Problem regarding a sample problem

[soumyadeep1997]

I am facing a problem regarding solving a stiff equation with all the optional inputs set as defaults. I am a materials engineer by profession and not a coding ninja so I apologise if my doubt seems trivial and irrelevant.
The equations are,
ydot(1)= 9y1 +24y2 +5cos t -(1/3)sin t; IV: y1(0)=(4/3)
ydot(2)= -24y1 -51y2 -95cos t + (1/3)sin t ; IV: y2(0) = (2/3)
Pls help and forgive my non expertise in these affairs. It is giving an output of:
at t= 1.0000e-01 y= -1.715510e+00 -1.431422e+00
at t= 2.0000e-01 y= -5.686007e+00 5.102048e-01
at t= 3.0000e-01 y= -8.586810e+00 2.012284e+00
at t= 4.0000e-01 y= -1.061630e+01 3.102939e+00
at t= 5.0000e-01 y= -1.195806e+01 3.871763e+00
at t= 6.0000e-01 y= -1.275005e+01 4.386789e+00
at t= 7.0000e-01 y= -1.309458e+01 4.697948e+00
at t= 8.0000e-01 y= -1.307155e+01 4.843806e+00
at t= 9.0000e-01 y= -1.274265e+01 4.853635e+00
at t= 1.0000e+00 y= -1.215648e+01 4.749991e+00
at t= 1.1000e+00 y= -1.135248e+01 4.550703e+00
at t= 1.2000e+00 y= -1.036358e+01 4.270213e+00
This is wrong result which you can verify from the original solution.

[ChrisRackauckas]

Do you lower the tolerances and check if it gets closer to the true solution? Numerical integrators have numerical error which has to be controlled but is never zero.

[soumyadeep1997]

I did that, also changed step size, the issue may be that custom jacobian cannot be inputted in this. I am at total loss regarding the solution. Thank you for your suggestion though sir.

[dilawar]

You are doing something odd; may be you can post your program.

I get the following solution at t=1 with rtol=1e-4 and atol=1e-6.

at t= 1.0000e+00 y= -1.210928e+01 4.722848e+00

I solved your system with xppaut as well and these results are close. XPPAUT file is here https://github.com/dilawar/playground/blob/32c28f6a118f1994052cce7759cffa56041fbf7a/XPPAUT/sys.ode .