The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> High-Speed I/O Design >> BER measurement
https://designers-guide.org/forum/YaBB.pl?num=1376061559

Message started by raja.cedt on Aug 9^th, 2013, 8:19am

Title: BER measurement
Post by raja.cedt on Aug 9^th, 2013, 8:19am

Hello all,
Can any one please explain how BER measured in real time? For example attached is one real time bath tub curve for TIA, in that paper he gave BER from that curve directly and repeated the curve for different inputs. I will be very happy if some one explain clearly.

THanks,
Raj.

Title: Re: BER measurement
Post by boe on Aug 9^th, 2013, 12:41pm

Raja.cedt,

Quote:

Can any one please explain how BER measured in real time?

Usually you stimulate your system with pseudo-random data & compare the received data with the data sent.
Does that help?
- B O E

Title: Re: BER measurement
Post by ywguo on Aug 14^th, 2013, 12:50am

Hi Raj.

Since it usually take very long time to test the BER with PRBS, BER is often estimated, calculated with jitter. If the total jitter is not less than one UI, there must me bit decision error. And the total jitter is split into deterministic jitter and random jitter. The deterministic jitter is the flat part in the bathtub curve, and the random jitter is the skirt-like part of the bathtub curve.

Since the random jitter is Gaussian distribution, then the probability of one clock edge outside the limits is calculated with erfc() function. That means the probability of bit error.

1/2×erfc(√2×α)=BER

For example, α=14 for BER=10^-12. To ensure BER=10^-12, Total jitter = DJ + 14×RJ must be less than one UI.

Best Regards,
Yawei

Title: Re: BER measurement
Post by raja.cedt on Aug 14^th, 2013, 2:43am

Dear ywguo,
Thanks for the reply. However i didn't understand The deterministic jitter is the flat part in the bathtub curve, and the random jitter is the skirt-like part of the bathtub curve. Could you please how?

Thanks,
Raj.

Title: Re: BER measurement
Post by Ken Kundert on Aug 14^th, 2013, 6:30pm

The deterministic jitter is large but its amplitude is bounded. Thus, the largest excursions come from the random jitter, whose amplitude is not bounded. If there were only deterministic jitters the walls of the bathtub would eventually go vertical because the amplitude of the deterministic jitter is bounded. The fact that walls have a slope to them rather than becoming purely vertical stems from the random noise (again, because its amplitude is unbounded). This is explained in http://www.designers-guide.org/Analysis/bang-bang.pdf.

-Ken