Mailing list archives : pcb-rnd

ID:4247
From:Robin Laing <Me...@TelusPlanet.net>
Date:Mon, 6 Jul 2020 10:02:39 -0600
Subject:Re: [pcb-rnd] Off topic (maybe) - Spice simulator
in-reply-to:4233 from gene glick <ge...@optonline.net>
Hello Gene,
 
Modelling is just that, an estimate of what will happen.  We have lots 
of experience with models in our work but not electronic models.
 
My experience goes before models and had a great electronics teacher 
that taught lots of quick rules for circuit analysis.  I have used that 
for most of my career as my experience started before there were 
simulations.
 
I am not directly involved with this project due to COVID-19, further 
out of the loop than I should be, all I do know is that the model is 
being used for timing calculations and I feel they are going in the 
wrong direction but that is not my choice.
 
Like you, I have experienced the "boom on test" of a design.
 
Robin
 
 
On 29/06/2020 17:40, gene glick wrote:
> Maybe you should ask why do they want such perfect models? No matter 
> what, your models will only approximate reality - so you just need to 
> know how close is good enough. Simulation really should only confirm 
> what you already know. When it doesn't work as expected, you ask why, 
> and maybe you missed something - sometimes it's something important. If 
> you rely on someone's model, you are trusting that it's 
> okay....sometimes, they are not. It's up to you, the engineer, to know 
> what your circuit does.
> 
> Have you ever heard of an engineer, looking over a charred, smoking mess 
> that should have been a circuit, and hear them say "it looked good in 
> simulation" :D Happens a lot!
> 
> 
> 
> On Fri, Jun 26, 2020 at 4:45 PM Robin Laing <MeSat@telusplanet.net 
> <mailto:MeSat@telusplanet.net>> wrote:
> 
>     I wish I could beat this into some peoples brains.  They expect models
>     to be perfect all the time.
> 
>     I have been told that if the model isn't working, then we have to look
>     at different tools to model it better.  I think that is part of the
>     issue in this case.  Simulation says one thing, real world says
>     something else.  In our case, the real world is the problem and we do
>     need to be closer to the simulation.  I am thinking that the real world
>     is related to the way that the circuit was constructed as a big part to
>     the issues.
> 
>     Robin
> 
> 
> 
 
 

Reply subtree:
4247 Re: [pcb-rnd] Off topic (maybe) - Spice simulator from Robin Laing <Me...@TelusPlanet.net>