[SI-LIST] Re: HSPICE - adding jitter to ethernet serial link

  • From: Scott McMorrow <scott@xxxxxxxxxxxxx>
  • To: hassan@xxxxxxxx
  • Date: Thu, 16 Feb 2006 12:16:45 -0500

Hassan,
I'd like to correct your understanding, and that of many on this list.  
It is a practical and theoretical impossibility to include Rj in a PWL 
stimulus.  By definition, a PWL stimulus is deterministic, and Rj is 
non-deterministic.  It is not possible to approximate a 
non-deterministic system with deterministic functions.  
Non-deterministic Rj is unbounded, which means that there are no 
statistical bounds on it's extent over time.  It must be treated with 
stochastic (statistical) methods. In real systems, with gaussian Rj, 
there is absolutely no bound on maximum total jitter. A statistical 
probability exists that you will measure any amount of jitter.  The 
question becomes "what is the probability of a particular amount of 
jitter occurring?"  This is normally stated as a Bit Error Rate. 

10 ^-12 BER is often a common measure used in compliance specifications, 
and is at approximately the +/- 7 sigma confidence interval. 
Essentially, to achieve 10^-12 BER confidence, the 1 sigma Rj value must 
be multiplied by 14.  A quick calculation on a 3.125 Gbps backplane 
channel, shows that 10^-12 BER would result in one error per link every 
320 seconds.  If we look at a backplane 100 total links, we would expect 
to see a link error every 3.2 seconds.  Clearly if this were the case, 
those of us who design backplane based systems would be out of a job, 
since this error logging rate would be unacceptable. Chris Cheng made 
this point about half a year ago.

Practically, we would like to have error free links.  For this to occur, 
we need a significantly high enough BER to preclude errors within the 
operational lifetime of this system.  So, lets say that we have the same 
backplane and need to achieve an error rate of one error per 10 years 
per link, or 3.15*10^8 seconds. At 3.125 Gbps, that constitutes 
9.855*10^17 total bits/link, for a BER of 1e^-18.  The statistical 
confidence interval for this is a little bit larger than +/- 9 sigma.

Knowing the 1 sigma Rj value of the transmitter is extremely important 
to complete a back-of-the-napkin calculation.  If we start with zero 
deterministic jitter, then total jitter over a 10 year sampling interval 
would be dependent solely on the Rj of the transmitter, as follows

Rj    Total Jitter @ 10^-18 confidence

500 fS           9 ps
1 ps            18 ps
2 ps            36 ps

In actuality, measured jitter is the linear superposition of Rj and all 
sources of Dj, and should be lower than the above numbers.  Essentially 
the trajectory of every potential edge transition is "smeared" by the 
probability density function of Rj, on an edge by edge basis, and 
overlapped in time to create the composite probability density 
function.  The summation of these results is the total probability 
density function.  However, since I am a practical engineer, and leave 
the mathematics to Ransom and Al, there is a quick worst case bounded 
calculation that one can make.  If we assume that all of the Dj 
trajectories are worst case, that is the worst case jitter pattern 
repeats over and over again, then we can use simulations (or 
measurements) of the worst case Dj as a basis for a worst case total 
jitter calculation.

Absolute worst case jitter is therefore.  Dj + Rj @ 10^-18 BER.  Using 
our little table above, at 2 ps Rj, we get

Absolute worst case jitter @ 10^-18 BER = 36 ps + Dj(measured or simulated)

This method will be highly conservative and should over predict the 
actual worst case jitter over time, since we are doing a simple addition 
and not a convolution of probability density functions. 

However, we can make some simple observations.  Given a particular Rj 
and BER target, we can compute a worst case eye jitter UI impairment.  
Using the same table as above, we have

Absolute worst case Unit Interval (UI) impariment @ 10^-18 BER
Rj        10^-18 jitter  3.125Gbps 6.26Gbps 10Gbps

500 fS      9 ps          .03 UI   .06UI    .09 UI
1 ps       18 ps          .06 UI   .11UI    .18 UI
2 ps       36 ps          .11 UI   .23UI    .36 UI

Clearly, as we approach faster bit rates, and a shorter UI, knowing the 
value of Rj is extremely important.  Our contention at Teraspeed 
Consulting is that accurate measurement of Rj is the key to 
understanding the limits of high bandwidth channel performance.



Regards,

Scott


Scott McMorrow
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



Hassan O. Ali wrote:
> Steve,
>
> So that we're all on the same page, let me summarize what I've said so far:
>
> - In an HSPICE simulation you can inject into a link a PWL stimulus that 
> includes source 
> RJ and DJ, and obtain at the end of your link, a time-domain output 
> eye-diagram that 
> includes the effects of source RJ and DJ as well as the DJ introduced by the 
> link. I've 
> not suggested that such a simulation will give you a BER number.
>
> - With a known data pattern, and given BER requirements, I think it is 
> possible to 
> obtain the worst-case output eye-pattern (i.e. the worst-case eye-closure) 
> through 
> HSPICE simulation with just a few bits - i.e. not through days of simulation. 
> I 
> suggested that the simulation may need strategic assignment of maximum RJ on 
> the bit 
> stream. Of course that needs to be confirmed. 
>
> Best regards.
>
> Hassan.
>  
>
> On Feb 15, steve weir <weirsi@xxxxxxxxxx> wrote:
>   
>> Hassan, using a PN pattern to emulate RJ doesn't make sense to 
>> me.  In my perverse universe, lots of post processing is needed to 
>> convert an HSPICE result into an estimate of BER, most notably an 
>> accounting for the effects of RJ. Given the paradox that we desire 
>> both:  near zero BER, and accurate BER values, I think it is hard 
>> enough to get valid probability stats on systems running full 
>> speed.  In simulation, I see it as impossible outside of 
>> post-processing.  And yes, accurate RJ numbers are critical to doing 
>> that post-processing reasonably well.
>>
>> Are you aware of an accurate way to get BERs based on the 
>> probabilistic effects without post processing?  If you do, can you 
>> share a description?
>>
>> Regards,
>>
>>
>> Steve.
>> At 06:00 PM 2/15/2006, Hassan O. Ali wrote:
>>     
>>> Chris,
>>>
>>> There are so many ways to skin a cat!
>>>
>>> High-speed chip makers do specify Tx RJ and DJ (jitter generation) 
>>> numbers and also Rx
>>> jitter tolerance numbers. As a matter of fact, various high-speed 
>>> interface standards
>>> call for the specification of those jitter characteristics. System 
>>> designers use those
>>> numbers while evaluating the performance over their links.
>>>
>>> The jitter tolerance number is a measure of the Rx performance. An 
>>> HSPICE simulation may
>>> not easily give you a BER number but it gives you an eye-pattern at 
>>> the end of the link
>>> which you can use to determine if it meets the far-end eye-pattern 
>>> expected by the
>>> receiver.
>>>
>>> Best regards.
>>>
>>> Hassan.
>>>
>>>
>>> On Feb 15, "Chris Cheng" <Chris.Cheng@xxxxxxxxxxxx> wrote:
>>>       
>>>> I am a firm believer of divide and conquer. The same reason you don't =
>>>> simulate your Tx/Rx PLL jitter while you are doing your prop delay =
>>>> analysis. They are separate components of a BERT budget. You are asking =
>>>> for trouble and unnecessary complexity if you try to do it all in the =
>>>> same time in SPICE or IBIS.
>>>> Besides, who really knows what his Tx Rj is anyways ?
>>>> Further more, without knowing your CDR jitter transfer, how do you know =
>>>> the BERT is really that bad AFTER your CDR recover the clock that is =
