as far as i know daisy just provides a single feto-maternal dataset, which limits the possibilities of usage (although even at physionet there is only limited data available; particularly the CINC2013 dataset would be big but as far as i remember it does not contauni a seperate maternal channel).
If you are working on separating maternal and fetal ECG you should consider to use the simulated data we provide via this website:
http://fernandoandreotti.github.io/fecgsyn/
The provided data contains rather realistic simulations on varying signal quality and physiological changes, which have been used elsewhere so that you can compare your methods. Moreover the provided tools even allow you to create your own scenarios.