Book covering basic physics/electronics relevant to computer science?

[Member 4194593]

What! You mean that everything my daddy taught me when I was 7 years old wasn't the whole truth? Of course he was often heard to opine "Gimmy my horse". He didn't get along well with technology.

[Barry H]

Try Code by Charles Petzold.

[makkak]

I am a 3rd year EE major at UC San Diego. I also do programming. I've been on both sides of the fence and can tell you that you don't need an understanding of electronics on the transistor level to program. I find it helpful to know the gate level and everything up, especially the von neumann architecture. But if you are curious about the magic underneath, here are some books I would recommend: - Semiconductor Device Fundamentals, Pierret - Microelectronic Circuits, Sedra and Smith - Digital Integrated Circuits, Rabaey et al The first will give you a good understanding of how the ideal device equations are derived from first principles. The second will introduce you to all the major devices: diodes, bjt's, MOS, CMOS and how to use them on a circuit level. This book leans towards analog circuits like amplifiers. The third is probably what you want. Goes over the CMOS manufacturing process, the MOS transistor, interconnects, combinational logic, sequential logic, arithmetic building blocks, and memory structures. But if I were you, I would just read wikipedia because electronics texts are long, dry, and hard to read.

[werD]

+1 I was just going to post that. "Code The Hidden Language of Computer Hardware and Software" by Charles Petzold [^]is a great book that goes from the ground up. I would highly recommend it to any level of programmer. In the book, he consructs a binary adding machine out of telegraph relays, it is pretty informatiuve to say the least :D

DrewG, MCSD .Net

[Jason Denton]

I was going to suggest that book as well. It's a great read and very entertaining. It's not so much a physics book, but it does start out very simple with basic logic gates and progresses to explain the circuitry and logic of how computers work. I enjoyed this book so much that I’m going to read his latest book on Alan Turing. Linky[^]

[sibrowne]

Not a book, but the Royal Institution Christmas Lectures this year were about microprocessors. They were broadcast on Channel 5 in the UK and look like they are still available to view online here[^]. I think it's UK only though. The first couple covered the basics like how transistors work.

[MattBeavis]

I would also vote for "The Art of Electronics" great book, I originally started out studying electronics and then some how fell into more computer related programming/scripting/application packaging and support based work. The book is one of the few books that still has pride of place on the shelf and it ocassionally gets thumbed through even 15+ years after its purchase (OMG!!! I hadnt realised how old it was... :) )

[mi5ke]

The basics of computer operation do not rely on electronics - there are many ways to make logic gates, with many different technologies. Electronics is just the handiest way to do it - size, speed, cost are all optimized. But things change, and who knows how we'll be making them in the future. For example, imagine making logic gates with water, pipes and valves - just like your home plumbing system. If we have a piece of piping with two valves in series, then water can only flow through the pipe if both valves are open - this is essentially an AND gate. If the two valves had been in parallel, then opening either of them would allow water to flow - this is an OR gate. Making an inverter (NOT gate) is trickier - imagine having a valve in the pipe which works when it is turned the opposite way from normal. The cunning thing is now to make the valves be operated by water-pressure, rather than by hand. Now the output of one bit of plumbing can control another bit of plumbing. Since any computer can be built from logic gates, you end up being able to build a computer - very large, very slow... and a bit wet. Transistors used in digital electronics are just doing the same thing. (Most other people here have been descibing "bipolar transistors", which are never used these days.) Everything uses CMOS transistors these days, which are very close to water valves descibed above. The 5 volt power supply is equivalent to the water supply used above. The more volts, the greater the pressure. (Too many volts and you get a leak... well, an arc, or a failure of some sort.) The flow of electricity is equivalent to the flow of water, above. CMOS transistors just act as valves - but which can be controled by voltage (pressure) supplied from another part of the circuit. Their operation is a bit more like treading on a garden hose pipe, but the principle is the same: pressure (voltage) controls flow (current). When you design a CMOS transistor circuit, you generally try to avoid constantly running current - instead you attempt to route voltage (pressure), or lack of it, around the place, rather than having to deal with flowing currents. Why? Well, the pipes can be thinner, the supply doesn't have to provide much water, and pressure changes travel quickly. In most CMOS circuits, current only flows in tiny amounts, and only during some transition - the slight amount of water needed to pressurize the next valve down the line, for example. This tells you why power consumption of circuits increase as they are used at higher speeds

[Brian W King]

Boylstad. Search for it. Its not an easy read. Thats about as dumded down as you can get and still be able to apply the physics behind the reality. Electircal theory has to come first, then you can build the fundamentals of gate theory on top of that. Without the electrical theory down pat though, gate theory is NOT going to make any sense. I was weaned on Boylstad and came out with an understanding that I haven't seen in many other people, including masters and doctorates.

