Experiment: Commissioning the Enlarger

Purpose

To turn an enlarger — one you built a head for, or one you bought, or one that came with the darkroom — into an instrument with a certificate: a written statement of what it was measured to do, on what date, at what settings, by whom.

The hypothesis. At its working aperture and magnification, this enlarger delivers an image that is evenly illuminated within a stated amount, uniformly sharp into all four corners, and repeatable from one exposure to the next within a stated fraction of a stop.

The control. The contact printing frame and its light box, whose uniformity has already been mapped and whose worst point is already a number in your notebook. It is the right control because it does everything the enlarger does except have a lens: no conjugates, no field angle, no three planes to keep parallel. Anything that shows up on the enlarger and not on the contact printer belongs to the lens, the alignment or the head — and anything that shows up on both is the paper, the developer or your technique, and this page hands it straight to the break/fix page at the end of the part.

The one variable that changes. The enlarger, and within it exactly one thing at a time: first the aperture with everything else held, then the magnification with everything else held. Same paper, same box, same developer, same dilution, same temperature, same agitation, same session, same reader. A uniformity sheet from Tuesday and one from Thursday are not two measurements of an enlarger; they are one measurement of an enlarger and one of a Thursday.

What you get at the end is four numbers and a signature. Everything else on this page exists to make those four numbers mean something.

The four acceptance tests, in the order they have to be run

1. 1A1 — Alignment, by mirror null in two axesPASS: the reflected image of the lens aperture sits concentric with the aperture itself to better than a fifth of its diameter, at both the negative stage and the easel, and still does after the head has been moved and locked again.
2. 2A2 — Uniformity, as density on paperPASS: the map is symmetric — opposite corners within 0.04 in density of each other — and stopping down two stops improves the corners. The magnitude is recorded rather than passed, for the reason the optics page gives.
3. 3A3 — Sharpness into the cornersPASS: all four corners resolve the same finest detail as the centre at the working aperture, read with a loupe on a dry print. A corner that fails must be diagnosed, not averaged.
4. 4A4 — Exposure repeatabilityPASS: the standard deviation of ten identical exposures is below half the smallest exposure increment you actually use — about 3 per cent if that increment is a twelfth of a stop.
5. 5Then the flare hunt, and the certificateFlare is measured rather than passed, because it changes the grade you print at. The certificate records all of it with a date, and is filed beside the sensitometer and densitometer certificates.

The order is not a preference. Alignment corrupts every measurement downstream of it, uniformity is needed before a corner's softness can be read honestly, and there is no point measuring the repeatability of a machine that has failed the first three.

Learning objectives

By the end of this session you should be able to:

• align three planes by a null method, and say why a null is more sensitive than a reading;
• state the alignment criterion as an angle, derive the mirror test’s sensitivity from the geometry, and show that the test is stricter than the tolerance;
• expose and read a uniformity map in density units, and explain why a mid-grey is the most sensitive detector available to you;
• separate lens falloff, source non-uniformity, misalignment and mechanical vignetting by changing one thing at a time;
• read four different corner failures apart by which corners fail and how;
• convert an exposure scatter into fractions of a stop and decide whether it matters;
• run three flare tests and say what each one excludes;
• write an instrument certificate that claims exactly what was measured and no more;
• apply the whole procedure to a bought enlarger, which is where most readers will start.

Prerequisites

Enlarger optics and design, which derives the alignment tolerance, the falloff arithmetic and the density-against-percentage conversion this page applies. None of that is re-derived here.

The safelight fog test, passed, with a safe working time written on the lamp. Every sheet in this session spends minutes on an easel under that lamp, and an unmeasured safelight would put a second variable into every reading.

Calibrating the densitometer, if you have built one. What this page needs from it is only its reflection mode’s internal consistency, not its absolute scale — see the note under Equipment.

Making a room dark, passed and dated, because the flare tests at the end cannot distinguish a leaking room from a leaking lamphouse.

Safety classification

Level A, and every one of the rubric’s criteria holds. The chemistry is the ordinary print sequence at working-strength dilutions in trays, in a ventilated room, with dedicated utensils; nothing is heated; no substance here is worse than irritant in these quantities; and the waste streams are the ones the household route already handles.

No electrical work is performed on this page at all. If the enlarger is mains-powered and unconverted, it is used as the certified appliance it is: plugged in, switched by its own switch or by a bought certified timer, and never opened. If it carries the LED head from the previous page, everything you touch is at 12 or 24 V.

What is not a hazard here, and why. No powder is opened and no concentrate is decanted: the three baths are made up to working strength before the session and handled with tongs. Metol and hydroquinone are the substances on the page that deserve respect, and the reason they earn no higher letter is the operation rather than the substance — the ruling of 5 September 2026 is explicit that the operation sets the level. Pouring a bought concentrate at 1+9 into a dish is this session; weighing the same agents out of a jar is a different one with different controls.

Two mechanical things are worth naming even though neither raises the level. An enlarger head high on its column is a mass at the end of a lever and the whole machine will tip if the baseboard is loaded unevenly or the head is unlocked — so the head is locked before hands come off it, and the baseboard stays clear. And an unconverted tungsten lamphouse gets hot enough to burn: it is switched off and left to cool before anything is adjusted inside it, and the alignment work in A1 is done with the lamp off wherever a torch will do instead.

Hazards

Developer on the skin. Metol is a recognised skin sensitiser and the control is that hands do not go in the dish: tongs, one pair per bath, and gloves.

Stop bath and fixer, irritant to the eyes at working strength. Eye protection, and sheets go into the dish edge-first rather than dropped.

Cross-contamination, which is a hazard to the experiment rather than to you and wrecks it completely. One dish per solution, each marked, each with its own tongs, chemistry in matching containers — because a trace of fixer in the developer gives inconsistent results, and on this page an inconsistent result is indistinguishable from an enlarger that fails A4.

Working in the dark with liquids, for three hours. Floor clear before the lights go off, trays in a fixed order that does not change, everything electrical outside the splash zone on a residual-current-protected socket.

The heavy head, the loaded column and the hot lamphouse, all as declared above.

Required PPE

Nitrile gloves whenever a dish is in reach, and a second clean pair for handling dry sheets when reading them, because a fingerprint on a uniformity map is a density difference you did not intend.

Safety spectacles with side protection for the whole wet half of the session.

Nothing optical, and nothing respiratory. There is no UV source on this page and no filter to specify; the eye protection that matters is against splashing. Ventilation is a control here (see below) but a respirator is not among the controls, because working-strength paper chemistry in open dishes at 20 °C produces no vapour that one would address.

Ventilation

An openable window or an extractor running for the whole session, which is the Level A control and is owned by Part II and by the room build. Nothing on this page changes it.

