curve was measured with an 0.3uS flat top to demonstrate the ultimate throughput of the system to exceed 100KHz. This produced a 122keV peak with a resolution of 3KeV (not shown) , which was very stable in FWHM and in position with increasing count rate. While of academic interest here, such a performance trade-off may be valuable elsewhere, where throughput is more important than resolution. Figure 6 shows that the system resolution performance meets the coaxial detector warranted 1KHz count Figure 6 rate performance specifications of 870 eV at 2uS shaping (4uS risetime) even at 30 kHz. The resolution is seen to degrade towards shorter shaping times due to the charge collection time in the coaxial detector crystal being longer than the integration time.
Figure 7 shows the variation in 122 keV resolution with count rate and rise time.
In general, the shorter shaping times give poorer resolution and better stability as would be expected.
For the 4 S data, even at the point of maximum throughput of 28 kHz at 75 kHz input rate (Figure 5), Figure 7 shows that the resolution at 900 eV is only 20 eV worse than the detector warranted value at 30 kHz input rate.
Figure 8 shows peak shift versus count rate. The stability for short shaping times is better, the tradeoff being resolution.