How does a voltage increase help?

[JDCentral]

I understand that increasing the voltage on a CPU or stick of RAM will sometimes yield a better overclock.

How/why does this work?

 

[theMan]

voltage is like the water pressure in a pipe, the more there is, the faster, the water comes through. so, basically, the more voltage you have, the faster the information can move through the chip.

 

[JDCentral]

mmkay... that's what I thought.

But I thought there might be a 'better' explanation for it.

How does this effect the gates and transistors and stuff? Aren't they only built to handle a certain voltage?

If you increase the speed of the water, you're probably going to increase the pressure, as well.

 

[The Pentium Guy]

Originally posted by: JDCentral
If you increase the speed of the water, you're probably going to increase the pressure, as well.

Exactly. If you increase the voltage too much, the system will crash (or it might break )

 

[JDCentral]

But.. in the case of the Athlon XP Mobile, things are quite different...

 

[The Pentium Guy]

What's different? I'm sure the system will either crash, break, or shut itself down automaticaly to prevent the cpu from frying

 

[paadness]

Let me explain it to you the easy way:

A water pump, the water flow is dependant on the the pump's Horse Power. The more HP, the more pressure and more water flow.

Talking about transistors, here is some info to help u:

Silicon based transistor need atleast 0.7 Volts to make the current flow. Below that, a potential barrier exists which does not allow current to flow.

So if u find ur CPU run on 1.3 V so know it must be consuming some current. Increasing the Voltage lets say to 1.5 V will further increase the current flow.

Yes transistors have a limit. Its like a water pipe which can burst if too much pressure is applied. Note that even on the prescribed voltage the transistor may already be on the limit.

First, u will always loose energy coz all of the input cannot be converted into output, this lost energy comes out as heat.

There is also leakage current in transistors so that will further add to the heat.

Then u ask why it works?

Suppose a water line has 5 connections at full pressure, if u add 2 more, a total of 7, the overall pressure will decrease. So u will need to increase pressure to get back to the previous state.

Same for CPU's, u order more work by increasing FSB, that has to be compensated by increasing voltage, as..

V = I R

For power dissapation,

P = I(square) x R

 

[Jeff7181]

If you understand how a transistor with it's 3 "legs" work it's even easier to understand how voltage effects it. But basically, it takes a fraction of a second for the transistor to allow current flow once it's turned on. If you switch the transistor on and off so fast that it turns off before current has a chance to flow, the transistor may produce a 0 instead of a 1, and then you have an error. If you increase the voltage, which is electrical pressure, there's more force causing current flow, so it takes less time for it to turn on. Make sense?

*EDIT* As far as how it effects the gates and whatnot, there is ALWAYS some current leakage... meaning, a small amount of current flows through the transistors even when they're off, but not enough for them to be considered on. More voltage = more current leakage because there's more pressure. Like a water hose with a leak, increase the pressure and more water leaks out. I've seen people say that most CPU's can still handle about 2.0 volts before immediate failure... meaning... you actually break the transistors. I'm not saying go ahead and run your new FX-55 on 2.0 volts though, lol.

I'll draw a picture to better illustrate what I'm saying...

 

[The Pentium Guy]

Originally posted by: paadness
Let me explain it to you the easy way:

A water pump, the water flow is dependant on the the pump's Horse Power. The more HP, the more pressure and more water flow.

Talking about transistors, here is some info to help u:

Silicon based transistor need atleast 0.7 Volts to make the current flow. Below that, a potential barrier exists which does not allow current to flow.

So if u find ur CPU run on 1.3 V so know it must be consuming some current. Increasing the Voltage lets say to 1.5 V will further increase the current flow.

Yes transistors have a limit. Its like a water pipe which can burst if too much pressure is applied. Note that even on the prescribed voltage the transistor may already be on the limit.

First, u will always loose energy coz all of the input cannot be converted into output, this lost energy comes out as heat.

There is also leakage current in transistors so that will further add to the heat.

Then u ask why it works?

Suppose a water line has 5 connections at full pressure, if u add 2 more, a total of 7, the overall pressure will decrease. So u will need to increase pressure to get back to the previous state.

Same for CPU's, u order more work by increasing FSB, that has to be compensated by increasing voltage, as..

V = I R

For power dissapation,

P = I(square) x R

Thanks for that :thumbsup:

 

[RichUK]

Originally posted by: Jeff7181
If you understand how a transistor with it's 3 "legs" work it's even easier to understand how voltage effects it. But basically, it takes a fraction of a second for the transistor to allow current flow once it's turned on. If you switch the transistor on and off so fast that it turns off before current has a chance to flow, the transistor may produce a 0 instead of a 1, and then you have an error. If you increase the voltage, which is electrical pressure, there's more force causing current flow, so it takes less time for it to turn on. Make sense?

*EDIT* As far as how it effects the gates and whatnot, there is ALWAYS some current leakage... meaning, a small amount of current flows through the transistors even when they're off, but not enough for them to be considered on. More voltage = more current leakage because there's more pressure. Like a water hose with a leak, increase the pressure and more water leaks out. I've seen people say that most CPU's can still handle about 2.0 volts before immediate failure... meaning... you actually break the transistors. I'm not saying go ahead and run your new FX-55 on 2.0 volts though, lol.

I'll draw a picture to better illustrate what I'm saying...

Wo thats deep, nice one Jeff7181, thats helped me understand a bit more clearly the voltage loads on the transistors ..

 

[JDCentral]

Article I found that is helpful : http://www.modthebox.com/overclock103.shtml

 

[Jeff7181]

Again... the numbers I pulled out of my ass...

Here is a diagram of an NPN transistor. With an NPN, current is applied to the base leg to switch it on. With PNP, the rest is basically the same, but the base leg is grounded to switch it on. I believe both types are used in processors.

*EDIT* (maybe one of the Intel employees can elaborate more on the actual voltages... like how much voltage is applied to the base leg of an NPN transistor in a processor and how many volts = "on" and stuff like that)