Build: The Enclosure and Exposure Stage

Purpose

To build the box that turns the module from the previous page into an instrument, and then to prove two things about it that cannot be argued from a drawing: that it is dark, and that it is even.

Part VI already built a light-tight box and settled how joints are made to keep light out. That work is cited here rather than repeated: overlap at least eight millimetres, clearance about half a millimetre, one right-angle turn minimum and two where you can get them, every surface in the passage matt black. Nothing on this page improves on that rule and nothing here contradicts it.

What is different is the second requirement. A camera has one job with light: keep it out except through the hole. This box has two, and they pull against each other. It must keep room light out, and it must spread its own light across 127 mm of wedge to within a few per cent — which is why the interior of this box is not all one colour, and why the acceptance test at the end is a map rather than an inspection.

Learning objectives

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

• set out an enclosure from the two numbers that actually constrain it — the wedge’s length and the uniformity target — rather than from the size of the box you happen to own;
• say why the chamber below the diffuser is painted white and the space above it matt black, and what each colour is doing;
• build a stage whose registration puts the wedge and the film in the same place, in the dark, without looking;
• run a fog test that measures light-tightness on film rather than inspecting it by eye, and say why the inspection is still worth doing first;
• map the illumination across the film area, express the worst deviation in log H, and compare it with the 0.02 log H the design lesson set;
• decide from your own map whether the instrument needs a second diffuser, a deeper chamber or a narrower illuminated field, and say what each of those three costs;
• write a build record that a stranger could use to rebuild the same geometry.

Prerequisites

The LED light source and exposure timer, built, bench-tested and with its warm-up curve measured. A box around an uncharacterised lamp is a box around an unknown.

Sensitometer design principles, for the three numbers this build is dimensioned from: the wedge is 127 mm long, the uniformity target is 0.02 log H end to centre, and a bare small source would have to be about 420 mm from the film plane to meet it unaided.

Part XIII’s step-wedge lab, because the first exposure this box makes is processed exactly as that page processes one, and its control strip is what the fog test is read against.

Safety classification

Level A. The classification rubric puts a procedure at Level A when the substances handled are at most irritant or harmful in home-darkroom quantities, when nothing is heated above 50 °C, when no mains-voltage construction or modification takes place, and when the waste is spent developer or silver-bearing fixer that can be collected. Every one of those holds. The work is cutting, taping, gluing and one ordinary film development; the electricity is the USB module already built and certified supply already chosen.

The criterion that had to be argued. Level B names high-brightness LED sources among its criteria, and the previous page was classified Level B partly for that reason. This page is classified A because the emitter it houses spends almost the whole session either switched off or behind an opal panel, and because the thing being built is the enclosure that the rubric’s own Level B control for an energetic source asks for. The judgement is not unconditional and the page states the condition: during the uniformity map the lamp runs with the lid off, and for that step the previous page’s rule stands unchanged — run at the lowest current that gives a usable reading, keep the emitter below eye level, and do not look into it. If you would rather hold the whole session at Level B, nothing on this page is made worse by doing so.

What is not a hazard here, and why. There is no concentrated reagent anywhere in this session. The only chemistry is one film development at working dilutions — developer at 1+1, a stop bath at 1+19, fixer at 1+4 — handled in a few hundred millilitres, with the hazard and control set out on Part XIII’s lab page and not restated here. There is no vapour-producing step, because the two products that would have made one, solvent adhesive and spray paint, are designed out in favour of tape, PVA and black card. There is no mains work: the box has no switch in a mains circuit and no lamp of its own beyond the module. And there is no dust hazard on the recommended route, because nothing is sawn or sanded — the plywood option reintroduces that and is the one route that carries Part VI’s wood-dust controls with it.

Naming the absences is how the assessment is made rather than asserted: the same silver bromide in a different session is a Level B weighing, and what holds this page at A is the quantity, the dilution and the fact that every energetic component arrives already built and already tested.

Hazards

Cuts. Three light passes through 5 mm foam board, not one heavy one. HSE’s principles, written for kitchens but general: keep the blade sharp, because a blunt blade needs force and force is what slips; cut on a stable surface; put the tool away the moment you set it down; never try to catch a falling one. Cut with the stroke moving away from the hand holding the work. For a cut, NHS guidance is pressure to stop the bleeding, rinse clean, cover with a sterile dressing, and no attempt to remove anything embedded.

Eye injury from flying particles, if you drill the cable entry or the button hole. OSHA state the principle plainly: eye or face protection where there is a hazard from flying particles, with side protection where there is a hazard from flying objects. Spectacles on for every hole.

Adhesive. PVA is the adhesive for this build; it is a water-based emulsion and the control is ordinary care and washed hands. If you reach for cyanoacrylate, PubChem’s aggregation of the ECHA notifications for ethyl 2-cyanoacrylate carries H302, harmful if swallowed, and H315, causes skin irritation, both under the signal word Warning; it also bonds skin, and the tube’s own label is the authority for that. Keep the work below the level of your face.