[tom1443]

I'm a EE with an MS in Comp Sci who works as an embedded programmer. Honestly after the first year of college practically nobody except research scientists and chip developers give a rat's ass about the physics of electronics. Well, maybe that is not totally true but close. If your purpose though is to learn some practical electronics that can help you be a better programmer, I would get one of those science kits that you can get at Radio Shack that lets you breadboard a bunch of circuits. They usually come with an experiment book and all the components including resistors, capacitors, leds, potentiometers, timer/counters, and op-amps, transistors, and logic gates and will give you a good start on the basics. I bought one for my teenage son and he loved it. The one I bought had a built in power supply, voltmeter and current meter. Unless you want to interface complex external devices to a computer yourself that is probably all you need. Once you feel good about that I'd be looking for really good computer architecture book. Here you can learn about machine language design and how it relates to the hardware, memory layouts, bus architecture, chips selects and chip select logic, keyboard scanning, and the like. That will get you to the point where you can read a computer processor manual or just about any chip manual and understand it. After that I'd be thinking about buying a single board computer. There are a bunch of them available in $100-$200 range that have some built in I/O and prototype areas. They let you program on a PC and download to the SBC via RS232. Stick with an 8051 based board, it will be simple to learn and lots of free tools are available, plus it is a pretty widely used chip still. I'd stay away from the stamp processors at first, they are a more specialized type of processor. Now you can play around with some assembly (or even C code) and wiggle some lights or read some switches. That will marry up the electronics with the programming and put you well on the way to be a robotics programmer.

[DiscoJimmy]

A really good basic overview without getting too complex would be the Radio Shack series on electronics by Forest M.Mimms III - they're really basic, with lots of illustrations, but they give you a good grounding of how electricity, basic components, and logic gates work.

[Charles Pikell]

:thumbsup:I would also highly recommend the Feynman Lectures, easy and comprehensive for both electronics and physics. Charles Pikell

[Charvak Karpe]

dan neely wrote:

Unfortunately transistors work based on quantum mechanical principles. According to Classical Physics they won't work. The closest you're going to get in laymans terms is "When voltage is applied to the control lead MAGIC HAPPENS and it connects the input and output leads. When the voltage is removed from the control the connection is broken."

Modern 45 nm transistors may have quantum mechanical principles, but I think regular NPN transistors and MOSFETs have classical operation. 1. Pure silicon does not conduct because there are no floating electrons. 2. By "doping" silicon with impurities to the left or right on the periodic table, you create atoms in the lattice structure with extra free electrons or missing electrons (known as holes). 3. Look up the operation of a P-N junction. It's something about how current can only flow from the positively doped silicon to the negatively doped silicon because there are no excess electrons in the P-type silicon to start current flow the other way. If you make the junction too small or raise the voltage too high, you get quantum tunneling of electrons across the P-type silicon, but that's not an issue with transistors that you can see. 4. MOSFETs are fun. That stands for Metal Oxide Semiconductor Field Effect Transistor, which explains their function. They have an N-P-N arrangement of silicon with a layer of oxide covering the trio and then a layer of metal above that. If you try to pass current across through the N-P-N silicon, the N-P reverse diode junction blocks it. But if you apply a positive voltage to the metal at the top, the whole thing acts as a capacitor, pulling a layer of electrons out of the P-type silicon to create a temporary N-type channel right up against the oxide layer. This allows current to flow across because now you have some N-N-N above the N-P-N. Martin, I highly commend you for your curiosity and desire to learn the basics of physics and how they power our computers. The knowledge itself is not necessary to be a coder, but that inquisitive mentality is what makes someone an excellent problem solver. Sort of changing the subject, but still on the topic of physics basics, how do people feel about the current level of knowledge about conservation of energy, flows of heat, etc.? How many people realize that energy saving light bulbs don't help in places that are simultaneously running electric heaters? Or, why is running air conditioner in reverse bet

[Alan Balkany]

Very interesting, especially the parts about applying physics to daily activities. You should write some articles on this! If you did, I'd be interested in reading them.

