Error in PN series for angular momentum flux along Schwarzschild eccentric orbit at 3PN

[nielsw2]

It looks like there is error in the 3PN angular momentum flux for eccentric orbits. It does not seem to match the paper -- Eq. (5.6) of https://arxiv.org/abs/2005.03044. You can also see the issue by comparing the circular limit of the eccentric PN series to the circular orbit series -- see the below code. In this limit the difference seems to only appear in the 3PN term with all other terms matching through at least 9PN.

<< PostNewtonianSelfForce`  
CircFluxPN=PostNewtonianExpansion[{"/Schwarzschild/Circular/Flux/Infinity","!-l"}];
eccLFlux=PostNewtonianExpansion[{"/Schwarzschild/Eccentric/Flux/LInfinity"}];
Normal[(32/5 y^(7/2))^-1 y^(-3/2 ) CircFluxPN["Series"]+O[y]^9];
Normal[ReleaseHold[(32/5 y^(7/2))^-1 eccLFlux["Series"]+O[y]^9/.highestPower->2]]/.e->0;
% - %% // Simplify

which gives the result

(379223 y^3)/2520

[nielsw2]

A sign error was found by @cmm0052 in the 3PN Ldot flux. This has been fixed in commit 5f3edad.

I have tested the changes and the circular limit now agrees with the circular PN series. I also checked that the eccentric result tends to the numerical value.

I will close this issue now.