diff --git a/CAMPAIGN.md b/CAMPAIGN.md
new file mode 100644
index 0000000..06717ba
--- /dev/null
+++ b/CAMPAIGN.md
@@ -0,0 +1,218 @@
+# Drain campaign: spend 2,638 points, then close the account
+
+Decision (2026-07-18): use the full points balance on southern imaging, then cancel
+the membership. Constraints and policy facts that shape the plan:
+
+- Balance 2,638 points; Plan-40 renews 12 Aug 2026 (GBP 32). Cancel BEFORE renewal.
+- Points cannot be refunded, and cannot be used without an active subscription
+  (support article 142963). So: drain first, cancel last, in that order.
+- **Cancellation target: 10 Aug 2026**, two days before renewal. Cancellation is
+  submitted via MyAccount/Cancel.aspx and requires explicit user confirmation on the
+  day; Claude must not submit it unprompted.
+- Rates: see the rate card in state/NOTES.md. Billing is per minute of imaging time.
+- Moon: bright nights roughly 24 Jul - 2 Aug (full ~29 Jul); dark again 5-12 Aug.
+  Full-moon week includes free access promotions; paid draining focuses on dark and
+  grey nights, narrowband carries the bright week.
+- Scope availability is volatile (much of Chile offline as of 18 Jul); re-check the
+  status line on go.itelescope.net before every session and substitute freely.
+
+## Session ledger
+
+Update after every run: date, scope, target, imaging minutes, points charged
+(from account/history.aspx), cumulative spend.
+
+| Date | Scope | Target | Plan file | Imaging | Est. pts | Actual pts | Notes |
+|---|---|---|---|---|---|---|---|
+| 2026-07-19 | T33 | 47 Tuc | ngc104-test.txt | 17 min | 0 (free) | DID NOT RUN | reservation 749228 expired unused: no images, no logs on T33 (checked 20 Jul); grid shows reschedule=No despite the flag being set at booking |
+| 2026-07-19 | T59 | NGC 6744 | ngc6744-t59.txt | 180 min | ~240 (0% moon disc. until 21 Jul) | **381** | SUCCESS (txn 851148): ran 00:05:22-03:29:56 SSO, Time Used 3h24m32s, 0% moon disc. Full dataset on scope server: L 15x300s BIN1, RGB 7x300s each BIN2; previews checked, round stars, galaxy clean: no refund case. Effective rate ~112 pts/hr of WALL time (not the 80/hr imaging-time estimate: see NOTES rate correction) |
+| 2026-07-20 | T33 | NGC 6752 | ngc6752.txt | 17 min | 0 (free) | DID NOT RUN | reservation 749229 expired unused: no images, no logs (checked 20 Jul 20:30 UTC, after window) |
+| 2026-07-21 | T32 | Centaurus A | cena-t32.txt | 120 min | ~350 (x1.6 model) | **160** | SUCCESS (txn 851344): ran 18:50:24-21:16:19 SSO local (08:50-11:16 UTC), Time Used 2h25m55s, 25% moon discount. Images present: /images/qisback/T32/NGC5128 on t32.itelescope.net:8032. Came in at 0.46x the x1.6 model, i.e. LESS than the original estimate: see the rate correction below |
+
+Cumulative spend: 541. Balance after (verified GetAccountStatus 2026-07-21
+11:40 UTC): **2,097**.
+
+**Rate correction (2026-07-21, supersedes the 2026-07-20 x1.6 warning).** Two
+paid runs now bracket the question, and both bill EXACTLY the plan's imaging
+minutes, not wall-clock:
+
+| Scope | Plan imaging | Wall (Time Used) | Billed | Points per imaging hour |
+|---|---|---|---|---|
+| T59 | 180 min (36x300s) | 3h24m32s | 381 | **127** (0% moon) |
+| T32 | 120 min (24x300s) | 2h25m55s | 160 | **80** (25% moon) |
+
+381 = 3.00 h x 127 and 160 = 2.00 h x 80, both to the point. So billing is on
+imaging time as originally documented, and the wall-clock theory is dead. What
+is wrong is the rate card in NOTES.md: it lists T59 at 80/hr (actual 127) and
+T32 at 146/hr (actual 80 even after the 25% moon discount, so a base near
+107/hr). The two look transposed, but two samples cannot prove that. **Do not
+apply a blanket multiplier.** Derive each scope's effective rate from its first
+actual bill, and treat any not-yet-observed scope's card rate as +/-60%.