>>>> used to capture your data ?
>>>>
>>>> -----Original Message-----
>>>> From: si-list-bounce@xxxxxxxxxxxxx
>>>> [mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of Hassan O. Ali
>>>> Sent: Wednesday, February 15, 2006 3:05 PM
>>>> To: si-list@xxxxxxxxxxxxx
>>>> Subject: [SI-LIST] Re: HSPICE - adding jitter to ethernet serial link
>>>>
>>>>
>>>>
>>>> I don't know how the two methods would compare. What do you think?
>>>>
>>>> Best regards.
>>>>
>>>> Hassan.
>>>>
>>>>
>>>> On Feb 15, "Chris Cheng" <Chris.Cheng@xxxxxxxxxxxx> wrote:
>>>>         
>>>>> =20
>>>>> If the channel is indeed linear, what's the difference between not =
>>>>>           
>>>> doing =3D
>>>>         
>>>>> Rj at all and simulate your PRBS, get the minium eye opening (Dj) then =
>>>>>           
>>>> =3D
>>>>         
>>>>> add a seperate Rj convoluted with impulse response ?=3D20
>>>>> =20
>>>>> -----Original Message-----
>>>>> From: si-list-bounce@xxxxxxxxxxxxx
>>>>> [mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of Hassan O. Ali
>>>>> Sent: Wednesday, February 15, 2006 2:18 PM
>>>>> To: si-list@xxxxxxxxxxxxx
>>>>> Subject: [SI-LIST] Re: HSPICE - adding jitter to ethernet serial link
>>>>> =20
>>>>> =20
>>>>> Chris,
>>>>> =20
>>>>> The HSPICE approach allows you to see the effect of source RJ as well =
>>>>>           
>>>> as =3D
>>>>         
>>>>> both source and=3D20
>>>>> channel-induced DJ. I think the convolution you're referring to =
>>>>>           
>>>> doesn't =3D
>>>>         
>>>>> consider source=3D20
>>>>> DJ.
>>>>> =20
>>>>> I'm thinking out loud here, but I think in simulation, if the input =
>>>>>           
>>>> =3D
>>>>         
>>>>> pattern is known, it=3D20
>>>>> doesn't need many bits to simulate a case of worst-case jitter (or =3D
>>>>> eye-opening) for a=3D20
>>>>> particular BER requirement.=3D20
>>>>> =20
>>>>> For example, for a BER of 10E-14, the max pk-to-pk RJ is approx. 15.3 =
>>>>>           
>>>> =3D
>>>>         
>>>>> times RJ(rms). To=3D20
>>>>> simulate the worst-case total jitter, In the PWL source you could use =
>>>>>           
>>>> =3D
>>>>         
>>>>> that amount of max=3D20
>>>>> pk-to-pk RJ right at the bit where you expect the worst-case ISI e.g. =
>>>>>           
>>>> at =3D
>>>>         
>>>>> a lone "1" that=3D20
>>>>> follows a long run of 0's. In other words, I think you can =
>>>>>           
>>>> strategically =3D
>>>>         
>>>>> define source=3D20
>>>>> RJ and DJ amounts along your bit stream and cover the entire =
>>>>>           
>>>> eye-pattern =3D
>>>>         
>>>>> voltage/time=3D20
>>>>> bounds in just a few bits.  Anyone out there has done this?
>>>>> =20
>>>>> Best regards.
>>>>> =20
>>>>> Hassan.
>>>>> =20
>>>>> =20
>>>>> =20
>>>>> On Feb 15, "Chris Cheng" <Chris.Cheng@xxxxxxxxxxxx> wrote:
>>>>>           
>>>>>> =3D20
>>>>>> If your channel response is linear. How does a separate analysis =
>>>>>>             
>>>> with =3D
>>>>         
>>>>> Rj only=3D20
>>>>> convoluted with the impulse response of the channel different from =
>>>>>           
>>>> what =3D
>>>>         
>>>>> you are doing ?=3D20
>>>>> How many bit do you think is need to simulate in SPICE a BERT of =
