# RE: [SI-LIST] : Differential termination

From: Peterson, James F (FL51) ([email protected])
Date: Fri Aug 18 2000 - 11:57:13 PDT

Selva,

These questions are ones that everyone asks in their first PECL (or ECL)
design (500 or 50?, Vtt, etc.?). I believe Johnson and Graham give a pretty
good answer in their book ("High Speed Digital Design" pg. 51-54).

best regards,
Jim Peterson
Jim Peterson
[email protected]
Honeywell Space Systems
M/S 934-5
13350 U.S. Hwy 19 N.
Clearwater FL, 33764-7290
Office : 727-539-2719
Fax : 727-539-2183

-----Original Message-----
From: selvaraj subramanian [mailto:[email protected]]
Sent: Friday, August 18, 2000 5:15 AM
To: [email protected]
Subject: [SI-LIST] : Differential termination

Hello,

I have a doubt in differential termination of PECL signals.

First let me explain the scenario. All my questions are for high speed
operation.

A PECL driver is driving a PECL receiver. Since the output stage of a PECL
is an emitter follower, a pull down resistor(Rpd) is necessary. Assume the
characteristic impedance of traces are 50 ohms.

Let me explain different cases

1. I can place the resistor near the driver itself and its value can vary
anywhere from 50 ohms to 500 ohms(as per the motorola recommendations).
Usually high value of resistor is used to reduce the power dissipation and
to get more swing. This is same if the you are treating the signals as
single ended or differential. This can be done only if the driver and the
receiver are so close. Otherwise, I can use the case 3.

Question: How can I determine the value of pull down resistor?

2. If I place the resistor at the far end(near the receiver), the value
of the resistor is made equal to the characteristic impedance of the trace
to terminate and this resistor can be connected to ground. If it is
connected to ground, the power dissipation will be heavy. So as per the
motorola recommendations, I can connect that resistor to Vtt and power
dissipation can be reduced. I believe Vtt = Vcc - 2.0V (I don't know how
this equation is derived). This is same if the you are treating the signals
as single ended or differential

Question: How this equation Vtt = Vcc-2.0V is derived?

3. I can place the pull down resistor near the driver and if the trace
between the driver & receiver is more, once again I have to terminate the
trace using a 100 ohms resistor across the true and false signals.

All the above cases will work correctly.

But my doubts are..

1. Which termination is better and will provide less power dissipation?

2. Let me change the case 2 for few more questions.

I can place two 50 ohms in series across the true and false signals.
From the middle of two 50 ohms resistor, I can connect an another 50 ohms
resistor to ground. This resistor provides the Vtt(1.3V) for LVPECL.

My questions are.

1. What will be the effective termination seen by each trace in this
configuration?
2. What will be the effective termination seen by the driver when you are
considering them in differential mode?
3. My doubt is whether we have to connect a capacitor in parallel with the
third resistor connected to ground for AC coupling?
4. If I connect in this configuration, the current flow through each trace
is through two 50 ohms and so I doubt the termination seen by each trace is
100 ohms. So effectively 100 ohms cannot terminate 50 ohms trace. Is this
ok?
5. What is the difference in providing a Vtt through the third resistor &
providing a Vtt through seperate power supply?

6. What is the difference between Vtt and Vbb(bias expected by the
receivers)? Are they both same or not?

regards,
Selva

************************************************************
S. SELVARAJ
Senior Engineer - VLSI/System Design - Global R&D Solutions
Wipro Technologies
"Samrudhi Bhavan"
37, Castle Street
Ashok Nagar
Bangalore-560 025, India
Tel: 91-80-5367297/98 Ext:206
Fax: 91-80-5576032
E-mail: [email protected] <mailto:[email protected]>
www.wipro.com <http://www.wipro.com>
The World's First SEI CMM Level 5 Software Services Company
************************************************************

**** To unsubscribe from si-list or si-list-digest: send e-mail to
[email protected] 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
****

This archive was generated by hypermail 2b29 : Wed Nov 22 2000 - 10:51:05 PST