Two things about this particular session make ventilation worth restating rather than assuming. It is long — three hours with three open dishes is more air than a printing session — and it is dark, so the usual warning that a room feels stuffy arrives late. Set the extractor before the lights go off and leave it.

Materials

Material	Quantity	Notes
Photographic paper, the grade and surface you print on	16 to 20 sheets, 8 × 10 in, one box	The whole session is a comparison, so a second box is a second variable. Variable-contrast RC is the course’s default and dries fastest, which matters when everything is read dry
A negative with fine detail right into the corners	1	A brick wall, a hedge, a page of newsprint photographed square on. It must have detail at the edges, which most pictures do not
A clear, fixed-out piece of the same film	1	For the flare test’s bright surround, and for the capped-lens control
A small first-surface mirror, or an ordinary mirror tile	1, about 60 × 60 mm	Ordinary mirror glass gives two reflections a millimetre apart; at this scale that is a nuisance rather than an error, and a first-surface mirror removes it
A mirror cut to sit in the negative carrier	1, optional	The neat way to check the negative stage against the lens board
Opaque black card	2 pieces, one A4 and one 25 mm square	The flare patch and the masking card
Black cloth, matt, about 1 m²	1	The room-reflection test
Shims: brass or aluminium foil, drink-can metal, thin card	an assortment	Alignment correction. Cut them oversize and trim
Adhesive labels and a chinagraph pencil	a few	Sheets are identified in the dark

Chemicals

All three baths are made up to working strength before the session, so that nothing is measured in the dark, and one mixing serves the whole three hours so that no sheet meets a fresher bath than another.

Chemical	Quantity	Form
Paper developer — a bought concentrate at its stated dilution, or D-72 made up separately	1 L working solution	Working-strength dish solution at 20 °C, ILFORD’s own dilution and time for the paper in use
Acetic acid or citric acid stop bath	1 L working solution	Working-strength dish solution
Fixer — sodium thiosulfate or an ammonium thiosulfate rapid fixer at its stated dilution	1 L working solution	Working-strength dish solution, non-hardening, within its published sheet capacity
Water for the wash	as the paper’s sheet requires	RC washes in minutes; fibre base does not, and a fibre session is a longer session

The developing agents — metol and hydroquinone — and the sulfite and carbonate are listed because the page depends on them, not because they are weighed here. A bought concentrate is the same chemistry in a bottle.

!One bath, twenty sheets, and the capacity that decides it

The single largest threat to this session is not the enlarger; it is a developer that has done more work by sheet eighteen than it had by sheet two. Check the published capacity on the maker’s sheet for the volume you have mixed, and if twenty sheets of 8 × 10 exceeds it, mix more solution rather than replenishing part way through. Every comparison on this page assumes the chemistry was the same for every sheet, and the moment it is not, a uniformity difference and a developer difference become the same measurement.

Equipment

The enlarger, on a solid bench, with its column locked. A negative carrier. Your working lens. A grain focuser — the one tool on this page worth buying rather than making, because focusing on the grain rather than on the image is what makes the focus independent of what the camera did. A timer that can be read in the dark without emitting light. Three dishes and three pairs of tongs, marked. A thermometer. A graduate. A loupe, 8× or 10×. A tape measure and a steel rule. An easel, or a sheet of glass and four register marks.

Where you have built them: Part XV’s reflection densitometer for reading the maps, and its photodiode head for the repeatability measurement at the easel.

📓Why this page can use a densitometer whose absolute scale is unestablished

Part XV is explicit that its reflection scale is relative to a white tile of a stated serial number, that the absolute reflectance of that tile is unknown, and that the scale’s absolute linearity has not been established. That sounds disqualifying and is not, for one reason: every number on this page is a difference.

A corner-to-centre ΔD, a difference between two apertures, a difference between opposite corners — each is a subtraction of two readings taken minutes apart on the same instrument against the same tile, so the tile’s unknown reflectance cancels exactly and only the internal consistency matters. That is precisely the figure Part XV does quote, in three flavours, and the one to compare your ΔD values against is the middle one: sample replaced between readings.

What you may not do with these numbers is compare them with somebody else’s, or with a manufacturer’s specification. The certificate says so in its own words.

Estimated cost

Cost band ££, and every penny of it is paper and chemistry. The mirror, the shims, the card and the cloth come to almost nothing; the loupe and the grain focuser are equipment you keep.

The planner prices the consumables and almost nothing else here. It has no line for a grain focuser, a first-surface mirror, an easel or shim stock, and its gap register already names a small spirit level among the things it could not price — which matters on this page only as a warning, because A1 does not use one.

The one large figure the planner does carry is the enlarger itself: a single listing for a new machine at £1,549, dated 5 September 2026, with the file’s own note that it is the least representative number in it, because almost every darkroom enlarger comes from the second-hand market and a used price is a transaction rather than a listing. This page is what makes a second-hand machine safe to buy, which is a genuine economy the planner cannot show.

Estimated consumables cost

One run of this page is the full commissioning: alignment, two uniformity apertures, two magnifications, the corner print, the repeatability series and three flare tests.

Consumed	This session	Sourced price	Cost this session
Photographic paper, variable-contrast RC, 8 × 10 in	18 sheets — 4 uniformity maps, 2 corner prints, 1 control, 3 flare sheets, and 8 for finding exposures and for the repeatability series	£33.50 for 25 (£1.34 a sheet) to £96.18 for 100 (£0.96), dated 5 September 2026	£17.28 to £24.12
Paper developer, working strength	1 L of 1+9, that is 100 mL of concentrate	£10.52–£20.03 per 500 mL to 1 L of concentrate	£1.05 to £4.01
Stop bath, working strength	1 L of 1+19, that is 50 mL of concentrate	£10.66–£12.18 per 500 mL of concentrate	£1.07 to £1.22
Fixer, working strength	1 L of 1+4, that is 200 mL of concentrate	£21.05–£25.98 per 1 L of concentrate	£4.21 to £5.20
Wash water	about 20 L	Not priced by the course	—
Shim stock, black card, black cloth, labels	as needed	None. The planner carries no line for any of them	—
One sacrificed negative, fixed out clear	1 frame	Not priced separately	—

The priced rows give £23.61 to £34.55 for one run, and it is a floor rather than a total. Three of the seven rows carry no dated price. The paper figure also assumes you find each exposure in one attempt, which you will not on the first commissioning: budget 25 sheets the first time and 15 every time after, because the exposures are written down and only the maps have to be repeated.

Equipment is excluded on purpose. The enlarger, the lens, the loupe, the grain focuser, the mirror, the densitometer, the trays and the tongs are not consumed and a page that quietly counted them would corrupt the only figure the consumables calculator has.