>>>>>           
>>>> 10e-14 =3D
>>>>         
>>>>> or -18 ?
>>>>>           
>>>>>> =3D20
>>>>>> ________________________________
>>>>>> =3D20
>>>>>> From: si-list-bounce@xxxxxxxxxxxxx on behalf of Hassan O. Ali
>>>>>> Sent: Wed 2/15/2006 9:02 AM
>>>>>> To: si-list@xxxxxxxxxxxxx
>>>>>> Subject: [SI-LIST] Re: HSPICE - adding jitter to ethernet serial =
>>>>>>             
>>>> link
>>>>         
>>>>>> =3D20
>>>>>> =3D20
>>>>>> =3D20
>>>>>> A piece-wise linear (PWL) source implemented as below in an HSPICE =
>>>>>>             
>>>> =3D
>>>>         
>>>>> subcircuit may be
>>>>>           
>>>>>> used to simulate RJ and DJ effects on the output.
>>>>>> =3D20
>>>>>> The total jitter (J) factors (e.g. 0.512, 0.221, etc.) in this case =
>>>>>>             
>>>> =3D
>>>>         
>>>>> were generated with
>>>>>           
>>>>>> a random number generator. The bit pattern is random but is supposed =
>>>>>>             
>>>> =3D
>>>>         
>>>>> to mimic an 8B/10B
>>>>>           
>>>>>> encoding where the maximum run-length (maximum number of consecutive =
>>>>>>             
>>>> =3D
>>>>         
>>>>> 1's or 0's) is 5.
>>>>>           
>>>>>> The pattern length was arbitrarily chosen as 160 bits.
>>>>>> =3D20
>>>>>> Mind you, this is just one arbitrary way of including DJ and RJ =3D
>>>>>>             
>>>>> effects. I'm sure more
>>>>>           
>>>>>> elaborate ways exist.
>>>>>> =3D20
>>>>>> Best regards.
>>>>>> =3D20
>>>>>> Hassan.
>>>>>> =3D20
>>>>>> =3D20
>>>>>> ************************************************
>>>>>> * HSPICE PWL voltage source
>>>>>> * Parameter description:
>>>>>> * TR - Rise/Fall time of the PWL source
>>>>>> * UI - Unit Interval time (Bit width =3D3D 1/BitRate)
>>>>>> * HI - Voltage value representing bit "1"
>>>>>> * LO - Voltage value representing bit "0"
>>>>>> * JD - Peak-to-peak deterministic jitter
>>>>>> * JR - Peak-to-peak random jitter
>>>>>> ************************************************
>>>>>> .SUBCKT pwl_source node1 node2
>>>>>> +TR=3D3DTRval UI=3D3DUIval HI=3D3DHIval LO=3D3DLOval
>>>>>> +JD=3D3DJDval JR=3D3DJRval
>>>>>> .param J=3D3D'(JD+JR/2)'
>>>>>> vsrc node1 node2 PWL
>>>>>> + 0.0    'HI', '0.512*J+UI*1'    'HI',
>>>>>> + '0.512*J+UI*1+TR'   'LO', '0.221*J+UI*2'    'LO',
>>>>>> + '0.221*J+UI*2+TR'   'LO', '0.868*J+UI*3'    'LO',
>>>>>> + '0.868*J+UI*3+TR'   'LO', '0.358*J+UI*4'    'LO',
>>>>>> + '0.358*J+UI*4+TR'   'LO', '0.379*J+UI*5'    'LO',
>>>>>> + '0.379*J+UI*5+TR'   'HI', '-0.003*J+UI*6'    'HI',
>>>>>> + '-0.003*J+UI*6+TR'   'LO', '0.2*J+UI*7'    'LO',
>>>>>> + '0.2*J+UI*7+TR'   'LO', '0.238*J+UI*8'    'LO',
>>>>>> + '0.238*J+UI*8+TR'   'LO', '0.197*J+UI*9'    'LO',
>>>>>> + '0.197*J+UI*9+TR'   'LO', '-0.424*J+UI*10'    'LO',
>>>>>> + '-0.424*J+UI*10+TR'   'HI', '-0.608*J+UI*11'    'HI',
>>>>>> + '-0.608*J+UI*11+TR'   'HI', '-0.47*J+UI*12'    'HI',
>>>>>> + '-0.47*J+UI*12+TR'   'LO', '-0.124*J+UI*13'    'LO',
>>>>>> + '-0.124*J+UI*13+TR'   'LO', '-0.094*J+UI*14'    'LO',
>>>>>> + '-0.094*J+UI*14+TR'   'LO', '-0.657*J+UI*15'    'LO',
>>>>>> + '-0.657*J+UI*15+TR'   'HI', '0.48*J+UI*16'    'HI',
>>>>>> + '0.48*J+UI*16+TR'   'LO', '-0.052*J+UI*17'    'LO',
>>>>>> + '-0.052*J+UI*17+TR'   'LO', '0.208*J+UI*18'    'LO',
