Book a short-exposure core set for Cen A, and make it a campaign habit #16

Merged

laurence merged 1 commit from hdr-core-booking into main

2026-07-21 13:42:18 +01:00

laurence commented

2026-07-21 13:42:10 +01:00

Owner

Why

Processing the 21 July Centaurus A data revealed that the nucleus is saturated. A single 300 s luminance sub reads 65313 ADU in the core, so the deep stack is clipped there and no processing recovers it. The only fix is more data at shorter exposures.

What was booked

plans/hdrcore-t32.txt — an exposure ladder covering a factor of 60 below the deep subs:

Filter	Frames	Exposure
Luminance	8	60 s
Luminance	8	15 s
Luminance	8	5 s
Red / Green / Blue	5 each	30 s

All BIN2, matching the deep stack's 0.5376"/px so the frames register without rescaling. 18.2 minutes of imaging, roughly 25 points.

Uploaded to T32 and verified byte-identical against the copy in this repo, then booked for 22 Jul 18:50-19:50 SSO local (08:50-09:50 UTC) — the same window in which the deep run actually succeeded, T32 having a 35 degree elevation floor and Cen A being an evening target from there. Verified in Reservation/MySummary.aspx; quota now 7:05 of 8:00.

The habit this establishes

CAMPAIGN.md now records that bright targets need a short set budgeted alongside the deep run, at about a sixth of the deep run's cost. NGC 6744 was already shot without one, and 47 Tuc and the SMC in queue #5 will certainly need one — a globular cluster core is the fastest thing there is to saturate. The rule of thumb used here: shortest sub = deep sub / 60.

Tools used

PowerShell Invoke-WebRequest for the reservation POST sequence and verification; curl multipart upload to the T32 ACP server's aupload.asp plus a round-trip diff of the uploaded file; Python (astropy) to measure the saturation level that motivated the whole thing.

Note

The image-processing pipeline that found this lives in the user's Downloads\NGC5128\20260721\stacked\scripts folder with a METHODS.md alongside, not in this repo, since it operates on image data rather than campaign state.

## Why Processing the 21 July Centaurus A data revealed that the nucleus is saturated. A single 300 s luminance sub reads **65313 ADU** in the core, so the deep stack is clipped there and no processing recovers it. The only fix is more data at shorter exposures. ## What was booked `plans/hdrcore-t32.txt` — an exposure ladder covering a factor of 60 below the deep subs: | Filter | Frames | Exposure | |---|---|---| | Luminance | 8 | 60 s | | Luminance | 8 | 15 s | | Luminance | 8 | 5 s | | Red / Green / Blue | 5 each | 30 s | All BIN2, matching the deep stack's 0.5376"/px so the frames register without rescaling. 18.2 minutes of imaging, roughly **25 points**. Uploaded to T32 and verified byte-identical against the copy in this repo, then booked for **22 Jul 18:50-19:50 SSO local** (08:50-09:50 UTC) — the same window in which the deep run actually succeeded, T32 having a 35 degree elevation floor and Cen A being an evening target from there. Verified in `Reservation/MySummary.aspx`; quota now 7:05 of 8:00. ## The habit this establishes CAMPAIGN.md now records that bright targets need a short set budgeted alongside the deep run, at about a sixth of the deep run's cost. NGC 6744 was already shot without one, and 47 Tuc and the SMC in queue #5 will certainly need one — a globular cluster core is the fastest thing there is to saturate. The rule of thumb used here: shortest sub = deep sub / 60. ## Tools used PowerShell `Invoke-WebRequest` for the reservation POST sequence and verification; `curl` multipart upload to the T32 ACP server's `aupload.asp` plus a round-trip diff of the uploaded file; Python (astropy) to measure the saturation level that motivated the whole thing. ## Note The image-processing pipeline that found this lives in the user's `Downloads\NGC5128\20260721\stacked\scripts` folder with a METHODS.md alongside, not in this repo, since it operates on image data rather than campaign state.

laurence added 1 commit 2026-07-21 13:42:10 +01:00

Book a short-exposure core set for Cen A, and make it a campaign habit 7bf834a184

Processing the 21 Jul Centaurus A data showed the nucleus is saturated: a
single 300 s luminance sub reads 65313 ADU in the core, so the deep stack
is clipped there and the core cannot be recovered by processing. The only
fix is more data at shorter exposures.

plans/hdrcore-t32.txt is that data: L 8x60s, 8x15s and 8x5s plus RGB
5x30s, all BIN2 so the frames register against the deep stack without
rescaling. The ladder spans a factor of 60 below the 300 s subs, which
covers the unknown true peak of the nucleus, and the RGB gives the core
its colour. 18.2 minutes of imaging, about 25 points.

Uploaded to T32 and verified byte-identical, then booked for the evening
of 22 Jul, 18:50-19:50 SSO local, which is the same window in which the
deep run actually succeeded (T32 has a 35 degree elevation floor and Cen A
is an evening target). Verified in MySummary; quota now 7:05 of 8:00.

The wider point is recorded in CAMPAIGN.md: bright targets need a short
set budgeted alongside the deep run. NGC 6744 has already been shot
without one, and 47 Tuc and the SMC in queue #5 will certainly need one,
a globular cluster core being the fastest thing there is to saturate.