opus-banner
previous | back | home | next
CACD Group
updated 2004.05.12

Nutmeg improvements

  • plot command uses the same syntax as in the original Berkeley verion of SPICE
    Since version 2.01 plots with multiple scales can be created: plot v1 v2 vs vscl1 v3 v4 vs vscl2 v5 v6 The above plot command would display vectors v1 and v2 against vscl1, v3 and v4 against vscl2 and v5 and v6 against default plot scale. In case 'vs' options are ommited, the vectors are plotted against the default scale of the active plot.

    Colors for the plot window can be set by setting colorn variables.
    Example:
      set color0 = r255g255b255 set background color to white
      set color1 = r000g000b000 set grid and text color to black
      set color2 = r000g255b000 set the first vector color to green
      set color3 = r255g000b000 set the second vector color to red
      plot v(1) v(2) the result is graph with the white background and black grid; v(1) is green and v(2) is red
      unset color2 unset the first vector color; the default will be used; also color3 variable will be ignored now
    Default plot type can be set using the plottype variable. Available types are: normal, point and comb.

    set plottype=normal Default line width can be 0, 1 or 2. Set it with: set linewidth=1 Default plot window width and height in pixels: set plotwinwidth=360
    set plotwinheight=360
    If you want to show the info frame (coordinates) in plot windows use: set plotwininfo Default vector identification mode is now manual identification. Manual identification can be initiated by right-clicking in the plot window. To make automatic identification the default (as in version 2.0 and below), enter: set plotautoident Note that all of these settings are valid only for plot windows opened after these setting have been chosen by the appropriate set/unset commands.
  • The spec command written by Anthony Parker, Macquarie University, was added.
  • DC analysis syntax in one dimension is now: dc [middle middle_val] parameter1 start1 stop1
      [dec|oct|lin] step1
    and in two dimensions: dc parameter1 start1 stop1 [dec|oct|lin] step1
      [parameter2 start2 stop2 [dec|oct|lin] step2]

    Middle value works only in one dimensional dc sweep, in two dimensional sweep it is ignored. If middle_val is given and it is between start1 and stop1 then dc analysis is performed in two steps. First dc analysis from middle_val to start1 is performed in backward direction. Then dc analysis from middle_val to stop1 is performed. The results are concatenated together. This approach is useful if the nodesets for the middle_val are given. In this case the speed and convergence are better.

