Question: Computation of Lauricella function

Hi All,

I am trying to compute the Lauricella function of type F_A [a; 1,...,1;  b1,...,bn;  x1,...xn] using the Burchnall–Chaundy expansion which basically reduces it to the product of Gauss hypergeometric functions. Convergence of the series is guaranteed as |x1| + ... |xn| < 1 always. This approcah seems to give correct result when |x1| + ... |xn|  is around 0.9 or less, and any increase above 0.9 results in rapid deviation from the expected value as observed from the output (obtained result is less than the expected value by some orders). Is it an issue due to the convergence or something else?.  Can anybody comment on the above issue as my obesrvation above is by repeated trials only. Any views on how I can take care of this issue or atleast increase the range of accurate result?

Also is there any specfic values for Lauricella functions available anywhere so that I can check the accuracy of my computation because as of now I am using some crude method of checking the accuracy of the output.

Thanks a lot in advance.

Best regards,

Vish