Prev: Newbie question, please...
Next: Micro controller and printer
From: Bret Cahill on 29 Jun 2010 18:54
> >> >Is there any circuit simulator that accounts for overload and failure
> >> >of components?
>
> >> >Bret Cahill
>
> >> One of the dorkier simulators would put little animated flames over
> >> the symbols of parts that dissipated too much power.
>
> >> But that's a gimmick. It's hard for Spice to understand your intent,
> >> application, heatsinking, and duty cycles.
>
> >It would have to be set up with inputs for current, wattage or voltage
> >limits.  Usually in reliability you have one distribution curve for
> >strength and another for expected loading over the lifetime of a
> >structure.
>
> >Another much more daunting but related undertaking would be to take
> >the electronics / mechanics analogy much further, i.e., determining
> >the electronic analogues to Coriolis forces, angular momentum, moment
> >of inertia in beam bending, etc.
>
> Those of us who are electronics design engineers don't work by
> analogy. We use real facts and hard numbers.

This isn't about a solution to a specific problem.

It's a fishing expedition for new ideas.

> There are accepted ways to calculate MTBF of electronic systems, and
> accepted ways to stress and test actual working hardware. A good EE
> can design electronics that has multiples longer MTBF than the
> MIL-HDBK-217 or Bellcore calculations suggest. A bad designer will
> often come in much, much worse.

The goal here would be new methods and new devices.

> >The principle difference is few care if someone burns up a chip but
> >just about everyone cares if an RC building or prestressed concrete
> >overpass comes down.  Even that difference fades in power production.
>
> >A serious attempt at such a project would probably yield a lot of
> >valuable insights and innovations long before it was completed.
>
> >Instead of using SPICE for simple mechanical impedance problems a
> >generalized application could be used for either.
>
> >A single click would convert mechanical to electronic and vice versa..
>
> None of that last part makes sense to me.

You draw a low pass fliter on the program, say a capacitor to ground,
and the mechanical equivalent, say a spring and damper, is
automatically drawn up as well.

A couple issues may complicate things:

A lot of the analogues probably wouldn't "map" one to one and anyone
using it would need to know which analogue to specify.

The hardest part would be formalizing the math for all those messy
mechanical situations that are "merely omitted" from many engineering
programs. The WTC "pancaked" down because the architect didn't think
joints were glamorous enough for his time.

Some researchers at MIT took a better look at the joints _after_ 9/11.


Bret Cahill









From: Bret Cahill on 29 Jun 2010 18:59
> >> >Is there any circuit simulator that accounts for overload and failure
> >> >of components?
>
> >> >Bret Cahill
>
> >> One of the dorkier simulators would put little animated flames over
> >> the symbols of parts that dissipated too much power.
>
> >> But that's a gimmick. It's hard for Spice to understand your intent,
> >> application, heatsinking, and duty cycles.
>
> >It would have to be set up with inputs for current, wattage or voltage
> >limits.  Usually in reliability you have one distribution curve for
> >strength and another for expected loading over the lifetime of a
> >structure.
>
> >Another much more daunting but related undertaking would be to take
> >the electronics / mechanics analogy much further, i.e., determining
> >the electronic analogues to Coriolis forces, angular momentum, moment
> >of inertia in beam bending, etc.
>
> That's what analog computers were made for. Convert the mechanical
> values into their electrical equivalents (analogues), set up the
> circuitry, let it run and plot the outputs. The operational amplifiers
> did the mathematics (operations). Nowadays digital computers do it
> faster with more accuracy.

Different analogy. The goal here isn't just calculating but new
devices and methods of dealing with older problems.


Bret Cahill


From: Jamie on 29 Jun 2010 20:01
Bret Cahill wrote:
> Is there any circuit simulator that accounts for overload and failure
> of components?
>
>
> Bret Cahill
>
>
The "Work Bench" Program? I seem to remember my kid using that in school
and it would generate some nice effects if you made a mistake in the
circuit that would otherwise destroy something..

Jamie

From: John Larkin on 29 Jun 2010 22:30
On Tue, 29 Jun 2010 18:21:35 -0300, YD <ydtechHAT(a)techie.com> wrote:

>Late at night, by candle light, Bret Cahill <BretCahill(a)peoplepc.com>
>penned this immortal opus:
>
>>> >Is there any circuit simulator that accounts for overload and failure
>>> >of components?
>>>
>>> >Bret Cahill
>>>
>>> One of the dorkier simulators would put little animated flames over
>>> the symbols of parts that dissipated too much power.
>>>
>>> But that's a gimmick. It's hard for Spice to understand your intent,
>>> application, heatsinking, and duty cycles.
>>
>>It would have to be set up with inputs for current, wattage or voltage
>>limits. Usually in reliability you have one distribution curve for
>>strength and another for expected loading over the lifetime of a
>>structure.
>>
>>Another much more daunting but related undertaking would be to take
>>the electronics / mechanics analogy much further, i.e., determining
>>the electronic analogues to Coriolis forces, angular momentum, moment
>>of inertia in beam bending, etc.
>>
>
>That's what analog computers were made for. Convert the mechanical
>values into their electrical equivalents (analogues), set up the
>circuitry, let it run and plot the outputs. The operational amplifiers
>did the mathematics (operations). Nowadays digital computers do it
>faster with more accuracy.
>
>- YD.

You can use Spice to simulate relatively simple mechanical and thermal
systems, but real-world mechanical systems have complex geometries
that need finite-element methods, and thermal systems involve
diffusion, neither of which Spice is especially good at.

Spice can model simple thermal systems with the equivalents...


1 farad = 1 gram of aluminum

1 amp = 1 watt of heat flow

1 ohm = 1 degree C per watt

1 volt = 1 degree C

1 second = 1 second

which is good to 5% maybe.


Worse sims are electromagnetics, where Maxwell's equations have to be
solved in space and time. The worst are chaotic nonlinear things, like
weather.

John



From: Bret Cahill on 30 Jun 2010 23:32
> > Is there any circuit simulator that accounts for overload and failure
> > of components?
>
> > Bret Cahill
>
> The "Work Bench" Program? I seem to remember my kid using that in school
> and it would generate some nice effects if you made a mistake in the
> circuit that would otherwise destroy something..

Only in power generation -- megawatts -- does the destruction of
anything in a circuit amount to a hill of beans.

You can fry computer chips all day long and never attract more
attention than Al Gore in a dust devil.


Bret Cahill


First  |  Prev  |  Next  |  Last
Pages: 1 2 3
Prev: Newbie question, please...
Next: Micro controller and printer