Bit error
A bit error (also called a bit kipper ) is an error in a single bit . Instead of the correct value "1", this bit then has the wrong value "0" or vice versa. This bit is then called toggled. Bit errors can occur in digital data transmission or in a data memory .
With digital data transmission
In digital data transmission, a bit error occurs when the value of a received bit does not correspond to the value of the sent bit. Bit errors can be caused by a disturbed, e.g. B. too noisy , transmission channel arise.
With data storage
A bit error is present in data memories when the value present in a memory cell does not correspond to the value originally to be stored. Bit errors in data memories can arise from permanent or volatile errors. Permanent errors arise from defective memory modules or their control circuits. Fleeting errors occur randomly due to external influences, e.g. B. glitches on.
Error detection and correction
A single bit error can be detected in a bit sequence with a parity bit , but it cannot be corrected.
A system can be assured to be able to recognize a certain maximum number n bit errors per certain amount of data (for example a maximum of 2 bit errors per 32-bit data word with ECC memory ). In general, it is then possible that it can automatically correct up to (n - 1) bit errors - the bit errors are then without consequence.
If there are n bit errors, no correction is possible; the information is lost (at this point) - but it is noticed. More complex procedures can now try to recover the information; for example, they can ask the sender to retransmit the information.
In the case of> n bit errors, bit combinations can result which the system incorrectly evaluates as correct, as well as combinations which are recognized as incorrect, but are "automatically corrected" in a different (incorrect) bit pattern that does not correspond to the original one.
There are complex procedures with error-correcting codes for error detection and error correction .
Effects
Bit errors that could not (automatically) be corrected or noticed:
- They can have no consequences if they occur in “empty areas” of a data carrier, or in less relevant or highly redundant data. (A short crack in a phone call is "simply overheard", a wrongly colored pixel in a photograph is "simply overlooked", a wrong letter in a text is also mostly uncritical.) Therefore, many data processing devices have no bit error detection (e.g. graphics cards, Sound cards, game consoles).
- Bit errors can change important data and bring (computer) systems into unforeseen states. If this is noticed, the system at a higher processing level (or the user himself) can recalculate the information, request anew or react in some other way. If the data / program change is not noticed, this can lead to incorrect (calculation) results and subsequently to incorrect decisions. A minimum level of protection against bit errors is therefore usually prescribed for workplace computers in companies, administration and research.
See also
Individual evidence
- ↑ ITU-T G.701 General Aspects of digital Transmission Systems, Vocabulary of digital Transmission and Multiplexing, and Pulse Code Modulation (PCM) Terms, 03-1993
- ↑ Communication technology, transmission - switching - processing, E. Herter / W. Lörcher, Hanser, 6th edition 1992
- ↑ Semiconductor circuit technology, U. Tietze / Ch. Schenk, Springer