RE: [SI-LIST] : implementing delay on board

About this list Date view Thread view Subject view Author view

From: Ken Wu (ken@force10networks.com)
Date: Wed Apr 11 2001 - 11:56:39 PDT


Saeen,
Good observation.
The delay depends on the line length as well as the loadings you put
at the end of your T-line. I bet you put a capacitive loading at the
end of T-line which is the reason causing the delay different from 165ps
as you expected. If you change your capacitive loading to 50ohm (match
your transmission line impedance), you will see the result you expected.
As a matter of fact, the input impedance from your driver side (looking into
the T-line) will be 50ohm only if your loading is 50ohm. If your loading
is a capacitor, then the input impedance will vary depending on the
combination
of line length and capacitor value. You can find the equations to calculate
the
input impedance from any microwave book. With a varying loading, you cannot
get
a linear delay result as you expected.

Regards,

Ken Z. Wu
Force10 Networks, Inc
1440 McCarthy Blvd
Milpitas, CA 95035
Tel: (408)571-3530
Email: ken@force10networks.com

-----Original Message-----
From: Saeen Malik [mailto:saeens@hotmail.com]
Sent: Wednesday, April 11, 2001 10:09 AM
To: si-list@silab.eng.sun.com
Subject: [SI-LIST] : implementing delay on board

Hello SI gurus,

I want to delay some clocks relative to data on my board to take care
of setup and hold time requirements of the receiver. I am using
Hspice and the topology is

Driver ---- Transmission line ---- Receiver

The source impedence of the driver is pretty close to the impedence of
the transmission line, therefore, no series termination is required.
(as per my simulations using Hspice). The risetime of the driver is
approximately 600 ps (0-100%). The propdelay of the transmission line
comes out to be 165 ps/inch.

I am using Hspice version 2000.2 and am using ideal transmission line.
(T elements). My understanding was, if the length of the transmission
line is longer than twice the rise time, any extra delay I add should
appear as pure delay (neglecting any degradation in risetime).
But, when I do a spice run on trace lengths from 10 inch to 25 inch
and measure the the delay between waveforms of different lengths,
the results do not appear uniform. For example, I expect to see
the delay between a 15 inch tline and 16inch tline to be 165ps, but it
is not so. Sometimes it is more and sometimes it is less (across the
length sweep 10-25inch). What could be the reason for this????

One more thing, the signal has a 50% duty cycle and I am measuring the
delay after 10 cycles (just to let the initial transitions die out).
The frequency is 125MHz.

Any inputs would be really appreciated...

Saeen

_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com

**** To unsubscribe from si-list or si-list-digest: send e-mail to
majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****

**** To unsubscribe from si-list or si-list-digest: send e-mail to
majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Thu Jun 21 2001 - 10:11:32 PDT