs more useful than BJTs in semiconductors?
It may be a dumb question, but since all EE classes are closed to non-EE majors at the university, I have to learn basic electronics by osmosis rather than in the classroom. 🙁
It may be a dumb question, but since all EE classes are closed to non-EE majors at the university, I have to learn basic electronics by osmosis rather than in the classroom. 🙁
I should have been more specific. As a freshman ECE major, we are being told that diodes have a voltage drop of .7, and the Vbe for fwd active transistors is .7 (well, we model them as linear devices for simpler circuit solving. Its supposed to be a good approximation for the actual exponential (?) curve).
I was wondering if the Vbe, Vce(sat), etc are constant for all the transistors across the die. We're dealing with only npn BJT's for now, so I'm probably off base somewhere.
I was wondering if the Vbe, Vce(sat), etc are constant for all the transistors across the die. We're dealing with only npn BJT's for now, so I'm probably off base somewhere.
CTho, this is a pretty complex subject. My answer should go something along the lines of "BJT's work this way" and "FET's work this other way" but that's going to want pictures and a fair amount of background. This is going to take a fair amount of typing. I'd better hand this off to the experts.
Here's a diagram and some text of a BJT's structure and operation (Google search "BJT operation", for the curious).
And here's one from Harvard on MOSFET's (Google search "MOSFET operation").
Remnant, the trick with using the same transistors to do different things depends on the way you hook them up. Essentially there's low-power/slower style CMOS 'static' logic, and then there are a wide variety of higher-power faster other types of circuitry. But they all use either nFET's or pFET's in CMOS.
Here's a diagram and some text of a BJT's structure and operation (Google search "BJT operation", for the curious).
And here's one from Harvard on MOSFET's (Google search "MOSFET operation").
Remnant, the trick with using the same transistors to do different things depends on the way you hook them up. Essentially there's low-power/slower style CMOS 'static' logic, and then there are a wide variety of higher-power faster other types of circuitry. But they all use either nFET's or pFET's in CMOS.
You should be able to if it's all satic. But there are leakage currents that drain dynamic nodes, and if you don't refresh them after a while, you will lose the state. So if the clock controls the refreshing, then you can freeze the state for a short while before some charges leak out.