Outgassing, and what it might do to film. FOMA instruct that unexposed film be stored out of reach of harmful vapours and gases, which is a manufacturer saying that vapours can damage sensitised material. Whether a particular paint or adhesive outgasses enough to fog a strip shut in a box with it is not documented anywhere this course has found, so what follows is a precaution and not a finding: leave the box open in a warm airy place until every surface is hard and odourless, and make the first strip that goes into it an unexposed control. That control is the light-leak test below, so the precaution costs you nothing you were not already doing.

Heat, at the lamp end. The heatsink from the previous page still reaches a temperature that will burn a finger held on it, and it is now inside a box where you cannot see it. Leave the module’s own heatsink proud of any wall it is fixed to, and never close the box on a lamp that has just been run at full current for the drift test.

The chemistry of the first exposure is Part XIII’s, at Part XIII’s dilutions, and its hazards are set out there.

Required PPE

Safety spectacles with side protection, for all knife work and for every hole drilled. The knife is the reason, not the drill: a blade that skates off a steel rule travels at eye height more often than anyone expects.

A steel rule and a self-healing cutting mat, which are protective equipment on this page and not conveniences. A plastic rule lets a blade climb its edge and into the hand holding the work; a steel one stops it.

Nitrile gloves and eye protection for the processing step only, as the Level A controls require whenever solutions are handled. The glove page is the authority on what the published permeation data do and do not support.

Nothing else is specified, and the reason is stated rather than assumed. No respirator, because the recommended route has no sanding, no sawing and no spray — the two products that would have justified one have been replaced. No splash protection beyond the Level A pair, because nothing here is more concentrated than a working bath. Take the plywood option and both sentences stop being true, and Part VI’s controls apply instead.

Ventilation

For the construction: an open window is sufficient, and the reason is that the recommended materials produce no vapour. PVA is water-based, tape produces nothing, and black card is card.

If you paint. Water-based matt black acrylic is a household product and is used as its label directs. If you use an aerosol of any kind, that is an outdoor operation, out of the wind, with nobody downwind of you, and the box is then left to cure outside for a day before anything photographic goes near it. The course’s preference is card, and the reason is given under stage 3.

For the processing step, the ventilation is Part XIII’s and the standard is HSE’s for manual film development: general ventilation with a through draught.

Parts

The recommended route is cut and taped board, because it needs no workshop, holds dimensions well enough for a 0.02 log H target and can be rebuilt in an evening when the uniformity map says the chamber should be deeper. A plywood box is better in every way except the one that matters most at this stage, which is that you will want to change it.

Two other routes are named because a reader will already own one of them. A moulded ABS project box, the kind sold for electronics with a gasketed lid and cable glands, is light-tight out of the packet and needs no joint made — but the ones at hobby prices are rarely 300 mm deep, so it usually becomes the lamp end with a board tube above it: bought light-tightness where the cable enters, made geometry where the wedge sits. A plywood box is the durable answer and the right one to build second, after the map has settled the chamber depth. The dimensions and the acceptance test below do not change either way.

Part	Quantity	What it does	Substitutes and notes
Black foam board, 5 mm, A1 or two A2 sheets	1	The walls, floor and lid	3 mm black mountboard works and is stiffer per millimetre; corrugated plastic sheet works and is not black inside its flutes, which matters at a cut edge. Part VI recorded that this course has not measured whether black board is light-tight in the thicknesses sold, and that finding stands: the box is proved on film, not on the material’s reputation
White card or white-faced board, A3	1	Lines the chamber below the diffuser	Ordinary matt white card. Not gloss: a gloss wall makes a specular image of the emitter that lands on the diffuser as a bright patch
Black card, A3	1	Lines the space above the diffuser, and the lid	Matt black card, or matt black acrylic paint on the board itself if you would rather paint than line
Opal acrylic sheet, 2 to 3 mm, cut about 190 × 130 mm	1	The diffuser	Opal acrylic is sold for light panels and signage. The course has not obtained a supplier’s transmission or diffusion figure for it and quotes none; the diffuser is specified below by what it must do and is accepted by measurement. Opal glass is the traditional part and is what ISO 5-2 names for the analogous job at the reading end of the chain
Float glass, 3 mm, about 150 × 120 mm, edges taped	1	Holds wedge and film in contact by its own weight	The same piece Part XIII’s lab already asks for. Tape every edge; an untaped 3 mm edge is a blade
Closed-cell foam, 5 mm, offcut about 150 × 120 mm	1	The pressure pad in the lid	Draught-excluder tape doubled over does the same thing. It must be even, or it tilts the glass
Aluminium or plastic angle, 10 mm, two lengths of 150 mm	2	The registration fences	Two strips of 3 mm board glued down work as well and are easier to trim to a fit
Black gaffer or photographic tape, 50 mm	1 roll	Every joint, inside and out	This is the structural fastener of the recommended route, not a repair
Hook-and-loop tape, self-adhesive	300 mm	Holds the lid down evenly	Two small cabinet catches are better and cost more
PVA adhesive	a little	The fences and the linings	Priced in the planner
Cable grommet or a rubber blanking plug, 8 mm	1	The cable entry	An offcut of foam pushed into an oversize hole is the version that costs nothing and works
Self-adhesive labels	a few	Serial number, wedge number, orientation marks	Written before assembly, not after
Film, from the same emulsion batch as the rest of the part	3 strips	The fog test, its control, and the first exposure	The batch number goes in the build record
A 21-step transmission wedge	1	The ruler	Stouffer T2115 or equivalent, 21 steps at a nominal 0.15 to a maximum density of 3.05, on a piece 12.7 × 127 mm. Its serial number is part of the instrument from this page onwards

