NOT A BUG: Wrong interaction point smearing

EDIT: 2022-07-13, this is not a bug - see comments for explanation.

Dear Support,

We in our team have found that the Gauss interaction point smearing works wrongly in MPDRoot.
This smearing is enabled by default on lines 77-81 of the runMC.C macro
https://git.jinr.ru/nica/mpdroot/-/blob/dev/macro/mpd/runMC.C

     // smearing of beam interaction point
     primGen->SetBeam(0.0,0.0,0.1,0.1);
     primGen->SetTarget(0.0,24.0);
     primGen->SmearGausVertexZ(kTRUE);
     primGen->SmearVertexXY(kTRUE);

But instead of pure Gaussian shape we observe a peak around 24 cm in the distribution of the Z coordinate of the interaction point.

This problem is observed with different model interfaces: HSD, UrQMD etc, but not with the BOX generator.

How can we resolve it and obtain pure Gauss?

Best regards, Viktar