[SI-LIST] Re: Jitter measurement floor on different high bandwidth oscilloscopes
- From: "Alfred P. Neves" <al@xxxxxxxxxxxxx>
- To: <wolfgang.maichen@xxxxxxxxxxxx>, "'Heyfitch'" <heyfitch@xxxxxxxx>
- Date: Thu, 10 Jun 2010 11:06:22 -0700
An interesting case of RJ data dependency is the following:
Histogram and estimate PDF of a simple RJ dominated process with only clock
like pattern (DJ=0), 10101010.... you should have a simple Gaussian PDF
described by mean, 1-sigma if you measure statistically significant sample
of single period jitter.
Now, create a simple data pattern with RJ dominated jitter, say 2^7-1 and
run this pattern through a channel with significant loss where the pattern
has significant power spectra density ( below 1/2 data rate, or nyquist).
Now, each UI transition has a specific RJ pdf associated with it, and in
every practical case the 1-sigma of each pdf varies from UI boundary to UI
boundary suggesting RJ pattern dependency. The estimated 1-sigma not only
varies with UI but, on the average, has increased from the simple clock.
When the UI's (unit intervals) were 400psec I believe it was usually safe to
assume statistical independence, but moving on to UI's of 40psec I believe
this isn't true, especially considering the ratios of RJ to TJ for
10E-12BER.
Alfred P. Neves <*)))))><{
Hillsboro Office:
735 SE 16th Ave.
Hillsboro, OR, 97123
(503) 718 7172 Business
(503) 679 2429 Mobile
Main Corporate office:
Teraspeed Consulting Group LLC
121 North River Drive
Narragansett, RI 02882
(401) 284-1827 Business
(401) 284-1840 Fax
http://www.teraspeed.com
Teraspeed is the registered service mark
of Teraspeed Consulting Group LLC
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of wolfgang.maichen@xxxxxxxxxxxx
Sent: Thursday, June 10, 2010 10:37 AM
To: Heyfitch
Cc: prasad; si-list@xxxxxxxxxxxxx; si-list-bounce@xxxxxxxxxxxxx;
wolfgang.maichen@xxxxxxxxxxxx
Subject: [SI-LIST] Re: Jitter measurement floor on different high bandwidth
oscilloscopes
Hello Vadim,
you are right, RMS assumes the two contributions are uncorrelated. For
scope sample time jitter and ADC (and amplifier chain) input noise this is
probably a pretty good assumption, but depending on the design of a
particular scope there could of course be some correlation or
anticorrelation. E.g. both could be caused by noise on the power supply
to the sampling circuitry (and that noise could be non-Gaussian e.g. if
caused by a switching power supply). That's one reason I tried to caution
right away that in a short post one can only scratch the surface. I did
not intend to imply that pure timing jitter and voltage noise jitter
always add up that way.
As far as ISI (data dependent) jitter is concerned, personally I normally
consider it voltage noise translated to jitter, if one broadens the
meaning of noise to "voltage deviation from the 'normal' waveform". RMS
addition rests on the assumption that both contributors are not only
independent of each other, but also that both have Gaussian distribution.
Especially the latter assumption can only be a more or less crude
approximation for ISI since ISI always comes from a limited number of
preceding bits. In an eye diagram ISI shows up as discrete bands of
transitions, whereas random jitter causes continuous broadening of the
transitions. The mathematically sound approach is of course convolution of
all jitter contributors (this still assumes independence of the jitter
contributors, as well as linearity of the system, so is not universally
applicable either), and RMS summation is simply the convolution result for
two Gaussian distributions.
Wolfgang
Heyfitch <heyfitch@xxxxxxxx>
Sent by: si-list-bounce@xxxxxxxxxxxxx
06/10/2010 08:31 AM
To
wolfgang.maichen@xxxxxxxxxxxx
cc
prasad <hariprasad.palli@xxxxxxxxx>, si-list@xxxxxxxxxxxxx,
si-list-bounce@xxxxxxxxxxxxx
Subject
[SI-LIST] Re: Jitter measurement floor on different high bandwidth
oscilloscopes
Wolfgang,
you stated that the pure timing jitter and translated voltage noise jitter
add up in RMS fashion. This would imply that the two are completely
uncorrelated. Is this always true? What comes to my mind is the ISI jitter
of a non-clock (i.e. non-periodic) signal caused by a band-limited
channel.
Would such jitter be interpreted as voltage or timing jitter or a
combination of both? If the latter, would the pure timing and voltage
components be considered uncorrelated?
Thanks.