Tools

Tool	For	Notes
Craft knife with fresh blades, steel rule, cutting mat	Every cut	Three light passes beat one heavy one through 5 mm board
Engineer’s square or a set square you have checked	Setting out	Check it against a factory-cut sheet edge; a square that is not square puts a taper in the chamber
Steel tape or rule reading to 1 mm	Every dimension	The dimensions go in the build record as built, not as drawn
Drill and an 8 mm bit, or a hole punch	Cable entry and button	Spectacles on
Lux meter, or the module’s monitor photodiode on a flying lead	The uniformity map	Whichever you use, use the same one for the whole map. Absolute accuracy does not matter here; consistency across the grid is the entire measurement
The module from the previous page, with its serial terminal	Exposures and monitor readings	Bring the firmware’s warm_up and expose routines to the bench with the box
A torch that fits inside the box	The dark-adapted inspection	Any small torch; brightness is not the point
The wet bench of Part XIII’s lab	The first exposure and the fog test	Tank, graduates, thermometer, water bath

Estimated cost

Cost band ££, and almost all of it is the wedge and the film. The box itself is board, tape and one piece of opal acrylic.

This page quotes prices only where the laboratory planner carries a dated one. It carries none for foam board, black card, opal acrylic, tape or hook-and-loop — the planner’s own gap list already names the enclosure and camera-build hardware as unpriced — and it carries none for the 21-step wedge either, which the planner records as a gap because no United Kingdom supplier price has been confirmed. What is durable, and is on this page instead of a price, is the specification: the chamber depth, the panel size and the acceptance criterion outlast any quotation.

Estimated consumables cost

Most of this session is capital. The box, the glass, the fences and the wedge are all bought once and used for years. What one run of this page actually consumes is three strips of film and one set of processing solutions, plus tape and glue.

A working figure the table needs and no manufacturer publishes. A 36-exposure cassette yields about twelve 135 mm strips once the leader and the fogged tail are discarded. That is the course’s own arithmetic from the frame pitch, not a figure from a film maker, and it is what every fraction of a cassette below is computed from — check it against your own bulk loader before you plan a batch around it.

Consumed	This session	Sourced price	Cost this session
Film, from the part’s emulsion batch	3 strips of 135 mm, about a quarter of a 36-exposure cassette	£6.37–£11.40 per one 35 mm roll, 36 exposures	£1.59–£2.85
D-76 or ID-11 stock, from Part VIII’s mixing lab	150 mL, diluted 1+1 and discarded after one batch	Costed in Part VIII’s mixing lab	—
Rapid fixer concentrate	60 mL, to make 300 mL at 1+4	£21.05–£25.98 per 1 L of ammonium thiosulfate concentrate, diluted 1+4 for film	£1.26–£1.56
Stop bath concentrate	15 mL, to make 300 mL at 1+19	£10.66–£12.18 per 500 mL of citric acid concentrate, diluted 1+19	£0.32–£0.37
Wetting agent	1.5 mL, for 300 mL at 1+200	£28.70 per 1 L of concentrate, diluted 1+200	£0.04
Black foam board and card	about one A1 sheet and two A3	None. The planner names enclosure and camera-build hardware as a price gap	—
Black tape	about 4 m	None. The planner carries no line for it	—
PVA adhesive	a few millilitres	£11.99–£14.29 per 5 L, which prices a whole shelf rather than this session	—
Opal acrylic, 2 to 3 mm	one piece about 190 × 130 mm	None. The planner carries no line for it	—

The priced rows come to £3.21 to £4.82 for one run of this session, at the ranges dated 5 September 2026 in the laboratory planner. That is a floor and not a total: four of the nine rows carry no dated price, are counted as nothing here, and none of them is free. Equipment stays out of the table on purpose — the glass, the fences, the meter and the wedge are not consumed. A second run, if the map sends you back to a deeper chamber, spends the board and the tape again and nothing else.

Alternative route

This page needs total darkness twice: once to load the fog-test strips, once to load the first exposure. A panchromatic film has no safelight, and FOMA say so in a line for FOMAPAN 100 — infrared light, or total darkness.

