matrix mis-match

Asked by sanjiv kumar

sanjiv kumar (view profile)

on 12 Sep 2019
Latest activity Answered by Guillaume

Guillaume (view profile)

on 12 Sep 2019
Accepted Answer by Guillaume

Guillaume (view profile)

M= 100; % gain for Si and 50 for Ge
R_d= 75; % responsivity for Si and 35A/W for Ge
q= 1.6*10.^-19; % charge
F_noise = 0.5; % nosie figure 0f Si and 0.95 for Ge
I_d = 15 * 10.^-9; % dark Current for Si and 700nA for Ge
B_e = 5*10.^9; % bandwidth for Si and 0.2GHz for Ge
k_B = 1.38*10.^-23; % boltzmann constant
T = 298 ; % tamperature in kelvins
R = 10*10.^3; % APD load resistance
k = -40:20:80;
K_w = 1*10.^-3 * 10.^( k / 10);
Pr1 = (K_w*0.75*0.75*(exp(-0.003*25./cos(pi/36))) .* 0.01 .* cos(pi/36) ./ ((2*pi*25.^2)*(1-cos(pi/3))));
I_p= sqrt(M*R_d*Pr1); %average photocurrent
I_short= sqrt(2*q*I_p*M.^2*F_noise);
I_dark = sqrt(2*q*I_d*M.^2*F_noise*B_e);
I_thermal = sqrt((4*k_B*T/R)*B_e);
I_noise1 = sqrt(I_short1 + I_dark + I_thermal);
SNR1 = I_p.^2*M^2/I_noise1;
SNR_db1 = 1.*log(SNR1);
figure
plot(k,SNR_db1,'-ro')
Error using /
Matrix dimensions must agree.

Adam Danz (view profile)

on 12 Sep 2019
We don't have the value for "I_short1" so we cannot run your code.
However, if I give it a scalar value, the code runs without error.

Guillaume (view profile)

on 12 Sep 2019

There doesn't appear to be any logic behind the choice of * vs .*, / vs ./, etc. If you don't understand the difference between the two and just try one or the other until it works, you're not going to go very far. Or if you don't know what is vector or scalar in your code, similarly, you'll have problems.
Few of the dotted operations in your expressions actually need to be dotted. But if you chose to dot the scalar expressions, then you should dot them all. Be consistent.
Also, note that a more concise way to write constants of the form 1.6*10^-19 (I removed the unnecessary dot) is 1.6e-19
As Adam said, you haven't told us what I_Short1 is, but it must be non-scalar for you to get an error. If so, your expressions should probably be:
SNR1 = I_p.^2 * M^2 ./ I_noise1;
Of course, I_noise1 must be a row vector with the same numbers of elements as K (7 elements).
By the way, why not:
K = -4:2:8; %why have it -40:20:80
K_w = 1e-3 * 10.^K; %if you're going to divide it by 10 on the next line