    Parameter can be specified in the following manner:
    Instance parameters: @inst[param]    ... example is '@r1[resistance]'
    @inst           ... '@r1' would mean r1 default
                        parameter (resistance)
    inst            ... 'r1' would mean r1 resistance, 'v1' would
                        mean v1 dc voltage
    @inst[param][i] ... i-th component of vector parameter param of
                        instance inst
    Model parameters: @@mod[param]    ... example is '@rm[tc1]' ... tc1 parameter for
                        model rm
    @@mod           ... '@rm' would mean default parameter for model
                        rm (in case rm is a resistance model this
                        would be rsh)
    @@mod[param][i] ... i-th component of vector parameter param of
                        model mod
    Temperatures: @@@temp  ... circuit global temperature in degrees celsius
    @@@tnom  ... global nominal temperature in degrees celsius for
                 instance and model parameter recalculation
    @@@tempk ... circuit global temperature in Kelvins
    @@@tnomk ... global nominal temperature in Kelvins for instance
                 and model parameter recalculation
    In case dec/oct/lin is ommited, step is the linear step size. In case dec/oct is used, step is the number of points per decade/octave. In case lin is used, step is the number of points in range between start and stop (for 10, 11 points are evaluated including start and stop point).
    Examples:
    dc @r1[resistance] 1m 1meg dec 10 ... sweep resistance for
                                          instance r1 from 1m to
                                          1meg, 10 points per decade
    dc @@rm[rsh] 50 90 lin 100        ... sweep rsh for model rm from
                                          50 to 90 with 100
                                          equidistant steps
    dc @@@temp -20 50 0.01            ... sweep global temperature
                                          from -20 to 50 (degrees
                                          celsius), step is 0.01
  • Improved let syntax. LHS can now be specified using the following syntax: let plot.vector[index1]...[indexn] = expression Examples: let dc2.a[3]=5      ... vector a in plot dc2, assign 5 to 4th
                            element
    let dc2.a=vector(5) ... assign a vector with elements 0..4 to
                            vector a in plot dc2
    An arbitrary expression can be used for index.
    If plot name is ommited, the referenced vector is created or modified in current plot.
    Instance/model parameters can be altered using the let command: let @r1[resistance]=5000 ... set parameter resistance for
                                 instance r1 to 5000
    let @@rm[tc1]=0.05       ... set model parameter tc1 for model rm
                                 to 0.05
    The syntax for parameters is the same as with the DC analysis. @@@temp and @@@tnom are not supported yet.
    Index into a vector parameter can be an arbitrary expression: let @v1[coeffs][2*1]=0.01 ... set 3rd element of parameter coeffs
                                  for instance v1 to 0.01.
  • Extended syntax for vectors in expressions: let c=dc1.a*2    ... take vector a in plot dc1 and mutiply it by
                         2, assign the result to vector c in current
                         plot
    let c=dc1.a[2]*2 ... vector a in plot dc1, take 3rd element,
                         mutiply by 2 and assign the result to vector
                         vector c in current plot
    let c=a[2][3]*2  ... a is a 2D vector, indices into this vector
                         are 2 and 3
    Another example: let c=@r1[p][3]   ... This will take the 4th element from
                          vector @r1[p].
    Assigning into a vector element for an instance/model parameter: let @r1[p][3]=0.05 Of course such vectors can be referenced in other plots too: let a=dc1.@r1[p] Note again the use of braces in the above example when accessing a vector element: let a=(dc1.@r1[p])[3] Instance/model parameters can be used in expressions. Some examples: let c=@r1[resistance]
    let c=@@rm[tc1]
    let c=@v1[coeffs]      ... assign coeffs vector parameter value
                               of instance v1 to c
    let c=@v1[coeffs][2*1] ... assign coeffs vector parameter's 3rd
                               element (from instance v1) to vector c
  • To construct a vector from its components let a=(10;20;30)
    let b=(12;22;32;a;a*2;90)
    b would now be a vector with the following elements: 12 22 32 10 20 30 20 40 60 90.
  • Display simulator information (analyses, devices, (user defined) nodes): siminfo [analyses] [devices] [nodes] [all] Code models are marked with '++'
  • Load code models and user defined nodes from a .cm file: cmload filename.cm
  • New command nameplot renames the current plot. The following command sequence performs an op analysis and names the plot with the results firstop: op
    nameplot firstop
  • Keywords next and previous when using setplot set the current plot to the next/previous plot. Examples: setplot next
    setplot previous
  • Functions min() and max() return the smallest/largest component of a vector. For complex vectors the smallest/largest component is determined by comparing the absolute value of the components.
    Function sum() returns the sum of the components of a vector. Examples: * Real vector
    let a=(1;2;3;4)
    * Complex vector
    let b=((0,-1);(2,1);(0,0))

    * min() example
    * c <- 1
    let c=min(a)
    * c <- (0,0)
    let c=min(b)

    * max() example
    * c <- 4
    let c=max(a)
    * c <- (2,1)
    let c=max(b)

    * sum() example
    * c <- 10
    let c=sum(a)
    * c <- (2,0)
    let c=sum(b)
  • Variable badcktfree forces the simulator to release a circuit if any errors are found in the circuit. It is usually set in the spinit file.
  • Variable badcktstop prevents the simulator from running the script in the .control block if any errors are found in the circuit. It is usually set in the spinit file.
  • Netlist parsing completely rewritten. Improved error reporting. Added .lib clause. Example: File lib1.lib:

    .lib section1
    r1 1 2 1
    r2 2 3 1
    .endl
    .lib section2
    r1 1 0 1
    r2 3 0 1
    .endl


    File netlist:

    Test netlist
    * Include section1 from lib1.lib
    .lib 'lib1.lib' section1
    * Include section2 from lib1.lib
    .lib 'lib1.lib' section2
    .end
  • Let now allows vector length to be modified. Example: let a=(1;2;3)
    let a=(1;2)
    let a=(2;3;4;5)
    * a now contains (2;3;4;5)
  • To create a plot and name it use setplot new and then nameplot. The first vector you create in that plot becomes the default scale. setplot new
    nameplot plot10
    let a=(2;4;5;9)
    let b=(1;2;3;4)
    * a is the default scale for plot10
  • Vector operator [|low,high] selects a range from a vector that corresponds to the range low <= default_scale_value <= high on the default scale. Example: setplot new
    nameplot plot10
    let a=(2;4;5;9)
    let b=(1;2;3;4)
    * Select values for b for which the scale value is
    * between 2.5 and 5.5
    let x1=b[|2.5,5.5]
    * Select values for b for which the scale value is
    * between 4 and 5
    let x2=b[|4,5]
    * Both x1 and x2 contain now (2;3)
  • Graph tagging. A graph is automatically tagged as plotn if no tag is specified (iplotn for the plots generated by iplot commands). n is automatically set by Nutmeg and is unique for a plot in a SpiceOpus session. The plot tag is displayed in the title bar of the graph as tag: graph_name
    To create and tag a graph, use plot create tag vectors: * Create a graph, tag it plota and display
    * vectors a and b vs default scale
    plot create plota a b
    In case a graph with the given tag exists, the new graph is tagged with it. This makes it impossible to refer to the older graph.
    To append some more data to a graph, use plot append tag vectors: * Appends vector c to graph tagged plota
    plot append plota c
    To append without autoscaling and repainting the graph, use plot quickappend tag vectors: * Quickappends vector c to graph tagged plota
    plot quickappend plota c
    After the last quickappend call autoscale on the updated graph by typing plot autoscale tag: * Autoscales graph tagged plota
    plot autoscale plota
    In case there are scaling and graph type commands in the plot append clause, they are applied to the graph. If the graph doesn't exist (the user has closed it), the graph is recreated and tagged with the given tag.
    A graph can be closed with plot destroy tag: * Close graph plota
    plot destroy plota
  • Nutmeg function floor() returns an integer vector with largest components not exceeding the components of the operand. ceil() returns an integer vector with smallest components not below components of the operand. round() rounds the vector components to the nearest integer. In case the operand is complex, these functions operate on real and imaginary parts independently.
  • Vector operator [%index] returns an element with the corresponding index from the vector if the index is integer. If the index is real, linear interpolation is used to calculate the return value. Example: setplot new
    nameplot plot10
    let a=(2;4;5;9)
    let b=(1;2;3;4)
    * x1 becomes 4
    let x1=a[%1]
    * x2 becomes 2.5 (in the middle between components
    * b[1] (2) and b[2] (3))
    let x2=b[%1.5]
  • Nutmeg command cursor manipulates cursors. A cursor is a vector of length 1 (vector component) that represents a real index into a vector. To initialize a cursor, create a vector: * Creates two cursors, cleft and cright
    let cleft=0
    let cright=0
    To move a cursor to the left edge of a vector, use: * Moves cleft to the left edge and cright to the right edge
    * of vector 'time'
    cursor cleft left time
    cursor cright right time
    To move a cursor to a certain level on a vector, use: * Moves cleft towards right until it reaches
    * level 0.5 on vector pulse
    cursor cleft right pulse 0.5
    * Moves cright towards left until it reaches
    * level 0.5 on vector pulse
    cursor cright left pulse 0.5
    Cursors are real indices into vectors. Since time is the scale for vector pulse, the time difference between cleft and cright is the pulse width (if vector pulse contains only one pulse with bottom level 0 and top level 1). Pulse width can be obtained by: * See explanation of the [%index] vector operator
    let pulsewidth=time[%cright]-time[%cleft]
    To put it in short: the syntax for the cursor command is: * Initialize a cursor
    cursor vec_el left|right
    * Move a cursor
    cursor vec_el left|right vector_expression level_expression