There is a route for a reader with no darkroom, and it is the same one Part XIII gives. A changing bag is enough for every loading step on this page, because nothing here needs a room: the strips are short, the box is small, and the wedge sandwich can be assembled inside a bag on a flat surface with a little practice. Practise the assembly in daylight with a scrap strip until your hands know the fences by feel, exactly as ILFORD tell a beginner to practise loading a spiral before the first real film.

The orthochromatic route also stands: run the whole part on ILFORD ORTHO Plus under a deep red safelight, take that film’s own development times, and choose it for the whole part rather than for one session, because the strips are meant to be compared with each other.

What has no alternative is the fog test itself, which must be done in the room and with the lighting the instrument will actually live with. A box tested in a changing bag has been tested against no light at all, which is the one condition it will never meet in use.

Waste streams

The construction produces card offcuts, tape and a little cured PVA, which are ordinary household waste. The first exposure produces the three streams Part XIII’s lab already sets out — spent developer, spent fixer and its first rinse as silver-bearing waste, and rinse water — each labelled per the container SOP and routed per the general waste SOP. The chemistry of those streams and the reason the answer is jurisdictional are on the disposal page, and local regulation governs what you may actually do with them, so check your local regulations before the first pour.

Build

Nine stages, about three hours for the box and its inspection. The film tests that follow it take a further session, because they need a development.

🛠The two numbers everything else follows from

127 mm of wedge, and 0.02 log H of permitted fall-off from the centre of it to the ends.

Everything below is an attempt to buy the second with geometry rather than with money. Convert the target once and keep it in your head: 10^0.02 = 1.047, so 0.02 log H is 4.7 per cent of illuminance. That is the number the lux meter has to see.

Stage 0 — Measure what you are housing (10 minutes)

Before any board is cut, put the module on the bench and write down five dimensions: the footprint of the heatsink, its height with the LED on top, the length of the tail from the board to the emitter, the diameter of the USB cable, and the width of the button you intend to bring outside. A box designed around a drawing of a module and built around the module itself are different boxes.

Write them in the build record now. Stage 8 will thank you.

Stage 1 — Set out the box (20 minutes)

The course’s reference dimensions, and where each comes from.

Dimension	Value	Why
Interior plan	180 × 120 mm	The wedge is 127 mm long; 180 mm leaves 26 mm of margin at each end so the field is not being used out to its corners
LED to diffuser	210 mm	The chamber depth. Long enough that the emitter’s own beam has spread across the whole panel before it arrives
Diffuser to film plane	45 mm	Short, deliberately: the diffuser is now the source, and a source close to the film is a large source in angular terms, which is what flattens the field
Interior height	about 300 mm	210 + 45, plus room under the LED for the heatsink
Exterior	about 200 × 140 × 330 mm	Board thickness and the lid

🧮Why 210 mm, when the design lesson said 420

The design lesson worked the fall-off for a small source and got a hard answer: at 127 mm of wedge and a 0.02 log H target, a bare emitter has to be about 420 mm away. Put this box’s numbers into the same arithmetic. Half the wedge is r = 63.5 mm and the source-to-film distance is d = 255 mm, so

tan θ = 63.5 ÷ 255 = 0.249, θ = 14.0°, cos θ = 0.970

and the fall-off at the end of the wedge is −4 log₁₀(0.970) = 0.053 log H, which is 13 per cent of the illuminance — nearly three times the target and a third of a wedge step.

So the bare geometry fails, and the box is built anyway. The diffuser is what closes the gap: an opal panel 180 mm long, 45 mm below the film, is not a small source at all, and each point of it illuminates the ends of the wedge from a different angle. The flattening is real and large. It is also not calculable from anything this course has read, because it depends on the panel’s scattering distribution, the chamber’s wall reflectance and the panel’s size together — three quantities the course has no figures for.

That is why stage 8 is a measurement and not a check. The design is an argument that the target is reachable; the map is the evidence that it was reached. If your map says 0.04 log H, the honest response is a deeper chamber or a second diffuser, not a smaller claim.

The instrument in section, bottom to top

Drawn to a readable thickness rather than to scale: the two air spaces are 210 mm and 45 mm and the emulsion pair is about two tenths of a millimetre. The colours of the two linings are the whole idea of the drawing. Layer depths are drawn to be readable, not to scale: on real film the base is roughly a hundred times the emulsion, and drawn honestly the emulsion would vanish. Any thickness given in the labels is the real one.

Stage 2 — Cut and assemble the chamber (35 minutes)

Cut four walls and a floor to the interior plan, allowing for the board thickness at the corners so that the interior comes out at 180 × 120 mm. Butt the corners, tape them inside and out, and check the diagonals: a box 2 mm out of square across 180 mm has a wall leaning by two thirds of a degree, which is invisible and of no consequence to the map, but a box 10 mm out has a tapering chamber and a uniformity map with a gradient in it that no diffuser will remove.

Cut a shelf ledge for the diffuser at 210 mm above the emitter’s face — four strips of board glued inside the walls, not a rebate cut into them. The panel then drops in and lifts out, which you will want in stage 8.