+
+## Planned allocation (~2,600 points, adapt as scopes/weather allow)
+
+**Phase 1, dark nights now - 23 Jul (~1,030 pts):**
+- T59 NGC 6744 LRGB, 3 h, ~240 (scheduled 19 Jul)
+- T32 Centaurus A LRGB, 2 h, ~292: early UK-afternoon window while it is high
+- T8 Corona Australis dust complex LRGB, 2 h, ~260
+- T70 (Chile) Scorpius tail / Prawn super-wide LRGB, 2 h, ~238
+
+**Phase 2, bright nights 24 Jul - 2 Aug (~440 pts, narrowband only):**
+- T59 NGC 6188 in Ha/OIII/SII 5nm, 3 h total, ~240
+- T8 Ha on the Cat's Paw / War and Peace field, 1.5 h, ~195
+- Plus any full-moon free promotions for clusters (cost 0, not part of the drain)
+
+**Phase 3, dark nights 3 - 10 Aug (~1,130 pts):**
+- T8 SMC + 47 Tucanae in one frame, 3 h LRGB, ~390: the signature shot, SMC well
+  placed in the second half of the SSO night by August
+- T8 or T59 Large Magellanic Cloud attempt if it clears the elevation limits late
+  in the window; otherwise Tarantula on T32, 2 h, ~292
+- Remainder (~450): best-of list: repeat the strongest target for more depth, or
+  Chile scopes (T71 106/hr, T73 224/hr) if they return from maintenance: Tarantula
+  and NGC 1313 preferred. Aim to land the balance under ~50 points.
+
+## Booked reservations (all with plan attached + reschedule flag)
+
+| ID | Scope | Night of (SSO/obs local) | Slot | Plan | Est. pts | Status (2026-07-20) |
+|---|---|---|---|---|---|---|
+| 749227 | T59 | 19 Jul | 00:05-03:45 | ngc6744-t59.txt | ~240 | COMPLETED, billed 381 |
+| 749228 | T33 | 19 Jul | 00:05-00:50 | ngc104-test.txt | 0 | EXPIRED unused (no run, no auto-reschedule) |
+| 749229 | T33 | 20 Jul | 00:05-00:50 | ngc6752.txt | 0 | EXPIRED unused (no run, no auto-reschedule) |
+| 749230 | T8 | 26 Jul | 21:35-23:55 | catspaw-t8.txt | ~146 (25% moon disc.) | still booked |
+| 749302 | T32 | 21 Jul | 18:50-21:50 | cena-t32.txt | ~350 (x1.6 rate, 25% moon disc.) | COMPLETED, billed 160. EVENING slot: T32 has a 35 deg elevation floor and Cen A culminates ~17:10 before dark, sinking to 11 deg by the old 00:35 slot (which could never have run). Ran 18:50-21:16, finishing 34 min early (plan complete, not a fault); images landed under /images/qisback/T32/NGC5128 |
+| (id not captured) | T8 | 22 Jul | 21:45-00:45 | cra-t8.txt | ~250 | booked + MySummary-verified 2026-07-21 11:45 UTC. This is queue #2 finally placed: see the resolution note below. Runs 11:45-14:45 UTC 22 Jul |
+| (id not captured) | T32 | 22 Jul | 18:50-19:50 | hdrcore-t32.txt | ~25 | **CANCELLED 2026-07-21 ~15:00 UTC, never ran, 0 points spent.** Booked to fix a saturated Cen A nucleus; the nucleus is not saturated (see the correction below). Deleted from MySummary with the other three reservations verified intact afterwards; quota back to 6:05/8:00 |
+| 749314 | T33 | 21 Jul | 00:05-00:50 | ngc104-test.txt | 0 (free) | booked + grid-verified 2026-07-21 08:55 UTC. 47 Tuc first-light retry, third attempt. Runs 14:05-14:50 UTC 21 Jul. Reschedule flag again No (free scopes never rebook) |
+
+The T33 first-light test is now on its THIRD booking (749314); the two prior free
+reservations expired unused. If this one also fails to run, the problem is not the
+booking but the auto-start on free scopes: escalate to walk-in (POST plan.asp per
+plans/README.md) or drop the free-imaging objective.
+
+## Booking queue (blocked by the rolling 8-hour reservation allowance)
+
+Book in order as quota frees (a session completing releases its hours). Exact
+New.aspx parameters; all plans already uploaded and verified on their scopes.
+Verify every booking via Edit.aspx?id= probing (sequential ids), NOT the grid.
+
+| # | Scope id | start | end | start sel | end sel | Plan | Est. pts |
+|---|---|---|---|---|---|---|---|
+| 1 | BOOKED as 749302 (see table above); the original 00:35 slot was an error, Cen A is an EVENING target (35 deg floor on T32) | | | | | cena-t32.txt | ~350 |
+| 2 | BOOKED 2026-07-21 for the night of 22 Jul, 21:45-00:45 (see table above); the 01:35 slot was blocked by hajo and low anyway | | | | | cra-t8.txt | ~250 |
+| 3 | GRAS059 | 2026-07-25T20:35:00 | 2026-07-25T23:55:00 | 8:35 PM | 11:55 PM | ngc6188-haoiii-t59.txt | ~120 (disc.) |
+| 4 | GRAS059 | 2026-07-27T20:35:00 | 2026-07-27T23:05:00 | 8:35 PM | 11:05 PM | ngc6188-sii-t59.txt | ~90 (disc.) |
+| 5 | GRAS008 | 2026-08-08T02:35:00 | 2026-08-08T06:05:00 | 2:35 AM | 6:05 AM | smc47tuc-t8.txt | ~390 |
+| 6 | GRAS032 | 2026-08-09T03:05:00 | 2026-08-09T06:05:00 | 3:05 AM | 6:05 AM | tarantula-t32.txt | ~365 |
+| 7 | (T70, blocked: server unreachable) | night TBD before 5 Aug | ~2.5h | | | scorpius-t70.txt (in scratch, re-upload needed) | ~238 |
+| 8 | WITHDRAWN: booked then cancelled 2026-07-21, premise was wrong (see below) | | | | | hdrcore-t32.txt | 0 |
+
+### CORRECTION (2026-07-21): the Cen A nucleus was never saturated
+
+An earlier entry here claimed the 21 Jul Cen A data came back with a saturated
+nucleus, and booked a short-exposure core set on that basis. **The claim was
+wrong and the booking has been cancelled without running, at a cost of zero
+points.**
+
+The 65313 ADU figure came from taking the maximum pixel inside a 300 x 300 px
+box centred on the frame. That measures the brightest thing near the middle of
+the image, which is not the same as measuring the galaxy: it was a foreground
+star 128 px (69 arcsec) from the nucleus, sitting inside the box. Median
+filtering the inner 800 x 800 px to remove stars leaves the galaxy peaking at
+**1944 ADU**, about a thirtieth of the clip level, with **not one star-free
+pixel above 30000 ADU**. The core looked blown in the first processed image
+because of the display stretch, whose white point was set by field stars, and
+re-rendering the same data with a second tone curve scaled to the galaxy
+recovered it fully.