>>>>>> + '0.208*J+UI*18+TR'   'HI', '-0.15*J+UI*19'    'HI',
>>>>>> + '-0.15*J+UI*19+TR'   'HI', '-0.097*J+UI*20'    'HI',
>>>>>> + '-0.097*J+UI*20+TR'   'LO', '0.089*J+UI*21'    'LO',
>>>>>> + '0.089*J+UI*21+TR'   'LO', '-0.535*J+UI*22'    'LO',
>>>>>> + '-0.535*J+UI*22+TR'   'LO', '0.096*J+UI*23'    'LO',
>>>>>> + '0.096*J+UI*23+TR'   'LO', '-0.433*J+UI*24'    'LO',
>>>>>> + '-0.433*J+UI*24+TR'   'LO', '0.603*J+UI*25'    'LO',
>>>>>> + '0.603*J+UI*25+TR'   'LO', '-0.119*J+UI*26'    'LO',
>>>>>> + '-0.119*J+UI*26+TR'   'HI', '-0.847*J+UI*27'    'HI',
>>>>>> + '-0.847*J+UI*27+TR'   'HI', '-0.234*J+UI*28'    'HI',
>>>>>> + '-0.234*J+UI*28+TR'   'HI', '-0.056*J+UI*29'    'HI',
>>>>>> + '-0.056*J+UI*29+TR'   'LO', '-0.535*J+UI*30'    'LO',
>>>>>> + '-0.535*J+UI*30+TR'   'LO', '-0.564*J+UI*31'    'LO',
>>>>>> + '-0.564*J+UI*31+TR'   'HI', '-0.191*J+UI*32'    'HI',
>>>>>> + '-0.191*J+UI*32+TR'   'LO', '0.16*J+UI*33'    'LO',
>>>>>> + '0.16*J+UI*33+TR'   'LO', '-0.552*J+UI*34'    'LO',
>>>>>> + '-0.552*J+UI*34+TR'   'LO', '0.277*J+UI*35'    'LO',
>>>>>> + '0.277*J+UI*35+TR'   'HI', '-0.491*J+UI*36'    'HI',
>>>>>> + '-0.491*J+UI*36+TR'   'LO', '0.463*J+UI*37'    'LO',
>>>>>> + '0.463*J+UI*37+TR'   'LO', '-0.166*J+UI*38'    'LO',
>>>>>> + '-0.166*J+UI*38+TR'   'HI', '-0.592*J+UI*39'    'HI',
>>>>>> + '-0.592*J+UI*39+TR'   'HI', '0.667*J+UI*40'    'HI',
>>>>>> + '0.667*J+UI*40+TR'   'LO', '-0.296*J+UI*41'    'LO',
>>>>>> + '-0.296*J+UI*41+TR'   'LO', '-0.11*J+UI*42'    'LO',
>>>>>> + '-0.11*J+UI*42+TR'   'HI', '-0.608*J+UI*43'    'HI',
>>>>>> + '-0.608*J+UI*43+TR'   'HI', '-0.351*J+UI*44'    'HI',
>>>>>> + '-0.351*J+UI*44+TR'   'LO', '0.818*J+UI*45'    'LO',
>>>>>> + '0.818*J+UI*45+TR'   'HI', '-0.365*J+UI*46'    'HI',
>>>>>> + '-0.365*J+UI*46+TR'   'LO', '0.201*J+UI*47'    'LO',
>>>>>> + '0.201*J+UI*47+TR'   'LO', '-0.247*J+UI*48'    'LO',
>>>>>> + '-0.247*J+UI*48+TR'   'LO', '0.446*J+UI*49'    'LO',
>>>>>> + '0.446*J+UI*49+TR'   'HI', '0.329*J+UI*50'    'HI',
>>>>>> + '0.329*J+UI*50+TR'   'LO', '0.706*J+UI*51'    'LO',
>>>>>> + '0.706*J+UI*51+TR'   'HI', '-0.165*J+UI*52'    'HI',
>>>>>> + '-0.165*J+UI*52+TR'   'LO', '0.302*J+UI*53'    'LO',
>>>>>> + '0.302*J+UI*53+TR'   'LO', '0.03*J+UI*54'    'LO',
>>>>>> + '0.03*J+UI*54+TR'   'HI', '0.152*J+UI*55'    'HI',
>>>>>> + '0.152*J+UI*55+TR'   'LO', '-0.255*J+UI*56'    'LO',
>>>>>> + '-0.255*J+UI*56+TR'   'HI', '-0.237*J+UI*57'    'HI',
>>>>>> + '-0.237*J+UI*57+TR'   'LO', '0.014*J+UI*58'    'LO',
>>>>>> + '0.014*J+UI*58+TR'   'LO', '0.64*J+UI*59'    'LO',
>>>>>> + '0.64*J+UI*59+TR'   'LO', '-0.162*J+UI*60'    'LO',
>>>>>> + '-0.162*J+UI*60+TR'   'LO', '0.311*J+UI*61'    'LO',
>>>>>> + '0.311*J+UI*61+TR'   'LO', '0.33*J+UI*62'    'LO',
>>>>>> + '0.33*J+UI*62+TR'   'LO', '-0.301*J+UI*63'    'LO',
>>>>>> + '-0.301*J+UI*63+TR'   'HI', '-0.513*J+UI*64'    'HI',
>>>>>> + '-0.513*J+UI*64+TR'   'LO', '0.102*J+UI*65'    'LO',
>>>>>> + '0.102*J+UI*65+TR'   'LO', '-0.695*J+UI*66'    'LO',
>>>>>> + '-0.695*J+UI*66+TR'   'LO', '-0.308*J+UI*67'    'LO',
>>>>>> + '-0.308*J+UI*67+TR'   'LO', '-0.444*J+UI*68'    'LO',
>>>>>> + '-0.444*J+UI*68+TR'   'LO', '0.387*J+UI*69'    'LO',
>>>>>> + '0.387*J+UI*69+TR'   'LO', '-0.031*J+UI*70'    'LO',
