Use uncalibrated energy as waveform max instead of actual max for timepoints

[hervasa2]

We should take care to get all PSD parameters from normalized waveforms. Here the normalization factor (E) should be the same across the processing chain (A/E, QDrift/E, ...) and equal to a filtered maximum of the waveform in ADC.
Therefore in:

LegendDSP.jl/src/dsp_icpc.jl

Lines 138 to 142 in acb4c56

	t10 = get_threshold(wvfs, wvf_max .* 0.1; mintot=config.kwargs_pars.tx_mintot)
	t50 = get_threshold(wvfs, wvf_max .* 0.5; mintot=config.kwargs_pars.tx_mintot)
	t80 = get_threshold(wvfs, wvf_max .* 0.8; mintot=config.kwargs_pars.tx_mintot)
	t90 = get_threshold(wvfs, wvf_max .* 0.9; mintot=config.kwargs_pars.tx_mintot)
	t99 = get_threshold(wvfs, wvf_max .* 0.99; mintot=config.kwargs_pars.tx_mintot)

wvf_max should be replaced by e_cusp or e_trap