+
+**The habit that IS worth keeping, in corrected form:** before booking extra
+time to fix a saturated core, *measure whether it is saturated* - median-filter
+the inner region to remove stars first, and check where the frame's bright
+pixels actually are. The check costs a minute. 47 Tuc and the SMC (queue #5)
+may still genuinely saturate, a dense cluster core being the fastest thing there
+is to clip, so run the same measurement on their data before assuming either
+way. plans/hdrcore-t32.txt stays in the repo (and on T32) as a ready-made
+ladder if one of them turns out to need it.
+
+### Session 2026-07-21 ~08:55 UTC (booking round run early, before Cen A ran)
+
+Done this session: free T33 47 Tuc retry booked as 749314 (grid-verified). NOT
+done, both correctly deferred:
+- **Cen A verify**: 749302 had not run yet (starts 08:50 UTC, this session ran at
+  08:48). Verify after 11:50 UTC.
+- **cra-t8 (#2)**: blocked twice over. (1) Quota: Cen A holds 3h until it
+  completes ~11:50 UTC; with Cat's Paw (2h20m) + Cen A (3h) + T33 (45m) held,
+  under ~1h55m is free, not enough for a 3h slot. (2) Conflict: user hajo holds
+  T8 00:25-02:25 SSO local (res 749207), overlapping our planned 01:35-04:35.
+
+**cra-t8 resolution options** (pick with a live grid re-check when quota frees):
+- (a) Shift to 02:25-05:25 after hajo. Downside: Corona Australis culminates
+  ~22:30 SSO local, so by 02:25-05:25 it is 4-6.5h past meridian and low; poor.
+- (b) PREFERRED: move cra-t8 to the NEXT night, 2026-07-23T01:35-04:35 (night of
+  22 Jul SSO), same clock time, avoids hajo, keeps decent altitude. Re-check that
+  night's T8 grid first. One night's slip is fine inside the drain window.
+- (c) Earlier slot 21:25-00:25 (before hajo) is astronomically best but starts
+  11:25 UTC, before Cen A releases quota at 11:50; not bookable in time.
+
+The scheduled 12:13 UTC job (ae3af497) fires after Cen A completes: it should
+verify Cen A + bill, then book cra-t8 by option (b) unless the grid has changed
+(hajo may cancel, opening option (a)/original). At corrected (x1.6) rates the
+full remaining queue runs ~2,530 vs 2,257 balance (~270 over): once actual bills
+for Cen A and CrA land, trim Phase 3 (shorten Tarantula #6 or the SMC slot)
+to land the balance near zero.
+
+### Session 2026-07-21 ~11:40 UTC (Cen A verify + cra-t8 booked)
+
+Ran ahead of the 12:13 UTC scheduled job, after Cen A had finished at 11:16 UTC.
+
+- **Cen A verified**: completed, billed 160, NGC5128 images on the scope server.
+  Ledger and rate model updated (the x1.6 model is withdrawn: see the rate
+  correction above).
+- **cra-t8 booked**, but on a better slot than the deferred option (b). Probing
+  New.aspx for the original 01:35-04:35 returned "Conflicting reservation. Start
+  time must be later than 2:25 AM" (hajo, as predicted), confirming that option
+  (a) was all that remained on the night of 21 Jul, and it puts Corona Australis
+  3.5-6.5 h past the meridian. Instead the run went to the EVENING of the next
+  night: **2026-07-22T21:45 - 2026-07-23T00:45 SSO local**, which brackets the
+  ~22:30 culmination, so the whole 115 min of imaging happens near the zenith.
+  This is option (c)'s geometry on option (b)'s night; the reason (c) was
+  unbookable earlier (quota held by Cen A) had expired by the time Cen A
+  completed. Grid probing showed no conflict on either side of the slot.
+  Verified in Reservation/MySummary.aspx; quota now 6:50 / 8:00.
+- Moon on the night of 22 Jul is ~57% illuminated, the last usable night before
+  the bright-week narrowband phase. LRGB on a dust complex under a half moon is
+  a compromise, but the alternative is slipping a Phase 1 broadband target into
+  the bright week, which is worse.