    The cursor position is stored in vec_el (either a real vector of length 1 or a real vector element of the form vector_name[element_index]).
    The following example measures the rise time of a signal named pulse using time for scale: * Create scalar c1 and c2 (cursors)
    let c1=0
    let c2=0
    * Get the bottom and the top level of the pulse
    let bottom=min(pulse)
    let top=max(pulse)
    * Move c1 and c2 to the left edge of vector pulse
    cursor c1 left pulse
    cursor c2 left pulse
    * Move c1 to 10% level between bottom and top
    cursor c1 right pulse bottom+0.1*(top-bottom)
    * Move c2 to 90% level between bottom and top
    cursor c2 right pulse bottom+0.9*(top-bottom)
    * Calculate difference between c1 and c2 using time for scale
    let trise=time[%c2]-time[%c1]
    * Display result
    echo Rise time $&trise
    One can specify at which crossing of the signal the cursor should stop. * This will stop at 4th crossing of 0.5 in v(10) when moving towards right
    cursor c right v(10) 0.5 4
    Finally only crossings of patricular type can be counted. * This considers only crossings where v(10) is rising
    cursor c right v(10) 0.5 4 rising
    * This considers only crossings where v(10) is falling
    cursor c right v(10) 0.5 4 falling
    * This considers all crossings
    cursor c right v(10) 0.5 4 any
    If the number of intercepts is left out, it is assumed to be 1. If the type of crossing is left out all crossings are cosidered (just like with 'any').
  • Transient waveforms for independent voltage and current sources can be changed. For the purpouse of the explanation consider the following netlist: Transient source netlist
    v1 1 0 dc 0 pulse 0 1V 1ms 1us 1us 2ms 4ms
    r1 1 0 1
    To change the pulse width for v1 from 2ms to 3ms, type: let @v1[coeffs][5]=3m
    * This would also do the trick. Note that the following
    * assignment changes the type of the source to PULSE no matter
    * what the current type is whereas the previous assignment (using
    * the coeffs parameter) changes only the 6th parameter to 3m no
    * matter what type of transient source is currently set for v1.
    let @v1[pulse][5]=3m
    To change the transient waveform for v1 from pulse to sin with parameters 0 1 1kHz, type: let @v1[sine][5]=(0;1;1k)
    * or
    let @v1[sin][5]=(0;1;1k)
    The same goes for the independent current sources.
  • First iteration voltages for finding dc or initial transient solution can be set by nodeset command. Similarly the initial conditions for transient analysis can be set by ic command. Both commands have the same syntax: nodeset [delete | <plot_name> |
         v(<node_name_1>) = [value_1] [v(<node_name_2>) = [value_2]
         [v(<node_name_3>) = [<value_3>] ... ]]]
    and ic [delete | <plot_name> |
         v(<node_name_1>) = [value_1] [v(<node_name_2>) = [value_2]
         [v(<node_name_3>) = [<value_3>] ... ]]]
    Examples:
    Without any arguments commands nodeset and ic list defined nodesets or initial conditions, respectively. nodeset
    Nodesets:
    v(1)    = 1.000000e+001
    v(3:x1) = 5.000000e-001
    or ic
    Initial conditions:
    v(2)    = 3.000000e+000
    v(6:x1) = 1.000000e-001
    If one wants to delete nodeset for node 1 (or initial condition for node 6:x1), and at the same time set additional nodeset for node 4 to 3V (or initial condition for node 7:x2 to 2V), then one has to issue: nodeset v(1)= v(4)=3V or ic v(6:x1)= v(7:x2)=2V All nodeset voltages (or initial conditions) can be set to an arbitrary analysis results. Commands: nodeset op1 and ic tran3 set all nodeset voltages to values of vectors in op1 plot (results of operating point analysis) and all initial conditions to final values of vectors in tran3 plot (results of transient analysis).
    All nodeset or initial conditions can be removed by: nodeset delete or ic delete commands respectively.
  • Nutmeg function unwrap() unwraps the phase vector (a real vector of phase angles). It uses the current units setting (degrees or radians).
  • Arithmetic substitution enables you to replace the expression in the curly braces by its value prior to the execution of the Nutmeg command. All valid nutmeg functions and operators may be used in the expression. Avoid spaces. Expansion rules depend on the type of the result (real, complex, scalar, vector)
    The expansion rules are as follows {expression} -> result

