Subject: FW: [IBIS-Users] EW UK article Critical about vendor IBIS models
From: Roy Leventhal (Roy.Leventhal@ieee.org)
Date: Wed Oct 02 2002 - 07:36:24 PDT
-----Original Message-----
From: Roy Leventhal [mailto:Roy.Leventhal@ieee.org]
Sent: Friday, September 27, 2002 8:49 PM
To: Angulo, John; 'ibis-users@eda.org'
Subject: RE: [IBIS-Users] EW UK article Critical about vendor IBIS models
Sorry,
This is old news. I presented a paper on these issues at the Cadence User
Group meeting in September, 2000 and again at IPC Designers Learning
Symposium in July, 2001. The results I presented are basically the same and
are a synopsis of my experience screening some 800+ IBIS models for 3Com in
the 3-1/2 years that I was there. Prior to either presentation I gave a talk
at the IBIS Committee Summit at PCB East in 1999 or 1998 on these same
issues.
The new news is that there is an IBIS subcommittee working on making the
IBIS spec more readable and the IBIS parser more effective at catching
problems.
The really significant news would be that the semiconductor industry cared
enough to clean up their act. It took the Japanese quality revolution to
make our semiconductor industry change its ways. Most of them did not
survive that challenge as independent, vertically integrated USA entities.
In that revolution it took ultimate consumers buying from companies whose
products did not break, cost less and incorporated new useful features
quicker.
Based on my recollections of simulating amplifier strings, when I had to
obtain SPICE models things were not any better. Simulation was just coming
into real use and SPICE models were not used by the process engineers to
maintain process quality and consistency. Then, two things happened. The
Japanese ate our lunch over the quality issue and the semiconductor industry
had to respond with more attention paid to process control.
By hearsay evidence, or so I've been told, the semiconductor industry now
makes much more intensive use of SPICE, HSPICE in particular, to control
quality. Supposedly then, the quality of SPICE models has improved because
the process engineers now have a payoff in their accuracy, and I hope,
completeness. But, you would have to present some data to me that would
support the idea that SPICE models are any better. I'd say that the evidence
for SPICE improvement is thin based on the fact that almost all IBIS models
are produced by simulating the SPICE model. I would guess that (clamp) diode
resistance (a SPICE parameter) is not critical to process control based on
the number of IBIS clamp curves with magnitudes of E23 that I have seen.
This is a direct result of not modeling that parameter in the SPICE model.
There is no payoff in producing and maintaining accurate IBIS models in
anything the semiconductor industry is interested in, save one - turning a
profit.
Prior even to the quality and availability issue over SPICE models was the
exact same problems regarding data sheets. The JEDEC Committee was simply
ignored by the semiconductor industry when the former tried to enforce
discipline regarding 1N - - - - and 2N - - - - data sheets. At that point
the semiconductor suppliers simply issued their own data sheets and ignored
the JEDEC Committee and its Standards.
When you think about it, there is no inherent reason for the semiconductor
industry to produce good, plentiful data sheets, IBIS models, SPICE models,
Applications Notes or any other "knowledge collateral" without a direct
connection to the profit motive. Do you think for a minute that your
company's Purchasing Department would, or would not, place an order based on
the availability of any or all of the items I've just listed?
Nothing will change until the semiconductor vendors start loosing or gaining
sales over the IBIS model quality issue. Second, it must be made plain to
them that the lack of models and their poor quality is the cause of loosing
sales. Third, loosing sales will not happen until particular technologies
have more than one/multiple suppliers available as alternative sources to
users. Plus, the bottom line impact of line yield problems, stretched out
development times due to excessive debugs of (unsimulated) physical boards,
etc., must be known to a user company's money people.
Regards,
Roy
-----Original Message-----
From: owner-ibis-users@server.eda.org
[mailto:owner-ibis-users@server.eda.org]On Behalf Of Angulo, John
Sent: Friday, September 27, 2002 3:49 PM
To: 'ibis@eda.org'; 'ibis-users@eda.org'
Subject: [IBIS-Users] EW UK article Critical about vendor IBIS models
The following is posted on behalf of Mike Ventham at Quantic EMC. Please
send all private replies to ventham@quantic-emc.co.uk only.
-----------------------------------------
IBISians,
I thought you might be interested in this article.
http://www.electronicsweekly.co.uk/issue/dailyview.asp?vpath=/articles/2002/
09/25/daily06.htm
Any comments?
P.S. Apologies if you get two copies by being on both IBIS lists. There is
no merged list to ensure everyone interested in IBIS only gets 1 copy.
Regards
Mike
________________________________________________________________
| Mike Ventham - Vice-President Engineering, |
| Quantic EMC Inc. Headquarters |
| Croft House, Chilcompton, 191 Lombard Ave., Winnipeg,|
| Somerset, UK, BA3 4JA Manitoba, Canada R3B 0X1 |
| Tel: 44 (0)1761 232191 Tel: (204) 942 4000 |
| Fax: 44 (0)7974 141685 Fax: (204) 957 1158 |
| Mobile: 44 (0)7971 553260 |
| Email: ventham@quantic-emc.co.uk http://www.quantic-emc.com |
|------------------------------------------------------------------
|For help or to subscribe/unsubscribe, email majordomo@eda.org
|with just the appropriate command message(s) in the body:
|
| help
| subscribe ibis <optional e-mail address, if different>
| subscribe ibis-users <optional e-mail address, if different>
| unsubscribe ibis <optional e-mail address, if different>
| unsubscribe ibis-users <optional e-mail address, if different>
|
|or email a written request to ibis-request@eda.org.
|
|IBIS reflector archives exist under:
|
| http://www.eda.org/pub/ibis/email_archive/ Recent
| http://www.eda.org/pub/ibis/users_archive/ Recent
| http://www.eda.org/pub/ibis/email/ E-mail since 1993
|------------------------------------------------------------------
|For help or to subscribe/unsubscribe, email majordomo@eda.org
|with just the appropriate command message(s) in the body:
|
| help
| subscribe ibis <optional e-mail address, if different>
| subscribe ibis-users <optional e-mail address, if different>
| unsubscribe ibis <optional e-mail address, if different>
| unsubscribe ibis-users <optional e-mail address, if different>
|
|or email a written request to ibis-request@eda.org.
|
|IBIS reflector archives exist under:
|
| http://www.eda.org/pub/ibis/email_archive/ Recent
| http://www.eda.org/pub/ibis/users_archive/ Recent
| http://www.eda.org/pub/ibis/email/ E-mail since 1993
This archive was generated by hypermail 2b28 : Wed Oct 02 2002 - 07:52:10 PDT