+- Reservation id NOT captured: the confirm response carries only
+  `ModalStatic.result("OK")` and MySummary does not print ids. Nothing depends
+  on the id unless the booking must be edited or deleted, in which case find it
+  by probing Edit.aspx?id= around the current sequence (749314 was issued
+  ~08:55 UTC on 21 Jul).
+- Budget: 2,097 left, 19 nights to 10 Aug. Remaining queue at card rates is
+  roughly 1,300, which UNDER-spends by ~800. With the x1.6 model withdrawn, the
+  next planning session should add work rather than trim it: candidates are a
+  second narrowband night on T59 (observed 127/hr, the only scope with a
+  trustworthy rate), or extending the Phase 3 SMC and Tarantula runs, both of
+  which have the sky time to absorb longer plans.
+
+Substitution rules: if T71 returns to service, move #2 and #7 to T71 (faster
+f/2.8 optics, cheaper 106 pts/hr; re-cut the plans for 1.55"/px). If T70's
+server returns, book #7 as-is. Check the go.itelescope.net status line before
+each booking round. After #6, allocate the remaining balance (~450 pts) to a
+final deep session on the best-performing scope, sized to land under 50 pts.
+
+**10 Aug: verify balance is near zero, download ALL data from data.itelescope.net,
+then (with explicit user confirmation) cancel via MyAccount/Cancel.aspx.**
+
+## Operational notes
+
+- Sessions are now RESERVED, not walk-in: upload the plan to the scope, then book
+  it at reservations.itelescope.net with the plan attached (procedure in
+  plans/README.md); iTelescope auto-starts the plan at reservation start and the
+  reschedule flag retries weather losses. Claude sessions monitor rather than
+  drive. T59 moon discount correction: 0% until 21 Jul, 25% from 22 Jul.
+- Cron jobs live only in an open Claude session; if a day's session has no live
+  session to fire it, the user prompts "run today's drain session" and any session
+  picks it up from this file plus state/TODO.md.
+- Bad-data protection: iTelescope refunds points for weather-ruined runs on
+  request; check each session's images before marking the ledger final.
+- Final data sweep before cancellation: everything on data.itelescope.net AND the
+  per-scope /images/qisback/ folders.
diff --git a/plans/README.md b/plans/README.md
new file mode 100644
index 0000000..71c823e
--- /dev/null
+++ b/plans/README.md
@@ -0,0 +1,100 @@
+# Observing plans and the T33 session runbook
+
+ACP observing plans for iTelescope scopes, plus the procedure to run one on T33
+(the free Siding Spring scope). Credentials are NOT in this repo: they live in
+the local claudetemp itelescope folder (creds.txt, `username=`/`password=` lines).
+
+## Plan format (ACP)
+
+See [ngc104-test.txt](ngc104-test.txt). Directives, then a target line:
+`Name<TAB>RA_decimal_hours<TAB>Dec_decimal_degrees`. Counts, intervals, binning
+and filters are parallel comma-separated lists. `#shutdown` parks safely at the
+end (iTelescope's own one-click plans do the same).
+
+## T33 endpoints (base http://t33.itelescope.net:8033, HTTP auth = iTelescope creds)
+
+| Purpose | Path |
+|---|---|
+| Status JSON (times, obsStat, obsOwner) | /ac/SystemStatus/asystemstatus.asp |
+| Console text | /ac/SystemStatus/aconsread.asp |
+| Scope connect control | /ac/SystemStatus/ascopeconn.asp |
+| Shutter/roof status | /ac/SystemStatus/ashutterctl.asp |
+| My plan folder (list/upload/delete) | /plans/qisback/aindex.asp, aupload.asp, adelfile.asp |
+| Run a plan ("Acquire Images") | POST /ac/Plans/plan.asp |
+| My logs | /logs/qisback/aindex.asp |
+| Finished data | data.itelescope.net |
+
+Upload: multipart POST to `/plans/qisback/aupload.asp`, fields `File1` (the file)
+and `DestPath=/plans/qisback`. Verified working 2026-07-18; content round-trips
+exactly.
+
+Run: the web UI posts the plan select to `/ac/Plans/plan.asp`; the select's value
+is the WINDOWS path `C:\ProgramData\ACP\ACP Web Data\Doc Root\plans\qisback\<file>`.
+Field names seen: `plan`, optional `VPhot`, `AutoLogoff`. NOT yet exercised live
+(observatory was closed, daytime); the first live session must confirm whether a
+scope-connect POST (ascopeconn.asp) is required first, and capture the exact
+response shape.
+
+## The 47 Tuc free-session test (ngc104-test.txt)
+
+- Target: NGC 104 / 47 Tucanae, RA 0.4015 h, Dec -72.0814 deg (circumpolar at SSO).
+- L 4x120 s, RGB 2x90 s each, Bin2: about 17 min exposure, about 24 min with
+  overhead, inside the daily free 30 minutes.
+- Uploaded to /plans/qisback/ on 2026-07-18 and verified.