Vadim Heyfitch
On Thu, Jun 10, 2010 at 1:34 AM, <wolfgang.maichen@xxxxxxxxxxxx> wrote:
> Hello Prasad
> jitter performance analysis of high-end scopes is decidedly non-trivial,
> but to get you started:
>
> As for random jitter measurement floor, there are three parameters that
> are important:
>
> (1) the scopes sample clock jitter (i.e. when the samples are actually
> taken vs. their ideal position)
> (2) the sampler's noise
> (3) the signal's slew rate
>
> The measured random jitter can be thought of two components, first pure
> random timing jitter, second noise that gets translated into timing
> jitter. A perfectly noise-free signal (and noise-free scope) can still
> show random timing jitter. On the other hand, noise on the signal (or in
> the sampling circuit) always gets translated into timing jitter, equal
to
> voltage noise divided by the slew rate. So a large and/or very fast
rising
> signal (corresponding to a large slew rate) will be less affected by a
> given amount of noise than a signal with slow slew rate, but the effect
> will never be zero. Both components (pure timing jitter and translated
> voltage noise) add up in RMS fashion.
>
> So if you have two scopes, one (A) with low jitter floor (sampling
timing
> jitter) but high noise floor, and another (B) with high jitter floor but
> low noise floor), it can depend on the signal to be measured which one
> will produce the lower measured jitter number. If the signal has a high
> slew rate (making it less sensitive to noise) then scope (A) has an
> advantage, otherwise scope (B). So in reference to your question which
is
> better, the answer is the common "it depends", in this case it depends
on
> the signal's slew rate.
>
> Wolfgang
>
>
>
>
>
>
>
> prasad <hariprasad.palli@xxxxxxxxx>
> Sent by: si-list-bounce@xxxxxxxxxxxxx
> 06/09/2010 08:57 PM
>
> To
> si-list@xxxxxxxxxxxxx
> cc
>
> Subject
> [SI-LIST] Jitter measurement floor on different high bandwidth
> oscilloscopes
>
>
>
>
>
>
> Hi every one....
>
> i am evaluating high bandwidth oscilloscopes (12GHz) from different
> vendors. I was looking the data sheets of them. One of the
> them(DSO91204A) has very good noise floor compared to others. Though
> its a good thing for me but when it comes to the jitter measurements ,
> the lowest jitter that can be measured on that is dependatnt on the
> slew rate of the signal ,which is actually true(since the voltage
> noise will have a second order effect on the timing of the signal).
> But when i looked at one more vendor (SDA13Zi) the noise floor is poor
> compared to other. In which case the lowest jitter that can be
> measured (jitter measurement floor)should be higher than earlier. But
> if you look at the datasheet, they have specified a fixed value for
> this which is very less .
> My question is , if the noise floor is high in the second box how
> would the jitter measurement floor be less?
> second one is , since the timing noise(jitter) is dependatnt on slew
> rate, how a fixed value is given in datasheet?
>
>
> please help me understand. Am i missing some other factor here?
> Welcome all your suggestions and ideas...
>
>
> thanks in advance...
> prasad
>
> h
>
> On 09/06/2010, colin_warwick@xxxxxxxxxxx <colin_warwick@xxxxxxxxxxx>
> wrote:
> > (Note: I sent this info to Hermann off-list but he suggested it might
be
> of
> > general interest. Send flames to me, not Hermann, if it isn't.)
> >
> > In ADS the implementation is:
> >
> >
> > "Fast" corner
> > (a) the max values are selected for all the I-V data (Pullup,
Pulldown,
> > Power Clamp a Ground Clamp) and for the waveform data (Ramp, Rising
> Waveform
> > and Falling Waveform), and
> > (b) the min values are selected for all R, L, C, delay and TT data.
> >
> >
> > "Slow" corner is the reverse obviously
> > (a) the min values are selected for all the I-V data (Pullup,
Pulldown,
> > Power Clamp a Ground Clamp) and for the waveform data (Ramp, Rising
> Waveform
> > and Falling Waveform), and
> > (b) the max values are selected for all R, L, C, delay and TT data.
> >
> >
> >
> > -- Colin
> >
> > -----Original Message-----
> > .
> > .
> > .
> >
> > Any feedback from the tool vendors how they implemented this selection
?
> >
> > Thanks and Regards
> >
> > Hermann
> >
> > EKH - EyeKnowHow
> > Hermann Ruckerbauer
> > www.EyeKnowHow.de
> > Hermann.Ruckerbauer@xxxxxxxxxxxxx
> > Veilchenstrasse 1
> > 94554 Moos
> > Tel.: +49 (0)9938 / 902 083
> > Mobile: +49 (0)176 / 787 787 77
> > Fax: +49 (0)3212 / 121 9008
> > ------------------------------------------------------------------
> > 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
> >
> > Old (prior to June 6, 2001) list archives are viewable at:
> > http://www.qsl.net/wb6tpu
> >
> >
> >
>
> --
> Sent from my mobile device
> ------------------------------------------------------------------
> 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
>
> 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
>
> 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
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
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
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
