Wrong solution using solve with a logarithmic

Hello,

I wanted to solve an equation containing a log function. The solution given by Matlab is not correct, until I multiply the logarithm by 10.

I'm aware of the difference between log and log10 and want to use the natural logarithm.

f_1 = 10* (eta_p/(b_pstern * g * eps_min))*(log(W_0/(W_0-B1)))== s*1000  ;
W_K_1 = double(solve(f_1, B1))

In the equation, everything is numeric except for B1. When I type the exact same equation into my calculator, I get the correct answer. The "10*" is the correction factor I use for the "correct" solution.

Am I using the log function wrong or is there some other error?

[dpb]

Would have to have a complete test case including constants to do any diagnositics...

[Matthew Worker]

Thank you for your response. The code with all the variables would be as following :

syms B1

f_1 = (6314642578216891*log(-35316/(B1 - 35316)))/536870912 == 300000 % With *10
f_2 = (5051714062573513*log(-35316/(B1 - 35316)))/4294967296 == 300000  % Without *10

ans = double(solve(f_1, B1))
ans2 = double(solve(f_2, B1))

f_1 is the corrected formula, as it offers the correct solution. f_2 is the formula that I need to use, but f_2 outputs the wrong solution. If I type f_2 into my calculator, I get the solution that Matlab outputs for f_1.

I hope you can follow my explanation!

Answer 1

[Walter Roberson]

The outputs cross-check properly

format long g
syms B1
f_1 = (6314642578216891*log(-35316/(B1 - 35316)))/536870912 == 300000 % With *10

f_1 = 

f_2 = (5051714062573513*log(-35316/(B1 - 35316)))/4294967296 == 300000  % Without *10

f_2 = 

f1_sol = solve(f_1)

f1_sol = 

f2_sol = solve(f_2)

f2_sol = 

double(f1_sol)

ans = 
          889.379424789019

subs(f_1, B1, f1_sol)

ans = 

lhs(ans)-rhs(ans)

ans = 

0

double(f2_sol)

ans = 
          7950.69092580946

subs(f_2, B1, f2_sol)

ans = 

lhs(ans)-rhs(ans)

ans = 

0