>>>>>> + '-0.031*J+UI*70+TR'   'HI', '-0.599*J+UI*71'    'HI',
>>>>>> + '-0.599*J+UI*71+TR'   'LO', '-0.071*J+UI*72'    'LO',
>>>>>> + '-0.071*J+UI*72+TR'   'LO', '0.077*J+UI*73'    'LO',
>>>>>> + '0.077*J+UI*73+TR'   'HI', '0.663*J+UI*74'    'HI',
>>>>>> + '0.663*J+UI*74+TR'   'HI', '-0.526*J+UI*75'    'HI',
>>>>>> + '-0.526*J+UI*75+TR'   'LO', '0*J+UI*76'    'LO',
>>>>>> + '0*J+UI*76+TR'   'LO', '0.193*J+UI*77'    'LO',
>>>>>> + '0.193*J+UI*77+TR'   'LO', '0.479*J+UI*78'    'LO',
>>>>>> + '0.479*J+UI*78+TR'   'LO', '-0.055*J+UI*79'    'LO',
>>>>>> + '-0.055*J+UI*79+TR'   'HI', '0.161*J+UI*80'    'HI',
>>>>>> + '0.161*J+UI*80+TR'   'LO', '0.404*J+UI*81'    'LO',
>>>>>> + '0.404*J+UI*81+TR'   'LO', '-0.249*J+UI*82'    'LO',
>>>>>> + '-0.249*J+UI*82+TR'   'HI', '-0.255*J+UI*83'    'HI',
>>>>>> + '-0.255*J+UI*83+TR'   'LO', '-0.668*J+UI*84'    'LO',
>>>>>> + '-0.668*J+UI*84+TR'   'HI', '-0.481*J+UI*85'    'HI',
>>>>>> + '-0.481*J+UI*85+TR'   'LO', '0.553*J+UI*86'    'LO',
>>>>>> + '0.553*J+UI*86+TR'   'LO', '0.16*J+UI*87'    'LO',
>>>>>> + '0.16*J+UI*87+TR'   'HI', '0.392*J+UI*88'    'HI',
>>>>>> + '0.392*J+UI*88+TR'   'LO', '-0.056*J+UI*89'    'LO',
>>>>>> + '-0.056*J+UI*89+TR'   'LO', '0.267*J+UI*90'    'LO',
>>>>>> + '0.267*J+UI*90+TR'   'LO', '0.384*J+UI*91'    'LO',
>>>>>> + '0.384*J+UI*91+TR'   'LO', '0.79*J+UI*92'    'LO',
>>>>>> + '0.79*J+UI*92+TR'   'LO', '-0.734*J+UI*93'    'LO',
>>>>>> + '-0.734*J+UI*93+TR'   'LO', '-0.006*J+UI*94'    'LO',
>>>>>> + '-0.006*J+UI*94+TR'   'HI', '-0.01*J+UI*95'    'HI',
>>>>>> + '-0.01*J+UI*95+TR'   'HI', '-0.039*J+UI*96'    'HI',
>>>>>> + '-0.039*J+UI*96+TR'   'LO', '0.648*J+UI*97'    'LO',
>>>>>> + '0.648*J+UI*97+TR'   'HI', '0.099*J+UI*98'    'HI',
>>>>>> + '0.099*J+UI*98+TR'   'LO', '0.409*J+UI*99'    'LO',
>>>>>> + '0.409*J+UI*99+TR'   'HI', '0.808*J+UI*100'    'HI',
>>>>>> + '0.808*J+UI*100+TR'   'HI', '0.458*J+UI*101'    'HI',
>>>>>> + '0.458*J+UI*101+TR'   'LO', '0.709*J+UI*102'    'LO',
>>>>>> + '0.709*J+UI*102+TR'   'LO', '0.489*J+UI*103'    'LO',
>>>>>> + '0.489*J+UI*103+TR'   'HI', '0.309*J+UI*104'    'HI',
>>>>>> + '0.309*J+UI*104+TR'   'LO', '0.637*J+UI*105'    'LO',
>>>>>> + '0.637*J+UI*105+TR'   'LO', '0.225*J+UI*106'    'LO',
>>>>>> + '0.225*J+UI*106+TR'   'LO', '-0.052*J+UI*107'    'LO',
>>>>>> + '-0.052*J+UI*107+TR'   'HI', '-0.1*J+UI*108'    'HI',
>>>>>> + '-0.1*J+UI*108+TR'   'HI', '0.288*J+UI*109'    'HI',
>>>>>> + '0.288*J+UI*109+TR'   'LO', '-0.219*J+UI*110'    'LO',
>>>>>> + '-0.219*J+UI*110+TR'   'LO', '-0.129*J+UI*111'    'LO',
>>>>>> + '-0.129*J+UI*111+TR'   'LO', '0.035*J+UI*112'    'LO',
>>>>>> + '0.035*J+UI*112+TR'   'LO', '0.131*J+UI*113'    'LO',
>>>>>> + '0.131*J+UI*113+TR'   'HI', '-0.23*J+UI*114'    'HI',
>>>>>> + '-0.23*J+UI*114+TR'   'LO', '0.24*J+UI*115'    'LO',
>>>>>> + '0.24*J+UI*115+TR'   'LO', '0.684*J+UI*116'    'LO',
>>>>>> + '0.684*J+UI*116+TR'   'LO', '0.027*J+UI*117'    'LO',
>>>>>> + '0.027*J+UI*117+TR'   'HI', '0.351*J+UI*118'    'HI',
>>>>>> + '0.351*J+UI*118+TR'   'LO', '-0.562*J+UI*119'    'LO',
>>>>>> + '-0.562*J+UI*119+TR'   'LO', '0.831*J+UI*120'    'LO',
>>>>>> + '0.831*J+UI*120+TR'   'LO', '-0.524*J+UI*121'    'LO',
>>>>>> + '-0.524*J+UI*121+TR'   'HI', '-0.041*J+UI*122'    'HI',
