ARA Testbed Efficiency (Rough Cut)

1. ARA Testbed Efficiency(Rough Cut) November 29th 2010 Ben Rotter

2. Deep Antenna Trigger Efficiency: Rough Measurement • Established a rough estimate of the deep channels trigger efficiency using the level one scalar rates. • Level one scalars are 100ns coincidence triggers • L0-L5: 3 out of 4 channels, geometry based • L9: 3 out of 8 channels • The level two (global trigger) is currently an OR for all level one scalars.

3. 500Hz Input pulse 2.5V DAC range (4096 values) ~0.5mV per DAC step Centered around 1.5V • Level one scalar rate measurements from various signal insertion SNR values. • DAC value refers to channel DISC08. Other channels are shifted linearly to align the noise vs channel singles rate scans. • Plateau at 500Hz shows correct relationship between pulse frequency and scalar counter • Max data transfer rate from pole to north is ~2Hz, so low noise rates are needed. • SNR was measured on the peak to peak output of the receiver box with filtering on the oscilloscope

4. 500Hz Input pulse • Similar plot to previous one, but with noise rate subtracted from signal rate • System is 100% efficient at 500Hz for these plots • System is more efficient at lower thresholds / higher noise rates • From these plots we can pick DAC levels with acceptable noise rates and plot the efficiency for each SNR

5. The L0 through L5 level one triggers have a 50% efficiency rate at an SNR of 4 with a 1hZ noise rate The efficiency improves slightly at higher noise rates, to a 50% point at SNR 3.8. As expected the 3 out of 8 trigger scheme is more efficient at lower SNRs. All 8 channels are receiving the same signal, so we expect a factor of two greater than the sum of the other L1s. (56 combinations for 3 out of 8, compared to only 24 with the other L1s) This is still just a rough estimate, and needs to be refined (better averaging over the trigger rate scans, with actual data once the PPS sync pulse from the rubidium clock is integrated into firmware).

6. Shallow Antenna Trigger Efficiency: Rough Measurement • Only three facets for Shallow antennas • Each antenna has two discriminators, one for a positive going and one for a negative going signal. • Two scalars per channel count the postive rate and a “OR” between the two. • 4 Level 1 scalars, 3 facets and 1 “3 out of 8” • global rate OR includes these level one scalars

7. for OR the positive DAC settings were held steady at 100 (to force those scalars to zero) Downward slope of green line at ~2100 corresponds to BAT01 Negative Unclear why upward slopes at 2000 and 2020 are not aligned, but all channels have similar effect. OR not coded correctly?

8. Lining up the singles rates with a noise input Actual DAC level (for +) = {center}+{offset} – {count} Actual DAC level (for OR) = {center} – {offset} + {count}

9. Channel singles rates after normalization “OR” should be around twice the size of “positive” since it combines two rate counters Note that channels singles rates are in units of 100Hz and no information about values less than 100Hz is gained from this plot

10. Level One curves all return similar rates at equal “normalized” DAC settings Noise extends out to very high thresholds, has odd “knee” effect. This is a problem

11. Efficiencies for batwings are comparable to discones, but extended noise tail makes triggering with batwings impossible without also having a high noise rate. 50% efficient with a noise rate of 1Hz at an SNR of 10. For comparison, discones do that at an SNR of 4

13. Extra slides:

14. Solved double counting issue in trigger rate readout • Caused by 1Hz rate being read out too quickly after setting DACs • Rates are still basically correct, and should all be shifted slightly to the right

15. A closer look at the improvement of efficiency at different noise rates. This is for the L0 trigger