Fit the module to the floor so that the emitter faces straight up the axis and its heatsink stands clear of the floor on spacers. A heatsink flat against a board floor has half its area doing nothing.

Stage 3 — Two interiors, and why they are different colours (15 minutes plus drying)

This is the stage that separates this box from Part VI’s camera, and it is worth ten seconds of thought before the glue comes out.

Wall’s 1912 instruction — that the interior of every camera should be coated with a dead black to prevent reflection and consequent fog — is exactly right for a camera and exactly wrong for the chamber of a sensitometer. A camera wants every non-image ray absorbed. The chamber below the diffuser wants the opposite: a ray thrown sideways by the emitter has not yet reached the film, and if a white wall returns it to the panel it arrives as illumination rather than as flare. A black chamber throws away most of the lamp and keeps the fall-off it was trying to fix.

Above the diffuser, Wall is right again. Light in the upper space that reaches the film off a wall has not passed through the diffuser and has not been graded by the wedge, so it is flare in the strict sense: a veil added equally to every step, which lifts the toe and flattens the curve. The upper walls are lined matt black for exactly the reason Part VI’s camera was.

So: white below the diffuser, black above it. Line rather than paint if you can, because a lining can be replaced when the map disappoints and a painted wall cannot, and because a card lining introduces no vapour into a box that will hold film. If you paint, use water-based matt acrylic, two thin coats, and leave the box open until it is hard and odourless.

✕A gloss white chamber is worse than a black one

Matt white scatters. Gloss white reflects, and a gloss wall makes a specular image of the emitter — a bright secondary source at a definite place — which lands on the diffuser as a hot patch and appears on your uniformity map as a bump that moves when you move the lamp. The wall you want is the one that has no direction in it.

Stage 4 — The diffuser (10 minutes)

The panel and the space beneath it are the diffuser chamber, and the panel is the only part of it you can get wrong in a way that cannot be fixed later.

Cut or order the opal acrylic to about 190 × 130 mm so that it sits on the ledge with a few millimetres of overlap all round and no gap at the edges. Light through a 1 mm gap at the edge of a diffuser is a line source at the exact place your uniformity is already weakest.

Peel both protective films. Handle it by the edges from now on: a fingerprint on a diffuser is a low-density patch that will appear in the map and then in every strip you ever expose.

Which face up does not matter optically for a symmetrical opal sheet, and does matter for the record. Mark one face, and always fit it the same way, so that if the map changes next year you can rule the diffuser out.

Stage 5 — The exposure stage and its registration (30 minutes)

The stage is a flat platform at the top of the upper space, with two fences at right angles. It exists so that the answer to “where was the wedge?” is the same every time without anybody looking.

The exposure stage, exploded, with the two fences that do all the work

1. 1Platform, with a 140 × 45 mm aperture — the only opening between the upper space and the film
2. 2Long fence, rear — sets the across-film position for both film and wedge
3. 3Short fence, left — sets where step 1 falls; the whole log H axis hangs off it
4. 4Film, emulsion up, against both fences — 35 mm strip, or a 101.6 × 127 mm sheet
5. 5Wedge, emulsion down, against both fences — 12.7 × 127 mm; never base down
6. 6Cover glass, 3 mm, taped edges — contact by weight; the lid foam evens the pressure

Drawn exploded; in use the four upper items are in contact. Both the film and the wedge locate against the fences, never against each other.

Three rules govern the stage and each of them is a mistake somebody has already made.

Both parts register against the fences, never against each other. If you locate the wedge on the film, then a film strip 2 mm out of place moves the wedge 2 mm with it and the whole exposure ladder shifts along the strip. Fences make that impossible.

Emulsion to emulsion, wedge face down. Turned over, the wedge’s polyester base sits between the two emulsions and light spreads inside it before it reaches the film, so every step boundary softens and the densities at the ends of the steps are no longer the densities of the steps. Mark the wedge’s emulsion side on its edge with a label, because in the dark you will not be able to tell.

Contact by weight, not by clamping. The glass presses hard enough. A lid screwed down onto foam that is 1 mm proud at one end tilts the glass, opens a gap at the other, and scratches the emulsion of a film that cost more than the box.

Cut the platform’s aperture to 140 × 45 mm — long enough for the wedge with margin, wide enough for a 35 mm strip, and no larger, because every square centimetre of aperture beyond that is light going past the film into the lid.

The sheet-film case needs one extra part and no extra aperture. A 4 × 5 inch sheet is 101.6 × 127 mm, so it lies on the same platform with its 127 mm dimension along the long fence and the wedge in the same place as before; the aperture already covers everything the wedge covers, and the rest of the sheet stays unexposed, which conveniently becomes a base-plus-fog area the size of a postcard. What it needs is a removable second fence set 101.6 mm back from the long one, so a sheet locates in that axis too — removable, because a fixed block in the way is what makes a 35 mm strip go in crooked in the dark. Cut a shallow notch in the front edge at the step-1 end, deep enough to get a fingernail under a strip: lifting one off a flat platform with gloved fingers in the dark is the surest way to fingerprint an emulsion.

