beamforming in preprocessing stage fails due to sens chanunits being different

[zits69]

Hi!

After normalization i run D = osl_inverse_model(D,p.template_coordinates,'pca_order',50);

Error msg:
Running 'Prepare data'
Done 'Prepare data'
Running 'Define sources'
computing surface normals
Failed 'Define sources'
Error using ft_scalingfactor (line 181)
cannot convert T to fT/mm

I've looked at the BF.mat and inside the data.MEG.sens.chanunit and some are 'T' and some are 'T/m' but it's not consistent in any way (ie: the T/m are peppered seemingly randomly).
The D object that goes in has the expected units {'fT'} {'fT/mm'}.

Could this have to do with the fact that i entered the pipeline at the normalization step with my own already preprocessed data (same order of steps as in the HMM example here)? The normalization runs fine and my D does have an 'inv' field so I can't see what may be missing...

Thanks
z