LAPACK different outputs when using solvers for banded matricies

I've been stuck at this for hours and hoping someone can figure out what I am missing. I am solving Ax=B firstly using DGESV which I am 99% sure is correct. Then I am puting A into a banded form and using DGBSV and getting a different output.

Here is my expected inputs and outputs where A is the first 9x9 matrix B is -r 1x9 and x is sk 1x9.

And this is what my usage of DGBSV looks like A=JBtemp,B=-r,x=SK:
[(https://i.stack.imgur.com/hiVTJ.png)

We can see that the x values dont match and here is my code usage of the two functions.

call dgesv(n,1,J0,n,ipiv,r,n,info)

call dgbsv(n,KL,KU,1,JBtemp,KU+1+KL+KL,ipiv,r,n,info)

This is a non-problem.

Both results are correct - yes, really! - for the data that you have given them. I checked that with Gaussian elimination.

In your B vector (what you label -r) in the first example it ends

50.1899...
70.9...
50.19...

but in the second example it ends

50.1899...
-70.9...
-50.19...

Note the change from positive to negative for the last two entries of B.