Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

BER (with No FEC) vs Pre-FEC BER vs Post FEC BER

2,728 views
Skip to first unread message

krishna G

unread,
Jan 14, 2016, 3:39:53 AM1/14/16
to
Hi,

Can you please explain the difference between BER (with No-FEC) , PRE-FEC BER and Post-FEC BER??

As per my understanding , PRE-FEC BER is the BER at the I/p of the FEC Decoder and Post FEC BER is the BER at the o/p of FEC Decoder. But when I saw the graph of BER vs OSNR , I see curves for No-FEC,PRE-FEC,POST-FEC and asymptotic..

Can you explain the difference elaborately.

Regards
Krishna

Huub van Helvoort

unread,
Jan 19, 2016, 5:39:13 AM1/19/16
to
Hello Krishna,
PRE-FEC BER are the bit errors caused by attenuation, ageing,
temperature changes of the optical fiber.
PRE-FEC indicates that the signal on the optical fiber is FEC
encoded.
The FEC decoder will recover the original signal, but depending
on the PRE_FEC BER it will succeed to recover the original signal
completely without errors.
Or, if the BER on the fiber is too high, the recovered signal will
contain bit errors.
If the signal was FEC encoded the remaining bit errors after the
decoder are called POST_FEC BER.

The NO_FEC BER are the bit errors detected when no FEC coding is
used on the optical fiber.

See also: https://goo.gl/cuhbjj

Best regards, Huub.




--
reply to hhelvooort with 2 'o's
================================================================
http://www.van-helvoort.eu/
================================================================
Always remember that you are unique...just like everyone else...

Ramakrishna Pratapa

unread,
Jun 21, 2016, 9:05:56 AM6/21/16
to
Just wanted to add a point here on the asymptotic nature of the curves:

As the OSNR reduces, the BER increases and FEC facilitates error correction depending on the coding gain the FEC algorithm can provide. Beyond the FEC limits, even minor OSNR degradation could mean significant performance loss which is seen in the asymptotic nature of the curve.
0 new messages