How does ADN2812 jitter could not tolerate 0.067 UIp-p synchronizing with 15 MHZ data stream; could ADN2913 tolerate the 0.067 UIp-p at 15MHZ?
How does ADN2812 jitter could not tolerate 0.067 UIp-p synchronizing with 15 MHZ data stream; could ADN2913 tolerate the 0.067 UIp-p at 15MHZ?
Dear HNL,
The target applications of both ADN2812 and ADN2913 are to support SONET and SDH telecommunications.
With their superior jitter performance, both ADN2812 and ADN2913 support not only targeted applications, but also many other data link applications.
In the datasheets, we present the typical Jitter Tolerance performances specified by GR-253-CORE, at the specified signal rates, respectively.
We haven't such jitter tolerance performance measured at your interested 15Mbps because
1. what is the industry application at this rate?
2. what is the jitter spectrum defined for this 15Mbps applications? and
3. our lab has a valid jitter evaluation kit supports SONET and SDH jitter performance.
If you agreed, let's use both device jitter tolerance performance data at OC-3, 155.52Mbps to derive the possible 15Mbps jitter tolerance performance.
ADN2812:
ADN2913:
Let's use the stringent subset for derivation:
OC-3 30Hz 50 UIp-p
300Hz 24 UIp-p
6500Hz 3.5 UIp-p
65kHz 0.8 UIp-p
1.3kHz 0.61 UIp-p
Let's assume the 15Mbps signal is the 10th harmonic of OC-3 signal, so we can easily estimate a jitter tolerance performance when CDR process this 15Mbps signal by using OC-3 jitter spectrum as:
OC-3 30Hz 5 UIp-p
300Hz 2.4 UIp-p
6500Hz 0.35 UIp-p
65kHz 0.08 UIp-p
1.3kHz 0.061 UIp-p
In your application, the 15Mbps signal source has strong jitter component at kHz?
Of course, this simple quick derivation lead to a jitter tolerance performance estimation when used both CDRs to process the 15Mbps signal. To be more precisely, you'd better to use one CDR evaluation board, apply the 15Mbps signal to the board for a real application evaluation.
Hope these are useful to you.
Best Regards,
Dongfeng
Thank you for your response Dongfeng.
Your Q. 1. What is the industry application at this rate? Answer:The
aplication is a tape transport for linear tape open technology (LTO).
Present design runs from 4 MHZ to 24 MHZ. So, I am thinking of evaluating
the ADN2913 if it can tolerate more jitter than ADN2812.
2. What is the jitter spectrum defined for this 15 Mbps
application?Answer:I don't know the spectrum but I measured the jitter in
the time domain with an oscilloscope to be 4.5 ns peak to peak period
jitter and calculated 4.5ns/(1/15MHZ) = 0.067 UI. see picture in attachment.
Your comment to scale the the jitter tolerance of OC-3 to the 10 th
harmonic is OK . I think you meant 130KHZ instead of 1.3KHZ.
Your Q: Does the 15Mbps signal source have strong jitter component at KHZ?
Answer: Yes, at about 50 KHZ. (the rotation of a capstan motor).
I already evaluated the ADN2812 and it could not maintain lock.
Do you know if the ADN2913 transfer bandwidth and the sample phase
adjustent can help to tolerate more jitter than the ADN2812?
Thanks,
HNL
On Mon, Nov 7, 2016 at 7:03 AM, Dongfeng <
Dear HNL,
Sorry for my typo, it should be 1.3MHz instead of 1.3kHz.
Did you lock on to an input clock signal or a burst mode PRBS type of modulated data signals?
Both ADN2812 and ADN2913 supports continuous mode, PRBS type, NRZ encoded data signals. If your input data signal is a burst mode signal, both ADN2812 and ADN2913 would lost lock easily.
Best Regards,
Dongfeng