开云体育

--- In LTspice@..., "sergio" <thetosk@> wrote:

Hello,

I kindly ask for an advice to choose the correct parameter to insert in the Box-Muller formula using the "rand" function.
In particular I need to have a white noise in the band of interest (called BW in the param def below), the RMS I called Vrmsn1 due to R1 in the BW.

my question is:

which value do I need to put to multiply "time" ? I use 2*BW is it correct?

.param Vrmsn1 = sqrt(4*kB*T*R1 *BW)

V = {Vrmsn1}*SQRT(-2*LN(1E-8+rand(time*2*BW)))*SIN(2*PI*rand(time*2*BW+1879))


using an fft with nfft points the average of the noise spectrum seen in the FFT should be:

10*log10( 4*kB*T*R1 *BW /(nfft/2) )

I cannot get this correspondence (maybe I miss some point!)



From an old post #40603 (that refers to an older one #5598) I read:

---
There are pseudo random number generator functions
available in behavioral sources: white(), rand(),
and random(). You would have to filter their
response to make it white or pink. Their frequency
domain spectral output is a specific function I do
not document, not white or pink (given sufficient
statistics it will increase in frequency and then
roll off). --Mike
---

that means the spectrum of "rand" is not really white ...

any suggestion to match the AC:
10*log10( 4*kB*T*R1 *BW /(nfft/2) )

with the time domain (just a voltage source with the box-muller
formula as above) is appreciated.

kind regards
Sergio

Hello Sergio,

V = {Vrmsn1}*SQRT(-2*LN(1E-8+rand(time*2*BW)))*SIN(2*PI*rand(time*2*BW+1879))

.param Vrmsn1=1
This voltage has the RMS value of 1V, but it has some power
beyond the frequency BW. When I filter it with a 7nd order
Butterworth filter with bandwidth BW, I got 438mVrms.
The spectrum has still a little bit energy above BW. So the
corresponding power density may be still a little bit less
between 0Hz and BW.

Best regards,
Helmut

Hi,

I uploaded the files into files/temp folder.

hope someone can give me a hint

Thanks in advance

Resetpin

--- In LTspice@... <mailto:LTspice%40yahoogroups.com>,
"resetpin" <resetpin@...> wrote:

Hello,

I used Filter Pro from Texas instruments to get a Filter design from

my specifications. They are:

Gain: 5V/V (13,97dB)
Part: Ideal Opamp: Order 4 Stages: 2
Corner frequency Attenuation: 10,979dB
Allowable PassBand Ripple 1dB
Center Frequency: 1,75 kHz
Stopband Attenuation: -30dB
Passband Bandwidth: 100Hz
Stopband Bandwith: 600Hz

Filterpro proviced a schematic and Values for the parts and a

Gain-Frequency plot.

When I try to redo this in LTSpice the frequency corner freqency of

the filter is at the right positon but the gain is about -40db.

I wonder where the differences are and would like to show the

LTSpice *.asc File compared with the Texax Results but don't know how
to post them here in the group.

Maybe someone is willing to help me out and show me the differences

or the problem.

Thanks in advance

Resetpin

Hello,

I used Filter Pro from Texas instruments to get a Filter design from my specifications. They are:
Gain: 5V/V (13,97dB)
Part: Ideal Opamp: Order 4 Stages: 2
Corner frequency Attenuation: 10,979dB
Allowable PassBand Ripple 1dB
Center Frequency: 1,75 kHz
Stopband Attenuation: -30dB
Passband Bandwidth: 100Hz
Stopband Bandwith: 600Hz

Filterpro proviced a schematic and Values for the parts and a Gain-Frequency plot.

When I try to redo this in LTSpice the frequency corner freqency of the filter is at the right positon but the gain is about -40db.

I wonder where the differences are and would like to show the LTSpice *.asc File compared with the Texax Results but don't know how to post them here in the group.

Maybe someone is willing to help me out and show me the differences or the problem.

