Maker Pro
Maker Pro

PSpice and series of capacitors

E

effepe

Jan 1, 1970
0
PSpice get crazy to determine voltages when I have a series of capacitors:

<http://img213.imageshack.us/img213/1521/ex231yg7.gif>

After 'googled' a bit, I tried to fix the limitation of floating pin
(seems that PSpice can't determine DC voltage in such situations) by
adding two shunt resistors (R2 and R3) of huge value (1 TOhm).

In option tab, I enabled GMIN (1E-12 S) too.

Is there a possibility to use PSpice or I have to switch to LT
SwitcherCAD III ?

Still was wondering how a software like PSpice can fall down so poorly
with a simple circuit like this.

Thanks in advance and my apologies for my (surely poor) English ;-)
 
C

Charlie Edmondson

Jan 1, 1970
0
effepe said:
PSpice get crazy to determine voltages when I have a series of capacitors:

<http://img213.imageshack.us/img213/1521/ex231yg7.gif>

After 'googled' a bit, I tried to fix the limitation of floating pin
(seems that PSpice can't determine DC voltage in such situations) by
adding two shunt resistors (R2 and R3) of huge value (1 TOhm).

In option tab, I enabled GMIN (1E-12 S) too.

Is there a possibility to use PSpice or I have to switch to LT
SwitcherCAD III ?

Still was wondering how a software like PSpice can fall down so poorly
with a simple circuit like this.

Thanks in advance and my apologies for my (surely poor) English ;-)
You are going to see the same problems in LTSpice, and just about any
other spice there is. Every node needs to have a DC path to ground for
the differential equations to make sense.

And then, you blew it by putting in a ridiculous resistance like a
teraohm. Usually, 100MEG is more than sufficient, occasionally a full
1G if you need really small shunt currents, but this is a simulator with
a realisitic dynamic range. There aren't real 1T resistances out there,
and they blow up the simulators when you have 1T resistances, and then
really small resistances. Get real, get some experience, and then you
can critize the simulators for real problems...

Charlie
 
H

Helmut Sennewald

Jan 1, 1970
0
effepe said:
PSpice get crazy to determine voltages when I have a series of capacitors:

<http://img213.imageshack.us/img213/1521/ex231yg7.gif>

After 'googled' a bit, I tried to fix the limitation of floating pin
(seems that PSpice can't determine DC voltage in such situations) by
adding two shunt resistors (R2 and R3) of huge value (1 TOhm).

In option tab, I enabled GMIN (1E-12 S) too.

Ading resistors in parallel to the capacitors is OK.
GMIN doesn't help anything in this case.
Is there a possibility to use PSpice or I have to switch to LT SwitcherCAD
III ?

I don't know whether there is an option in PSPICE to do that automatically.
LTspice adds a little bit of conductance to avoid this problem.
Still was wondering how a software like PSpice can fall down so poorly
with a simple circuit like this.

Thanks in advance and my apologies for my (surely poor) English ;-)

SPICE requires a DC-path to every node. This is
not a limitation. It just remembers you about a
possible weakness of your design.

Even in a real circuit, you can't predict the DC-voltage
on the capacitors. It's detrmined by the insulation
resistance of the capacitors which may vary many
decades over production lots.
And last but not least it's bad design practice to
connect capacitors in series without an additional
resistor in parallel.

Best regards,
Helmut
 
E

effepe

Jan 1, 1970
0
Charlie Edmondson ha scritto:
Usually, 100MEG is more than sufficient,

Also using 100MEG, PSpice puts out exotics voltages.
 
E

effepe

Jan 1, 1970
0
Helmut Sennewald ha scritto:
And last but not least it's bad design practice to
connect capacitors in series without an additional
resistor in parallel.

It's just an exercise assigned as homework.
 
J

Jim Thompson

Jan 1, 1970
0
Helmut Sennewald ha scritto:


It's just an exercise assigned as homework.

Bad professor, bad professor ;-)

Or maybe GOOD professor thinning the class down to a manageable size
;-)

...Jim Thompson
 
Q

qrk

Jan 1, 1970
0
Bad professor, bad professor ;-)

Or maybe GOOD professor thinning the class down to a manageable size
;-)

...Jim Thompson

I say GOOD professor for beating in to the students heads about DC
floating nodes and the proper way to deal with them in Spice and real
life. One can learn from bad design examples.
 