>>>>>> + '-0.041*J+UI*122+TR'   'HI', '0.675*J+UI*123'    'HI',
>>>>>> + '0.675*J+UI*123+TR'   'HI', '-0.035*J+UI*124'    'HI',
>>>>>> + '-0.035*J+UI*124+TR'   'HI', '0.083*J+UI*125'    'HI',
>>>>>> + '0.083*J+UI*125+TR'   'LO', '-0.247*J+UI*126'    'LO',
>>>>>> + '-0.247*J+UI*126+TR'   'HI', '-0.372*J+UI*127'    'HI',
>>>>>> + '-0.372*J+UI*127+TR'   'LO', '-0.44*J+UI*128'    'LO',
>>>>>> + '-0.44*J+UI*128+TR'   'HI', '-0.861*J+UI*129'    'HI',
>>>>>> + '-0.861*J+UI*129+TR'   'LO', '0.162*J+UI*130'    'LO',
>>>>>> + '0.162*J+UI*130+TR'   'HI', '-0.228*J+UI*131'    'HI',
>>>>>> + '-0.228*J+UI*131+TR'   'HI', '0.606*J+UI*132'    'HI',
>>>>>> + '0.606*J+UI*132+TR'   'HI', '-0.002*J+UI*133'    'HI',
>>>>>> + '-0.002*J+UI*133+TR'   'LO', '0.385*J+UI*134'    'LO',
>>>>>> + '0.385*J+UI*134+TR'   'HI', '0.012*J+UI*135'    'HI',
>>>>>> + '0.012*J+UI*135+TR'   'LO', '0.14*J+UI*136'    'LO',
>>>>>> + '0.14*J+UI*136+TR'   'LO', '-0.286*J+UI*137'    'LO',
>>>>>> + '-0.286*J+UI*137+TR'   'HI', '-0.124*J+UI*138'    'HI',
>>>>>> + '-0.124*J+UI*138+TR'   'HI', '0.132*J+UI*139'    'HI',
>>>>>> + '0.132*J+UI*139+TR'   'HI', '0.65*J+UI*140'    'HI',
>>>>>> + '0.65*J+UI*140+TR'   'LO', '0.422*J+UI*141'    'LO',
>>>>>> + '0.422*J+UI*141+TR'   'HI', '-0.173*J+UI*142'    'HI',
>>>>>> + '-0.173*J+UI*142+TR'   'HI', '-0.701*J+UI*143'    'HI',
>>>>>> + '-0.701*J+UI*143+TR'   'HI', '0.674*J+UI*144'    'HI',
>>>>>> + '0.674*J+UI*144+TR'   'LO', '-0.063*J+UI*145'    'LO',
>>>>>> + '-0.063*J+UI*145+TR'   'LO', '0.697*J+UI*146'    'LO',
>>>>>> + '0.697*J+UI*146+TR'   'HI', '-0.464*J+UI*147'    'HI',
>>>>>> + '-0.464*J+UI*147+TR'   'HI', '0.241*J+UI*148'    'HI',
>>>>>> + '0.241*J+UI*148+TR'   'LO', '0.158*J+UI*149'    'LO',
>>>>>> + '0.158*J+UI*149+TR'   'HI', '0.725*J+UI*150'    'HI',
>>>>>> + '0.725*J+UI*150+TR'   'LO', '-0.715*J+UI*151'    'LO',
>>>>>> + '-0.715*J+UI*151+TR'   'HI', '0.379*J+UI*152'    'HI',
>>>>>> + '0.379*J+UI*152+TR'   'HI', '-0.341*J+UI*153'    'HI',
>>>>>> + '-0.341*J+UI*153+TR'   'LO', '0.387*J+UI*154'    'LO',
>>>>>> + '0.387*J+UI*154+TR'   'HI', '-0.158*J+UI*155'    'HI',
>>>>>> + '-0.158*J+UI*155+TR'   'LO', '0.637*J+UI*156'    'LO',
>>>>>> + '0.637*J+UI*156+TR'   'LO', '-0.883*J+UI*157'    'LO',
>>>>>> + '-0.883*J+UI*157+TR'   'LO', '-0.083*J+UI*158'    'LO',
>>>>>> + '-0.083*J+UI*158+TR'   'HI', '-0.087*J+UI*159'    'HI',
>>>>>> + '-0.087*J+UI*159+TR'   'HI', '0.573*J+UI*160'    'HI',
>>>>>> .ends
>>>>>> =3D20
>>>>>> =3D20
>>>>>> =3D20
>>>>>> =3D20
>>>>>> =3D20
>>>>>> On Feb 15, "Chris Cheng" <Chris.Cheng@xxxxxxxxxxxx> wrote:
>>>>>>             
>>>>>>> May be its just me but Rj only makes sense when you are trying to =
>>>>>>>               
>>>> =3D
>>>>         
>>>>> extract a BERT=3D20
>>>>> curve
>>>>>           
>>>>>> which seems to be a MATLAB job.
>>>>>>             
>>>>>>> If you have your Tx Rj, channel model and jitter transfer of your =
>>>>>>>               
>>>> =3D
>>>>         
>>>>> CDR, the rest of=3D20
>>>>> the
>>>>>           
>>>>>> work is straight forward.
>>>>>>             
>>>>>>>> At 10:40 AM 2/7/2006 -0800, Ali Burney wrote:
>>>>>>>> =3D20
>>>>>>>>                 
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> I am running gigabit Ethernet link simulation using HSPICE and =
