Circuits that shouldn't work...?

[Ken Smith]

What are "good results" for that (useless?) circuit?

Results that match what really happens (except for the actual smoke) are
all that I ask. If a circuit would pass 1000A through 1N4007 and destroy
it, I want the spice to tell me about the 1000A.

 

[Ken Smith]

try a standard full bridge from 120 60 Hz into a 5000 uF cap and 240 ohms
load.

---------+------------+-------
! ! ! !
--- --- ! !
^ ^ ! !
! ! --- \
+--(V)----+ --- /
! ! ! \
--- --- ! !
^ ^ ! !
! ! ! !
---------+------------+-------

You mean like this? I tried it. It works fine. What did you expect me
to see?

 

[Jeroen Belleman]

What are "good results"[...]

Results that match what really happens (except for the actual smoke) are
all that I ask. If a circuit would pass 1000A through 1N4007 and destroy
it, I want the spice to tell me about the 1000A.

Well, it *does* show you that you're trying to put 1kA through a 1N4007.
It just doesn't tell you that there's anything wrong with doing so. Nor would
I want it to. It's up to the user to judge what's reasonable. You can't just
switch your brain off and believe everything the simulator tells you.

Perhaps Saber is the simulator for you. That one does have thermal
effects in its models. You still can't switch your brain off though...

Jeroen Belleman

In an evolving man-machine system, the man wil get dumber faster
than the machine gets smarter.
[ by Scott Guthery, SEN Jan 87, p. 20 ]

 

[Ken Smith]

What are "good results"[...]

Results that match what really happens (except for the actual smoke) are
all that I ask. If a circuit would pass 1000A through 1N4007 and destroy
it, I want the spice to tell me about the 1000A.

Well, it *does* show you that you're trying to put 1kA through a 1N4007.
It just doesn't tell you that there's anything wrong with doing so.

So you agree with me about what spice will do.

Nor would
I want it to. It's up to the user to judge what's reasonable.

Go back and re-read what I said and you will find that you are agreeing
with me on this too. So I really don't see why you then said:

 

[Joseph2k]

---------+------------+-------
! ! ! !
--- --- ! !
^ ^ ! !
! ! --- \
+--(V)----+ --- /
! ! ! \
--- --- ! !
^ ^ ! !
! ! ! !
---------+------------+-------

You mean like this? I tried it. It works fine. What did you expect me
to see?

It produces wrong results for me. Of course i can calculate what it should
do by hand very easily and test circuits produce the result my hand calcs
say it should be. I guess i will have to build up a set of fairly
universal PDF's and post it on ABSE. Take me about a week to get around to
all that.

 

[YD]

Fouls up how?

I've modeled:


---->!----
! !
(V) --- 0.1F
! ---
GND !
GND

With no problem.


I've also modeled:


K1
------ -------------+-----------+---------
) ( L2 ! !
) ( ! --- C1
) ( ! ---
L1 ) +-- GND ! !
------ ( ! GND
( !
( !
-------->!---


Where (IIRC):

L1 = 100mH + 2R
L2 = 1mH + 0.1R
K1 = 0.95
C1 = 1000uF + 0.01R

... and got good results.

You shouldn't, if the diode in the upper leg of the secondary is
really missing.

- YD.

 

[colin]

It produces wrong results for me. Of course i can calculate what it should
do by hand very easily and test circuits produce the result my hand calcs
say it should be. I guess i will have to build up a set of fairly
universal PDF's and post it on ABSE. Take me about a week to get around to
all that.

LTspice works very well for me, ive used other simulators too,
however you have to remember the results you get are only as precise as the
device models,
timesteps and lack of parasitic elements etc allow.

you can set many parameters in LTspice wich trade off accuracy for
simulation speed.

you can add parasitic elements and problems that some rectifier circuits
seem to have will disapear,
theres a very good ltspice group on yahoo wich I remember reading about this
very thing,
although ive never experienced it.
its not a problem with the simulation but a problem with what you are trying
to get it to simulate,
if you could leave out parasitic elements in real life you would probably
see all sorts of problems !

certainly start up of high Q oscillators is one area that requires the most
attention, for me anyway,
but there is a excelent help info for this on the ltspice group.

I wouldnt waste time trying to find and post circuits that dont simulate
well,
but try and learn how to tune the simulation to get acurate results.

simulation does save a lot of time and wasted components/pcbs etc and does
allow you to find a far more
optimum solution but does in itself I find take a lot of time sometimes
messing about with the
parameters and parasitics to get accurate results But within a realistic
simulation run time.

its not a substitute for knowledge of electronics in fact it can require a
much greater
understanding to realise why the results arnt as expected.

If you have a specific problem with a circuit you cant get round with simple
advice from the
examples/tutorials/educational samples, its best to post the circuit file to
the yahoo ltspice grp.

Colin =^.^=

 

[Ken Smith]

It produces wrong results for me. Of course i can calculate what it should
do by hand very easily and test circuits produce the result my hand calcs
say it should be. I guess i will have to build up a set of fairly
universal PDF's and post it on ABSE. Take me about a week to get around to
all that.

My ISP doesn't carry ABSE. Could you post your LTspice schematic here.
It is just ASCII text. I'd really like to see what the problem is.

 

[Ken Smith]

K1
------ -------->!---+-----------+---------
) ( L2 ! !
) ( ! --- C1
) ( ! ---
L1 ) +-- GND ! !
------ ( ! GND
( !
( !
-------->!---

[....]
You shouldn't, if the diode in the upper leg of the secondary is
really missing.[/QUOTE]

Diode missing! I don't see no diode missing

 

[Winfield Hill]

Ken Smith wrote...

K1
------ -------->!---+-----------+---------
) ( L2 ! !
) ( ! --- C1
) ( ! ---
L1 ) +-- GND ! !
------ ( ! GND
( !
( !
-------->!---

[....]
You shouldn't, if the diode in the upper leg of the secondary
is really missing.

Diode missing! I don't see no diode missing [/QUOTE]

It's missing the transformer's wiring resistance, and may be
missing its leakage inductance, both of which dramatically
affect the charging of C1, and it's missing C1's esr as well.

 

[Ken Smith]

[....]

It's missing the transformer's wiring resistance, and may be
missing its leakage inductance, both of which dramatically
affect the charging of C1, and it's missing C1's esr as well.[/QUOTE]


From my post where I said it worked fine.
==Where (IIRC):
==
== L1 = 100mH + 2R
== L2 = 1mH + 0.1R
== K1 = 0.95
== C1 = 1000uF + 0.01R


I don't think I've tried it without the extra resistances in there, but
I'll bet it will ring on the capacitances in the diodes etc.

 

[Winfield Hill]

Ken Smith wrote...

From my post where I said it worked fine.

Sorry, I didn't see that.

Where (IIRC):

== L1 = 100mH + 2R -- too good
== L2 = 1mH + 0.1R -- likely too good
== K1 = 0.95 -- likely too good
== C1 = 1000uF + 0.01R -- yep, too good

Did you vet it against bench measurements?

 

[Ken Smith]

Ken Smith wrote...

Sorry, I didn't see that.


Did you vet it against bench measurements?

The (IIRC) was because this was more than a year back. The ESR value for
the capacitor was just a guess, but the other numbers, I really used[1],
were from measuring the transformer.

[1] vs. what I remember them as.

One problem with bench testing was that the sinewave would start when I
closed the switch, not at some determined phase angle. It looked close
enough for me to trust the spice stuff on it.

The real circuit had an extra diode in it to isolate the filtered output
from a higher current unfiltered output.