Loss of Excitation Settings

[mls1]

We're reviewing existing settings for a loss of excitation protection scheme and have found that the offset and mho circles were developed based on unsaturated reactances rather that saturated values. This doesn't make sense to us since system voltage should be near nominal. Would this be done in the case of a week GSU where generator side voltage may get low if the unit is absorbing significant vars? There is no explanation as to why it was set this way so we are looking to see if this is a common practice. It makes the circles bigger so it is more conservative.