Tony,
I meant to add to my last post that my observations on the sample boards
were from the Altium SMT versions.  One is a rather mediocre board that was
done with 8 layers and no planes so the power and ground were routed in with
the signals.  This could account for the poor power and ground routes.  The
other board was the benchmark board and while it did look better, it still
had some problems as well.  I have copied my previous post to the DXP forum
below for reference.  As more people who have used Situs post their comments
on their experiences as you did, it will be helpful in understanding if
Situs is better or not on large SMT boards.  You mentioned that your board
was a simple two layer board.  Was it surface mount or through hole?  My
autorouting experience with through-hole designs has always been good, but
that is not where the majority of the work is these days for me.  Most of
the boards I get these days are double sided surface mount designs that are
fairly dense.  The autorouter just doesn't do a good enough job for me on
such projects.  With all the fanfare from Altium about Situs, I was looking
forward to something better than what I saw.
-------------------------------
Previous message posted to DXP forum:
I admit I have only tried the autorouter with minimal training with
disastrous results but I thought that was probably my ignorance in setting
it up properly.  I then went to load the demo boards that came with DXP and
viewed the surface mount designs autorouted by Altium to see what someone
who knows how to use Situs could produce.

Rather than describe what I found, I will just point to a few areas out of
several:
Project:  C:\Program Files\Altium\Examples\PCB Auto-Routing\PCB
Auto-Routing.PrjPCB (open the SMT board)
1.  GND routing on bottom layer between R288 & R31
2.  VDD routing on bottom layer at C140
3.  8 layers and no power planes?
4.  GND routing near U2 pin 34
5.  U94 pin 27 takes an unnecessarily long path to it's destination.
6.  VDD routing near C60
7.  Numerous acid traps throughout the design.

Project:  C:\Program Files\Altium\Examples\PCB Benchmark\PCB
Benchmark.pcbdoc
8.  routing of R435 pin 2
9.  routing of U16 pins 38 & 40 to R336 pin 2.
10.  U59 pin 13 & pin 12.
In general "PCB Benchmark.pcbdoc" is better than the "PCB
Auto-Routing.PrjPCB" but there are still several areas that need
improvement.

If this is the best that a trained Situs user can produce, I am not
impressed at all.  I wonder what the yields of these designs would be in a
real production environment.

-------------------------------
----- Original Message -----
From: "Tony Karavidas" <[EMAIL PROTECTED]>
To: "'Protel EDA Forum'" <[EMAIL PROTECTED]>
Sent: Friday, August 30, 2002 10:15 PM
Subject: Re: [PEDA] Service Pack 7 vs DXP issues


