0
\$\begingroup\$

When designing a transmission link, are there any standardized/typical objectives for maximum bit error rate? Apart from some papers speaking about values like \$10^{-10}\$ or \$10^{-12}\$ for optical fiber I not only failed to find something comprehensive, but I found hardly anything at all. Just some measurements/simulations.

I'm mostly interested in values for wireless links, but anything about wired connections is appreciated as well.

\$\endgroup\$

1 Answer 1

Reset to default
1
\$\begingroup\$

When designing a data transmission link, you design it to meet the "requirements". If the link raw-data error rate is likely high then you either retransmit several times and meet the "requirements" by shear numbers or you design in "packet retransmit protocols" in order to meet the "requirements".

Raw bit error rates can be quite high but, by using the appropriate protocol this can be significantly reduced.

As to whether there is a "standard" that all data links adhere to regarding data-payload errors I would say not - if you are designing something then you are designing to a set of requirements (or a spec) and this takes priority.

\$\endgroup\$
2
  • 1
    \$\begingroup\$ A complication when trying to characterize error rates for things like wireless links is that bit errors in the raw data link are often not independent; if they were, and if 99% of 256-bit packets went through with no bit errors, that would suggest a raw error rate of less than 0.005%, which would in turn suggest that more than 99.999999% of packets would go through with at most eight bit errors, so adding forward error correction sufficient to recover eight bit errors would ensure reliable delivery. Of course, wireless links don't work that nicely. \$\endgroup\$
    – supercat
    Commented Dec 3, 2014 at 16:02
  • \$\begingroup\$ I was measuring a visible light wireless link. Before creating the transmitter, I was told that the signal should be detectable 50 cm away from the the receiver. I wanted to have something more measurable than detectability and analysed the eye pattern, measuring its Q factor (which you can easily turn into BER). Then I thought about "drawing a line", above which the BER is acceptable, but didn't want it to seem that I adjusted that value so that it's still good at the desired distance, 50 cm. That's why I was looking for standards/typical values. \$\endgroup\$
    – hryghr
    Commented Dec 3, 2014 at 22:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.