Enlarger alignment and uniformity check

Purpose

To confirm against written criteria that the enlarger’s three planes are still parallel and its uniformity is still what it was, and to leave a dated line on the machine’s certificate saying so.

When to use it

After the head is knocked or moved, the column or lens board unlocked, a lens or carrier changed, a head refitted, or an enlarger bought second-hand — then at the interval on the certificate.

This is the re-check. Corner sharpness, focus and filter behaviour, exposure repeatability and the flare tests belong to commissioning the enlarger, and a machine that has not been through that page has nothing for this one to compare against. The tolerance and the falloff arithmetic are derived on the optics page and not repeated here.

Before you start

You should already have a safelight with a measured safe working time from the safelight test, and the enlarger’s certificate to hand: a map with nothing to compare it against is half a measurement.

On the bench: a first-surface mirror about 60 mm square, and a second cut to sit in the carrier if yours takes one; shim stock — foil, drink-can metal, thin card; four sheets of your usual paper from one box, each notched at the same corner so you know in the dark which way round it is; a reflection densitometer if you built one; the three baths at working strength and 20 °C in marked dishes, one pair of tongs each.

PPE and room, as the commissioning page sets them: nitrile gloves whenever a dish is in reach and a clean second pair for dry sheets, safety spectacles for the wet half, and the window open or the extractor running before the lights go off. Let an unconverted tungsten lamphouse cool before touching inside it; never open a mains-powered head.

Procedure

1. Write the configuration down first: enlarger, lens, focal length, carrier, working aperture, working magnification, negative format. Every number below belongs to those.
2. Set the head to the working height and lock it; set the lens to the working aperture. If it is the built LED head, start the warm-up its record specifies.
3. Take the carrier out, lay the mirror flat on the baseboard reflecting face up, and light the head.
4. Look down the axis through the open stage, move your eye until the iris and its reflected image are as nearly concentric as you can get them, and judge. PASS: the reflection is displaced by less than a fifth of the iris diameter.
5. Turn the mirror through 90° and repeat step 4. A tilt has two components, and one view can null one while missing the other. Repeat once more a stop down, where the test is stricter.
6. Put the second mirror in the carrier, reflecting face down, and repeat steps 4 and 5 looking up the axis from beside the baseboard — the negative stage against the lens board.
7. Where a plane fails, shim under the low one of the two fixings it pivots on, tighten to the torque you actually use, and re-run steps 4 to 6. Trim the shim after it passes, not before.
8. Raise the head 100 mm, lower it, lock it and run step 4 again. An alignment that does not survive being moved is a coincidence.
9. Carrier back in and empty, working aperture and magnification. Find with a test strip the exposure that lands on a mid-grey — not light, not dark, the middle.
10. Expose one whole sheet, process it with the rest of the session, and dry it fully. Every sheet here is read dry.
11. Read nine positions — the centre, four corners and four edge midpoints — and write each as ΔD = D(point) − D(centre).
12. Repeat steps 9 to 11 two stops down, same magnification, exposure adjusted to land on the same mid-grey.
13. Apply PASS 1, symmetry: opposite corners, and opposite edge midpoints, agree within 0.04 in density.
14. Apply PASS 2, response to aperture: the worst corner-to-centre ΔD is smaller two stops down.
15. Record the worst corner-to-centre ΔD at the working settings, in density and in stops. That one is recorded, not passed: the course sets no limit on it, and the optics page says why.
16. Empty the dishes to their labelled containers — developer and stop by general chemical waste, fixer and first wash by silver-bearing waste, after the disposal caveat. Return unused paper to its sealed packet.
17. Add today’s dated line to the enlarger’s entry in the calibration record, and file the sheets with it.

How you know it worked

• The null passed in two axes at both stages, and again after the head was moved and re-locked.
• Two nine-point maps exist, at two apertures, from one box of paper in one session.
• All four opposite-pair differences are inside 0.04, and the worst corner improved on stopping down.
• The worst ΔD is a number with its aperture and magnification beside it, not a verdict.
• The certificate line says what was measured at which settings, never “aligned”.

If a step fails

Step 4, the reflection is doubled. Ordinary mirror glass reflects from both surfaces, about a millimetre apart here. Null on the brighter one consistently, or get a first-surface mirror.

Step 6, no mirror fits your carrier. Check both stages against the baseboard and rely on transitivity, writing down that you did: two errors in the same direction cancel in that test and do not cancel in a print.

Step 8, it passed and now fails. A collar that does not lock, or a column that flexes. Push gently on the head and re-measure the projected image with a rule: a column that creeps is a repair, not an alignment fault.

Step 13, symmetry fails. Stop and go back to step 4. Asymmetry has no geometric floor, so it is misalignment, an off-centre source, or something mechanical in the path. If the null still passes, rotate the carrier 180° and re-map — a pattern that follows the carrier is the carrier.

Step 14, the corners do not improve on stopping down. A mechanical vignette rather than optical falloff. Look up through the lens from the corner of the easel: is the whole aperture visible from there? Suspect the carrier aperture, a filter drawer, a bellows fold or a lens board recess.

Step 15, you have no densitometer. Cut the sheet into its nine squares, shuffle them and sort them by eye against a neutral surround. If you cannot reliably order them, the field is even to the limit of your eye — a real finding, weaker than a number. Symmetry survives almost intact. Write down which test you did.

Any step, the gradient follows the paper rather than the easel. That is development, not the enlarger: rotate the sheet 180° in the dish and re-run one map. The break/fix page owns that test.

What to record

One dated block on the enlarger’s calibration record: the configuration from step 1; the null result in two axes at both stages, and separately the result after the head was moved; the two sets of nine ΔD values; the worst opposite-pair difference against 0.04; the worst corner ΔD at each aperture, in density and in stops; the reading method, with the densitometer’s tile serial or the words “visual, sorted by eye”; the paper and its batch; and what you shimmed. File the sheets with it: a number whose sheet has been thrown away is an assertion.