Stage 6 — Lid, seal and pressure (20 minutes)

The lid is Part VI’s problem, solved Part VI’s way. Build a shallow tray that slips over the top 10 mm of the box: overlap at least 8 mm, clearance about half a millimetre, two right-angle turns in the passage, every surface in the passage matt black. That geometry keeps light out at a fit no craft knife can guarantee, which is the whole reason it exists.

Glue the foam pad to the underside of the lid so that it lands centrally on the cover glass. Fit the hook-and-loop or the catches so that the lid closes to a stop rather than to a squeeze: you want the same pressure every time, and a catch that can be pulled tighter is a variable you have introduced by hand.

🛠Where the baffles go, and where they must not

Part VI put a baffle inside the camera body — a ledge computed to intercept stray light without cropping the picture. Nothing like it may go inside this instrument’s light path. A ledge below the diffuser casts a shadow on the panel, which is a hole in your uniformity map; one above it shades the ends of the wedge, which is the exact error the geometry was designed to avoid. This box is the one place in the course where the reflex to baffle is wrong.

All the baffling therefore happens at the entries. The lid’s overlapping tray is a baffle: two right-angle turns, matt black, no straight line from outside to the film. The cable hole is baffled by its plug and by the loop of slack behind it. And the upper space is matt black throughout, which is the same job done by absorption instead of geometry. The map is the test: baffle the entries and it is unchanged, baffle the path and it is not.

Stage 7 — Cable, button and indicator, for a box worked in the dark (15 minutes)

Everything on the outside of this box has to be findable by touch, because it will be operated in a room where you cannot see it.

The cable entry goes in a side wall of the lower chamber, below the diffuser, and is fitted with a grommet or a foam plug. Below the diffuser, because a hole above it is a hole aimed at the film. Bring the cable in with a loop inside the box, taped to a wall, so that a tug on the outside lands on the tape and not on the module.

The button goes on the outside, where a hand can find it, and is a different shape from anything else on that face. A recessed button and a proud one are two different things in the dark; two identical buttons are a mistake waiting for a strip.

The indicator is on the outside, and this is a deliberate choice. The instrument needs a way to say “the lamp is on” that does not put a light source in a room where panchromatic film is out of its wrapper, and the course cannot state a brightness at which an indicator leaves a panchromatic emulsion unmarked, because no manufacturer figure for it exists in anything it has read. So the indicator sits outside, recessed, shielded and pointed at a wall away from the loading area — and the light-leak test below is run with the indicator lit, so the question is answered on film rather than by assertion.

Mark the outside of the box, in raised tape you can feel: which end is step 1, which face is the front, and where the lid latches.

Stage 8 — The build record (15 minutes)

An instrument without a record is a box. Photograph the interior before the lid goes on, the stage with a wedge and a scrap strip in place, and the outside with its markings. Then write down, in the lab notebook:

• the interior plan and the two air spaces, as built and measured with a rule, not as drawn;
• the diffuser’s material, thickness and which face is up;
• the wedge’s serial number, and whether it is a calibrated part;
• the module’s serial, its drive current and its firmware version, copied from the previous page’s label;
• the date, and a serial number for the instrument itself.

That last item sounds like bureaucracy and is not. From the calibration experiment onwards, every curve this course asks you to plot is quoted against an instrument certificate, and a certificate names the instrument it belongs to.

Testing and calibration

Two acceptance tests, in this order, and the second one is the one this page exists for. Both tables are blank on purpose: the course has not built and measured this enclosure, and printing a uniformity figure it has not measured would be inventing a measurement.

Test 1 — Is it dark? First by eye, then on film

The inspection, which is necessary and not sufficient. Kodak’s darkroom guidance gives the method and it transfers directly: sit in the dark for five minutes so your eyes adapt, and then look. Do it in three passes. Torch inside the closed box, looking at the seams from outside — light coming out of a joint is far easier to see than light going into one. Then torch outside, sweeping the box at 300 mm from every angle including from below, while you look into it with the lid off. Then repeat with the lid on and the cable in place, because the cable entry is the leak this design is most likely to have. Tape every glow you find, and re-run the pass that found it.

The light-leak test, which is the actual test. Your dark-adapted retina does not integrate; a strip of film does. A leak too faint to see in five minutes can still fog a strip over the ten minutes the box sits on the bench with the room lights on while you set up.

1. In darkness, load two strips from the same emulsion batch. One goes on the stage in the closed box. The other goes straight back into its light-tight tin: it is the control, and it establishes the base-plus-fog this test is read against.
2. Leave the box closed, lamp unplugged, indicator lit, in the brightest place the instrument will ever be used, for ten times the longest continuous period it will ever sit there loaded. If a session takes three minutes from loading to exposure, that is thirty minutes.
3. Process both strips together, in the same tank, with the developer, time and temperature of Part XIII’s lab — D-76 or ID-11 at 1+1, at 20.0 °C, for the film’s own published time. Kodak’s own sheet says the 1+1 dilution is made just before use and discarded after one batch, so both strips must go in together or the second sees a different developer.
4. Read the two strips against each other on a light box, edge to edge and touching.

Reading	Control strip	Box strip	Difference
Base plus fog, by eye against the control
Any local darkening, and where on the strip
Verdict: pass, or the leak’s location

The criterion is a difference you cannot see when the two strips touch. That is a coarse test and the page says so: it resolves perhaps 0.02 to 0.03 of density by eye, and Part XV’s densitometer will resolve better and should be used to repeat it once it exists. A difference you can see is a leak, and its position on the strip usually names it — darkening at one end points at the nearest joint, an even veil points at the lid or the cable entry.

Test 2 — Is it even? The uniformity map

This is the acceptance test for the whole build.

1. Put the module on, run warm_up for the time your own warm-up curve established, and leave it on for the whole map. A lamp switched off and on between grid points measures its own drift, not the box.
2. Take the lid and the glass off. Work in a darkened room so that the meter sees the instrument and not the ceiling.

The grid to map, and the one it is tempting to map instead

1. 1Fifteen positions, five by three — 0, 32, 63.5, 95 and 127 mm along; rear, middle and front across
2. 2The centre reading — every other point is expressed as a deviation from this one
3. 3The three-point map — what most builders take; it cannot see a tilt across the width or a hot spot between points
4. 4The sensor itself — a lux-meter head about 20 mm across averages over a patch wider than the wedge; the fine structure is not recoverable by this method

Take the grid, not the line. The extra twelve readings cost five minutes and are the difference between finding a tilt and discovering one six months later on a strip.

3. Put the sensor at the film plane and read it at fifteen positions: five along the 127 mm axis at 0, 32, 63.5, 95 and 127 mm from the step-1 fence, and three across at the middle and both edges of the 35 mm strip. Record the raw reading, not a percentage.
4. Convert each to a deviation from the centre reading:

Δ log H = log₁₀(E_centre ÷ E_point)

Deviation of one grid point, in log exposure

E_centre is the reading at the middle of the field and E_point the reading at that grid position. A positive value means the point is darker than the centre and its steps will read low.

Position along the wedge, mm	Reading, centre row	Reading, near edge	Reading, far edge	Worst Δ log H
0
32
63.5 (centre)
95
127

The acceptance criterion is 0.02 log H, worst point to centre, from the design lesson, and it is an eighth of a nominal wedge step.

📓Read the shape of the map, not only its worst number

A map that is symmetrical, high in the middle and low at both ends, is geometry: the chamber is too short or the panel too small, and the fixes are a deeper chamber, a bigger panel or a narrower illuminated field.

A map that is lopsided, falling steadily from one end to the other, is not geometry. It is a box out of square, an emitter off the axis, or a diffuser that is not sitting flat on its ledge. Fix the mechanics; no amount of diffusion will remove a tilt.

A map with a bump in it is a hot spot: a gloss patch on a chamber wall, a gap at the edge of the diffuser, or the emitter’s own beam pattern showing through a panel that is too thin. Cree publish a viewing angle for their emitters — 110° for a white part — but no near-field intensity distribution, so the course quotes nothing about the shape of the beam and treats the bump as something to be diffused away and then re-measured.

If the map fails, the three fixes and what each costs. A deeper chamber is free in money and costs illuminance, which the design lesson already showed you have to spare at the film end but not at the paper end. A second diffuser, 30 mm below the first, is the most effective single change and costs more illuminance still. Masking the field down to 140 mm of aperture costs nothing and buys nothing if the fall-off is already inside the wedge’s own length — it only helps when the map shows the last centimetre falling off a cliff.

Re-map after every change. A uniformity map is cheap and an uncertain instrument is not.

Test 3 — The first exposure

The instrument’s first real output, and it is not a calibration: it is a sanity check that the whole chain works before the next page spends a session measuring it.

Warm up. In darkness, load a strip against the fences, wedge emulsion down, glass on, lid closed, and give the design exposure of 0.5 s. Process it with an unexposed control from the same batch at the conditions above — for FP4 Plus at EI 125 in D-76 at 1+1, ILFORD publish 11 minutes in a spiral tank at 20 °C with intermittent agitation.

Then answer four questions before writing anything down. Does the strip show a usable range of steps, the clear end nearly black and the dense end still separable from base-plus-fog? Are the step boundaries sharp — soft means the wedge went in base-down or the contact was poor? Is the tone even along each step, or is one end lighter, which is your uniformity map appearing on film? And is the control clean?

Do it promptly. ILFORD tell you to process exposed film as soon as practical, and a first exposure left in a drawer for a fortnight tests the box and your latent image together with no way to tell them apart.

