Ensemble Time Correlated Single Photon Counting (eTCSPC) measurements were performed using either an IBH-5000U (IBH, Scotland) or a Fluotime 200 (Picoquant, Germany) system. The excitation source of the IBH machine were a 470 nm diode laser (LDH-P-C470, Picoquant, Germany) operating at 10 MHz for donor excitation. The excitation and emission slits were set to 2 nm and 16 nm, respectively. The excitation source of the Fluotime200 system was a white light laser (SuperK extreme, NKT Photonics, Denmark) operating at 20 MHz at 485 nm excitation and emission slits set to 2 nm and 5 nm, respectively. Additionally, in both systems, cut-off filters were used to reduce the amount of scattered light (>500 nm for donor emission). For green detection, the monochromator was set to 520 nm.
 
All measurements were conducted under magic angle conditions (excitation polarizer 0 degree, emission polarizer 54.7 degree). In the IBH system, the TAC-histograms were recorded with a bin width of 14.1 ps within a time window of 57.8 ns, while the Fluotime200 was set to a bin width of 8 ps within a time window of 51.3 ns. Photons were collected up to a peak count of 100’000 corresponding in average to a total number of 30*10^6 photons. The instrument response function IRF (~230 ps FWHM for the IBH, ~ 150 ps for the Fluotime200) was collected under the same recording settings at the excitation wavelength of the sample without cutoff-filters using a scattering Ludox-dispersion, which yielded a comparable count rate as the later on measured samples. For the IBH system, it was needed was performed before each measurement session a reference measurement with a continuous light signal to account for the differential non-linearity of the counting electronics. 

The recorded uncorrelated photons yield a reference histogram that is ideally constant. After recording of this measurement, the average number of photons in each time-bin is calculated. Next, the measurement was smoothed by a window function using a Hanning-filter with a window-size of 17 bins. The smoothed decay histogram was normalized to the previously calculated average number of photons. Instead of correcting the experimental histogram the model function is multiplied by the smoothed and normalized reference histogram to preserve the Poissonian statistics of the measured fluorescence intensity histograms of interest.

Note: hGBP1-254-577 measurements refer to subensemble TCSPC measurements of the respective variant