As we understand it, the current plans will have available only the WAPP for E-ALFA at the end of 2004. We hope that longer range plans will get us the additional capacity we need to achieve our science goals as outlined in the ideal backend summary, but here is the minimum that we can expect to have on January 1, 2005.
See Jeff Hagen's notes page for more information on the (Wide Band Pulsar Processor) WAPP in general, and Jeff's handout which also gives more details. Here, we summarize the most relevant ones with a few astronomer-friendly units added for convenience (for those who might not think naturally in kHz yet). For each of the 7 beams, the maximum capability is given below. The last line contains what the ideal backend would deliver for comparison.| Bandwidth | # pols | # of spectral points | d(freq) | d(vel @ 1.4 GHz) | sampling level |
|---|---|---|---|---|---|
| 100 MHz | 1 | 2048 | 48.8 kHz | 10.6 km/s | 9 |
| " | 1 | 8192 | 12.2 kHz | 2.6 km/s | 3 |
| " | 2 | 4096 | 24.4 kHz | 5.3 km/s | 3 |
| 50 MHz | 1 | 4096 | 12.2 kHz | 2.6 km/s | 9 |
| " | 1 | 16384 | 3.0 kHz | 0.65 km/s | 3 |
| " | 2 | 2048 | 24.4 kHz | 5.3 km/s | 9 |
| " | 2 | 8192 | 6.1 kHz | 1.33 km/s | 3 |
| Ideal 200 MHz | 2 | 8192 | 24.4 kHz | 5.3 km/s | >= 8 bit (12?) |
It would be a shame to throw away one polarization!
9 level sampling helps a lot with RFI mitigation. In terms of sensitivity, the clipping losses are already negligible when one samples at 9 level, but higher sampling can prevent the spread of RFI across the spectrum. Jon Hagen reports that at 8-bit sampling, the spread is largely contained. This has been corroborated by studies carried out at Berkeley.
Hanning/other smoothing will degrade the spectral resolution.