Alternative route

Without a densitometer, the whole of A2 still works and it costs you precision rather than the result. Cut each uniformity sheet into its nine squares, shuffle them, and try to sort them into order by eye against a neutral surround. If you cannot reliably order them, the field is even to about the limit of your eye — a real finding, and a weaker one than a number. The symmetry criterion survives almost intact: comparing two opposite corner squares side by side is the easiest visual judgement there is, far easier than judging a gradient. Write down which test you did.

Without an enlarger of your own — a shared, hired or borrowed darkroom. Take this page with you and run it there. A shared darkroom’s enlarger is the archetype of a machine nobody has measured, and a commissioning certificate for it is the most useful thing you can leave behind. The result belongs to that machine on that day and it expires when somebody knocks the head, which is a fact worth writing on the certificate rather than a reason not to make one.

Without an enlarger at all. Three of the seven tests transfer to the contact printer unchanged, and they are worth running there: the uniformity map, the repeatability series and the flare tests. What does not transfer is alignment and corner sharpness, and their absence is the contact printer’s genuine advantage — two surfaces clamped together cannot be out of parallel and have no field angle to lose light at. If you print by contact, this page’s real content for you is the certificate: the idea that an instrument’s claims are dated measurements rather than assumptions.

Without a dark room at all. There is no route to this measurement without a space that can be made dark, because every one of its detectors is a sheet of photographic paper. What a reader without one can do is what the room page sets out: a temporary blackout for the length of one session. The tests do not need a permanent darkroom; they need three uninterrupted hours of darkness and a bench that does not move during them.

Waste streams

Three, unchanged from every printing session and owned elsewhere. Spent developer and stop bath go to their labelled containers under the general chemical waste procedure. Spent fixer and the first wash carry dissolved silver and go to silver-bearing waste for recovery, as Part XII teaches. Local regulation governs disposal and this page gives no jurisdiction-specific instruction.

The processed sheets are ordinary solid waste once dry — but they are the raw data behind the certificate, so keep them, dated, in the same envelope as the certificate. A number without the sheet it came from is an assertion.

Preparation

Do all of this in the light. Nothing here should be happening while you are trying to read a bubble or find a shim in the dark.

Prepare the machine. Bench solid. Column locked. Lens clean, checked off-axis with a torch for the haze and fungus the optics page describes. Carrier clean. If you have a converted head, start its warm-up now, because the first measurement is twenty minutes away and the head should be settled before it.

Choose and write down the working settings: the aperture two stops down from maximum — Rodenstock’s own recommendation for a six-element lens, and ILFORD’s f/8 — the magnification you print at most, and the negative format. These three go on every sheet and on the certificate, because a result without them is not a result.

Prepare the baths, one mixing for the whole session, in three marked dishes in a fixed order, at 20 °C, with the thermometer left in the developer and one pair of tongs per dish.

Prepare the sheets. Twenty from one box, each notched at the same corner so that in the dark you know which way up and which way round it is — which matters more here than anywhere else in the course, because a uniformity map read the wrong way round reverses its own diagnosis. One sheet goes straight back into its packet as the control.

Prepare yourself. Kodak ask five minutes of dark adaptation before looking for leaks and ten before a visual judgement. Sit them out; you will judge badly for the first ten minutes.

Procedure

A1 — Alignment, by mirror null, in two axes (40 minutes)

A spirit level tells you whether a plane is horizontal. It does not tell you whether two planes are parallel, and parallel is what matters — a head squared to a bench that leans is still aligned, and a head level on a bench that is level is still misaligned if the easel sits on a warped board. Worse, the tolerance the optics page derives is about 0.28° at the negative stage at f/5.6 and 4×, which over a 100 mm level base is a rise of half a millimetre — less than the width of a bubble’s line.

So the method is a null: two things are made to coincide, and coincidence is judged far more finely than a position is read.

The mirror null, and what a misaligned reflection looks like

1. 1Mirror flat on the easel, reflecting face up — ordinary mirror glass gives a second faint reflection; a first-surface mirror does not
2. 2Lens and iris, at the working aperture — the thing whose reflection you are judging
3. 3Eye at the open negative stage, on the axis — carrier out; look down the axis, not from beside it
4. 4A tilt of theta returns the beam at two theta — this doubling is why a null beats a reading
5. 5Aligned: reflected iris concentric with the real one — the pass
6. 6Out by a fifth of a diameter: the limit — a crescent on one side, edges nearly touching on the other
7. 7Clearly out: the reflection breaks the iris edge — shim and re-check

The right-hand panel is what you actually judge. Everything in the left-hand panel is there to explain why a displacement you can see corresponds to an angle you cannot measure any other way at home.

Lens board against easel. Remove the negative carrier. Lay the mirror flat on the baseboard, under the lens. Set the head to your working height and the lens to your working aperture. With the head lit — the focus mode of an LED head, or the lamp of a tungsten one, briefly — look down the axis through the open negative stage and find the reflection of the iris. Move your eye until the real iris and its reflection are as nearly concentric as you can get them, then judge.

Negative stage against lens board. Put the mirror in the carrier, reflecting face down, and look up at it from beside the baseboard along the axis; the same coincidence, the same judgement. If you cannot get a mirror into your carrier, the practical substitute is to check both stages against the baseboard and rely on transitivity — with the caution that two errors in the same direction can cancel in that test and will not cancel in a print.

Do both in two axes. Turn the mirror through 90° and repeat, because a tilt has two components and a single view can null one while missing the other.

🧮Deriving the pass criterion, and showing it is stricter than the tolerance

A mirror tilted by θ deviates a reflected beam by 2θ. The iris sits a distance v above the mirror, so its reflected image is displaced sideways by 2θv. Require that displacement to be smaller than a fifth of the iris diameter D = f/N:

θ < D ÷ (10 v) = 1 ÷ [10 N (1 + m)]

The mirror test's criterion, as an angle

because v = f(1 + m) and the focal length cancels — the same cancellation the optics page’s exposure derivation produced, and for the same reason.

At f/8 and 4×: θ < 1 ÷ (10 × 8 × 5) = 0.0025 rad = 0.14°. The tolerance the optics page derived for the negative stage at f/5.6 and 4× was 0.28°. So a mirror null judged to a fifth of an iris diameter is about twice as strict as the tolerance it is testing, which is what an acceptance test should be.

Two consequences worth having. The test gets stricter as you stop down, because a given displacement is a larger fraction of a smaller iris — so run it at the working aperture and then check at one stop smaller. And the test gets stricter at higher magnification, which is the same (1 + m) that made the negative stage the critical plane in the first place.

