matlab simultaneous equations question

Hey Guys,

I am new to Matlab and i am having trouble getting this set of simultaneous equations to work, i have 7 equations and 7 variables and when i run the program matlab does not return an answer but i am pretty sure that this equation should have an answer, here is my code any help would be appreciated.

syms i_ph I_01 I_02 R_s a_1 a_2 R_p;

%a_1 = ((n_1*25)/1.6e-19);
%a_2 = ((n_2*25)/1.6e-19);

g_1 = -4.75e-4 + i_ph+ I_01*(exp((4.75e-4*R_s)/a_1)-1)-I_02*(exp((4.75e-4*R_s)/a_2)-1)-((4.75e-4*R_s)/R_p);
g_2 = -4.8964e-4 + i_ph+ I_01*(exp((0.15+4.8964e-4*R_s)/a_1)-1)-I_02*(exp((0.15+4.8964e-4*R_s)/a_2)-1)-((0.15+4.8964e-4*R_s)/R_p);
g_3 = -5e-4 + i_ph+ I_01*(exp((0.3+5e-4*R_s)/a_1)-1)-I_02*(exp((0.3+5e-4*R_s)/a_2)-1)-((0.3+5e-4*R_s)/R_p);
g_4 = -3.5775e-4 + i_ph+ I_01*(exp((0.45+3.5775e-4*R_s)/a_1)-1)-I_02*(exp((0.45+3.5775e-4*R_s)/a_2)-1)-((0.45+3.5775e-4*R_s)/R_p);
g_5 = -5.1839e-4 + i_ph+ I_01*(exp((0.5282+5.1839e-4*R_s)/a_1)-1)-I_02*(exp((0.5282+5.1839e-4*R_s)/a_2)-1)-((0.5282+5.1839e-4*R_s)/R_p);
g_6 = -5.2284e-4 + i_ph+ I_01*(exp((0.5822+5.2284e-4*R_s)/a_1)-1)-I_02*(exp((0.5822+5.2284e-4*R_s)/a_2)-1)-((0.5822+5.2284e-4*R_s)/R_p);
g_7 = i_ph+ I_01*((exp(0.6483)/a_1)-1)-I_02*((exp(0.6483)/a_2)-1)-((0.6483)/R_p);

solution = solve(g_1,g_2,g_3,g_4,g_5,g_6,g_7,i_ph,I_01,I_02,R_s,a_1,a_2,R_p);
solution.i_ph;
double(solution.i_ph);

Answer 1

[Walter Roberson]

When you use floating point values like that, then solve() is going to convert them into rational fractions for the purpose of finding a solution. However, the fraction it constructs is going to be based upon converting the double precision representation of values such as 1.69e-19 into fractions. Most of those numbers do not have exact binary representations, so the fractions are not going to be what you would expect. For example, 4.75e-4 is probably not going to be converted to 475/10000. So even if you managed to get a solution from solve() based upon those converted fractions, it would be the solution to a different problem then you thought you were posing.

You would probably be better off trying to use fsolve() if you have the Optimization Toolkit.