Basics of Electricity, Circuits, Transistors, Processors, and CPU Temps

[paintboy12364]

wonderful post! Another vote for sticky here. if you have taken 8th grade physical science, all the formulas for volts, amps, etc. should already be drilled in your head. Thanks for taking the time to write all that down for everyone because I sure as hell was confused about transitors. I doubt many ppl will take the time to read it but it really gave me a feel about whats goin on behind an overclock. Thanks alot!

[undefeatedone]

Sorry I haven't been checking on this thread in awhile. Yeah, you'll find this in 8th grade technology class, but probalby not the transistor part, and since I know a lot of people tend to forget or not pay attention in school, I thought this would be nice to have on hand @ EOCF.

[Grandtete]

Yup its mosfet. The DC voltage is reduced by using a technique called Pulse Width Modulation (PWM). They turn the circuit on and off very fast. Now say you had a 12V input. If your circuit is on half the time and off the other half, then the average voltage is 6V. By switching the circuit much faster and altering the time on vs time off a set voltage can be generated. Capacitors are then used to smooth this pulsed current. When you change CPU voltage in the bios, the PWM circuits are altering the time on and off times and hense changing the voltage.

Please remember that Ohms law does not apply to many devices, such as bulbs, heater elements...and many semi-conductors!

This is because there resistance DOES NOT stay constant. Hense the voltage is not proportional current and therefore ohms law does not apply.

Ever heard of the term "Thermal runaway"? This is a term used to describe the way semi-conductors (including CPU's) can kill themselves, unless well cooled. The semi-conductor heats up and its resitance decreases, hense more current is drawn, hense the semiconductor heats up more etc etc.

Most semi-conductors reduce in resistance when heated up. Odd isnt it! (Can't be explained using the kinetic theory...)

Also, the person was correct that 70A would frazzle you. Fortunantly, the voltage is low, and the skins resistance relativly high which equates to a pathetic current and no death. If on the other hand you stick this low voltage into your salt rich body, then the resistance would be low enought to generate a significant current and stop the heart. There was one case of someone killing themself with a digital multimeter this way. This could be a myth, but the science is correct.

[(sin)morpheus]

Definetly a myth on being killed by a DMM. There is no way to get killed by a DMM. By the way, can you explain how a bulb's resistance doesn't stay constant and if it doesn't what law would you use instead of ohm's law?

Another note on the killed by DMM myth. There is one basic rule called kirchoff's law that prevents this from happening.

Here's a link on kirchoff's law http://en.wikipedia.org/wiki/Kirchhoff%27s_circuit_laws

Thermal runaway is not why cpus amoung other things kill themselves. They get killed for the same reason that a resistor dies which is from too much heat. When the cpu reaches a certain temperature the material begins to break down because it can't handle the heat.

The properties of silicon state that the hotter it gets, the closer it comes to becoming an insulator, the cooler it gets the closer it becomes to being a conductor. So, your thermal runaway theory doesn't work.

Here's a little blurb on thermal runaway for those interested. http://en.wikipedia.org/wiki/Thermal_runaway

[Grandtete]

Kirchhoff's laws are just the conservation of energy for electricity. This has nothing to do with whether a dmm could kill you. Most DMM's use 9V for resistance and continuants tests. In theory this could easily kill you, as the story says that the guy connected this 9V source across his heart (through his hands), UNDERNEATH his skin. The resitance under the skin is:

r = pl

resitance = resitivity X length... (in a 2d body lol). Now I dont know what the resistivity of the internal human body is, but I do know that is is small. There is a load of lovely salt and other electrolytes in the body. Electricity obviously conducts well, as your nerves function. Now, V = IR, you can see that it is easy for 9V to kill you, as the internal body resistance would have to be VERY large to prevent the heart from stopping.

Good page about that story here http://www.darwinawards.com/darwin/darwin1999-50.html

You have linked to the incorrect definition for thermal runaway. It is a large cause for semiconductor death. If it defined as (for power transistors):

"Power MOSFET thermal runaway. The on-resistance of a power MOSFET increases with temperature. With the resistance increase, the power dissipated for a constant on-current increases, the junction temperature increases, and the resistance further increases until the device is in thermal equilibrium with the heat removal system. If the heat removal system is inadequate, thermal run away occurs."