>>>>>>>>>                   
>>>> =3D
>>>>         
>>>>> have a=3D3D20
>>>>>           
>>>>>>>>> question on adding jitter to the transmitted data. Is there any =
>>>>>>>>>                   
>>>> =3D
>>>>         
>>>>> way to=3D3D20
>>>>>           
>>>>>>>>> add Gaussian jitter to the transmitted data?  I know that tools =
>>>>>>>>>                   
>>>> =3D
>>>>         
>>>>> such as
>>>>>           
>>>>>>>>>  =3D20
>>>>>>>>>                   
>>>>>>>> =3D20
>>>>>>>>                 
>>>>>>>>> Hyperlynx allow you to add transmit jitter.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Thanks and best regards,
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> A Burney
>>>>>>>>>                   
>>> ------------------------------------------------------------------
>>> 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:
>>> <a href='//www.freelists.org/webpage/si-
>>>       
> list'>//www.freelists.org/webpage/si-list</a>
>   
>>> For help:
>>> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>>>
>>> List FAQ wiki page is located at:
>>>                 <a 
>>> href='http://si-list.org/wiki/wiki.pl?Si-List_FAQ'>http://si-
>>>       
> list.org/wiki/wiki.pl?Si-List_FAQ</a>
>   
>>> List technical documents are available at:
>>>                 <a href='http://www.si-list.org'>http://www.si-list.org</a>
>>>
>>> List archives are viewable at:
>>>                 <a href='//www.freelists.org/archives/si-
>>>       
> list'>//www.freelists.org/archives/si-list</a>
>   
>>> or at our remote archives:
>>>                 <a href='http://groups.yahoo.com/group/si-
>>>       
> list/messages'>http://groups.yahoo.com/group/si-list/messages</a>
>   
>>> Old (prior to June 6, 2001) list archives are viewable at:
>>>                 <a 
>>> href='http://www.qsl.net/wb6tpu'>http://www.qsl.net/wb6tpu</a>
>>>
>>>       
>> ------------------------------------------------------------------
>> 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:
>> <a 
>> href='//www.freelists.org/webpage/si-list'>//www.freelists.org/webpage/si-
>>     
> list</a>
>   
>> For help:
>> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>>
>> List FAQ wiki page is located at:
>>                 <a 
>> href='http://si-list.org/wiki/wiki.pl?Si-List_FAQ'>http://si-
>>     
> list.org/wiki/wiki.pl?Si-List_FAQ</a>
>   
>> List technical documents are available at:
>>                 <a href='http://www.si-list.org'>http://www.si-list.org</a>
>>
>> List archives are viewable at:     
>>              <a href='//www.freelists.org/archives/si-
>>     
> list'>//www.freelists.org/archives/si-list</a>
>   
>> or at our remote archives:
>>              <a href='http://groups.yahoo.com/group/si-
>>     
> list/messages'>http://groups.yahoo.com/group/si-list/messages</a>
>   
>> Old (prior to June 6, 2001) list archives are viewable at:
>>              <a 
>> href='http://www.qsl.net/wb6tpu'>http://www.qsl.net/wb6tpu</a>
>>   
>>
>>
>>     
> ------------------------------------------------------------------
> 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
>   
>
>   


------------------------------------------------------------------
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: