Number of doublets is equal to nExp

[der4005]

I've run the DoubletFinder twice on my data, with and without prefiltering cells based on number of features. In both instances, the number of doublets identified is exactly equal to nExp (despite nExp being different in the two examples). Is this normal behavior? Are there ever instances in which DoubletFinder finds a number of doublets below or above nExp? It was also interesting that the identified doublets are fairly low on the nFeature & nCount distributions, which I wouldn't have expected:

I'm using a split-pool combinatorial barcoding protocol that has a much lower doublet rate than droplet-based protocols (https://support.parsebiosciences.com/hc/en-us/articles/360053107311-What-is-the-expected-doublet-rate). I used a 2% rate to calculate nExp, but I'm wondering if actual doublets in my data are lower than this. Is there a way to determine a reasonable nExp based on the analysis performed by DoubletFinder? For example, I checked the distribution of pANN:

and it doesn't have the same tail shown in this picture:

I'm not sure if this is meaningful though.