Shimming. Where a plane is out, find the two fixings it pivots on and put foil under the low one. Shim oversize, trim after. Tighten to the same torque you will actually use — a joint that is aligned loose and tightened afterwards is not aligned.

Then move the head and check again. Raise it 100 mm, lower it, lock it, and re-run the null. An alignment that does not survive being moved is not an alignment, it is a coincidence, and it will be gone by the third print. This is the step people skip and it is the one that decides whether the certificate means anything next month.

A2 — Uniformity, as density on paper (45 minutes)

Why paper and not a meter. Uniformity is measured with the detector the instrument exists to expose, and this instrument exists to expose paper. A photopic lux meter has a different spectral response from the emulsion, so it answers a slightly different question. More importantly, paper is the most sensitive detector you own for this job: at a mid-grey the paper’s curve is at its steepest, so a small difference in illuminance becomes the largest difference in density it will ever become. The optics page’s arithmetic puts it at about 0.04 in density for a 5 per cent difference in light on a middle grade — and 0.04 is the only figure this course has for a change that is detectable on paper, which ILFORD publish in the context of safelight testing.

1. Carrier empty, no negative. Working aperture, working magnification. Head warmed to its rule.
2. Find the exposure that gives a mid-grey with a quick test strip. Not light, not dark: the middle.
3. Expose a whole sheet. Process it with everything else, in the same bath at the same temperature for the same time, and dry it fully — a print read wet reads about a tenth of a density low, and drying is not uniform across a squeegeed sheet.
4. Read nine positions: the centre, the four corners, and the midpoint of each edge. Express every reading as a difference from the centre.

ΔD = D(point) − D(centre)

One grid point's deviation

5. Repeat at two stops smaller, at the same magnification, with the exposure adjusted to land on the same mid-grey.
6. Repeat at a second magnification, back at the working aperture, again landing on the same mid-grey.

The nine reading positions, and a worked pair of maps

1. 1Nine positions: centre, four corners, four edge midpoints — the centre is the reference and reads zero by definition
2. 2Symmetry test: opposite corners within 0.04 of each other — PASS. Asymmetry has no geometric floor, so it is a fault
3. 3Aperture test: corners improve on stopping down — PASS. Optics improve; metal does not
4. 4Magnitude: recorded, not passed — the cos⁴ reference alone gives about -0.20 at the corner at 4x

Worked figures, not measurements of any enlarger. What the pair of maps demonstrates is the two judgements, which are about symmetry and about response to aperture rather than about size.

!The two pass criteria, and the one figure that is recorded instead

PASS 1 — symmetry. Opposite corners must agree within 0.04 in density, and so must opposite edge midpoints. This is a real pass mark and it is defensible, because asymmetry has no geometric floor: cos⁴ falloff, a mixing chamber and a lens are all symmetric about the axis by construction, so anything that is not symmetric is misalignment, an off-centre source, or something mechanical in the way. If this fails, go back to A1 before doing anything else.

PASS 2 — response to aperture. The corner-to-centre difference must get smaller when you stop down two stops. This too is defensible, because optical falloff narrows the cone about the axis and a piece of metal does not care. If the corners do not improve, you have a mechanical vignette — a carrier aperture, a filter drawer, a bellows fold, a lens board recess — and it is found by looking up through the lens from the corner of the easel to see whether the whole aperture is visible from there.

RECORDED, not passed — the magnitude. The worst corner-to-centre ΔD at each aperture and each magnification goes on the certificate as a number, and the course sets no limit on it. The optics page explains why: the cos⁴ reference alone puts the corner of a 4× enlargement about a third of a stop down, which is about 0.20 in density, five times the figure a contact printer meets. A pass mark borrowed from an instrument with no lens would fail every enlarger ever made. What the number is for is comparison with itself: with the same enlarger next year, and with the decision about whether to burn the corners and by how much.

A3 — Sharpness into the corners (25 minutes)

Put the fine-detail negative in the carrier. Focus at full aperture with the grain focuser at the centre of the field — ILFORD’s own sequence, and it works because depth of focus is narrowest wide open, which is where the point of best focus is easiest to find. Then stop down to the working aperture and print. Dry the print and read all four corners with the loupe, plus the centre.

PASS: every corner resolves the same finest detail as the centre. If one or more does not, the diagnosis is in which corners fail and how, and you must diagnose rather than average.

Four corner failures, and how each one is drawn on the print

1. 1Misalignment — asymmetric: one corner soft, the opposite sharp, with a keystone. Recovers on stopping down
2. 2Field curvature, or the lens outside its scale range — symmetric: all four corners equally soft. Recovers on stopping down
3. 3The lens does not cover — symmetric, soft AND dark, worsening to the extreme edge. Does NOT recover
4. 4A buckled or popping negative — a band across the middle, in a different place on every print

Two questions separate all four: is it symmetric, and does it improve on stopping down. Only the third fails both.

What fails, and how	Diagnosis	Confirming test
One corner, or one edge, soft; the opposite one sharp	Misalignment. Softness that varies across the frame asymmetrically is the signature	A1 again. And look for keystone: measure the printed image’s two opposite edges with a rule
All four corners soft, symmetrically, and they improve markedly on stopping down two more stops	Field curvature, or the lens working outside its magnification range	Check the lens against its maker’s scale range. Rodenstock rate a simple three-element lens 2× to 8× and no further
All four corners soft and dim, with the softness continuing to the extreme edge and getting worse	The lens does not cover the format. Not curable by stopping down beyond a point, because coverage is a limit rather than an aberration	Compare the negative diagonal with the lens’s rated maximum film format
Softness in a band or a patch, in a different place on each print, sometimes appearing part way through an exposure	A buckled negative, or film popping	A glass carrier, or let the negative warm in the beam for a minute before final focusing
Everything soft, evenly, including the centre	Focus, or diffraction from too small an aperture. Not a corner fault at all	Refocus at full aperture; then check that N(1 + m) is not deep into the diffraction region

A4 — Focus, depth of focus and the filter check (15 minutes)

Three quick checks that use what A3 has already set up.

Focus with the paper in place. A sheet of paper is about 0.2 mm thick and the easel holds the emulsion at its top surface. Focus on a spare sheet in the easel, as ILFORD instruct, not on the bare baseboard.

Confirm the depth of focus is doing what you think. Focus at full aperture, print at the working aperture, and compare with a print focused at the working aperture. If they differ visibly, either your focus finder is not sitting flat or the alignment is worse than A1 said.

The filter check. If you print with a contrast filter below the lens, focus with the filter out, then swing it in and refocus, and note whether the focusing head moves. The optics page’s paraxial estimate says a thin gelatin filter shifts the focus by about 0.03 mm — negligible against a depth of focus of millimetres — and a 2 mm glass one by about 0.67 mm, which is not negligible at full aperture. No manufacturer figure for this shift was found in this course’s corpus, so this check is how you get one for your own filter.

