[ibis-macro] Re: Jitter Transfer Function the Jitter Model Discussion
- From: Gregory R Edlund <gedlund@xxxxxxxxxx>
- To: wkatz@xxxxxxxxxx
- Date: Fri, 23 Sep 2011 12:34:38 -0500
You hit the nail on the head, Walter.
Is the EQF function new to IBIS 5.1? Maybe I should make another drawing
with all the inputs and outputs of a garden variety DLL. Or maybe that
already exists? Can somebody clue me in?
Greg Edlund
Senior Engineer
Signal Integrity and System Timing
IBM Systems & Technology Group
3605 Hwy. 52 N Bldg 050-3
Rochester, MN 55901
From: "Walter Katz" <wkatz@xxxxxxxxxx>
To: <ibis-macro@xxxxxxxxxxxxx>
Date: 09/22/2011 03:26 PM
Subject: [ibis-macro] Re: Jitter Transfer Function the Jitter Model
Discussion
Sent by: ibis-macro-bounce@xxxxxxxxxxxxx
Greg, Mike,
I should have done a better job of what I meant by Tx Jitter is input to
the EQ block. By EQ block I meant the function “EQF” within the DLL that
has as its input the digital stimulus as defined in BIRD120 and whose
output the level driving the output stage . In 5.0, and in the current
Jitter BIRD, jitter is generated by the EDA tool on the stimulus input. The
“EQF” function has to maintain a stack of stimulus values at each bit_time
and then use logic (its EQ block) to set the level driving the output
stage, the “EQF” then determines the time that the signal level transitions
(where the jitter gets injected) on the waveform that gets passed the
output stage.
So Greg is correct. Tx Jitter is inserted at the output of the EQ block,
but it still has to be included as an input to the “EQF” function. If the
jitter needs to be a Jitter Spectrum, then the Jitter Spectrum would need
to be passed into the DLL which in turn is passed to the “EQF” function.
Bottom line; Greg thinks how the EQ silicon operates, and I was thinking
about how the EQ function in the DLL operates.
Walter
From: ibis-macro-bounce@xxxxxxxxxxxxx [
mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Mike Steinberger
Sent: Thursday, September 22, 2011 3:14 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: Jitter Transfer Function
All-
I strongly suggest that we be very precise in the terminology we use, and
so I offer the following two simplified definitions:
Jitter Spectrum: The frequency domain spectral density of the phase noise
on a clock, as compared to an ideal reference clock. To normalize out the
resolution bandwidth of the measurement equipment, this spectral density is
expressed with respect to the spectral density of the main spectral
component: dB Carrier per root Hz (dbC/root Hz).
Jitter Transfer Function: The ratio of the Fourier transform of the output
phase noise of a system divided by the Fourier transform of the input phase
noise for that system. This is a transfer function. It has a magnitude and
a phase. It's a mapping between an input and an output.
Both of these terms have been widely used in communications analysis for a
long time. For example the Bell Labs "Blue Book":
Members of the Bell Labs Technical Staff, Transmission Systems for
Communications, Fifth Edition, pg. 723-40, copyright Bell Telephone
Laboratories, Inc. 1982.
Thanks for your understanding and your cooperation.
Mike Steinberger
On 09/22/2011 12:57 PM, Walter Katz wrote:
All,
There have been several suggestions to enhance AMI to support a Jitter
Transfer function for the Tx input and Rx CDR reference clock. The Jitter
Transfer Function is simply a table containing two columns; Frequency and
Amplitude. This table can either be a Table in the .ami file or a pointer
to a file containing the table.
So the following specific questions:
Do we want the .ami file to contain a Jitter Transfer Function for
the Tx stimulus and the Rx reference clock?
If yes.
Do we want to add this to the Jitter BIRD (BIRD123), or create
a new BIRD?
Do we this to be implement as a Table in the .ami file and/or a
pointer in the .ami file to a file?
Walter
Walter Katz
wkatz@xxxxxxxxxx
Phone 303.449-2308
Mobile 720.333-1107
Other related posts:
