[SI-LIST] Re: Fibre channel interconnect margins
- From: "Chris Cheng" <Chris.Cheng@xxxxxxxxxxxx>
- Date: Wed, 5 Jul 2006 14:10:29 -0700
"Using the BER as the benchmark for acceptance (something like 10e-12)"
I started the whole discussion based on the above statement.
And I agree with what you say that no company in its right mind would =
ever reveal their BERT. But have you ever consider what those sigma will =
look like at PCIe Gen II or 10G ethernet (pick a BERT between 10e-12 to =
-19 as you pointed out). At 1 or 2G rate, you can live with a few ps of =
non-existing Rj but not at 10G, it will eat your margin alive.
-----Original Message-----
From: Paul Levin [mailto:levinpa@xxxxxxxxxxxxx]
Sent: Wednesday, July 05, 2006 1:50 PM
To: Chris Cheng
Cc: weirsi@xxxxxxxxxx; David Instone; si-list@xxxxxxxxxxxxx
Subject: Re: [SI-LIST] Re: Fibre channel interconnect margins
Dear Chris,
You've asked this question before and not received what you consider a=20
satisfactory answer. Clearly it has to be better than 1e-12; almost as=20
clearly, it has to be worse than 1e-19 (1 error after thirty years at 1=20
Gbps if my math serves me correctly.)
Have you ever considered that (in my humble opinion) no company in its=20
right mind would ever reveal what their number was since its competitors =
would then try to use that information against it?
Regards,
Paul Levin
Xyratex
_____________________
Chris Cheng wrote:
> Steve,
> I start off by asking the same question before, "does your company =
ship =3D
> products with 10e-12 error rate ?" and it seems it comes down to the =
=3D
> following answers :
>=20
> a) "Its a fact of life, I'll try to keep it to 10e-x13,14,15 or =
better."
>=20
> Well, I still want to know what that number is ? And when you have a =
=3D
> customer that has a peta byte of installation, how does that translate =
=3D
> into actual errors per second in your FCAL ?=3D20
>=20
> b) It may happen in some part of the link but overall the link can =3D
> tolerate it
>=20
> In particular,
>=20
>>If over any finite period of time the RJ causes the frequency=3D20
>>as seen by the PLL to change then the PLL will move the VCO, thus=3D20
>>creating a limit to the max observed RJ. If the RJ is =
distributed=3D20
>>so that the frequency does not have to change then the 'single=3D20
>>incremental interval' effect will apply.
>> Have we not then got a jitter distribution that is Gaussian in=3D20
>>form but with limits to the maximum deviations?
>=20
>=20
> I don't know about FCAL jitter tolerance spec but let's take the OC-xx =
=3D
> spec as a reference, what happen when the jitter spectrum passes ft ? =
Is =3D
> the jitter still bounded ?
> To take it out to a non-ideal world, how does one trade off jitter =3D
> tolerance (for example, by Rj and Dj) and jitter accumulation (for =3D
> example, by PLL supply noise) ? You only have one choice of loop =3D
> dynamics in your PLL and you can not help one without hurting the =
other. =3D
> And I would bet jitter accumulation from PLLVDD dwarfs the concern =
over =3D
> jitter tolerance. Just look at how elaborate those PLLVDD distribute =
=3D
> schemes we have (isolation traces, filters etc). Can you still say =
your =3D
> Rj is unconditionally bounded by such PLL's ?
>=20
> c) 10e-12 happens only in methodology, you are really running a BERT =
=3D
> much much lower, possibly beyond the life of the system or the =
component =3D
> or reasonable measurement technique
>=20
> Which, I think, is essentially what you are saying in the last =
sentence. =3D
> Do we really claim a 10e-12 simply because we got an erroneous Rj from =
=3D
> an instrument in a Dj dominated system, multiply the sigma to 14 or 15 =
=3D
> times per the BERT spec and scare ourselves to such error rate ?=3D20
> And before the instrument guys jump on me again, let me be clear that =
I =3D
> don't think the instrument is at fault but rather the problem lies on =
=3D
> the application (Dj dominate system measurement) that force the =3D
> instrument to report back a number that it cranked out based on some =
=3D
> fixed assumptions not suited for such situation.
>=20
> For those who claim a), my hats off to you because you are a braver =
man =3D
> than me to publicly say that.
>=20
> For those who claim b), I sure would like to know what kind of PLL you =
=3D
> have
>=20
> And my money will be on c).
>=20
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx
> [mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of steve weir
> Sent: Tuesday, July 04, 2006 3:40 PM
> To: David Instone
> Cc: si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: Fibre channel interconnect margins
>=20
>=20
> David, I did not regard it as an attack just an opinion that is=3D20
> different and worth discussing.
>=20
> The basis of our disagreement appears to be in the definition of=3D20
> bound. I look at things from the standpoint of electrical=3D20
> noise. Time interval in a timing circuit is the result of the=3D20
> magnitude of some electrical quantity, and is always causal, each=3D20
> event defining a new interval follows the previous. This means =
that=3D20
> noise effectively multiplies the interval by some factor 1/oo <=3D3D =
K=3D20
> <=3D3D oo. Jitter is still unbounded, but every incremental interval =
has=3D20
> a positive duration.
>=20
> So far we have been talking about noise in the oscillator itself.
>=20
> Now, let's see what a PLL does to this quagmire. If noise =
hammers=3D20
> the VCO then the PLL feedback loop applies gain to divide the =
effect=3D20
> of the noise. If we still believe in infinity, infinity divided =
by=3D20
> anything is still infinity. In practice will the oscillator stop =
for=3D20
> an unlimited time? It will only when it fails. On the other =
end,=3D20
> two successive pulses can occur essentially on top of each other.
>=20
> A receiver PLL will take a finite amount of time to realign within =
a=3D20
> fixed amount of phase to the jittered stream for the case of the=3D20
> oscillator event, and will take a different, much longer amount =
of=3D20
> time to align to the short term frequency offset that noise in =
the=3D20
> PLL error amplifier causes. The phase error between the source=3D20
> stream and the recovered clock in the latter case generally follows =
a=3D20
> classic 2nd order step response. The golden PLL is a PLL with=3D20
> specific frequency response and damping. Even if we have a PLL =
that=3D20
> uses N=3D3D1, the PLL only starts correcting after a timing error =
is=3D20
> already apparent. For a timing error of sufficient magnitude =
data=3D20
> moves outside the timing window, a data recovery error is =
guaranteed,=3D20
> and no PLL is going to prevent that. A nasty little problem that=3D20
> gets into systems is power supply noise coupled into the VCO =
and/or=3D20
> error amplifier by one means or another. For systems with high Ns =
it=3D20
> can get really ugly.
>=20
> On a slightly different tack, for a PLL using a PFD, the unit=3D20
> interval is that at the phase comparator input which is VCO/N or=3D20
> Fref. Noise whacking the error amplifier will push the VCO off=3D20
> frequency until new information arrives to get it back. If the =
noise=3D20
> jumps the VCO up it can take up to VCO/N cycles before we start=3D20
> correcting. If noise slows the VCO down, it will take at least =
one=3D20
> cycle of Fref to get it back.
>=20
> So, I think the only place that we are having any semantic trouble =
is=3D20
> on the notion of unbounded noise. While we likely will never see=3D20
> such a thing, the math really does tell us that an interval can =
go=3D20
> virtually to zero 1/oo, or last forever. I think the important =
part=3D20
> of this concept is that it says that random noise ( jitter ) will=3D20
> create data errors sooner or later. And I think doubt about that =
is=3D20
> where the discussion began. The tough issue is finding the =
actual=3D20
> random jitter. The value is often way overestimated because=3D20
> deterministic jittter that we have difficulty correlating gets=3D20
> incorrectly classified as RJ. People turn the crank on the math =
and=3D20
> conclude that their links are 10E-12 or 10E-14 when they are =
really=3D20
> more like 10E-20 from an RJ standpoint.
>=20
> Regards,
>=20
>=20
> Steve.
> At 06:35 AM 7/4/2006, David Instone wrote:
>=20
>>Steve,
>> Firstly, my initial response was in support of Alan's posting =
not=3D20
>>an attack on your reply to him. Your definition follows that of =
FC=3D20
>>and other serial standards. FC defines random jitter in FC-PI-3 as
>>
>>>jitter, random (RJ): Jitter that is characterized by a Gaussian=3D20
>>>distribution. Random jitter is
>>>defined to be the peak-to-peak value for a BER of 10-12, taken =
to=3D20
>>>be approximately 14 times
>>>the standard deviation of the Gaussian distribution.
>>
>>
>> So lets look at it both ways
>>
>>>That means that any single incremental interval can never have=3D20
>>>jitter of more than -(1UI-epsilon).
>>
>>If that jitter is all Gaussian then hasn't it been truncated, or =
do=3D20
>>we have to say that it's not RJ because it's bounded?
>>
>>
>>
>>>If on the other hand we want to integrate phase compared to some=3D20
>>>distant fixed timing reference, then a stream can theoretically=3D20
>>>precess total time interval error by an unbounded amount.
>>
>>FC measures jitter against a timing reference derived from a =
golden=3D20
>>PLL. If over any finite period of time the RJ causes the =
frequency=3D20
>>as seen by the PLL to change then the PLL will move the VCO, thus=3D20
>>creating a limit to the max observed RJ. If the RJ is =
distributed=3D20
>>so that the frequency does not have to change then the 'single=3D20
>>incremental interval' effect will apply.
>> Have we not then got a jitter distribution that is Gaussian in=3D20
>>form but with limits to the maximum deviations?
>>
>>Regards
>>Dave
>>
>>steve weir wrote:
>>
>>>David,
>>>
>>>I would just like to make certain that we are talking along the=3D20
>>>same lines here. The operation of the oscillator, no matter =
what=3D20
>>>its construction is causal. So the closest that any two events =
can=3D20
>>>occur is epsilon. That means that any single incremental =
interval=3D20
>>>can never have jitter of more than -(1UI-epsilon).