Then adopt the rule that makes the question go away: focus with the filter in place.

A5 — Exposure repeatability (20 minutes)

Repeatability, over ten identical exposures measured at the easel with the photodiode head, each one preceded by a full switch-off and switch-on so that the test includes the whole operating cycle rather than a convenient part of it. Take the mean and the sample standard deviation, and convert:

stops = log₂(1 + s)

Scatter in the units that decide

where s is the standard deviation as a fraction of the mean.

PASS: the standard deviation is below half the smallest exposure increment you actually use. That is the right form of the criterion because it compares the machine’s noise with the smallest deliberate change you make: an instrument whose scatter is bigger than your finest adjustment cannot be adjusted. Part XVII owns f-stop timing and this course has not yet fixed an increment, so you state yours. If you have not chosen one, use a twelfth of a stop, which makes the criterion a standard deviation below about 3 per cent of the mean.

Without a photodiode head, the same test runs on paper and costs more sheets: ten identical mid-grey exposures on ten sheets, processed together, read and compared. It is a coarser measurement — you are now measuring the enlarger and the processing together — and that is a limitation to write down rather than a reason not to do it.

A6 — The flare hunt (25 minutes)

Three tests, and each excludes something different. Run them in this order.

F1 — the capped-lens test, which excludes leaks. Lens capped, no negative, head on at printing current for four times your longest printing exposure, with a sheet on the easel half covered by a card. Process it with the session. PASS: no visible line between the halves and no reading above the control sheet. A fail means light reached the paper without going through the lens: the lamphouse, the bellows, the room, or a lighted dial — all four of which Kodak name explicitly among the things that fog paper.

F2 — the flare-patch test, which measures veiling glare. Put the clear, fixed-out film in the carrier so the whole field is at full brightness. Lay the 25 mm square of opaque black card in the middle of a sheet on the easel. Expose for the time that would give a good mid-grey through a real negative, remove the card, and process the sheet. The area under the card received no image-forming light at all; any density there is light that came from the bright surround and arrived by scattering inside the lens, the bellows and the room.

Read that density against the control sheet’s base-plus-fog. The difference is your flare, in density units, under the most punishing condition your enlarger will meet — which is exactly what a printer meets when printing a thin negative with large clear areas.

F3 — the room-reflection test, which asks whether your walls cost you black. Two identical sheets, both given an exposure well past the one that first reaches maximum black. One made with the room as it is; one with the black cloth draped over the baseboard surround, the wall behind the enlarger and anything else pale within a metre. Process together, dry, and read both.

PASS: the two agree within 0.04 in density. A difference larger than that says your room is measurably costing you maximum black, and the fix is matt black where the light lands rather than a change of paper.

📓What this page could not source, and what it does instead

No published account of the capped-lens, flare-patch or room-reflection tests was found in this course’s corpus. The three procedures above are built from the mechanism rather than copied from a manual: F1 is the same logic as the head’s own leak check, F2 is the standard way of measuring veiling glare — a known-zero area inside a bright field — and F3 is a controlled comparison with one variable.

The criteria attached to them are equally honest about where they come from. 0.04 in density is ILFORD’s published figure for a detectable change on paper, in a different context, pressed into service here because it is the only such figure the course has. It is a detectability threshold, not a quality standard, and it is used as one: F1 and F3 ask whether the effect is detectable at all, which is a question that threshold can answer.

And the honest gap: this course has not itself run F3 in a small light-walled darkroom and has no result to report. The research plan asked whether room reflections measurably reduce maximum black, and the answer here is a test rather than a finding. If you run it, the number you get is worth more to you than any figure this page could have printed.

A7 — Write the certificate (10 minutes)

Now, while the sheets are still on the bench and the settings are still on the enlarger. A certificate written a week later is a certificate written from memory.

Expected observations

A1. Most enlargers that have never been aligned are out. Expect one plane clearly off, twenty minutes on shims, and a first null that looks fine while the second, at 90°, does not — a tilt has two components.

A2. The map falls off towards the corners on every enlarger, and that is geometry rather than a fault. What you are looking for is symmetry and response to aperture. A stripe or a bright patch is a head fault rather than a lens one.

A3. A corner soft at the working aperture and sharp two stops down is alignment or field curvature; a corner that never comes right is coverage; a softness that moves between prints is the negative.

A5. A warmed LED head should scatter a few per cent. A tungsten head switched from cold every time will be worse, because the filament’s output climbs through the first part of every short exposure.

F2. Expect a measurable flare density. Every enlarger has some, and a zero means your reading was not sensitive enough rather than that your enlarger is perfect.

What is happening chemically

The whole of this page uses one chemical fact: a sheet of photographic paper turns a small difference in light into a large difference in density, and it does it most strongly in the middle of its scale.

Part XIII established the shape: a paper’s characteristic curve has a long, nearly flat toe where the highlights live, a short steep middle, and a hard ceiling at maximum black rather than a taper. The slope of that curve is what converts illuminance into density, so the same five per cent difference in light produces almost nothing on the toe, about 0.04 in density in the steep middle, and almost nothing again at the ceiling. That is the entire reason A2 exposes to a mid-grey rather than to a light tone or a dark one, and it is why the map would be useless if it were made at either end.

Underneath the curve is the ordinary development chemistry the course teaches elsewhere. A latent image is a few atoms of silver at a sensitivity speck; the developer reduces the whole grain that carries one and leaves the rest; the number of grains that cross that threshold in a given time is what the density is. The steep part of the curve is where a small change in exposure moves a large number of grains across the threshold at once, which is the same statement as “the slope is high” in a different vocabulary.

Two chemical confounders can imitate an optical result and both are controlled by procedure rather than by argument.

Uneven development produces a density gradient across a sheet that looks exactly like an illumination gradient. Continuous agitation in a dish large enough for the sheet to move, and the same agitation for every sheet, is the control. The test that separates it is the one the break/fix page owns: rotate the paper through 180° on the easel and see whether the pattern follows the paper or the room.

Dry-down. A print read wet reads lower than the same print dry, and the difference is not uniform across a sheet that has been squeegeed unevenly. Every sheet on this page is read dry, and dried the same way.

⚗Why the maximum-black test in F3 is the one chemical measurement here

Maximum black is a ceiling set by the paper rather than by the exposure: past a certain point more light makes no more density, because the silver available in the emulsion has been reduced. ILFORD publish 2.15 for their current Multigrade RC Deluxe against 2.05 for the previous generation on the same base, which is what makes it a usable reference at all — it is a property of a named material rather than of your session.

That is why F3 works as a test of the room. Both sheets were taken past the exposure that first reaches maximum black, so both should be sitting on the same ceiling, and any difference between them is light that arrived without carrying an image. It is the only measurement on this page where the paper’s chemistry supplies the reference rather than the detector.

Data to record

Everything in this table goes in the notebook during the session, not afterwards.

Record	Where it comes from	Why it is on the certificate
Enlarger, lens, focal length, carrier	Read off the machine	A certificate belongs to a configuration, not to a room
Working aperture and magnification	Your choice at Preparation	Every other number is conditional on these two
Alignment result, two axes, two stages, before and after moving the head	A1	The “after moving” line is what says the setting survives use
Nine ΔD values at the working aperture	A2	The map
Nine ΔD values two stops down	A2	The falloff-against-vignette discriminator
Nine ΔD values at the second magnification	A2	The lens-against-chamber discriminator
Worst-point ΔD in each case, and in stops	A2, converted	The number a future comparison uses
Corner sharpness verdict, four corners	A3	Pass or a named diagnosis; never “mostly fine”
Whether the filter shifts focus, and by how much	A4	Because no manufacturer figure exists
Ten exposure readings, mean, standard deviation, and the figure in stops	A5	The repeatability
The exposure increment you compared it against	A5	Without it the pass criterion is not reproducible
F1 verdict, F2 flare density, F3 difference	A6	Flare changes the grade you print at
Head warm-up rule applied, and whether it was	A1 to A6	A measurement made during warm-up has a gradient in time
Room leak test date, safelight test date	The room and safelight certificates	Two of this page’s assumptions, dated
Paper, developer, dilution, temperature, time	Preparation	The detector is a material and it has a batch
Date, and who did it		The whole point

Analysis

Turn the maps into two judgements and one number

Judgement 1, symmetry. For each of the two corner pairs and each of the two edge pairs, take the difference between opposite readings. All four differences must be within 0.04. Report the worst.

Judgement 2, response to aperture. Take the worst corner ΔD at the working aperture and at two stops down. The second must be smaller. Report both.

The number. The worst corner-to-centre ΔD at the working settings, in density and in stops. To convert, divide by the local slope of your paper’s curve to get Δlog H, then divide by 0.301:

stops = ΔD ÷ (γ × 0.301)

Density difference into stops

with γ ≈ 2 for a middle grade, from the optics page’s arithmetic. A worst point of 0.20 in density is about a third of a stop — which is what you burn back in, and now you know how much.

Read the second magnification

Compare the shape of the map at the two magnifications, expressed as a fraction of the sheet.

• Same shape, same magnitude: the lens and the geometry. Expected, and nothing to fix.
• Worse at the higher magnification: also expected, because the field angle grows with m at a fixed focal length, and cos⁴ grows with it.
• A stripe, a patch, or an off-centre bias that changes: the head’s mixing chamber. Go back to its build page’s stage 2.

The enlarger certificate

One page. Every later page that quotes an exposure from this enlarger cites it, and the calibration record set is where it lives — with the honest note that that worksheet’s own gap list still says the enlarger’s sheet does not exist yet, so for now this table is the sheet.

Field	What goes in it
Enlarger, and how it is identified	Make, model, and a mark you put on it yourself
Head	The donor’s own lamphouse, or the built LED head with its own record attached
Lens, focal length, maximum aperture, and its rated scale range	From the maker where you have it; “unknown” where you do not
Carrier	Glass or glassless; anti-Newton or not
Date, and by whom	And the date the next check is due
Working settings	Aperture and magnification, and the negative format
Alignment	Result in two axes at both stages, the criterion used, and whether it survived the head being moved
Uniformity	Worst corner-to-centre ΔD at two apertures and two magnifications, in density and in stops
Uniformity symmetry	Worst opposite-pair difference, against the 0.04 criterion
Safe working aperture range	The apertures at which the corners are acceptable to you, with the reason
Warm-up rule	From the head’s record, and the note that it was applied
Repeatability	Standard deviation in stops, the number of exposures, and the increment it was compared against
Flare	F1 verdict, F2 density above base-plus-fog, F3 difference
Reading method	Densitometer and tile serial, or “visual, sorted by eye”
Known limitations	Every gap on this page that applies to you, written out
Signature and notebook page	So the certificate points back at the raw data

✕Five claims this certificate must not make

“Aligned.” Say what was measured: “reflected iris concentric to better than a fifth of its diameter at f/8 and 4×, in two axes, at both stages, re-checked after moving the head”. An unqualified “aligned” is a claim about every setting, and you measured one.

“Uniform to 0.04.” Not unless you measured it, and you almost certainly did not, because the cos⁴ reference alone is five times that at the corner of a 4× enlargement. Write the number you got.

“Sharp corner to corner.” At one aperture, one magnification, with one lens, on one negative. All four qualifiers or none.

“Repeatable to 2 per cent.” Only with the number of exposures and the cycle they included. Ten exposures without switching off is a different measurement from ten with.

“Flare-free.” No optical system is. A zero on F2 means your reading method could not see it, and the honest sentence is “no flare density detectable above base-plus-fog by visual comparison”, which says what you did.

Applying all of this to a bought enlarger

Which is where most readers will start, and nothing changes except two things.

Run A1 first and expect to fail it. A machine that has been moved, boxed, carried and set up on a different bench has no reason to be aligned, and second-hand enlargers are frequently sold by people who never checked.

Add one test at the front: does it hold its settings? Set the head at a height, lock it, print, and then push gently down on the head and re-measure the projected image size with a rule. A column that creeps is a column whose focus creeps, and no amount of alignment fixes it. If it creeps, the fault is usually a worn locking collar or a dirty column, and both are cheap.

Then re-run the whole page after any move, knock, lens change or carrier change. The certificate belongs to a configuration and every configuration eventually changes.

Troubleshooting

Symptom	Likely cause	Test that distinguishes it
The mirror reflection is doubled and you cannot tell which to null on	Ordinary mirror glass reflecting from both surfaces	Use the brighter one consistently, or get a first-surface mirror. At this scale the two are about a millimetre apart
The null looks perfect in one axis and hopeless at 90°	Normal. A tilt has two components	Shim for one axis, re-check both, iterate
Alignment passes, then fails after the head is moved	A locking collar that does not lock, or a column that flexes	Push-test the head; re-tighten to a fixed torque
The uniformity map is asymmetric even after A1 passes	An off-centre source, or something mechanical in the path	Rotate the carrier 180° and re-map; a pattern that follows the carrier is the carrier
The map has a stripe parallel to nothing in particular	The mixing chamber, emitter pitch printing through	The head’s own T4, at two magnitudes
Corners do not improve on stopping down	A mechanical vignette	Look up through the lens from the corner of the easel: can you see the whole aperture from there?
Every sheet is a different overall density	A5 has already told you: repeatability	Check the warm-up rule was applied, and whether the timer or the lamp is the variable
A gradient that follows the paper rather than the easel	Development, not the enlarger	Rotate the sheet 180° in the tray. This is the break/fix page’s test
The control sheet itself is grey	Not the enlarger. Paper, storage or the developer	A fresh sheet from an unopened packet, processed alone
F2 gives a large flare density and the lens looks clean	The bellows, the room, or an internally shiny lens barrel	Repeat F2 with the black cloth in place; if it drops, it was the room
Everything passes and prints still look flat	Flare that F2 measured and you have not acted on, or a negative that is genuinely flat	Compare the projected range against the paper’s ISO(R) figure, remembering ILFORD state it as projected on the baseboard

Clean-up

Trays emptied into their labelled waste containers, rinsed, dried face down; tongs rinsed and hung, one per dish; thermometer rinsed and put away dry.

Then the machine. Lens cap on. Carrier out, blown clean, stored closed. Head lowered to the bottom of the column and locked. Leave the shims in and leave the fixings at the torque you tightened them to — the alignment you measured belongs to that state and nothing else. Wipe the baseboard, because a dried splash of fixer there is a permanent mark on the next print’s easel.

Storage

The certificate and its sheets go together, in one dated envelope, filed with the sensitometer and densitometer certificates. A summary whose raw data has been thrown away cannot be checked.

The chemistry: developer and stop are one-session solutions here and are not kept. Fixer may be kept within its published capacity, labelled and dated, under storage rotation.

The paper: the remainder goes back into its sealed light-tight packet in a cool dry place in its original packaging — ILFORD’s own instruction, and what makes the next commissioning comparable with this one.

Disposal considerations

Three streams, each described as chemistry and general practice rather than as a jurisdiction-specific instruction.

Spent developer is an alkaline solution of oxidised developing agents and sulfite. It goes to the labelled container under general chemical waste.

Spent stop bath is a dilute weak acid, largely spent. Same container, same procedure, and never mixed with the fixer container.

Spent fixer and the first wash carry dissolved silver as thiosulfato complexes, which is why they are collected rather than tipped: Part XII teaches the chemistry and the recovery, and silver-bearing waste is the procedure.

And the standing sentence, which is not a formality: local regulation governs what happens to all three. The disposal policy sets out what the course can and cannot tell you, and why the answer is jurisdictional.

Questions

1. Your mirror null is judged at f/4 and passes. A friend argues you should have judged it at f/11 because “smaller is more precise”. Using the criterion θ < 1 ÷ [10 N(1 + m)], say who is right and by what factor.
2. A uniformity map shows the top two corners at −0.05 and the bottom two at −0.19, at 4× and f/5.6. Stopping down two stops gives −0.03 and −0.11. Which pass criterion has failed, which has passed, and what do you do next?
3. Explain why this page can use a densitometer whose absolute reflection scale is admittedly unestablished, and name one number on the page that this reasoning would not justify.
4. Ten exposures give a mean of 3.86 arbitrary units with a standard deviation of 0.15. Convert to stops. Then say whether the enlarger passes if your smallest exposure increment is a sixth of a stop, and whether it passes if it is a twenty-fourth.
5. F2 gives a density of 0.11 above base-plus-fog under the black patch. F3 shows no difference between the two maximum-black sheets. What have you learnt about where the flare is coming from, and what would you check next?
6. The bottom-left corner of every enlargement is soft, and stopping down two stops nearly cures it. Give the diagnosis and justify it from depth of focus rather than from experience.
7. Read this certificate line and name one claim it does not support: “Uniformity: ±0.04 D. Alignment: good. Lens: 50 mm. Checked 2026-09-05.”
8. Your enlarger passes every test on this page and your prints are still a grade flatter than the same negatives printed on a friend’s machine. Both of you use the same paper and the same developer. Name the two most likely explanations and the measurement that separates them.

Further experiments

Measure your own paper’s mid-tone slope instead of using γ ≈ 2. Print a step wedge by contact, read it, and take the slope through the middle of the curve. Every density-to-stops conversion on this page becomes yours rather than an approximation, and Part XIII’s paper sensitometry has already given you the method.

Map the falloff against the cos⁴ reference. Compute the field angle at each of your nine positions from v = f(1 + m) and the position’s distance from the axis, take cos⁴ of each, convert to density with your own γ, and plot your measured map against it. Where your lens beats the reference, it is doing something clever with its pupil; where it does not, something else is going on. Rodenstock publish exactly this comparison for their own lenses, and you can now publish it for yours.

Settle the room-reflection question properly. F3 as written is a yes-or-no. Run it as a series instead: bare room, cloth on the baseboard only, cloth on the wall only, cloth on both. You will find out which surface is costing you the black, which is the difference between painting one wall and painting four.

Test whether alignment survives a year. Re-run A1 in twelve months without touching anything, and record what you find before you correct it. Nobody publishes a drift figure for a domestic enlarger and you would then have one.

Compare two lenses on the same machine. Everything except the lens is now measured, which makes your enlarger a test bench. Run A2 and A3 with a second lens and you have a controlled comparison of two lenses that no brochure can give you, because it is at your magnifications on your machine.

Recap

An enlarger is commissioned by four tests with criteria you can apply, and the criteria are chosen so that each one is defensible rather than convenient.

Alignment is a null rather than a reading, because a mirror doubles the angle and coincidence is judged far more finely than position. The criterion — the reflected iris concentric to better than a fifth of its diameter — works out at θ < 1 ÷ [10 N(1 + m)], which at f/8 and 4× is 0.14°, about twice as strict as the tolerance it tests. And it is only an alignment if it survives the head being moved.

Uniformity is measured as density on paper at a mid-grey, because that is where the paper’s curve is steepest and five per cent of light becomes 0.04 in density. Two things are passed and one is recorded: symmetry passes, because asymmetry has no geometric floor; response to aperture passes, because optics improve on stopping down and metal does not; and the magnitude is written down rather than judged, because the cos⁴ reference alone puts an enlarger’s corner five times below what a contact printer meets.

Corner sharpness is read with a loupe on a dry print and diagnosed rather than averaged, because misalignment, field curvature, insufficient coverage and a buckled negative fail differently and are cured differently. Repeatability is expressed in stops and compared with the smallest exposure increment you actually use, because an instrument noisier than your finest adjustment cannot be adjusted.