Thanks in advance

Resetpin

Hello RL,

Please try my example to see and understand my point.

Files > Temp > z_from_FFT_with_noise.asc

Run the TRAN simulation.
Plot V(z2) and I(V2)
FFT of V(z2) and I(V2) with 1048576 points
-> plot FFT of V(z2)/FFT of I(V2)

Best regards,
Helmut


--- In LTspice@..., legg@ wrote:

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

--- In LTspice@..., legg@ wrote:

Given a plot of a pink or white noise voltage, and a synchronized plot of the current induced in a partially reactive load -

Will the 'normalized' ratio of the FFT plots be an indication of Z?

I'm looking at the low frequency end of the FFT, to avoid sampling issues. By normalized, I mean that 1 volt would produce 1 amp with both plots resolving to 0dB at the minimum sampled frequency (the load being partially inductive in this example).

It seems much too simple......

RL

Hello RL,

It will only work when you filter the FFT-output, but the
the FFT-results can't be filtered in LTspice. Thus you have
to export the FFT-data and process them in an external program.
This method only work with a linear system and it's precision
may be somewhat limited due to group delay variation.

Best regards,
Helmut

Filter the FFT results? The source is, ideally, uniform in amplitude across the band; so ideally the FFT would be a flat (or at least a straight) line.....What does filtering an FFT output entail?

I am looking at a pink noise source that has been prefiltered to include only a decade or two, so a few assumptions can be made about any FFT output registering outside this region.

With both V and I being monitored synchronously, there are a hell of a lot of variables being weeded out, but the sample duration is limited and the resolution is fairly course (~8bit 2500 data points per variable). So even if the calculated Z were valid, it'd be a crude approximation at best. I'll post a few plots in temp to show what the results actually look like.

The Z in this case is a loudspeaker transducer coil, so it's static characteristics are fairly easily obtainable using a simple swept tone. Given the signal processing capabilities that are falling into our laps with LTspice and even the most modest digital scopes these days, some questions go begging for an answer.

I was also concerned about phase relationships and delays, hence the curiosity about the actual spot 'Z' produced by the calculation. If R is known, then the phase could be intuited. If L is independent of temperature, then a new R value could theoretically be winkled. If other things are known, then phantom R/L/C quantities, or functional shifts in the same could also be evaluated. It's probably already standard practice, somewhere, DSP101 or something FAIK.

As with any 'new' tool, there are applications that don't jump out at you, or get their own chapter in the manual. Knowing the limits could save some time fiddling about - so a reference or pointer could be worth a thousand words.

Doing this again, with a source that had an unpredictable frequency content......could be a very non-invasive sensing method that costs only software, which, as everyone knows, is free......(insert smiley face here).

RL

--- In LTspice@..., "beaudoin.christopher" <beaudoin.christopher@> wrote:

I am modelling a voltage noise source by defining the noise source with a PWL file. The circuit is very basic (two resistors no L's or C's) and I have setup a transient analysis using the time step equal to the time increment in the PWL file. In doing this, what I am finding is that the transient analysis is somehow modifying the statistics(I'm guessing by integration/interpolation)of the random signal defined in the PWL file. For this reason, I'd like to gain a better understanding of what this transient analysis is doing in this very simple case. For example, are there are time constants implicit to the solver that I should be aware of such that I should not expect to achieve a steady state case until some time "Tsteadystate" even if there are no Ls or Cs in the circuit? Any understanding you can offer would be greatly appreciated.

Best Regards,
Chris

Hello Chris,

LTspice dynamically changes the time step. You should simply
a small maximum time step in the .TRAN command. This ensures
that nothing will be lost. LTspice does a linear interpolation
between your defined PWL data points.

LTspice uses data compression by default when it saves the
results of the simulation. If you you want the best result,
you should disable data compression in this case. Please add
the following SPICE-directive to your schematic. It will
switch off data compression.

.options plotwinsize=0

Best regards,
Helmut