During experiments at CERN it has occurred sometimes that an excess of events has been announced for a very concrete energy range, but additional statistical analysis, or renewed experimental input, has smoothed down the initial deviations. This author is aware of three such events. In 1984 the experiment UA1 got preliminary excess in the range of 40 GeV \cite{UA1} (see also \cite{nature}). During the mid running of LEP 2, experiment L3 got some excess in the range of 68 GeV \cite{L3}. And finally, perhaps the more popular, during the last run of LEP 2 the ALEPH collaboration, and also L3, got excess signal in the range around 115 GeV \cite{aleph}. Our purpose here is to examine if all the three signals can be linked to an unlocated calibration problem. It could happen if an unforeseen non-uniformity in distributing energy causes some background events to accumulate, simulating to be a signal. This problem can happen if a naive nuclear mass model is used at some point in the physics calculations in detectors. We all known than nuclear magic numbers, causing microscopical corrections in mass and level calculations, occur at Z=28, 40, 50, 82 and 114, as well as at N=28, 50, 82, 126 and 284. It is a less know fact \cite{rivero} that the highly unstable nuclei, beta plus decaying, having these Z numbers have approximately, within a 2% or less, the same mass that the corresponding, beta minus, highly unstable having these N numbers. The corresponding nuclear masses are about 45, 68, 91, 115, 175 and 246 GeV. Thus if a calibration or a detector depends, say, on secondary decay from near proton drip-line nuclei, a mass model taking these masses into account should be used for the physics of the material. The same would happens with secondary decays near neutron drip line, but this line is rarely reached experimentally. If simple analytic approximations, such as Weisaker formula, are used instead, we can expect errors to happen related to these mass values. As all the CERN experiments share legacy code both for simulations and actual calibration of measurements, it could be that such approximation were hidden in some shared code. Of the four values in CERN reach, three of them correspond very accurately to the troubled event excesses, while the other is just under the peak of the Z0 particle, so that a error there should be masked under the huge quantity of real Z0 events. A second possibility is that CERN procedures are correct and all the three deviations correspond to physical events. This is a proposition a lot more unlikely to accept because it implies that nuclear data to be excessively sensible to particles beyond the nuclear scale. But it could be claimed on grounds of the two extant masses, 175 and 246 GeV, which do not correspond to CERN measures but are also commonly assocciated to high energy physics. 246 Ge could be any particle with a coupling such that its mass coincides with the vacuum expected value of the higgs field, and 175 GeV is the area of the top mass, or its associated mesons. Additional support for the physical reality of the excesses could be coming now from HERA \cite{carli} where some events have been reported about 40 GeV. A common bug between the code of HERA and CERN is more unlikely that between CERN experiments. \bibitem{UA1} UA1 Collaboration, {\it Associated production of an isolated, large-transverse-momentum lepton (electron or muon), and two jets at the CERN p\-p collider.} Phys. Lett. B. 147 (1984) p. 493-508 \bibitem{L3} P. Garcia-Abia, {\it Search for charged Higgs bosons in L3} {\tt http://arxiv.org/abs/hep-ex/0105057} \bibitem{aleph} ALEPH Collaboration, {\it Observation of an excess in the ALEPH search for the Standard Model Higgs boson} {\tt http://arxiv.org/abs/hep-ex/hep-ex/010 \bibitem{carli} T. Carli, D. Dannheim, L. Bellagamba {\it Events with Isolated Charged Leptons and Large Missing Transverse Momentum at HERA} {\tt http://arxiv.org/abs/hep-ph/0402012} \bibitem{rivero} A. Rivero {\it Distance to the drip lines} unpubl.