Then flare, in three tests that exclude three different things, and the certificate — which claims exactly what was measured, at what settings, on what date, and says in its own words what it does not claim.

Check your understanding

Question 1. Why is the mirror method used for alignment rather than a spirit level, given that a good level is cheap and quick?

• Because a spirit level cannot be read in the dark
• Because a level measures whether a plane is horizontal, while the requirement is that three planes are parallel to about a quarter of a degree - which over a 100 mm base is half a millimetre of rise, below a bubble line width, and which a level cannot see even in principle if the whole bench leans
• Because levels are not accurate enough to measure anything
• Because the mirror also measures the focus

Show the answer and why

Answer: Because a level measures whether a plane is horizontal, while the requirement is that three planes are parallel to about a quarter of a degree - which over a 100 mm base is half a millimetre of rise, below a bubble line width, and which a level cannot see even in principle if the whole bench leans

Two separate objections, and the first is the fatal one. Parallelism and horizontality are different properties: an enlarger on a bench that leans by two degrees is perfectly aligned if all three of its planes lean by two degrees, and a level would condemn it. The second objection is sensitivity: the derived tolerance is about 0.28 degrees at the negative stage, which is half a millimetre over 100 mm, and no bubble resolves that. The mirror method sidesteps both because it is a null between two things that must coincide, and because reflection doubles the angle it is looking for.

Question 2. A uniformity map at the working aperture reads: corners -0.19, -0.20, -0.18, -0.19; edge midpoints about -0.08; centre 0. What is the verdict?

• Fail: 0.19 is far outside the 0.04 the contact printer meets
• Pass on symmetry, since opposite corners agree to 0.01, and the magnitude of about -0.19 is recorded rather than passed because the cos⁴ reference alone accounts for roughly that much at the corner of a 4x enlargement
• Fail: the edge midpoints should be zero
• Cannot be judged without knowing the paper

Show the answer and why

Answer: Pass on symmetry, since opposite corners agree to 0.01, and the magnitude of about -0.19 is recorded rather than passed because the cos⁴ reference alone accounts for roughly that much at the corner of a 4x enlargement

This is the map of a perfectly ordinary, correctly aligned enlarger. What the numbers demonstrate is symmetry, which is the pass criterion that has a floor of zero: cos to the fourth falloff, a mixing chamber and a lens are all symmetric about the axis, so an asymmetry is a fault while a symmetric gradient is geometry. Applying the contact printer 0.04 figure to an enlarger would fail every enlarger ever made, which is a sign that the limit has been borrowed from the wrong instrument. The paper does matter for the conversion into stops, which is why the certificate records the paper.

Question 3. Ten exposures give a standard deviation of 4 per cent of the mean. Your smallest exposure increment is a twelfth of a stop. Does the enlarger pass A4, and what is the reasoning?

• Yes, because 4 per cent is a small number
• No: 4 per cent is log to base 2 of 1.04, which is 0.057 of a stop, against a criterion of half a twelfth, which is 0.042 of a stop - the machine scatters more than half your finest deliberate adjustment
• Yes, because repeatability only matters above 10 per cent
• It cannot be decided without knowing the paper grade

Show the answer and why

Answer: No: 4 per cent is log to base 2 of 1.04, which is 0.057 of a stop, against a criterion of half a twelfth, which is 0.042 of a stop - the machine scatters more than half your finest deliberate adjustment

The conversion is what makes the judgement possible: log2(1.04) = 0.057 stop. The criterion is expressed against your own finest adjustment because that is the only thing that makes an absolute figure meaningful - a machine whose noise exceeds your smallest deliberate change cannot be steered, and one whose noise is far below it is precise enough regardless of what the number looks like. Change the increment and you change the verdict: at a sixth of a stop the criterion is 0.083 and the same machine passes comfortably.

Question 4. The flare-patch test F2 gives 0.11 above base-plus-fog. What exactly has been measured?

• A light leak in the lamphouse
• Veiling glare: light from the bright surround that reached an area receiving no image-forming light, by scattering inside the lens, the bellows and the room - which is what lifts the shadows and takes away maximum black on a real print
• The paper base density
• The safelight fog

Show the answer and why

Answer: Veiling glare: light from the bright surround that reached an area receiving no image-forming light, by scattering inside the lens, the bellows and the room - which is what lifts the shadows and takes away maximum black on a real print

The design of the test is what makes the answer unambiguous. Under the opaque patch, no image-forming light can arrive at all, so any density there came from somewhere else, and the clear fixed-out negative guarantees the surround is as bright as the system will ever make it. A leak is excluded because F1 was run first with the lens capped, which is the point of running the three tests in order. Safelight fog is excluded by the control sheet and by the safelight test having already passed. The number matters because flare raises the shadows without touching the highlights, which is a contrast loss that a harder grade cannot repair.

Question 5. Your alignment passes, you print for a fortnight, and then every corner is soft again. Nothing has been dropped. What does the page say to suspect, and what did the original procedure do about it?

• The lens has developed haze
• A locking collar or column that does not hold, which is why A1 ends by moving the head, locking it and re-measuring - an alignment that does not survive being moved is a coincidence rather than a setting
• The paper has changed
• The head has drifted thermally

Show the answer and why

Answer: A locking collar or column that does not hold, which is why A1 ends by moving the head, locking it and re-measuring - an alignment that does not survive being moved is a coincidence rather than a setting

The re-check after moving the head is the step that separates a machine that is aligned from a machine that was aligned once. An enlarger is raised and lowered several times a session, and a collar that grips at one height and slips at another produces exactly this: correct at commissioning, wrong a fortnight later, with nothing having happened that anyone would remember. Haze is a real fault and it produces flatness rather than corner softness. Thermal drift changes exposure rather than focus, and it is the head record that catches it.

Question 6. What is the strongest reason this page insists that the certificate names the paper, the developer, the dilution, the temperature and the time?

• To make the record look thorough
• Because the detector in every measurement on the page is a sheet of photographic paper, and its slope is what converts illuminance into density - change the material or the processing and the same enlarger yields different numbers
• Because the chemistry is hazardous and must be logged
• Because the paper is the most expensive item

Show the answer and why

Answer: Because the detector in every measurement on the page is a sheet of photographic paper, and its slope is what converts illuminance into density - change the material or the processing and the same enlarger yields different numbers

Every ΔD on the page is an illuminance ratio multiplied by the local slope of a particular paper developed a particular way, so the paper and its processing are inside the measurement rather than beside it. A softer grade, a different developer or four minutes instead of two would give a different set of numbers for an unchanged enlarger, and a certificate that did not say which was used would be uncomparable with its own successor. This is the same discipline the densitometer certificate applies when it insists that a reflection density is quoted relative to a named tile.

Check my answersClear