[T10] Oscilloscope response for SAS-4 design validation

min-jie_chong at keysight.com min-jie_chong at keysight.com
Tue Jan 31 10:47:58 PST 2017


Hi fellow PHY T10 members,

This email is sent out to the broader audience as a follow-up to the T10 call last week regarding the oscilloscope's response that could affect SAS-4 testing.
We hope to continue the discussion in the next call to talk about this topic.


A proposal was approved to change the scope bandwidth requirement from 40 GHz to 33 GHz.
See the snippet that I extracted from one of the meeting minutes.

http://www.t10.org/cgi-bin/ac.pl?t=d&f=16-272r0.pdf

[cid:image001.png at 01D27BB6.7D60B5C0]

However, I think it's missing an important element - the response of the scope. In the OIF-CEI 3.1 spec, there are multiple references of a fourth-order Bessel-Thomson response, which I believe is what a typical equivalent time (sampling) scope response is. I also believe the OIF-CEI 3.1 spec was referencing to the equivalent time scope's response when it was written. Although the bandwidth specifies the -3 dB attenuation point in the equivalent time scope's response, the roll off is rather gradual and there's usable bandwidth beyond the -3 dB cutoff.

[cid:image002.png at 01D27BB6.7D60B5C0]

On the other hand, the real-time scope is what we typically used for SAS PHY measurements and it has a different response. It has a rather sharp roll off beyond the -3 dB bandwidth, with no usable bandwidth beyond that. In order to achieve the fourth-order Bessel-Thomson response of 33 GHz bandwidth, a higher bandwidth scope (e.g. 50 GHz) will be required. This contradicts to the intent of reducing equipment cost, but it does adhere to the OIF-CEI 3.1 standard.

[cid:image005.jpg at 01D27BB7.DD95A850]

I would like to table this for discussion at the SAS PHY meeting to see if we should carry forward the requirement from OIF-CEI 3.1 and explicitly mention this requirement, or define real-time scope response ((flat frequency response with a brick wall filter) for the SAS-4 spec. I would go with the latter because I don't believe the impact of a real-time scope response is significant to the results, as much of the 22.5 Gb/s signal higher frequency components are attenuated by the lossy channel.

Regards,
Min-Jie Chong
Keysight Technologies
719-590-2010

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.t10.org/pipermail/t10/attachments/20170131/cb664392/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 19782 bytes
Desc: image001.png
URL: <http://www.t10.org/pipermail/t10/attachments/20170131/cb664392/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.png
Type: image/png
Size: 25908 bytes
Desc: image002.png
URL: <http://www.t10.org/pipermail/t10/attachments/20170131/cb664392/attachment-0003.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image005.jpg
Type: image/jpeg
Size: 43373 bytes
Desc: image005.jpg
URL: <http://www.t10.org/pipermail/t10/attachments/20170131/cb664392/attachment-0001.jpg>


More information about the T10 mailing list