>>>
>>>If on the other hand we want to integrate phase compared to some=3D20
>>>distant fixed timing reference, then a stream can theoretically=3D20
>>>precess total time interval error by an unbounded amount.
>>>
>>>Regards,
>>>
>>>
>>>Steve.
>>>At 03:10 AM 7/4/2006, David Instone wrote:
>>>
>>>>Steve,
>>>> I didn't disallow an infinite time between events. I allow =
for=3D20
>>>>the time between events to be between 0 and infinity, but not =3D
>=20
> negative.
>=20
>>>>Thus if I'm measuring the time between edges and my reference I=3D20
>>>>can measure an infinite time between my reference and a =
following=3D20
>>>>edge but never more than 1 UI between the last edge and my =
reference.
>>>>That last edge could of course be from a edge that should have=3D20
>>>>occurred an infinite amount of time in the future, but from the=3D20
>>>>point of view of the measurement it's only 1 UI early.
>>>>Regards
>>>>Dave
>>>>
>>>>
>>>>steve weir wrote:
>>>>
>>>>>David, I disagree. It does not change causality. It changes =
the=3D20
>>>>>incremental delay between two events. Imagine for a moment =
that=3D20
>>>>>we have a simple relaxation oscillator as the basis of our=3D20
>>>>>VCO. In the presence of an infinitely large noise pulse, =
which=3D20
>>>>>is the limit for random noise, it takes an infinite amount of=3D20
>>>>>time for the ramp to reach the threshold. The next cycle will=3D20
>>>>>not begin untilt he current cycle completes. It may sound =
like=3D20
>>>>>something from Douglas Adams, but it really is mathematically =
and=3D20
>>>>>physically sound.
>>>>>
>>>>>Regards,
>>>>>
>>>>>Steve.
>>>>>At 01:50 AM 7/4/2006, David Instone wrote:
>>>>>
>>>>>>Because it makes for a nice simple clean definition. However, =
I=3D20
>>>>>>believe you have to take the real world into consideration.
>>>>>>Allowing the RJ to be really unbounded means that each edge in =
a=3D20
>>>>>>bit stream could be advanced or delayed by an infinite amount.
>>>>>>Taken to extremes this means that the order of edges could be =
=3D
>=20
> reversed.
>=20
>>>>>>This is obviously absurd, the measured time between edges can=3D20
>>>>>>reduce until it is zero, it cannot go negative. The time=3D20
>>>>>>between edges can of course go to +ve infinity, but that isn't =
a=3D20
>>>>>>bit error, the system has failed or been switched off.
>>>>>>steve weir wrote:
>>>>>>
>>>>>>>RJ really is unbounded by definition.
>>>>>>>
>>>>>>>Steve.
>>>>>>>At 09:46 AM 7/3/2006, Steven Kan wrote:
>>>>>>>
>>>>>>>
>>>>>>>>>Date: Fri, 30 Jun 2006 21:48:56 -0700
>>>>>>>>>From: Alan.Hiltonnickel@xxxxxxx
>>>>>>>>>Subject: [SI-LIST] Re: Fibre channel interconnect margins
>>>>>>>>>
>>>>>>>>>In fact, I think that companies DO ship products that toss a =
=3D
>=20
> random
>=20
>>>>>>>>>error approximately every 10e-xx or so. Why? Because the =3D
>=20
> statistical
>=20
>>>>>>>>>theory behind those errors is that random/Gaussian noise is, by
>>>>>>>>>definition, unbounded - errors are a fact of life, even if the =
=3D
>=20
> error
>=20
>>>>>>>>>rate is very low.
>>>>>>>>
>>>>>>>>I suppose we're way off in the weeds, here, but is the noise =3D
>=20
> actually
>=20
>>>>>>>>unbounded? Or does it just behave in a Gaussian-like manner=3D20
>>>>>>>>within the realm
>>>>>>>>of times/rates that matter for shipping product? I suppose =
if=3D20
>>>>>>>>I sat in my
>>>>>>>>chair for long enough, a truly unbounded system might cause =
a=3D20
>>>>>>>>gold bar to
>>>>>>>>pop into existence on my desk, but my empirical GBR (gold-bar =
=3D
>=20
> rate) is
>=20
>>>>>>>>currently 0.
>>>>>>>>
>>>>>>>>-----------------------------------------------------------------=
-=3D
>=20
>=20
>>
>=20
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>=20
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>=20
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>=20
> List FAQ wiki page is located at:
> http://si-list.org/wiki/wiki.pl?Si-List_FAQ
>=20
> List technical documents are available at:
> http://www.si-list.org
>=20
> List archives are viewable at: =20
> //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
> =20
>=20
>=20
>=20
------------------------------------------------------------------
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 FAQ wiki page is located at:
http://si-list.org/wiki/wiki.pl?Si-List_FAQ
List technical documents are available at:
http://www.si-list.org
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:
