开云体育

FDR happen in all manner of nonlinear devices from ferrite cores to dielectrics where losses are a function of frequency. Try to model a ferrite core losses ( both series and ||Rs) with typical data provided by manufactures you'll find loss Rs are prop to sqrt(f) . Skin effect also is a good example- in fact its hard to find an electronic device excempt from this behavior.

George

Sent: Fri, 16 Sep 2011 08:43:18 +0100
Subject: [LTspice] Re: Freqeucny Dependent resistor








Hi,

doesn't the skin effect represent a frequency dependent resistor?

Regards,
Schmocki

--- In LTspice@..., "Dave" <dave.g4ugm@...> wrote:

-----Original Message-----
From: LTspice@...
[mailto:LTspice@...] On Behalf Of John Fields
Sent: 16 September 2011 00:38
To: LTspice@...
Subject: Re: [LTspice] Re: Freqeucny Dependent resistor


On Thu, 15 Sep 2011 13:20:01 -0700, you wrote:

I believe it is impossible to have a physically realizable

resistor that is frequency >dependent and has no reactive component.

---
Carbon button microphone?

Well as that has a "?" I suppose its worth treating as a question, but I
guess its tongue in cheek. Its not frequency dependant as such, merely
pressure dependant. Its also easily simulated by feeding a "wav" file into a
voltage dependant resistor. It wouldn't be much use as a mic if it was
frequency dependant.

I think the only way to have a frequency dependant resistor is to have some
circuit that measures frequency and use the output of that to control the
voltage on Voltage Dependant Resistor, but that has lag as its hard to
measure frequency at an instant.

Of course frequency dependant components such as inductors and capacitors
derive their frequency dependence from the fact that the current flow in
them depends on the rate of change of voltage or visa versa. This property
has a dual effect, it gives the desired frequency dependence, but it also
adds the non-resistive behaviour.

Could you use some sort of gyrator circuit to produce the desired effect?

--
JF

--------------------------------------------------------------

This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom
they are addressed. If you have received this email in error
please notify the system manager. Please note that any views or
opinions presented in this email are solely those of the author
and do not necessarily represent those of the company. Finally,
the recipient should check this email and any attachments for
the presence of viruses. The company accepts no liability for
any damage caused by any virus transmitted by this email.


Gill Research & Development Ltd is a limited company registered
in England and Wales. Registered number: 3154453.
Registered office: The George Business Centre, Christchurch
Road, New Milton. BH25 6QJ


[Non-text portions of this message have been removed]

Hi,

doesn't the skin effect represent a frequency dependent resistor?

Regards,
Schmocki

--- In LTspice@..., "Helmut" <helmutsennewald@...> wrote:

--- In LTspice@..., "pindsen" <windven@> wrote:

Hi,

I need an BSS138 in my LTSpice simulation.
Since I can't find that component in the program itself, I need to make it.

I have done the following steps but LTspice won't simulate:

1) Downloaded a Zetex BSS138 spice subcircuit and placed it in a "BSS138.sub" file
2) The spice file is placed in c:&#92;program files&#92;LTC&#92;LTspiceIV&#92;lib&#92;sub
3) Included a nmos symbol (nmos.asy) in the design
4) Renamed the symbol to BSS138/ZTX
5) Add a spice directive command ".inc BSS138.sub"
6) Hit the run button
7) LTspice can't simulate and says "Can't find definition of model "bss138".

The Zetex spice model can be seen here:

*ZETEX BSS138 Spice Mosfet Subcircuit Last revision 11/91
*
.SUBCKT BSS138/ZTX 3 4 5
* Nodes D G S
M1 3 2 5 5 MOD1
RG 4 2 343
RL 3 5 6E6
D1 5 3 DIODE1
.MODEL MOD1 NMOS VTO=1.109 RS=1.474 RD=1.59 IS=1E-15 KP=0.597
+CGSO=23.5P CGDO=4.5P CBD=53.5P PB=1 LAMBDA=267E-6
.MODEL DIODE1 D IS=1.254E-13 N=1.0207 RS=0.222
.ENDS

Can anyone help?

Best regards
Carsten Wind
Denmark

Hello Carsten,

4a)
Ctrl-right-mouse-click on the placed symbol nmos.
Change Prefix:MN to Prefix:X

2) I always recommend to save model files in the folder of the
schematic.

Best regards,
Helmut

Hello Helmut,
Thanx for your answer.
I did what you proposed in 4a) and it worked :-) But what does it do? What is Prefix and what does a change from MN to X do?

Then, how should the .include statement look like if the model files is placed in the folder of the schematic?

Best regards
Carsten Wind

--- In LTspice@..., Hubert Hagadorn wrote:

I believe it is impossible to have a physically realizable
resistor that is frequency dependent and has no reactive
component.

In Laplace expressions LTspice will accept such a construct.
The key is to use the abs() function to convert the expression
to purely real. If desired this can be converted to purely
imaginary by multiplying by the sqrt(-1). For example:

V1 1 0 ac=1 ; .ac analyses always require an ac source
B1 2 0 V= V(1) Laplace 1/(1+sqrt(s/6.3)) ; half pole at 1Hz
B2 3 0 V= V(1) Laplace 1/(1+abs(sqrt(s/6.3))) ; real only version
B3 4 0 V= V(1) Laplace sqrt(-1)/(1+abs(sqrt(s/6.3))) ; imaginary

On Thu, 15 Sep 2011 13:20:01 -0700, you wrote:

I believe it is impossible to have a physically realizable resistor that is frequency >dependent and has no reactive component.

---
Carbon button microphone?

--
JF

Hi all: Okay, let me try this one. Take a transistor, bias it and have the collector and the emitter be the input and output. The signal enters the collector by a cap and the signal exits via the emitter. Somewhere, there is a frq. to Voltage IC that feeds the base of the transitor. Is this what you are looking for? Bill
To: LTspice@...
From: tony@...
Date: Thu, 15 Sep 2011 22:14:11 +0000
Subject: [LTspice] Re: Freqeucny Dependent resistor
































--- In LTspice@..., Christian Thomas <ct.waveform@...> wrote:

That's a pretty damning point!

But does it hold? You've squared up a vector quantity to get something that

by definition has no direction. Telling me that power is a scalar is surely

a starting point, not a proof that nothing reactive is there.

CT

On 15 September 2011 21:46, John Woodgate <jmw@...> wrote:

**

In message

<CANj54jz_C3C0wTdO5kkCRbVh3qDQsrda9qatksP5h1bX+ohtqA@...>,

dated Thu, 15 Sep 2011, Christian Thomas <ct.waveform@...>

writes:

But are you sure that radiation resistances are only real?

By definition: they are notionally responsible for the real power that

is radiated. Reactive elements cannot be responsible.

--

OOO - Own Opinions Only. Try www.jmwa.demon.co.uk and www.isce.org.uk

John Woodgate, J M Woodgate and Associates, Rayleigh, Essex UK

When I point to a star, please look at the star, not my finger. The star

will

be more interesting.

Christian,



I don't think John was stating that there is no reactive part, only that power cannot be dissipated in it.



Regards,

Tony


















[Non-text portions of this message have been removed]

Correction

"*Though I do remember nothing significant in phase angle within the audio
band*" should be no significant change in resistance. I don't remember what
the reactive elements were.

CT

Tony,

I missed this reply of yours earlier. I hadn't thought of the skin effect,
though given that is the result of penetration of fields, it must have some
reactive element, surely? I did actually look into this in some depth about
a year ago in relation to inductor design but more has flowed out than
remained. (Though I do remember nothing significant in phase angle within
the audio band. But to say it was near zero required about three
assumptions to be true which might have been a stretch.)

You also succinctly cover the radiation question that John does. ka = 2 is
considered the usual limit, incidentally, though on what grounds don't know.
ka = 1 might be flat and ka = 2 might give you a bandwidth. I have usually
regarded it as an experimental result - although the equations aren't that
difficult IIRC, there may be losses that leave the practical results a
little awry. But, as you say, plenty of reactive element in there.

Christian

Hi again Tony,

My starting question was whether radiation impedance actually held up as
only real - his example of a possibility. John then slightly moved the
goalposts in response by saying that they were by definition only real; look
no further than the power response (and we know that nothing can be
dissipated reactively). This was not so damning an argument as it looked, I
said, because power is not a vector quantity.

I don't care where the power goes. I just wanted to know if there was a
direction associated with the amplitude.

To keep this on track - perhaps Nigel can tell us whether this was what he
had in mind as a VCVS type f dependent resistor. I think we would know
about it if it could be done.

Christian

--- In LTspice@..., Michael Stuarts wrote:

The model for SW appears to missing. If anyone can help me create
a .model statement for schematic that would be great. I could not
figure out how to do it from the help file alone.

Why not? Do you have a reading comprehension problem? This is a
serious question because Help both clearly explains that you must
provide a model statement to define your switch and gives you a
link to an example schematic that you may run if you are too
unfamiliar with the correlation between the netlist notation
example provided and how it derives from the schematic.

Perhaps English is not your native language? Or perhaps you really
didn't read Help? I am curious because I really would like to
understand why Help (which seems perfectly fine to me) does not
seem to work for so many users (so there is no wrong answer on
your part other than a less than honest one).

Regards -- analogspiceman

PS: Here is the relevant section from Help.
-----------------------------------------------------------
S. Voltage Controlled Switch

Symbol Names: SW

Syntax: Sxxx n1 n2 nc+ nc- <model> [on,off]

Example:

S1 out 0 in 0 MySwitch

.model MySwitch SW(Ron=.1 Roff=1Meg Vt=0 Vh=-.5 Lser=10n Vser=.6)

The voltage between nodes nc+ and nc- controls the switch's
impedance between nodes n1 and n2. A model card is required
to define the behavior of the switch. See the schematic file
.&#92;examples&#92;Educational&#92;Vswitch.asc to see an example of a model
card placed directly on a schematic as a SPICE directive.
-----------------------------------------------------------

--- In LTspice@..., Christian Thomas <ct.waveform@...> wrote:

That's a pretty damning point!

But does it hold? You've squared up a vector quantity to get something that
by definition has no direction. Telling me that power is a scalar is surely
a starting point, not a proof that nothing reactive is there.

CT



On 15 September 2011 21:46, John Woodgate <jmw@...> wrote:

**


In message
<CANj54jz_C3C0wTdO5kkCRbVh3qDQsrda9qatksP5h1bX+ohtqA@...>,
dated Thu, 15 Sep 2011, Christian Thomas <ct.waveform@...>
writes:

But are you sure that radiation resistances are only real?

By definition: they are notionally responsible for the real power that
is radiated. Reactive elements cannot be responsible.

--
OOO - Own Opinions Only. Try www.jmwa.demon.co.uk and www.isce.org.uk
John Woodgate, J M Woodgate and Associates, Rayleigh, Essex UK
When I point to a star, please look at the star, not my finger. The star
will
be more interesting.

[Non-text portions of this message have been removed]

Christian,

I don't think John was stating that there is no reactive part, only that power cannot be dissipated in it.

Regards,
Tony

--- In LTspice@..., "michaelstuarts" <michaelstuarts@...> wrote:

The model for SW appears to missing.
If anyone can help me create a .model
statement for schematic. That would be great.
I could not figure out how to do it from
the help file alone.

Thanks,

//Michael

Hello Michael,

From the Help:

S1 out 0 in 0 MySwitch
.model MySwitch SW(Ron=.1 Roff=1Meg Vt=0 Vh=-.5 Lser=10n Vser=.6)

You don't say what your switch will be used for, but I'm betting that if you deleted the last three terms in the model above, it would do what you wanted.

If you're having trouble relating that to your schematic, do this:
1. Place a sw from the parts chooser, and change its value to MySwitch.
2. Add a SPICE directive, and paste in the .model text.

If you're trying to control it from something like TTL or CMOS, change Vt to 1.8V, or something between the logic levels.

Regards,
Tony

<snip>

I believe it is impossible to have a physically realizable resistor that is frequency dependent and has no reactive component. A transmission line comes close in that its resistance is almost constant for a range of frequencies, but as you know at very low frequencies, neglecting any series resistive components, its reactance is largely capacitive.

Hubert

</snip>
Hello Hubert,

You are of course correct: the only resistor possible that had no reactive component would also have zero size. Any conductor of finite size inevitably has both inductance and capacitance, even if it has no resistance, like a superconductor.

I think the salient point of the discussion really is whether we can legitimately model part of a system, in which the reactive part doesn't significantly change as a consequence - with a notional frequency-dependent resistor.

I think there are converse examples too, where the reactive part changes without a consequential change to the resistive part: a capacitor where the real part of permittivity reduces with frequency without significant change to the imaginary part. A good example of this is an FR4 PCB trace. It's capacitance reduces with frequency, whilst its resistance increases, albeit for a completely unconnected reason. I'm sure there are many more.

(I detect the red pen of the moderator twitching, so had better stop here.)

Regards,
Tony

That's a pretty damning point!

But does it hold? You've squared up a vector quantity to get something that
by definition has no direction. Telling me that power is a scalar is surely
a starting point, not a proof that nothing reactive is there.

CT



On 15 September 2011 21:46, John Woodgate <jmw@...> wrote:

**


In message
<CANj54jz_C3C0wTdO5kkCRbVh3qDQsrda9qatksP5h1bX+ohtqA@...>,
dated Thu, 15 Sep 2011, Christian Thomas <ct.waveform@...>
writes:

But are you sure that radiation resistances are only real?

By definition: they are notionally responsible for the real power that
is radiated. Reactive elements cannot be responsible.

--
OOO - Own Opinions Only. Try www.jmwa.demon.co.uk and www.isce.org.uk
John Woodgate, J M Woodgate and Associates, Rayleigh, Essex UK
When I point to a star, please look at the star, not my finger. The star
will
be more interesting.

[Non-text portions of this message have been removed]

The model for SW appears to missing.
If anyone can help me create a .model
statement for schematic. That would be great.
I could not figure out how to do it from
the help file alone.

Thanks,

//Michael

--- In LTspice@..., Philip Bellingham <rmhc78a@...> wrote:

Tony,

All valid points. I agree that creating SPICE models for s-parameter
characterized passive parts using only native SPICE passives (C, L, and R) is an
art and the native models are often inadequate even with the provision for
adding parasitics. The same is true for modeling active components. Therefore, I
agree that creating SPICE models that can be used for all types of analysis
(TRAN, AC, etc.) is a daunting task when the only data you have is s-parameters.
My experience has been that frequency domain characterization and simulation is
often sufficient. To that end, the utility, in our group files, that converts an
s-parameter file to a set frequency dependent controlled sources in SPICE,
creates models that are quite acceptable when performing AC network
analysis.??This does limit you to AC analysis, and for active devices you are
constrained to the bias conditions present??when the s-parameters were measured.


Regards,

???? - Philip

All,

By definition s-parameters are small signal linear AC parameters, and in most cases the parameters are a function of frequency. Usually, one is interested in only a single frequency or a mall range of frequencies. The work fine for their intended purpose.