+
+### When to run
+
+47 Tuc clears T33's 20 degree minimum elevation in the second half of the SSO
+night. SSO local = UTC+10; the good window is roughly 14:00-19:30 UTC (UK
+afternoon/evening). Sunrise at SSO was 06:14 local (20:14 UTC) at recon time;
+stop well before dawn.
+
+### Procedure
+
+1. GET asystemstatus.asp: confirm observatory open (obsStat) and no other user
+   (obsOwner n/a). If someone is on, wait and retry; T33 is first-come.
+2. Confirm the plan file is still in /plans/qisback/ (aindex.asp).
+3. POST plan.asp with the plan's Windows path to start acquisition.
+4. Poll aconsread.asp (console) every minute or two: slew, centering, filter
+   changes, exposures. The run self-terminates via #shutdown.
+5. Images land in /images/qisback/ on the scope server and then on
+   data.itelescope.net; logs in /logs/qisback/.
+6. Record the session outcome in state/NOTES.md.
+
+Abort path if something looks wrong: the web UI exposes an abort on the console
+page; locate it live before starting exposures (not yet mapped).
+
+## Reservations (verified working 2026-07-18)
+
+Reservations live at http://reservations.itelescope.net ("Reservations Pilot"),
+reached from each scope's /ac/MenuItems/reservation-edit.asp iframe as
+`?t=<scope-id>&u=<username>` (scope ids look like GRAS059 for T59, GRAS033 for
+T33; full list in the resources JSON on go.itelescope.net/Reservation/Overview.aspx,
+whose embedded DayPilot events also show everyone's bookings, roof/sun times, and
+the moon discount per night). No separate login: it trusts the u= parameter once a
+session cookie exists.
+
+Creating one is a three-step WebForms dance against
+`New.aspx?start=<ISO local>&end=<ISO local>&r=<scope-id>` (times are OBSERVATORY
+LOCAL): (1) GET the form, (2) POST tokens + DropDownStartTime/DropDownEndTime
+(e.g. "12:05 AM"/"3:45 AM") + PlanToRunListBox=<plan file> + RescheduleCheckbox
++ ButtonRefreshPlans=Refresh Plans, which re-renders with Confirm enabled,
+(3) re-POST the fresh tokens with the same fields + ReservationOKButton=Confirm
+Reservation. Success = response script `ModalStatic.result("OK")` AND the booking
+appearing in the grid: the modal also says "OK" on silent failure, so ALWAYS
+re-fetch the grid and look for the reservation id + username.
+
+Availability probing (2026-07-21): a plain GET of
+`New.aspx?start=...&end=...&r=...` is a cheap, side-effect-free conflict check.
+The form comes back with the two time dropdowns populated for +/-60 min around
+the requested times, and if the slot collides with someone else's booking it
+also renders "Conflicting reservation. Start time must be later than H:MM AM".
+No message plus a full pair of dropdowns means the window is clear. This beats
+reading the DayPilot grid, whose event divs do not survive a plain
+Invoke-WebRequest fetch (they are rendered client-side).
+
+Rules learned: maximum reservation length 4:00; recommended duration 1.5x total
+imaging time; the attached plan is auto-started by iTelescope at reservation
+start (no walk-in POST needed); RescheduleCheckbox auto-rebooks the identical
+timeslot on the first available night of the next 30 if the run does not happen.
+Delete/edit via Edit.aspx?id=<reservation id> from the grid.
diff --git a/plans/catspaw-t8.txt b/plans/catspaw-t8.txt
new file mode 100644
index 0000000..c541cfb
--- /dev/null
+++ b/plans/catspaw-t8.txt
@@ -0,0 +1,12 @@
+;
+; Cat Paw NGC 6334 in Ha on T8 - phase 2 narrowband
+; Ha 9x600s Bin1. 90 min imaging. Bright moon OK.
+;
+#largepreviews
+#tiff
+#count 9
+#interval 600
+#binning 1
+#filter Ha
+NGC6334	17.340000	-36.116667
+#shutdown
diff --git a/plans/cena-t32.txt b/plans/cena-t32.txt
new file mode 100644
index 0000000..94e44c3
--- /dev/null
+++ b/plans/cena-t32.txt
@@ -0,0 +1,12 @@
+;
+; Centaurus A (NGC 5128) on T32 for qisback - drain campaign phase 1
+; L 12x300s, RGB 4x300s each. 120 min imaging.
+;
+#largepreviews
+#tiff
+#count 12,4,4,4
+#interval 300,300,300,300
+#binning 2,2,2,2
+#filter Luminance,Red,Green,Blue
+NGC5128	13.424333	-43.019139
+#shutdown
diff --git a/plans/cra-t8.txt b/plans/cra-t8.txt
new file mode 100644
index 0000000..03663b3
--- /dev/null
+++ b/plans/cra-t8.txt
@@ -0,0 +1,12 @@
+;
+; Corona Australis dust complex on T8 for qisback - drain campaign phase 1
+; L 8x300s Bin1, RGB 5x300s each Bin2. 115 min imaging.
