[SI-LIST] Re: Waveform edge ringing

  • From: wolfgang.maichen@xxxxxxxxxxxx
  • To: "Nijagunamurthy, Hithesh (GE Intelligent Platforms)" <hithesh@xxxxxx>
  • Date: Tue, 23 Nov 2010 11:00:19 -0800

Hitesh,

VIH and VIL should indeed be specified in the datasheet. I'd be surprised
if the device lacks such fundamental specifications. Maybe unless it is a
well-defined logic family (e.g. LVTTL or CMOS) where the author assumes
that these thresholds are known (but it would at least state which logic
family it is).

What I meant with my statement is that you typically strive to strobe in
the center of the bit interval, i.e. put the clock edge(s) as far away from
the transition points of the data waveform. If you really have no idea what
setup and hold time for your part could be, then I'd assume maybe 25% of
your bit interval, leaving a 50% eye opening. That would put equal
requirement on clock and data timing accuracy (jitter). But that is a very
(!) rough guess and most likely conservative. Since you already do
simulations it will pay to dig down a bit to determine the true values.

Just my 2 cents

Wolfgang






                                                                           
             "Nijagunamurthy,                                              
             Hithesh (GE                                                   
             Intelligent                                                To 
             Platforms)"               <wolfgang.maichen@xxxxxxxxxxxx>     
             <hithesh@xxxxxx>                                           cc 
             Sent by:                  <si-list@xxxxxxxxxxxxx>,            
             si-list-bounce@fr         <si-list-bounce@xxxxxxxxxxxxx>      
             eelists.org                                           Subject 
                                       [SI-LIST] Re: Waveform edge ringing 
                                                                           
             11/23/2010 10:35                                              
             AM                                                            
                                                                           
                                                                           
                                                                           




Wolfgang,
Thanks for the reply.
I could not understand what you meant by "Data lines in general will have
less of a problem with
such waveforms unless the clock strobes it very close to the bit boundary
(which isn't a good thing to do anyway)".

Where can I find the hold time info for VIH and VIL? I quickly glanced thru
datasheet. But did not find anything.

-Hithesh


-----Original Message-----
From: wolfgang.maichen@xxxxxxxxxxxx [mailto:wolfgang.maichen@xxxxxxxxxxxx]
Sent: Wed 11/24/2010 12:00 AM
To: Nijagunamurthy, Hithesh (GE Intelligent Platforms)
Cc: si-list@xxxxxxxxxxxxx; si-list-bounce@xxxxxxxxxxxxx
Subject: Re: [SI-LIST] Waveform edge ringing


Hithesh,

first thing to note is that VIH and VIH are typically static thresholds.
I.e. the receiver is guaranteed to see a high only if the input stays above
VIH for sufficiently long time. For very short pulses going above VIH (or
below VOL), and without knowing the bandwidth / reaction time of the
receiver you do not know whether it already switched. Worst case for you
would be that the initial spike just barely toggles the clock input, then
when it drops and rises again (this time to full level) it could toggle it
again - the result being double-clocking (clock sees two rising edges where
there should only be one).

Also consider that what you have is only a simulation - unless your models
are pristine and hyper-accurate (a rare case) reality may look quite
different - the spike being larger, the peak or dip being more (or less)
extreme, and so on. But the waveform indicates a good potential for trouble
on the clock pin. Data lines in general will have less of a problem with
such waveforms unless the clock strobes it very close to the bit boundary
(which isn't a good thing to do anyway). Of course any such ringing eats
into your setup/hold margin.

Wolfgang







             "Nijagunamurthy,
             Hithesh (GE
             Intelligent                                                To
             Platforms)"               <si-list@xxxxxxxxxxxxx>
             <hithesh@xxxxxx>                                           cc
             Sent by:
             si-list-bounce@fr                                     Subject
             eelists.org               [SI-LIST] Waveform edge ringing


             11/23/2010 02:53
             AM






I am doing signal integrity simulation for PCI Address/Data bus signals
going from an FPGA to a Power PC.
The signals are daisy chained from the FPGA to 6 Power  PCs. The daisy
chain routing is 55 ohm impedance controlled.

The waveform edge has some ringing at VIH on the rising edge and VIL at
the falling edge. The ringing is only for one cycle(see pic).
Can this be counted as a valid waveform.
Let's say this is input to a clock buffer.
The part where I am confused is(on the rising edge) the waveform goes
above VIH, so now the output of the receiver IC should be High. Now if
the edge goes 200mV below VIH after say 1ns, and comes back to Vcc, will
the output of the receiver change?
There is similar behavior at VIL.

Now, let's say this is Addr or data line. Would the setup time would be
violated?

Image - http://www.imgplace.com/viewimg217/6307/69wave.jpg

Thanks.
-Hithesh


------------------------------------------------------------------
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:
http://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:

http://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:
http://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:
                                 http://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:
http://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:     
                http://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: