bit? an infinitesimal switch? physically?

Question

I know i should have asked this question a long time ago in my BCA while we were using bits as smallest storage units which have only two values generally, 0 & 1. But as I recently started reading Let us C by Yashwant Kanetkar to revise the language, a question formed into my mind. A question very simple but mind-bogglingly difficult to understand:
In physical form, what is bit?
An infinitesimally small on/off switch? I mean very old computers used big switches to store data. Switch on for 1; Switch off for 0.
If it is such a small switch, how is it that so so so... many switches in amounts of many GBs be stored on small smaaaaaal.. memory devices such as a 16GB memory card???

Answer 1

A bit is anything we can engineer.

A solenoid has been used for a bit in electromechanical computers.

A vacuume tube has been used to create a circuit that can store the on/ off value of a bit.



Modern memory marshalls small amount of electrons to create a bit.



Odd, from the present point of view, were storage system like mercury delay line or storing info on the surface of a CRT monitor.

Answer 2

Physically, a bit can be just about anything. It could be the voltage on a silicon wire, or the orientation of a brass gear, or the position of a billiard ball, you get the idea. All that matters is that the physical component is part of some larger process that can make use of it as if its state represents one of two possible values.



In most of the computers we use, bits in working memory (registers, RAM, process caches, and any other memory incorporated in the processor architecture) are represented using voltages on silicon wires. As I recall, a binary 0 is represented using high voltage and a binary 1 using low voltage, which is counterintuitive, but supposedly there's some engineering reason why that works better. Bits on a hard drive are represented using the orientations of ferromagnetic atoms.



None of these components is infinitesimally small (or even infinitely small, which I think was what you were trying to get at; 'infinitesimally small' sounds like the object is actually large). Our current understanding of physics suggests that there is a scale below which distance has no meaning anymore, and that a certain finite region of space, time and energy is required to store 1 bit of information or perform a binary operation on it. Although we have done a pretty good job of miniaturizing computer components, existing hardware is still nowhere near this physical limit.

Answer 3

You're on the right track, sort of...



A solid state memory actually stores a very tiny electrical charge to represent the state. That tiny physical space can hold a charge (1 state) or no charge (the other state).



On a magnetic memory device, the state is represented by the way a magnetic particle is orientated.



The smallest unit of charge is the electron. How big is an electron? Very small.



It's very good to let your mind form such questions and seek answers....

Answer 4

Not infinitesimally small there is a limit to how many switches can be stored on a single chip but yes that's basically what they are. They're very tiny electronic switches that can either be on or off.





Back in the old days they were electronic switches - huge glass tubes called valves - see image:



http://www.google.co.uk/imgres?imgurl=http://www.keith-snook.info/valve-data/Valves%2520for%2520sale%2520pics/Mullard-EF86-long-short.jpg&imgrefurl=http://www.keith-snook.info/valves-for-sale.html&h=556&w=1024&sz=111&tbnid=tVYAy-DM53WzJM:&tbnh=72&tbnw=132&zoom=1&usg=__bzzbFOgM_kCDpbmhQal3hIjJ3bE=&docid=P9QXUiu8GdEb6M&sa=X&ei=UNpEUuTyEIjAtQb3sYG4DQ&ved=0CEEQ9QEwBA



Then due to a better understanding of how crystals work the transistor was invented: This too is an electronic switch that serves the same purpose but much smaller. The transistors got smaller as the properties of semiconductors were better understood but a transistor is basically an electronic switch:



http://www.google.co.uk/imgres?imgurl=http://www.tandyonline.co.uk/media/catalog/product/cache/1/image/9df78eab33525d08d6e5fb8d27136e95/b/c/bc635-transistor.png&imgrefurl=http://www.tandyonline.co.uk/bc635-npn-transistor.html&h=599&w=599&sz=49&tbnid=K7OFUOoP6eKsDM:&tbnh=90&tbnw=90&zoom=1&usg=__Md1XP6atU3ok4UhpLYSfmvbZPVo=&docid=XyzrFjPa-kV8pM&sa=X&ei=RttEUt6XLoTMtQaEk4C4CQ&ved=0CEMQ9QEwAw



The above image is a single transistor but the actual piece of semiconductor is not very big here's the way it works:



Just take a semi-conductor (for example silicon) and dope it with a small amount of impurities either negatively charged (n-type) or positively charged (p-type). Sandwich the them together to make npn or pnp type transistor, a connecting wire goes from the edges and a third wire to the middle slice. Now if a current above a certain level hits the middle layer of the sandwich current can flow through the transistor else it can't - that's how a transistor works. Of course you can make the wires really thin if you're using photo-chemical etching so you can take a piece of silicon and etch the circuits so that they can be microscopically small.



here's more info:



http://www.usbmemorysticks.net/flash-drive-technology



http://en.wikipedia.org/wiki/Bipolar_junction_transistor



http://www.youtube.com/watch?v=aWVywhzuHnQ



hope this helps.



Edit:



As Modulo_function pointed out a bit doesn't necessarily have to be an electronic switch. When data is stored on a hard disk drive it's stored magnetically - north/south or south/north oriented tiny magnets. The can be transmitted as pulses of light - photons present = 1 photons absent = 0 or encoded onto radio signals, in fact anything that has two states can be used to store binary digits. Back in the early days of computing data was stored on punched paper where the presence or absence of a hole in the paper would represent the binary digit.

Answer 5

a bit is 1/8 of a byte in computer parlance or octet in networking parlance.

you basically have it down pat. it's binary - on or off, 1 or 0, high or low. for TTL, 5V or 0v (ground). you get the idea.



a bit is represented as lower case b and a byte is represented as upper case B. so in networking if you are looking at Mbps, that's Mega (SI) (1e6=1000^2=1e3^2) bits per second. also shown as Mb/s.

however... Megabytes today is shown in microsoft OS's and browsers as IEC units (2^20=2^10^2=1048576=1024^2) which is MebiBytes or MiB. so a browser unfortunately shows MB/s when it really means MiB/s or KB/s when it really means KiB/s.

with computer measurements, things are going more and more towards IEC units which are powers of 1024=2^10.



16GB memory card means 16*2^30*8 bits. but mind you, flash memory blocks in a memory card can be written up to about 3000-10000 times (Program/Erase cycles) depending on spec/design.

so their maximum amount of writes for the device is 16GiB * 2^30B/1GiB * 3000PEcycles



in a physical form, a bit is simply the binary state of something. it can be the state of a transmission line to/from a logic gate or of the output or input of a gate. for this you have to get into digital electronics. it can be the state of a parallel I/O line on an arduino board or PIC chip. it can be the state of a serial port data channel, such as found with SATA or USB, firewire, thunderbolt, etc.



it can be RF transmission state for WIFI or zigbee or cell signals or DECT 6.0. those send transmission signals and encode and modulate them in such a way (such as QAM) in order to be able to transmit and receive bits over the air wirelessly.



MLC flash memory has 3 states. TLC has 4 I think (or was it 8?), and SLC has 2 states for each cell (I suppose you could call it a bit because that's what it ends up being, actually in MLC's case it ends up being 2 bits since 2^2=4 or 2nd root of 4 is 2).



I would have to look up MLC and TLC's number of states to be sure, since that's not of great importance to my work.



I could go on and on....