+;
+#largepreviews
+#tiff
+#count 8,5,5,5
+#interval 300,300,300,300
+#binning 1,2,2,2
+#filter Luminance,Red,Green,Blue
+CrA-complex	19.030000	-36.955000
+#shutdown
diff --git a/plans/hdrcore-t32.txt b/plans/hdrcore-t32.txt
new file mode 100644
index 0000000..908c510
--- /dev/null
+++ b/plans/hdrcore-t32.txt
@@ -0,0 +1,17 @@
+;
+; Centaurus A (NGC 5128) HDR core set on T32 for qisback - drain campaign
+; Companion to cena-t32.txt (2026-07-21, 24x300s LRGB BIN2), whose nucleus is
+; saturated: the core reaches 65313 ADU in a single 300s luminance sub.
+; This is the short-exposure ladder to blend under that core.
+; L 8x60s, 8x15s, 8x5s spans a 60x range below 300s; RGB 5x30s gives the
+; nucleus colour. BIN2 throughout so the frames register to the deep stack
+; without rescaling. 18.2 min imaging total.
+;
+#largepreviews
+#tiff
+#count 8,8,8,5,5,5
+#interval 60,15,5,30,30,30
+#binning 2,2,2,2,2,2
+#filter Luminance,Luminance,Luminance,Red,Green,Blue
+NGC5128	13.424333	-43.019139
+#shutdown
diff --git a/plans/ngc104-test.txt b/plans/ngc104-test.txt
new file mode 100644
index 0000000..0241f79
--- /dev/null
+++ b/plans/ngc104-test.txt
@@ -0,0 +1,12 @@
+;
+; 47 Tucanae (NGC 104) free-session test plan for qisback
+; L 4x120s, RGB 2x90s each, Bin2. ~17 min exposure, ~24 min with overhead.
+;
+#largepreviews
+#tiff
+#count 4,2,2,2
+#interval 120,90,90,90
+#binning 2,2,2,2
+#filter Luminance,Red,Green,Blue
+NGC104	0.401500	-72.081389
+#shutdown
diff --git a/plans/ngc6188-haoiii-t59.txt b/plans/ngc6188-haoiii-t59.txt
new file mode 100644
index 0000000..4299bdd
--- /dev/null
+++ b/plans/ngc6188-haoiii-t59.txt
@@ -0,0 +1,12 @@
+;
+; NGC 6188 Fighting Dragons on T59, session A: Ha + OIII - phase 2 narrowband
+; Ha 6x600s, OIII 6x600s. 120 min imaging. Bright moon OK.
+;
+#largepreviews
+#tiff
+#count 6,6
+#interval 600,600
+#binning 1,1
+#filter Ha,OIII
+NGC6188	16.669167	-48.796667
+#shutdown
diff --git a/plans/ngc6188-sii-t59.txt b/plans/ngc6188-sii-t59.txt
new file mode 100644
index 0000000..864d194
--- /dev/null
+++ b/plans/ngc6188-sii-t59.txt
@@ -0,0 +1,12 @@
+;
+; NGC 6188 on T59, session B: SII + extra Ha - phase 2 narrowband
+; SII 6x600s, Ha 3x600s. 90 min imaging. Bright moon OK.
+;
+#largepreviews
+#tiff
+#count 6,3
+#interval 600,600
+#binning 1,1
+#filter SII,Ha
+NGC6188	16.669167	-48.796667
+#shutdown
diff --git a/plans/ngc6744-t59.txt b/plans/ngc6744-t59.txt
new file mode 100644
index 0000000..6e76e66
--- /dev/null
+++ b/plans/ngc6744-t59.txt
@@ -0,0 +1,13 @@
+;
+; NGC 6744 paid session on T59 for qisback
+; L 15x300s Bin1, RGB 7x300s each Bin2.
+; 180 min imaging = ~240 pts at Plan-40 rate 80/hr before moon discount.
+;
+#largepreviews
+#tiff
+#count 15,7,7,7
+#interval 300,300,300,300
+#binning 1,2,2,2
+#filter Luminance,Red,Green,Blue
+NGC6744	19.162833	-63.857500
+#shutdown
diff --git a/plans/ngc6752.txt b/plans/ngc6752.txt
new file mode 100644
index 0000000..ae0e494
--- /dev/null
+++ b/plans/ngc6752.txt
@@ -0,0 +1,12 @@
+;
+; NGC 6752 (Pavo globular) free-session plan for qisback
+; L 4x120s, RGB 2x90s each, Bin2. ~17 min exposure, ~24 min with overhead.
+;
+#largepreviews
+#tiff
+#count 4,2,2,2
+#interval 120,90,90,90
+#binning 2,2,2,2
+#filter Luminance,Red,Green,Blue
+NGC6752	19.181111	-59.984444
+#shutdown
diff --git a/plans/scorpius-t70.txt b/plans/scorpius-t70.txt
new file mode 100644
index 0000000..f74ed3d
--- /dev/null
+++ b/plans/scorpius-t70.txt
@@ -0,0 +1,12 @@
+;
+; Scorpius tail super-wide (Prawn/Cat Paw region) on T70 for qisback - phase 1
+; L 10x300s, RGB 4x300s each. 110 min imaging.
