[SI-LIST] Re: SDA bandwidth
- From: wolfgang.maichen@xxxxxxxxxxxx
- To: gypse <gypse@xxxxxxxxx>
- Date: Tue, 14 Oct 2008 09:36:21 -0700
Salut Jean,
regarding the bandwidth, you are correct to consider the rise time rather
than the bit period (although of course they are not completely unrelated,
since the rise time must alway be significantly shorter than one bit
period). Lacking additional information, assuming a rise time of 30 - 40%
of the bit interval is often a good guess, in your case that gives 30 -
40ps. The corresponding equivalent bandwidth is then approx. 8 - 11 GHz.
Of course there is still considerable signal energy above that limit, so a
better number is the knee frequency, 0.5/Tr, i.e. around 17 GHz for a 30ps
rise time signal. And at the same time, you'd get 30% attenuation of that
frequency if you chose your scope bandwidth equal to this frequency, which
is quite substantial, so you may want to choose a somewhat higher
bandwidth scope than that. Unfortunately even high-end real-time scopes
top out around 20 GHz today, so you'll be running into technical
limitation and may have to settle for less than ideal (ideal would be a
bandwidth of 3 - 6 times your highest frequency of interest). You can get
equivalent time sampling scopes (Agilent, LeCroy, and Tektronix all have
such scopes & sampling plugins with analog bandwidth up to 50 ... 100
GHz), but of course they can't do single-shot capture of a data stream. On
the upside, there is virtually no limit to the timing resolution this type
of scope can achieve (sub-ps).
As for timing resolution with a real-time scope, your assumption is far
too pessimistic. True, a 60 GS/sec scope will only give you 6 sample
points per 100ps interval, but then the scope does interpolation between
the curves. The simplest type would be linear interpolation ("connect the
dots with straight lines"), already that would give you much better
resolution for your edge crossings than 100ps/6. Modern scopes all use
better interpolation methods than that (sin(x)/x interpolation), which
will give you quite accurate edge crossing information and allow jitter
measurements down to single picoseconds and below. Another way to see it
is that the scope doesn't just measure low and high (like a logic analyzer
would), rather it gives you finely resolved voltage information at each
sample instant, so you have more information in the voltage direction
which then allows to determine the crossing point more precisely as well.
There are other differences between real-time scopes and equivalent time
sampling scopes that will be important for you - single shot capture
capability, influence of trigger jitter, ability to decompose timing
jitter into its components etc. As a shameless plug, my book "Digital
Timing Measurements" (available from Amazon.com, ISBN-10: 0387314180,
ISBN-13: 978-0387314181) talks at length about these considerations in a
(I hope!) easy-to-understand manner. There are also a lot of apps notes
out that discuss these topics.
Hope that helps
Wolfgang
gypse <gypse@xxxxxxxxx>
Sent by: si-list-bounce@xxxxxxxxxxxxx
10/14/2008 07:23 AM
To
si-list@xxxxxxxxxxxxx
cc
Subject
[SI-LIST] SDA bandwidth
Hi all,
Have you got an idea of the required features (bandwidth + sampling rate)
for a
SDA to analyse 10Gbits/s serial links ?
Bandwidth:
Thumbs rules for the bandwidth would lead me to choose 5 times the
fundamental
frequency to see the 5th harmonics ie 25 GHz. But high speed devices have
a rise
time of only 40ps, so the estimated bandwidth of such signals would rather
be
0.5/30ps = 12 Ghz ... So, what to choose ?
Sampling rate:
Am I wrong if I say that with a sampling rate of 60 Gsamples/s, I
will capture an eye with only 6 points and therefore I will be able to
measure
jitter with an accuracy of only 16 %. Should I need a higher sampling rate
?
Thanks.
Jean.
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List technical documents are available at:
http://www.si-list.net
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List technical documents are available at:
http://www.si-list.net
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