    real scalar -> value
    complex scalar -> (rvalue,ivalue)
    real vector -> (value1;value2;...)
    complex vector -> ((rvalue1,ivalue1);(rvalue2,ivalue2);...)
    Example: set the voltage source DC parameter for sources V1, V2, ..., V20 to 0. let counter=1
    while counter le 20
    let @v{counter}[dc]=0
    let counter=counter+1
    endwhile
  • Function integrate() calculates new vector which is defined integral of the argument on its scale. For instance if v(3) is a time depended voltage (its scale is time), then vector integ_v3 obtained by integ_v3 = integrate(v(3)) represents:
          
    Before the integration takes place the original vector is approximated by a polynomial. The degree of polynomial approximation can be set with polydegree variable.
  • By default parameterised subcircuits are reparsed every time an arbitrary subcircuit parameter is changed. This is not convenient in cases when one wants to change more subcircuit parameters one after another. In this case the subcircuit shall be parsed after every parameter change instead only at the end. To avoid such situations (which typically arise in optimisation loops) manual subcircuit reparsing can be invoked by setting manualscktreparse variable. If the variable is set then subcircuit is reparsed only when required by scktreparse command. The syntax of the command is: scktreparse [<sckt_name_1> <sckt_name_2> ...] [<@sckt_mod_name_1> <@sckt_mod_name_2> ...]
  • Function area() will return total area of the specified circuit in future. At the moment only MOSFET devices are taken into account. Therefore currently area occupied by MOSFETs is returned. The function takes one argument which represents circuit number. The circuit number can be obtained by setcirc command. If circuit number is zero then area of current circuit is calculated.
  • Command copyplot copies data from source to destination plot. If destination plot does not exist it is created. The syntax is: copyplot <source_plot_name> <destination_plot_name>
  • SpiceOpus names items in subcircuits from bottom to top (e.g. node1:sckt1:sckt2). That is different from some other Spices which use top to bottom naming (e.g. sckt2:sckt1:node1). This fact is noted in the header of the output raw file by line Sckt. naming: from bottom to top.
  • All current subcircuit parameter values can be saved or restored by scktparams command. A parameterset (set of subcircuit parameters) is saved by get and restored by set subcommand. A particular parameterset can be printed by print subcommand or removed by remove subcommand. If parameterset_name is not given with print or remove subcommands then list of available parametersets is printed or all parametersets of the subcircuit are removed, respectively. The syntax of the command is: scktparams <sckt_name|@sckt_mod_name> get|set|print|remove [parameterset_name]
  • Plot stack.
    To push a reference to a plot onto the plot stack, use pushplot <name>. To pop a plot from the plot stack, use popplot.
    You can refer to the plot on the top of the stack with __topstack. Similarly you can refer to the newest plot using __topplot. __topstack and __topplot can be used in expressions, on the left side of the '=' in let command, with the destroy command, and with the setplot command.
    Plots pushed onto the stack cannot be destroyed. They must be popped first.
    destroy all clears the plot stack and deletes all plots with the exception of the constants plot (const).
    To display the contents of the plot stack, use plotstack.
  • Commands for copying, moving/renaming and deleting files are: copy source_filename destination_filename
    move from_filename to_filename
    remove filename1 [filename2 ...]
  • timer(x) function returns the difference between the CPU time in seconds and x. And clock(x) function returns the difference between absolute time in seconds and x. CPU time means time spent for SpiceOpus application.
  • Command destroyto removes all plots created after the plot passed by an argument. Its syntax is: destroyto plot_name If current plot was removed by destroyto command then plot in argument becomes current one.
  • Command script implements procedures in Nutmeg. The syntax is: script argument [script_name [command_line]] The second item, argument, can be add, delete, list or run.
    • add adds new line (Nutmeg command) to the script (procedure). It needs script_name and command_line arguments. Lines are always added at the end, after the last one.
    • delete without script_name deletes all scripts. If script_name is specified then only that script is deleted. command_line argument is not needed.
    • list without script_name lists names of currently available scripts. If script_name is specified then commands in that script are listed. command_line argument is not needed.
    • run executes commands in specified script. It needs script_name argument.
previous | back | home | next