converting orbits with e>1 to KEP gives null values for covariances.

[Little-Ryugu]

When converting orbits with e > 1 to KEP using layup convert, the covariance matrix is not converted and is output as null values. Should we fix this or give a warning/error to the user?

[matthewholman]

Could you check is the conversion to COM (cometary) orbits works okay? KEP is assuming the orbit has e<1, so that there is a positive semimajor axis.

[Little-Ryugu]

Just checked there for all orbital formats. The covariance matrix only outputs nan values for BKEP and KEP. BCART, CART, COM and BCOM all work properly to convert the cov matrix.

I think the nans are created from this line in the code, when the semi-major axis is negative from e > 1 orbits

layup/src/layup/utilities/orbit_conversion.py

Line 619 in 52fb952

M = (epochMJD_TDB - tp) * jnp.sqrt(mu / a**3)

[matthewholman]

It seems that we aren't properly defending against BKEP and KEP orbits with a<0. We should be using BCOM and COM for those cases.

…

On Tue, Jul 15, 2025 at 2:11 PM Ryan Lyttle ***@***.***> wrote: *Little-Ryugu* left a comment (Smithsonian/layup#288) <#288 (comment)> Just checked there for all orbital formats. The covariance matrix only outputs nan values for BKEP and KEP. BCART, CART, COM and BCOM all work properly to convert the cov matrix. I think the nans are created from this line in the code, when the semi-major axis is negative from e > 1 orbits https://github.com/Smithsonian/layup/blob/52fb952ad985617844ecc96d30d239a459df6247/src/layup/utilities/orbit_conversion.py#L619 — Reply to this email directly, view it on GitHub <#288 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA5CVWODB2D2MBIPCK3Q4GT3IVU7RAVCNFSM6AAAAACBSDW6SCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTANZVGY3TMOBWHA> . You are receiving this because you commented.Message ID: ***@***.***>

-- Matthew J. Holman, PhD Senior Astrophysicist Center for Astrophysics | Harvard & Smithsonian 60 Garden Street, MS #51 Cambridge, MA 02138 (617) 496-7775