Configuration:

*The first value is the delay set on the pulse generator; the second one is the actual delay.

@Channel-5 result is not reported in this page.

No ND filter

Integration time = 6min for each data

Data processing: dead-time correction, background correction.

Note the x axis ranges for channel-1 and channel-2 are different.

Analysis: In this experiment, the configuration of channel-1 keeps the same so that we can use it as a reference to distiniguish the atmospheric variation. We have drew a conclusion in ND test#2 that the channel-1 response is basically linear. The only problem I see for channel-1 here is that the data1/data2 of analog signal has a positive signal. Is this due to the atmospheric variation? We already knew the UV power usually decays ~30% after the data acquisition. Possibly the UV power decays quicker at beginning. Therefore, probably the data1 channel-1 has a non-linear behavior at initial bins and data2 channel-1 signal becomes more linear after a power decay. Note the ND test#2 was taken after the gate delay test. The data1/data6 seems to suggest a strong power decay. But 291 ratio looks flatter than 285. Other channel-1 ratios of different data look good.

Compared with channel-1 to exclude atmospheric variation, the near-range analog signals of channel-2 with different gate delay look behaving similarly. Additionally, the analysis of the ratio of channel-1 against channel-2 also proves the channel-2 analog signal is good above ~500m.

Another issure is SIB. Does gating reduce the SIB? You might think so If looking at data1/data6 analog signal of channel-2. But the PC ratio looks good. I can't explain why the analog ratio has a steeply positive slope at far range in channel-2. (We already see analog signal has problem at far-range in ND test#2 or other tests when the signal is weak.) I have to say analog signal is not linear in the weak signal region.

Conclusion: 1) We might consider inserting a ND0.3 (~50% extinction) for channel-1.

2) For channel-1, data with different gate delays behave similarly. I would use a 1-1.5us delay for channel-2.

For more details, please check the PDF files: data1/data2, data2/data3, data3/data4, data4/data5, data5/data6, data1/data6.