> I have one contrary experience. I re-did a simple two layer board I did
> by hand years ago, and Situs did better than any previous version. In
> fact it looks pretty damn good...close to hand routing.
>
> All previous routers could manage 100%, but they looked like hell. Situs
> did not.
> The problem is I can do little boards by hand anytime. I'm hoping for
> Situs to step up to the plate for the LARGE boards that I don't want to
> do by hand anymore.
>
> Tony
>
>
> > -----Original Message-----
> > From: Rob Young [mailto:[EMAIL PROTECTED]]
> > Sent: Friday, August 30, 2002 5:58 PM
> > To: Protel EDA Forum
> > Subject: Re: [PEDA] Service Pack 7 vs DXP issues
> >
> >
> > I am not experienced in using the new Situs router but I have
> > tried it with disastrous results.  Since I thought I might
> > have been doing something wrong, I loaded the sample boards
> > routed at Altium by what I would have thought to be an
> > experienced Situs user.  I posted a list of sample areas to
> > look at on these boards a while ago on the DXP forum.  From
> > what I saw, the results were worse than my results in some
> > cases.  Perhaps I am too picky but I see no reason for 5 or
> > six power vias near each other all connected by short traces
> > on various layers to tie one pin to the power buss when just
> > one via would have done the job.  Several acid traps and
> > traces exiting pads at oddball angles with stairstepping.
> > After reviewing the sample boards, I decided not to mess with
> > Situs again until a few service packs have come out.
> > Overall, I did not see anything in the sample boards to
> > indicate that Situs is an improvement.
> >
> > Rob
> >
> > ----- Original Message -----
> > From: "Abd ul-Rahman Lomax" <[EMAIL PROTECTED]>
> > To: "Protel EDA Forum" <[EMAIL PROTECTED]>
> > Sent: Friday, August 30, 2002 4:57 PM
> > Subject: Re: [PEDA] Service Pack 7 vs DXP issues
> >
> >
> > > At 04:23 PM 8/30/2002 -0400, Michael Reagan (EDSI) wrote:
> > > >I probably would pay for SP7  with as long as they met the
> > long list
> > > >of
> > our
> > > >requirements also.
> > >
> > > I too have been put off by the complexities introduced with
> > DXP. If it
> > > had been, say, a bug fix for 99SE, at a corresponding
> > price, there is
> > > no doubt that I'd be using it. If the additional features had come
> > > without greatly confusing the user interface (for one used
> > to 99SE),
> > > I'd be wanting them too. We were long awaiting an improved
> > autorouter,
> > > and I had heard rumors about how good it was going to be, rivalling
> > > Specctra, etc.
> > >
> > > I haven't seen any comments on the DXP list on Situs except
> > for some
> > > information about the design rules it follows, which have not been
> > > much improved, i.e., there are apparently still plenty of
> > rules which
> > > are
> > ignored.
> > >
> > > Something is wrong.
> > >
> > > The unfortunate thing about the DXP release is that the work was
> > > invested, it would seem, in advance of a true marketing study, at
> > > least of one involving a sufficient number of existing users.
> > > Programming for a service pack is one thing, programming
> > for feature
> > > improvements may be something else. It is difficult, I'd
> > think, to go
> > > back, but it might not be
> > impossible.
> > >
> > > The theory behind the Client/Server architecture was that the
> > > individual modules were separately maintainable. How much
> > the Advanced
> > > PCB server was modified to make it into the DXP PCB server, I don't
> > > know.
> > >
> > > I would think that solid software management for a product
> > like Protel
> > > would involve continuously fixing bugs, as soon as
> > possible, releasing
> > > service packs regularly, and sometimes including feature
> > improvements
> > > -- gradually -- as part of the process. A maintenance model allows
> > > this,
> > which
> > > is probably one reason why Altium has gone that way. The
> > transition,
> > > however, has not been handled well. There should never have
> > been such
> > > a dead time with no service pack. SP7 should have been
> > released long
> > > ago.
> > >
> > > I can understand the argument that was probably put forth: since we
> > > are going to make all these major changes, we need to put all our
> > > effort into them instead of fooling around with code that
> > is going to
> > > become obsolete anyway. Yet this argument is one that keeps
> > software
> > > buggy on into eternity. There is a reason why organisms
> > only change a
> > > little DNA at a time! Make too many changes at once, nothing works
> > > well any more.
> > >
> > > So then you have to do all kinds of new software testing,
> > etc., to try
> > > to find the bugs that have been introduced with the
> > changes. Plus, a
> > > crucial part of the "organism" is the user. Confuse the
> > user, and the
> > > best
> > software
> > > becomes next to useless.
> > >
> > > But it might not be impossible to put together an SP7,
> > perhaps much of
> > > the coding has already been done and even tested to some
> > degree. I'd
> > > suggest a price of, say, $1K for it, fully appliable to DXP (or,
> > > perhaps, to ATS) when the user decides to go that way. Enough 99SE
> > > users might pay for an SP7 to make it worthwhile; it would generate
> > > good will among the users -- except for those who insist
> > that anything
> > > short of feature improvement should be free, period.
> > >
> > > As far as $2000 for the DXP upgrade, the fact is that a
> > truly improved
> > > autorouter would be worth $2K just by itself. Problem is,
> > in order to
> > > get it -- assuming that Situs is actually greatly improved
> > now or in
> > > the near future -- we have to move into a user interface that is
> > > sufficiently different to put many of us off. Unless Altium does
> > > something about this. Remember, the whole point of
> > Client/Server was
> > > to modularize the programs while permitting interaction.
> > >
> > > (While I was a Beta tester for DXP, events in my own life prevented
> > > me,
> > and
> > > thus far have continued to prevent me, from investing much
> > time in DXP
> > > either during Beta or subsequently. Perhaps the autorouter is truly
> > > magic, and it has simply escaped comment on the DXP list; in that
> > > event, I
> > presume
> > > that someone who knows better will enlighten us.)
> > >
> > >
> > >
> > >
> > **********************************************************************
> > > **
> > > * Tracking #: 80B2D86297784D429EB1D3578C179B77B45AEA09
> > > *
> > >
> > **************************************************************
> > **********
> >

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* To post a message: mailto:[EMAIL PROTECTED]
*
* To leave this list visit:
* http://www.techservinc.com/protelusers/leave.html
*
* Contact the list manager:
* mailto:[EMAIL PROTECTED]
*
* Forum Guidelines Rules:
* http://www.techservinc.com/protelusers/forumrules.html
*
* Browse or Search previous postings:
* http://www.mail-archive.com/proteledaforum@techservinc.com
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to