RE: [IBIS-Users] {Spam?} {Filename?} LVDS IBIS model - simulation issue

From: Tom Dagostino <tom@teraspeed.com>
Date: Wed Sep 18 2013 - 11:39:14 PDT
Andrey

 

I've modeled a lot of LVDS buffers in the past and this is definitely the
strangest looking pulldown IV curve I've ever seen.  And as Arpad says it
does not at all look like a reasonable result.

 

First, why are you loading your LVDS buffer with a 50 Ohm differential
termination?  Most I have seen are looking for 100Ohms.

 

Have you inverted the pulldown's polarity?  I would also expect to see the
pulldown curve to have positive current between ground and Vdd.

 

LVDS buffers have a feedback loop in them so stabilize the Vcm of the
buffer.  Usually this only affect the pullup IV curve but sometimes only the
pulldown.  This makes extracting the IV characteristics of the controlled
device difficult.  The feedback is fighting the curve tracer that is
measuring the IV characteristics.  While you are doing the IV sweep you are
by definition changing Vcm and the feedback is trying to correct for this.

 

BTW, Ip = -In.  One is sourcing current and the other sinking.  The sourcing
should have negative currents and the sinking positive.

 

Tom Dagostino

 

Teraspeed Labs

9999 SW Wilshire St.

Suite 102

Portland, OR 97225

USA

 

971-279-5325  Office

971-279-5326   FAX

503-430-1065  Cell

 

tom@teraspeed.com 

www.teraspeed.com <http://www.teraspeed.com/>  

 

Teraspeed Consulting Group LLC

16 Stormy Brook Road 

Falmouth, ME 04105

401-284-1827

 

From: owner-ibis-users@eda.org [mailto:owner-ibis-users@eda.org] On Behalf
Of Muranyi, Arpad
Sent: Wednesday, September 18, 2013 8:31 AM
To: ibis-users@eda.org
Subject: RE: [IBIS-Users] {Spam?} {Filename?} LVDS IBIS model - simulation
issue

 

Andrey,

 

I am not sure that your PD I-V curve is correct.

 

First of all, it should have a positive sign, but

its shape doesn't look right either.

 

Can you describe how you generated these I-V curves?

 

Thanks,

 

Arpad

========================================================

 

 

From: owner-ibis-users@eda.org [mailto:owner-ibis-users@eda.org] On Behalf
Of Andrey Babintsev
Sent: Wednesday, September 18, 2013 9:38 AM
To: ibis-users@eda.org
Subject: Re: [IBIS-Users] {Spam?} {Filename?} LVDS IBIS model - simulation
issue

 

Re-sending with updated attachment

Hi,

I am making IBIS models for LVDS (low voltage differential signal) IO
buffer. 
To model an LVDS buffer I created the following test circuit, assigned the
generated IBIS model to both buffers, set one buffer as inverted output.


Problem:
In the image below the YELLOW waveform is what I saw and the PURPLE waveform
is what I expected to see. 

YELLOW = IBIS model
PURPLE = Spice ( reference ) model

Here is combined pullUp (yellow) and pullDown (green) IV curves (actually it
is spice curves from which IV tables were derived):


Lets assume that simulator tries to find operational point for LVDS buffer
driving high. Then for the non-inverted buffer a point (Vp, Ip) will be
taken from pullUp table, for the inverted buffer a point (Vn, In) from
pullDown table, the following conditions should be satisfied: Ip = In, Vp -
Vn = Ip*50. On the image above it is seen that there are 2 points satisfying
the condition: Vp =~  1.45V and Vp =~ 2.2V. Seems that something like that
happened during simulation and simulator got to the second point. (BTW,
there are 60 points for the region 0.9 - 1.6V and 10 points for the region
1.6-2.5V in the IBIS model IV tables).

Question:
What can be done to resolve this situation? Should IV tables or another IBIS
structures be modified and how? I am making ibis ver. 4.2 models.

Thanks,
Andrey



On 09/18/2013 06:33 PM, Andrey Babintsev wrote:

Warning: This message has had one or more attachments removed
(fail.min.png). Please read the "EDA-Attachment-Warning.txt" attachment(s)
for more information.

Hi,

I am making IBIS models for LVDS (low voltage differential signal) IO
buffer. 
To model an LVDS buffer I created the following test circuit, assigned the
generated IBIS model to both buffers, set one buffer as inverted output.


Problem:
In the image below the YELLOW waveform is what I saw and the PURPLE waveform
is what I expected to see. 

YELLOW = IBIS model
PURPLE = Spice ( reference ) model

Here is combined pullUp (yellow) and pullDown (green) IV curves (actually it
is spice curves from which IV tables were derived):


Lets assume that simulator tries to find operational point for LVDS buffer
driving high. Then for the non-inverted buffer a point (Vp, Ip) will be
taken from pullUp table, for the inverted buffer a point (Vn, In) from
pullDown table, the following conditions should be satisfied: Ip = In, Vp -
Vn = Ip*50. On the image above it is seen that there are 2 points satisfying
the condition: Vp =~  1.45V and Vp =~ 2.2V. Seems that something like that
happened during simulation and simulator got to the second point. (BTW,
there are 60 points for the region 0.9 - 1.6V and 10 points for the region
1.6-2.5V in the IBIS model IV tables).

Question:
What can be done to resolve this situation? Should IV tables or another IBIS
structures be modified and how? I am making ibis ver. 4.2 models.

Thanks,
Andrey


-- 
This message has been scanned for viruses and 
dangerous content by  <http://www.mailscanner.info/> MailScanner, and is 
believed to be clean. 


-- 
This message has been scanned for viruses and 
dangerous content by  <http://www.mailscanner.info/> MailScanner, and is 
believed to be clean. 


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.




--------------------------------------------------------------------
|For help or to subscribe/unsubscribe, e-mail mikelabonte@eda-stds.org
|or ibis-request@eda-stds.org
|
|IBIS reflector archives exist under:
|
|  http://www.eda-stds.org/ibis/email_archive/ Recent
|  http://www.eda-stds.org/ibis/users_archive/ Recent
|  http://www.eda-stds.org/ibis/email/         E-mail since 1993
image001.png image002.png image003.png
Received on Wed Sep 18 11:41:31 2013

This archive was generated by hypermail 2.1.8 : Wed Sep 18 2013 - 11:42:12 PDT