+;
+#largepreviews
+#tiff
+#count 10,4,4,4
+#interval 300,300,300,300
+#binning 1,1,1,1
+#filter Luminance,Red,Green,Blue
+Scorpius-tail	16.916667	-40.000000
+#shutdown
diff --git a/plans/smc47tuc-t8.txt b/plans/smc47tuc-t8.txt
new file mode 100644
index 0000000..05dd554
--- /dev/null
+++ b/plans/smc47tuc-t8.txt
@@ -0,0 +1,12 @@
+;
+; SMC + 47 Tucanae in one T8 frame - phase 3 signature shot
+; L 8x300s Bin1, RGB 6x300s each Bin2. 130 min imaging. Dark of moon.
+;
+#largepreviews
+#tiff
+#count 8,6,6,6
+#interval 300,300,300,300
+#binning 1,2,2,2
+#filter Luminance,Red,Green,Blue
+SMC-47Tuc	0.650000	-72.500000
+#shutdown
diff --git a/plans/tarantula-t32.txt b/plans/tarantula-t32.txt
new file mode 100644
index 0000000..cf70646
--- /dev/null
+++ b/plans/tarantula-t32.txt
@@ -0,0 +1,12 @@
+;
+; Tarantula Nebula NGC 2070 on T32 - phase 3
+; L 10x300s, RGB 5x300s each. 125 min imaging. Pre-dawn rising target.
+;
+#largepreviews
+#tiff
+#count 10,5,5,5
+#interval 300,300,300,300
+#binning 2,2,2,2
+#filter Luminance,Red,Green,Blue
+NGC2070	5.643333	-69.100000
+#shutdown
diff --git a/state/NOTES.md b/state/NOTES.md
index 37fc86d..3a4323c 100644
--- a/state/NOTES.md
+++ b/state/NOTES.md
@@ -35,3 +35,46 @@ Credentials live OUTSIDE this repo (local claudetemp, never commit them).
 - Weather per site: Weather/Weather.aspx?site=UDRO|SSO|DSC|SRO|AC|EYE.
 - Account: Plan-40 (GBP 32.00 per 28 days), renewal 12 Aug 2026, subscribed
   since Apr 2019, balance 2,638 points at recon time.
+
+## Rate card (2026-07-18, points per imaging hour, billed per minute)
+
+Extracted from MyAccount/UpdatePlan.aspx: each scope row carries
+data-default-rate (dark-of-moon base); effective rate = base x (1 -
+AccessDiscount from DataService GetMyAvailablePlans) x (1 - moon discount,
+server-side table keyed on illumination, not extracted yet). Plan-40 discount
+is 10% on most scopes, 20% on a few, 0% on the newest (T25, T26, T73, T74).
+
+Plan-40 rates before moon discount: T2 50, T5 68, T8 130, T10 146, T11 161,
+T14 97, T17 135, T18 81, T20 89, T21 76, T24 186, T25 322, T26 394, T30 161,
+T32 146, T33 0 (free), T59 80, T68 0 (free), T70 119, T71 106, T72 202,
+T73 224, T74 295, T75 202, T80 119.
+
+Standouts: T59 at 80/hr is the value big scope (20" CDK, square 37 arcmin
+field, 900 s subs); T5 at 68/hr the cheapest paid; T26 at 394/hr the dearest.
+Billing is per minute of imaging (exposure) time, so overhead is not billed.
+
+**RATE CORRECTION (2026-07-20, from the first real bill):** the paragraph above
+underestimates. T59 session 851148: 180 min of exposure, wall clock 00:05:22 to
+03:29:56 (3h24m34s), history "Time Used" 3h24m32s = WALL time, billed 381 points
+at 0% moon discount. That is ~112 pts/hr of wall time or ~127 pts/hr of exposure
+time, vs the 80/hr the card predicted. So either Time Used (wall clock) is what
+gets billed, or the card's data-default-rate is stale. Consequence: estimate
+sessions at rate x reserved-slot-length (slots run ~1.15-1.5x imaging time), i.e.
+multiply old exposure-based estimates by ~1.6 until more bills confirm the model.
+
+**RATE CORRECTION 2 (2026-07-21, second bill; SUPERSEDES the x1.6 rule above).**
+T32 session 851344: 120 min of exposure, Time Used 2h25m55s, 25% moon discount,
+billed 160. 160 = 2.00 h x 80 exactly, and the T59 bill is 381 = 3.00 h x 127
+exactly. Both land on the plan's IMAGING minutes to the point, and neither lands
+on wall time. So the wall-clock theory is dead and the original "billed per
+minute of exposure" documentation was right; what is wrong is the card's
+per-scope numbers. Observed effective rates: **T59 127/hr** (card said 80),
+**T32 80/hr even after a 25% moon discount**, implying a base near 107/hr (card
+said 146). The pair looks transposed, but two samples cannot prove it. Rules:
+do NOT apply any blanket multiplier; derive each scope's rate from its first
+real bill; until then treat a card rate as +/-60%; and compute estimates from
+the plan's exposure total, never from the reserved slot length.
+Paid imaging needs no reservation when a scope is idle: same walk-in ACP flow
+as the free scopes, points just deduct. Reservations (to guarantee a slot) go
+through lookup.itelescope.online, which is Cloudflare-protected: Playwright
+needed.
diff --git a/state/PROJECT.md b/state/PROJECT.md
index 9139d4e..316170d 100644
--- a/state/PROJECT.md
+++ b/state/PROJECT.md
@@ -8,6 +8,11 @@
 A reference and review of the iTelescope.net remote telescope network: what each
 telescope is, what it is good at, and which one to book for a given kind of target.
