Prev: Bitwise manipulation of data and memory issue
Next: Problems with parfor, sim and non local workers
From: Achal on 29 Jun 2010 05:14
Hi,
I am working with the following code

format long;
clear;
%
T=300; % temperature
% constants
h=6.626068e-34; % planck's constant
me=9.109e-31; % electronic mass
k=1.386065e-23; % boltzmann's constant
q=1.602e-19; % electronic charge
na=1e+17; % acceptor concentration
nd=0; % donor concentration
tox=2e-7; % oxide thickness
vfb=-0.9; % flat band voltage
phit=k*T/q; % thermal voltage
vg=0:20e-3:2; % gate voltage
psis1(101)=0; %
%psis2(101)=0; %
%psis3(101)=0; %
phib = phit*log(na/10e+10);
x=0;
for i=1:101
psis1(i)=fsolve(@(x)surf_pot_eq1(vg(i),vfb,x,na,nd,tox,300),phib);
%psis1(i)=fsolve(@(psi_s)surf_p_eq1(vg(i),vfb,gamma,Na,Nd),phib);
%psis1(i)=fsolve(@(psi_s)surf_p_eq1(vg(i),vfb,gamma,Na,Nd),phib);
end
figure(1)
plot(vg,psis1);


func [eq1] = surf_pot_eq1(gate_v,fb_v,surf_p,n_a,n_d,t_ox,temp)

h=6.626068e-34; % planck's constant
me=9.109e-31; % electronic mass
k=1.386065e-23; % boltzmann's constant
q=1.602e-19; % electronic charge
Ec=1.12; % conduction band energy
Ev=0; % Valence band energy
ed=45.5e-3; % Ec - Ed
ea=44.39e-3; % Ea - Ea
me_n=1.08*me; % electron effective mass
me_p=0.56*me; % hole effective mass
nc=(2*(2*pi*me_n*k*T(i))^1.5)/(h*h*h)*1e-6; % conduction band density of states
nv=(2*(2*pi*me_p*k*T(i))^1.5)/(h*h*h)*1e-6; % valence band density of states
Kox=3.9; % dilectric constant oxide
Ks=11.7; % dilectric constant silicon
eo=8.854e-14; % permittivity of free space
phit = K*temp/q; % thermal voltage
cox=Kox*eo/t_ox; % cap per unit area
phit=k*T/q; % thermal voltage
Ei=Ec/2+k*temp*3/(4*q)*log(me_p/me_n); % intrinsic fermi-level as function of temperature
ef=fsolve(@(fe)fe_b(n_a,n_d,ea,ed,nc,nv,fe,temp),0);% fermi-level estimation for a particular temperature
phi_b=Ei-ef;
po=nv*exp((Ev-ef)/phit);
gamma=sqrt(2*q*Ks*eo*po)/cox; % body effect coeficient

eq1 = fb_v + surf_p + gamma*sqrt(phit*exp(-surf_p/phit) + surf_p - phit + exp(-2*phi_b/phit)*(exp(surf_p/phit) - surf_p - phit)) - gate_v;

the dimensions of vector x match with the dimensions of vector phib, but I still get this error

??? Attempt to execute SCRIPT surf_pot_eq1 as a function.

Error in ==> surf_p_eq_check>@(x)surf_pot_eq1(vg(i),vfb,x,na,nd,tox,300) at 23
psis1(i)=fsolve(@(x)surf_pot_eq1(vg(i),vfb,x,na,nd,tox,300),phib);

Error in ==> fsolve at 180
fuser = feval(funfcn{3},x,varargin{:});

Error in ==> surf_p_eq_check at 23
psis1(i)=fsolve(@(x)surf_pot_eq1(vg(i),vfb,x,na,nd,tox,300),phib);

any kind of help will be really appreciated
Thanks
Achal




ellieandrogerxyzzy(a)mindspring.com.invalid (Roger Stafford) wrote in message <ellieandrogerxyzzy-0902061139470001(a)pool1459.cvx3-bradley.dialup.earthlink.net>...
> In article <ef28597.0(a)webx.raydaftYaTP>, gsboris <gsboris(a)yahoo.com> wrote:
>
> > gsboris wrote:
> > >
> > >
> > > while solving set of nonlinear equations "index out of bounds
> > > numel(x)=2" shows up if I have more then 2 equations
> > Even set of > 2 linear equations gives that error
> >
> > function file:
> > function F = myfun(x)
> > F = [x(1) - x(2) + x(3);
> > 2*x(1) - x(2) + x(3);
> > 12*x(1) - x(3) + 20*x(3)];
> >
> > file calling for set of equations to solve it using fsolve:
> > options=optimset('Display','iter');
> > [x,fval] = fsolve(@myfun,x0,options);
> > return:
> > ??? Attempted to access x(3); index out of bounds because numel(x)=2.
> > Error in ==> myfun at 3
> > F = [x(1) - x(2) + x(3);
> > Error in ==> fsolve at 180
> > fuser = feval(funfcn{3},x,varargin{:});
> > Error in ==> mytry at 8
> > [x,fval] = fsolve(@myfun,x0,options);
> >
> > Any help out there?
> -----------------
> It looks as though you have called on 'fsolve' with the variable, 'x0',
> not set to a vector with three elements. 'x0' is supposed to be the
> vector that 'fsolve' starts with in its search for a solution and it
> should be the same size as the vector 'myfun' uses as input.
>
> Your 'myfun' looks to me like a set of homogeneous linear equations
> whose only solution would be x = 0. Is that what you intended?
>
> (Remove "xyzzy" and ".invalid" to send me email.)
> Roger Stafford
From: Achal on 29 Jun 2010 05:24
Hi,
I am working with the following code