Now this also applys to almost all transistors and semi conductors. Certain thermistors are an exception, for example positive temperature coefficient thermistors I believe. (Pritty sure NPN thermistors are the opposite).

[Grandtete]

Oh about the bulb. A bulb is said to not be ohmic as resistance changes as the bulb heats up. As the bulb fillament gets hotter its resitance increases due to the increased kinetic energy of the charge carriers. The graph looks like this:



A ohmic device (such as a perfect resistor) looks like this:



Notice the difference. In the resistor the current is proportional to the voltage. In the other current varies as voltage changes.

For the bulb, most often one uses graphs such as the one above to work out the figures you need. It also helps to work out the resistance of the fillament at 0V....just try ohms law for that!

A good thread about why a hot processor kills itself:

http://forum.pcstats.com/showthread.php?t=16419

(Adding to the link above, I also would think that cooling the chip would improve the signal quality due to the reduced movement of electrons. This would reduce the excess energy that the electrons have and thus reduce the radiation emmitted. Will have to check that though.) Interesting:

I = naqv

Current = number of charge carriers x cross section area x charge on a charge carrier x drift velocity.

Now that page said that drift velocity increased...cross sectional area and charge carrier stays constant so:

Current is proportional to K x drift velocity. (Where K is defined as a constant)

If drift velocity increases then current must also increase if K is to remain constant.

Therefore, as the processor heats up the current must be increasing. Therefore resistance must be reducing...as we can assume that voltage is remaing constant. (Controlled by external source)

This would suggest that processors decrease in resitance as they heat up which is the generally what happens with transistors, diodes etc.

[(sin)morpheus]

You didn't answer my question on what you use for non ohmic devices in place of ohm's law. A graph wasn't the answer I was looking for because you need to be able to get the graph somehow.

By the way, the reason kircoff's law relates to the DMM story is because of the 9V battery. Sum of the current going in is equal to sum of the current going out. There really is no way that a 9V battery has enough power to stop a heart.

[ether.real]

Awesome info on the MOSFETs. It all makes sense now. Another reason why I love this place!!!!

[Grandtete]

Yes you use graphs. You take numerous readings for voltage and current and plot the graph. These are done expermimentally. (Using a voltage controlled power source of a linear resistor as a potential divider). You then extrapolate the graph and read off the values you need. Most non ohmic components will have a graph such as that in the data sheet.

kirchoff's law is still satisfied. Its nothing to do with whether the circuit could work. Energy would be conserved...it always is.

The reason the 9V battery could produce a sufficient current (100milli-amps) to stop the heart is due to the change in magnetic flux. The wire creates a magnetic field around itself. This field then collapses, which generates a significant voltage to produce 100 mA in the body. Btw, those stun guns run on one 9v battery....

[Paxton407]

BUMP! Sticky the **** out of this.

[ether.real]

Undefeatedone.....wanna consolidate some of this new information into your original post?

[(sin)morpheus]

They do run off a 9V battery, but I don't believe it's just a normal off the shelf one.

[Grandtete]

Yes they are the same type. You only need a tiny current to cause nasty effects, above 10mA. If cause to deliever this kind of shock you need a high enough voltage to overcome the bodys natural skin resistance. Therefore, they step the voltage up (Somehow, buck boost perhaps) to a large value.

Also, most modern alkaline batteries can produce rather large currents. When I was testing commen AA brands in the lab, some could sustain currents of 2 or more amps for a good 30 seconds. (This turned out to be directly proportional to the amount of zinc powder in the cell.)

[(sin)morpheus]

Wow, I hate you now, you know more than me.

You should start a thread that builds on this called advanced electronics. I want to learn ****it!

[ether.real]

yes yes!!!!!

[Grandtete]

Pah, I hardly know anything. I am starting an electronic engineering course at university in a few months though...

Might post some electronics mods in the future.

[(sin)morpheus]

Well, you know more than us, so post a **** guide!

[undefeatedone]

Originally Posted by ether.real

Undefeatedone.....wanna consolidate some of this new information into your original post?

Sure, sum it up guys, since you guys are the experts at this here. I'll put it in the article and then quote or cite whoever writes it, your choice. Title it and make subcategorical titles if necessary.

[(sin)morpheus]

Well, I think what we were talking about goes just a little past what is considered basic electricity.

[ValueSize]

that may be true but this has been one of my fav reads so far on this forum