 
+Since 2026-07-18 the active objective is the drain campaign (CAMPAIGN.md): spend
+the account's 2,638 points on southern-sky imaging by 10 Aug 2026, download all
+data, then cancel the membership before the 12 Aug renewal. Cancellation needs
+explicit user confirmation on the day.
+
 ## Scope
 
 - In scope: reviews and spec tables for every telescope on the iTelescope network,
diff --git a/state/TODO.md b/state/TODO.md
index ffc7f9c..13a5890 100644
--- a/state/TODO.md
+++ b/state/TODO.md
@@ -13,9 +13,68 @@
   from lat 51.5°N), matched to Q62/X07 scopes, seasonal booking calendar, and a
   July-August starter campaign (T33 free, T71 narrowband, T73 galaxy, T8 LMC later).
 
+## Done (recent)
+
+- 2026-07-20: NGC 6744 on T59 SUCCEEDED (reservation 749227, txn 851148): full
+  LRGB dataset on the scope server (L 15x300s BIN1, RGB 7x300s each BIN2),
+  previews verified clean. Billed 381 pts (est was ~240: see the rate
+  correction in NOTES.md). Balance now 2,257. Ledger updated in CAMPAIGN.md.
+
+## Done (recent)
+
+- 2026-07-21 ~08:55 UTC: free T33 47 Tuc retry booked as reservation 749314
+  (grid-verified), third attempt at the first-light test, runs 14:05-14:50 UTC
+  21 Jul, costs 0.
+
+## Done (recent)
+
+- 2026-07-21 11:40 UTC: Cen A on T32 verified COMPLETE (billed 160, NGC5128
+  images on the scope server), and queue #2 (Corona Australis on T8) booked for
+  the night of 22 Jul, 21:45-00:45 SSO local, which brackets the target's
+  culmination instead of chasing it down the western sky. Rate model rebuilt
+  from two actual bills. Details in the CAMPAIGN.md session note.
+
 ## In progress
 
-- (nothing)
+- T33 first-light test on its THIRD booking (749314), which had NOT yet run as of
+  11:40 UTC 21 Jul: it starts 14:05 UTC. Verify after 14:50 UTC (images and logs
+  on t33.itelescope.net:8033 were still empty/stale at 11:40, as expected). The
+  two prior free reservations (749228, 749229) expired unused with reschedule=No;
+  if 749314 also fails to run, stop rebooking and either walk it in (POST
+  plan.asp per plans/README.md) or drop the free objective.
+- The T32 HDR core set (plans/hdrcore-t32.txt) was booked and then CANCELLED on
+  2026-07-21 without running: the Cen A nucleus is not saturated, the earlier
+  claim having measured a foreground star. See the correction in CAMPAIGN.md.
+  Zero points spent. The plan file remains available on T32 if a genuinely
+  saturated target turns up.
+- The 21 Jul Cen A data has been fully processed off-repo, in the user's
+  Downloads\NGC5128\20260721\stacked folder: masters, plate solution (0.27
+  arcsec on 237 stars), an LRGB composite, a globular cluster survey (289
+  candidates, 62% catalogue-confirmed, density falling 10.5x outward), surface
+  photometry (Sersic n = 4.27, Re = 400 arcsec, total G = 6.56, which matches
+  the catalogued brightness and so validates the photometric chain), and a
+  transient/moving-object search (both null, with measured limits). METHODS.md
+  there documents the lot. Nothing from that analysis needs to come back into
+  this repo, but two results affect FUTURE sessions and are recorded here:
+  **(a)** measure saturation properly before buying short subs, and **(b)** the
+  faint outer halo in that stack is limited by the sky-plane subtraction
+  (-17.9 ADU/px of real halo light absorbed), not by exposure time, so more
+  integration alone would not have gone deeper.
+- CrA on T8 runs 11:45-14:45 UTC 22 Jul. Verify after 14:45 UTC: images under
+  /images/qisback/T8/ on t8.itelescope.net:8008, then the actual bill from
+  account/history.aspx into the ledger. This is the first T8 bill, so it also
+  fixes T8's effective rate (card says 130/hr; trust nothing until observed).
+- DRAIN CAMPAIGN ACTIVE (see CAMPAIGN.md): spend the full balance by 10 Aug,
+  then cancel before the 12 Aug renewal (user confirmation required for the
+  cancel itself). 541 of 2,638 spent; 2,097 remain. The x1.6 estimate multiplier
+  is WITHDRAWN: billing is on plan imaging minutes, and per-scope rates must come
+  from actual bills (T59 127/hr, T32 80/hr observed; the NOTES rate card is
+  unreliable). On current numbers the campaign UNDER-spends by ~800: plan extra
+  or longer runs, do not trim. Reschedule flag never sticks (grid always shows
+  No): treat every reservation as one-shot and verify the morning after.
+  Download the NGC 6744 and NGC 5128 data from data.itelescope.net before the
+  10 Aug cancellation sweep. Update the CAMPAIGN.md ledger after every session
+  from account/history.aspx.
 
 ## Pending