Tools Backfocus Solver
Backfocus Solver
Nearly every corrector wants the sensor a fixed distance behind it, and being 2 mm out puts elongated stars in the corners of every frame you take. This works out the stack that gets you there from parts you already own, and tells you when it can’t.
Every figure below is re-keyed from a manufacturer page or drawing and carries a link to it. Where a number could not be confirmed, the row says so in the list before you pick it. There are no plausible-looking substitutes in this catalogue.
42 rows. 6 could not be confirmed against a manufacturer page or drawing and are marked in the list.
ZWO FF80 APO (built-in flattener)
Verified against the manufacturersource- Output thread
- M48x0.75
- Focus range
- up to 115.00 mm (minimum not published)
- Image circle
- 44 mm
"Maximum accessory connection (with the focuser fully retreated): 1. Imaging mode: 115mm (from the base of M48×0.75 male thread)" (135 / 155 / 168.5 from M54 / M68 / M76). Image-only manual, read by eye.
70 rows. 11 could not be confirmed against a manufacturer page or drawing and are marked in the list.
ZWO ASI2600MM Pro
Verified against the manufacturersource- Input thread
- M42x0.75
- Sensor
- 23.5 × 15.7 mm · 3.76 µm pixels
Between corrector and camera
8 rows, every figure confirmed against a manufacturer source.
Only manufacturer-supplied kits are catalogued so far. Generic aftermarket spacer sets are not here yet, and an invented set of lengths would produce a confident stack out of rings you may not own.
Nothing solved yet
Still needed: the spacers you own, whether an off-axis guider sits in the train, whether a filter drawer or wheel sits in the train.
Nothing is filled in on your behalf. Where a product ships in more than one configuration you are asked which one you own, because the two are not interchangeable and the difference is often larger than the tolerance.
Or open an example train
- Baader MPCC III + ZWO ASI2600MC ProA clean solve with the rings in ZWO's box
- Baader MPCC III + ZWO ASI2600MC DUOThe threads don't meet: which adapter do you have?
- ZWO ASI533MC ProZWO publishes two configurations for it: which is yours?
- Baader MPCC III + ASI2600MC Pro + ZWO M42 filter drawerZWO's own drawer chart, reproduced with 2 mm of filter glass
How this is computed
required[mm] = flattenerBackfocus[mm] + Σ( glassThickness[mm] × (n − 1) / n )
fixed[mm] = cameraFlangeToSensor[mm] + Σ components[mm]
residual[mm] = required[mm] − fixed[mm] − Σ rings[mm]
rings: spacers, extenders and adapters you own, each used at most as
many times as you have it, where every ring's thread fits the nextThe required distance is the corrector’s own published figure, measured from the face the manufacturer measured it from, which is why every row names that face: quote the same corrector from a different thread and the figure changes.
Filter glass shifts focus outward, so it makes the required distance longer, by t × (n − 1) / n. The familiar “one third of the thickness” is that expression at n = 1.5 and is wrong by about 2.5% for ordinary filter glass. Where a filter’s glass type is unpublished the solver assumes n = 1.52 and says so in the result.
Tolerance is the corrector’s published figure where it has one, and ±0.5 mm otherwise, the usual working figure for a 55 mm system. Faster systems have tighter tolerances.
Lengths and threads are solved together, because a ring of the right length can still have the wrong thread. ZWO’s ASI2600 Pro kit includes an extender that is M48 on one side and T2 on the other: an M48 corrector screws straight onto it, a T2 corrector needs the kit’s 0 mm ring first, and which of those is true decides the stack.
The stack is chosen fewest-pieces-first, because each extra joint adds a chance of tilt. When an exact stack exists that needs more rings, the result shows that one too. How the numbers are computed.
Source: docs/03-physics-reference.md § 5 · Agena, “A primer on back focus in astronomy” · ZWO, “The best solution of 55 mm back focal length”. Implemented in packages/physics/src/backfocus.ts, with the worked examples as unit tests.
Last updated