H

Hal Murray

Jan 1, 1970
0
I say GOOD professor for beating in to the students heads about DC
floating nodes and the proper way to deal with them in Spice and real
life. One can learn from bad design examples.

Especially if one learns something the hard way.

Suppose you are putting 2 caps in series to get a higher voltage
rating. What size resistors do you use to ensure a DC ballance?
 
M

Marra

Jan 1, 1970
0
I would rather use my brain than use a simulator.

At the end of the day a good engineer doesnt need a less than perfect
simulator.
 
J

Jim Thompson

Jan 1, 1970
0
I would rather use my brain than use a simulator.

At the end of the day a good engineer doesnt need a less than perfect
simulator.

You would never be a competitor anyway, so who gives a shit. You
certainly don't qualify for the moniker "good engineer".

...Jim Thompson
 
J

Joel Koltner

Jan 1, 1970
0
Marra said:
I would rather use my brain than use a simulator.

Well, your software looks like it comes out of the '70s or early '80s at best,
and at that point in time you could arguably still be productive without a
simulator.
At the end of the day a good engineer doesnt need a less than perfect
simulator.

True, but many a good engineer does feel that he needs a job, and for many
jobs today simulators are requisite. :)

This same discussion comes up every now and again on SED, Marra -- Google can
dig them up and you can read how many jobs have clearly demonstrable
productivity gains via the use of simulation.

---Joel
 
C

Charlie Edmondson

Jan 1, 1970
0
effepe said:
Charlie Edmondson ha scritto:



Also using 100MEG, PSpice puts out exotics voltages.

So, as the student, why is that? What is it about that arrangement of
resistors and capacitors causes those voltages? Learn and understand...

Charlie
 
J

Jim Thompson

Jan 1, 1970
0
Charlie Edmondson ha scritto:

LTSPice simulate it correctly.

LTspice adds the R's that everyone has mentioned. Likewise series R's
for inductors.

"correctly"?... you are an IDIOT.

...Jim Thompson
 
M

Mike Engelhardt

Jan 1, 1970
0
Jim,
LTspice adds the R's that everyone has mentioned.
Likewise series R's for inductors.

"correctly"?... you are an IDIOT.

I'm afraid you are probably the village I-word here, Jim.
You might first check out what LTspice actually does there,
instead of spinning rhetoric.

LTspice is probably more accurate then without such matrix
modifications if you are talking real world circuits. In
any case, it's certainly more accurate then PSpice's
addition of gmin to current sources, or PSpice's Gear
integration, both of which have contributed to design and
analysis errors of silicon, in one case, even by your good
self.

Regards,

--Mike
 
J

Jim Thompson

Jan 1, 1970
0
Jim,


I'm afraid you are probably the village I-word here, Jim.
You might first check out what LTspice actually does there,
instead of spinning rhetoric.

LTspice is probably more accurate then without such matrix
modifications if you are talking real world circuits. In
any case, it's certainly more accurate then PSpice's
addition of gmin to current sources, or PSpice's Gear
integration, both of which have contributed to design and
analysis errors of silicon, in one case, even by your good
self.

Regards,

--Mike

I guess I incorrectly assumed you added shunt resistance around
floating caps, since the OP said, "LTSPice simulate it correctly".

Floating caps are floating caps... aren't they???

...Jim Thompson
 
J

Jim Thompson

Jan 1, 1970
0
I guess I incorrectly assumed you added shunt resistance around
floating caps, since the OP said, "LTSPice simulate it correctly".

Floating caps are floating caps... aren't they???

...Jim Thompson

Now that I drove to the grocery store and back.

I don't quite understand your little outburst.

HOW can you "simulate it correctly" if it's truly a floating node?

You have to be doing at least some kind of surreptitious node setting,
or adding R to ground.

I will continue to call idiots as I see them ;-)

...Jim Thompson
 
J

Jim Thompson

Jan 1, 1970
0
Jim Thompson ha scritto:
you are in my kill file

You are ignorant beyond specification... and you will always be ;-)

...Jim Thompson
 
Top