Rick

In message <CANj54jz_C3C0wTdO5kkCRbVh3qDQsrda9qatksP5h1bX+ohtqA@...>, dated Thu, 15 Sep 2011, Christian Thomas <ct.waveform@...> writes:

But are you sure that radiation resistances are only real?

By definition: they are notionally responsible for the real power that is radiated. Reactive elements cannot be responsible.
--
OOO - Own Opinions Only. Try www.jmwa.demon.co.uk and www.isce.org.uk
John Woodgate, J M Woodgate and Associates, Rayleigh, Essex UK
When I point to a star, please look at the star, not my finger. The star will
be more interesting.

Tony,

All valid points. I agree that creating SPICE models for s-parameter
characterized passive parts using only native SPICE passives (C, L, and R) is an
art and the native models are often inadequate even with the provision for
adding parasitics. The same is true for modeling active components. Therefore, I
agree that creating SPICE models that can be used for all types of analysis
(TRAN, AC, etc.) is a daunting task when the only data you have is s-parameters.
My experience has been that frequency domain characterization and simulation is
often sufficient. To that end, the utility, in our group files, that converts an
s-parameter file to a set frequency dependent controlled sources in SPICE,
creates models that are quite acceptable when performing AC network
analysis.?This does limit you to AC analysis, and for active devices you are
constrained to the bias conditions present?when the s-parameters were measured.


Regards,

?? - Philip


________________________________
From: Tony Casey <tony@...>
To: LTspice@...
Sent: Thu, September 15, 2011 12:42:09 PM
Subject: [LTspice] Re: About impedance

?


--- In LTspice@..., Andy <Andrew.Ingraham@...> wrote:

Naive question. Why would you want to do that? If you have s-parameter
files, presumably they came from measurements of a real circuit. So why
ask Spice (or anything else) to generate a circuit?

Often you have s-parameters for a component (a transistor or MMIC),
which you get from the component vendor, and you want to use it in a
circuit.

Andy

Indeed. And for passive products, such as SAW filters or isolators, S-parameters
are the only data you're ever likely to get from vendors.

Also for capacitors and inductors, vendors such as Murata and ATC also provide
S-parameter data for use at frequencies at which the simple equivalent circuits
we're used to in SPICE are hopelessly inadequate to describe the performance.

Regards,
Tony




[Non-text portions of this message have been removed]

John

I put my caveat in for good reason ...

The way I thought about it was to think what I would do to design one. I
would need a reactive component somewhere, to do it easily, and then I would
cancel out that component's phase shift with an all-pass. I am fairly
certain that can't be done perfectly (though for the moment I can't quite
think why not for a first order original. I suppose because it has to add,
so you can never get a perfect time delay.)

But are you sure that radiation resistances are only real? I'm fairly
certain that acoustic radiation impedances aren't naturally real and that
it's only because drivers are mass controlled that they appear so. (I
suspect that means masses in the f = ma sense rather than an equivalent C.)

Christian



On 15 September 2011 20:32, John Woodgate <jmw@...> wrote:

**


In message
<CANj54jwDGEMhNQA5E=DD=cMmYJTwDN=mMCrX2FaMfB4YgiLxuw@...>,
dated Thu, 15 Sep 2011, Christian Thomas <ct.waveform@...>
writes:

In which case the answer needed is "No, you can't. Or at least you
can't have a full solution. (I think that must be right).

I'm not sure. We have real resistances that change with frequency -
radiation resistances for example, There are definitely resistances - at
any fixed frequency they are consistent with Ohm's and Joule's Laws.

So why can't I define a resistor R(f), such that R(f) = kf, or K*F(f)? k
and K have dimensions, of course, but that doesn't seem to be a problem.
--
OOO - Own Opinions Only. Try www.jmwa.demon.co.uk and www.isce.org.uk
John Woodgate, J M Woodgate and Associates, Rayleigh, Essex UK
When I point to a star, please look at the star, not my finger. The star
will
be more interesting.

[Non-text portions of this message have been removed]