[pg az]

Charvak Karpe wrote:

regular NPN transistors and MOSFETs have classical operation

While "regular" transistors are not flirting with the theoretical limits to their miniaturization, as you clearly essentially know to even talk about "lattice" or "free electrons", these concepts are intrinsically quantum mechanical, an NPN transistor isn't analogous to say a triode.

Charvak Karpe wrote:

empty bottle of syrup turned upside down

For an exciting video try the google query (( PITCH DROP EXPERIMENT )).

pg--az

[Dan Neely]

Charvak Karpe wrote:

Modern 45 nm transistors may have quantum mechanical principles, but I think regular NPN transistors and MOSFETs have classical operation.

Nope. The end effects look classical but the components they function on (eg holes) of all come from quantum effects. The continued shrinking of processes is adding quantum effects at at different level, such that the pseudo classical explanations that you get from crunching the quantum equations on larger transistors no longer apply.

This idea that particles could only contain lumps of energy in certain sizes
moved into other areas of physics as well. Over the next decade, Niels Bohr pulled
it into his description of how an atom worked. He said that electrons traveling
around a nucleus couldn't have arbitrarily small or arbitrarily large amounts of
energy, they could only have multiples of a standard "quantum" of energy.

Eventually scientists realized this explained why some materials are conductors of
electricity and some aren't -- since atoms with differing energy electron orbits
conduct electricity differently. This understanding was crucial to building a
transistor, since the crystal at its core is made by mixing materials with varying
amounts of conductivity.

http://www.pbs.org/transistor/science/info/quantum.html[^][^] http://www.madsci.org/posts/archives/2000-03/952639215.Ph.r.html[^]

Today's lesson is brought to you by the word "niggardly". Remember kids, don't attribute to racism what can be explained by Scandinavian language roots. -- Robert Royall

[destynova]

Interesting points! Have you seen the 'shouting at hard drives slows them down' video? :)

[pg az]

destynova wrote:

'shouting at hard drives slows them down' video?

I tried, with the Google query (( shouting hard drive windows )), but at least the first few links did not seem to say anything about Microsoft playing "catch up". Since this DTRACE-demo is talking about milliseconds, obviously ETW has the precision to give us this kind of information, but, correct me if I'm wrong, Microsoft does not expose their ETW-trace-points like a file of "Debug Symbols".

pg--az

[destynova]

I should have provided a link, like this one :)

[Edwin Smith]

I would beg to differ! An FET (the T means transistor) is essentially a triode and functions much the same as a vacuum tube triode although at a much lower voltage. the only difference between an NPN transistor and a vacuum tube triode is that the transistor requires a positive (+) to turn it on and a vacuum tube triode (and an FET) requires a negative (-) bias to turn it off. Otherwise both an NPN transistor and a vacuum tube triode are exactly analogous. The emitter is analogous to the cathode, the base is analogous to the grid and the collector is analogous to the plate. With an FET the source is analogous to the cathode (although there are both N-channel and P-channel FETs which have the opposite polarities) the gate is analogous to the grid and the drain is analogous to the plate. All of the mentioned devices have three connections which function in almost exactly the same way and intrinsically do so at the quantum level. Besides, intrinsically, triode means three electrodes. Both transistors and vacuum tubes have both Newtonian and quantum characteristics. However for the most part we use the smooth curves of the "real" world to describe their characteristics while any transfer of energy happens in distinct quanta. Both transistor and vacuum tube noise are quantum phenomena but they are usually expressed as DB in the analog world and mostly irrelevant in the digital world. All energy converted in a resistor can be expressed in the simple formula E=IR where E is volts, I is Amperes, and R is Ohms. Algebraically it is a linear function but at the quantum level Schrodinger's cat has the final say. You might just as well have said that everything is intrinsically quantum mechanical which adds nothing to the conversation. Diodes were discovered before quantum mechanics was even a theory and so were electricity and x-rays. There was great controversy as to whether light was composed of particles or waves because it has the characteristics of both. To learn basic electronics and Newtonian physics does not at all require any knowledge of quantum mechanics at all. I learned about donors and holes in semiconductors and electrons, neutrons, and protons in my electronics class and also noise but it really wasn't necessary to know to do design work with either transistors or vacuum tubes. In fact it was many years later that I even heard about quantum mechanics. It is commendable that somebody is curious enough to want to know electronics and physics. In fact a basic grounding in Physics should be ma