format long;
clear;
%
T=300; % temperature
% constants
h=6.626068e-34; % planck's constant
me=9.109e-31; % electronic mass
k=1.386065e-23; % boltzmann's constant
q=1.602e-19; % electronic charge
na=1e+17; % acceptor concentration
nd=0; % donor concentration
tox=2e-7; % oxide thickness
vfb=-0.9; % flat band voltage
phit=k*T/q; % thermal voltage
vg=0:20e-3:2; % gate voltage
psis1(101)=0; %
%psis2(101)=0; %
%psis3(101)=0; %
phib = phit*log(na/10e+10);
x=0;
for i=1:101
psis1(i)=fsolve(@(x)surf_pot_eq1(vg(i),vfb,x,na,nd,tox,300),phib);
%psis1(i)=fsolve(@(psi_s)surf_p_eq1(vg(i),vfb,gamma,Na,Nd),phib);
%psis1(i)=fsolve(@(psi_s)surf_p_eq1(vg(i),vfb,gamma,Na,Nd),phib);
end
figure(1)
plot(vg,psis1);


func [eq1] = surf_pot_eq1(gate_v,fb_v,surf_p,n_a,n_d,t_ox,temp)

h=6.626068e-34; % planck's constant
me=9.109e-31; % electronic mass
k=1.386065e-23; % boltzmann's constant
q=1.602e-19; % electronic charge
Ec=1.12; % conduction band energy
Ev=0; % Valence band energy
ed=45.5e-3; % Ec - Ed
ea=44.39e-3; % Ea - Ea
me_n=1.08*me; % electron effective mass
me_p=0.56*me; % hole effective mass
nc=(2*(2*pi*me_n*k*T(i))^1.5)/(h*h*h)*1e-6; % conduction band density of states
nv=(2*(2*pi*me_p*k*T(i))^1.5)/(h*h*h)*1e-6; % valence band density of states
Kox=3.9; % dilectric constant oxide
Ks=11.7; % dilectric constant silicon
eo=8.854e-14; % permittivity of free space
phit = K*temp/q; % thermal voltage
cox=Kox*eo/t_ox; % cap per unit area
phit=k*T/q; % thermal voltage
Ei=Ec/2+k*temp*3/(4*q)*log(me_p/me_n); % intrinsic fermi-level as function of temperature
ef=fsolve(@(fe)fe_b(n_a,n_d,ea,ed,nc,nv,fe,temp),0);% fermi-level estimation for a particular temperature
phi_b=Ei-ef;
po=nv*exp((Ev-ef)/phit);
gamma=sqrt(2*q*Ks*eo*po)/cox; % body effect coeficient

eq1 = fb_v + surf_p + gamma*sqrt(phit*exp(-surf_p/phit) + surf_p - phit + exp(-2*phi_b/phit)*(exp(surf_p/phit) - surf_p - phit)) - gate_v;

the dimensions of vector x match with the dimensions of vector phib, but I still get this error

??? Attempt to execute SCRIPT surf_pot_eq1 as a function.

Error in ==> surf_p_eq_check>@(x)surf_pot_eq1(vg(i),vfb,x,na,nd,tox,300) at 23
psis1(i)=fsolve(@(x)surf_pot_eq1(vg(i),vfb,x,na,nd,tox,300),phib);

Error in ==> fsolve at 180
fuser = feval(funfcn{3},x,varargin{:});

Error in ==> surf_p_eq_check at 23
psis1(i)=fsolve(@(x)surf_pot_eq1(vg(i),vfb,x,na,nd,tox,300),phib);

any kind of help will be really appreciated
Thanks
Achal




ellieandrogerxyzzy(a)mindspring.com.invalid (Roger Stafford) wrote in message <ellieandrogerxyzzy-0902061139470001(a)pool1459.cvx3-bradley.dialup.earthlink.net>...
> In article <ef28597.0(a)webx.raydaftYaTP>, gsboris <gsboris(a)yahoo.com> wrote:
>
> > gsboris wrote:
> > >
> > >
> > > while solving set of nonlinear equations "index out of bounds
> > > numel(x)=2" shows up if I have more then 2 equations
> > Even set of > 2 linear equations gives that error
> >
> > function file:
> > function F = myfun(x)
> > F = [x(1) - x(2) + x(3);
> > 2*x(1) - x(2) + x(3);
> > 12*x(1) - x(3) + 20*x(3)];
> >
> > file calling for set of equations to solve it using fsolve:
> > options=optimset('Display','iter');
> > [x,fval] = fsolve(@myfun,x0,options);
> > return:
> > ??? Attempted to access x(3); index out of bounds because numel(x)=2.
> > Error in ==> myfun at 3
> > F = [x(1) - x(2) + x(3);
> > Error in ==> fsolve at 180
> > fuser = feval(funfcn{3},x,varargin{:});
> > Error in ==> mytry at 8
> > [x,fval] = fsolve(@myfun,x0,options);
> >
> > Any help out there?
> -----------------
> It looks as though you have called on 'fsolve' with the variable, 'x0',
> not set to a vector with three elements. 'x0' is supposed to be the
> vector that 'fsolve' starts with in its search for a solution and it
> should be the same size as the vector 'myfun' uses as input.
>
> Your 'myfun' looks to me like a set of homogeneous linear equations
> whose only solution would be x = 0. Is that what you intended?
>
> (Remove "xyzzy" and ".invalid" to send me email.)
> Roger Stafford
From: Achal on 29 Jun 2010 05:42
I think I figured the problem with the code function was not properly defined. Issue with the syntax in defining the function.

Regards
Achal
 | 
Pages: 1
Prev: Bitwise manipulation of data and memory issue
Next: Problems with parfor, sim and non local workers