Troubleshooting

What you see	Likely cause	What to do
One end of every strip is consistently denser	In order of likelihood: a tilt in the map (box out of square, emitter off axis, diffuser not flat on its ledge); the wedge registering against one fence only, so step 1 lands differently each time; a leak at one end	Re-run the uniformity map, where a tilt shows directly. Then expose two strips with the wedge reversed end for end: if the dense end follows the wedge it is the wedge, and if it stays at the same end of the box it is the box
The fog test strip is veiled all over	Light entering through the lid seal or the cable grommet, or the indicator, or fog from the film’s own age or storage	Re-run with the indicator taped over. If it passes, the indicator is the leak. If it still fails, re-run with the box inside a dark cupboard: a pass there means the box, a fail means the film
Step boundaries are soft	The wedge went in base-down, or the glass is not making contact	Check the emulsion mark on the wedge. Then check the foam: a pad that is proud at one end lifts the glass at the other
Scratches along the strip	The glass is being slid over the film rather than lowered onto it, or grit on the platform	Lower the glass, never slide it. Wipe the platform and both glass faces before every session; a stage that lives in a dusty room needs a cover
The uniformity map changes between sessions	The diffuser has been refitted the other way up, the module has moved on its spacers, or the meter is being placed differently	This is why the diffuser has a marked face and the module is fixed rather than laid in. Check both, then check that you are putting the sensor down at the same height each time — a lux meter head that sits 5 mm high at one grid point has moved 10 per cent of the way to the diffuser
Illuminance far too high for paper work	The design has headroom for film and not for paper, as the design lesson’s factor of 130 predicted	Reduce the drive current first, since it is the adjustment that costs nothing. Only then add a fixed neutral filter, and if you do, it becomes part of the instrument and goes on the certificate
The lamp will not light with the lid closed and will with it open	A wire is being pinched by the lid	Look before you force it. The loop of slack taped inside the box exists for exactly this

Questions

1. Your uniformity map reads 41.0 lx at the centre of the field and 37.6 lx at the 127 mm end. Express the deviation in log H, compare it with the design target, and say by what factor the illuminance at the end would have to change to bring it inside.
2. The map is symmetrical and reads 0.031 log H at both ends. You may make the chamber 60 mm deeper or fit a second diffuser. Using the cos⁴ arithmetic for the bare-source part of the problem, estimate what the extra depth alone would buy, and then say why that estimate is a lower bound on what you will actually measure.
3. A builder lines the whole interior — chamber and upper space alike — in matt black, on the grounds that Part VI’s camera was black and it worked. Say what happens to the illuminance at the film plane and what happens to the uniformity, and which of the two they will notice first.
4. The fog test passes with the indicator taped over and fails with it lit. Give the two acceptable engineering responses, and say which one you would defend to somebody who wanted the indicator because they work in the dark.
5. Somebody proposes locating the wedge by lining it up with the edge of the film rather than against the short fence, since the film is already against the fences. Explain, in terms of the log exposure axis, what a 2 mm error in that alignment does to the strip, and whether it makes the curve wrong or only shifts it.
6. Your first exposure at 0.5 s gives a strip in which only the first six steps are separable and everything beyond step 7 is at maximum density. Say which quantity you would change, in which direction, and why the answer is not “expose for less time”.

Further experiments

Measure the wall reflectance you assumed. Line the chamber white, map it, then line the same chamber in matt black and map it again without changing anything else. The ratio of the centre readings is what the white lining is worth in illuminance, and the difference in the two maps’ worst deviations is what it is worth in uniformity. The course has no published figure for either, and yours would be a real measurement of your own instrument.

Find the depth at which the diffuser stops helping. Build the ledge as a set of four positions at 150, 180, 210 and 240 mm and map each. Plot worst deviation against chamber depth. The curve will flatten somewhere, and where it flattens is the point at which further depth is buying illuminance loss and nothing else — which is the design decision this page had to make from geometry alone.

Test the diffuser against a cheaper one. Two sheets of tracing paper, or a sheet of white translucent polypropylene, cost a fraction of opal acrylic. Map all three at the same chamber depth and the same drive current, and record the centre illuminance as well as the deviation: the honest comparison is uniformity at equal illuminance, because any diffuser can be made more even by throwing more light away.

Run the light-leak test in the sun. An hour of direct sunlight on the closed box is a far harsher test than a lit room and takes no longer. A box that passes it will never be your problem again, and if it fails, the pattern on the strip is a map of the leak.

Check your understanding

Question 1. Your uniformity map reads 41.0 lx at the centre of the field and 37.6 lx at the far end of the wedge. What is the deviation in log H, and does the box pass?

• 0.008 log H — a comfortable pass
• 0.038 log H — a fail, by nearly a factor of two on the target
• 0.083 log H — a fail, and about half a wedge step
• 3.4 log H — the map is meaningless

Show the answer and why

Answer: 0.038 log H — a fail, by nearly a factor of two on the target

log₁₀(41.0 ÷ 37.6) = log₁₀(1.0904) = 0.0376, so about 0.038 log H against a target of 0.02. It fails, and it fails by a factor of 1.9 rather than by a hair, which tells you the fix is structural rather than cosmetic. Work the other direction to see what would pass: 0.02 log H means a ratio of 10^0.02 = 1.047, so the end would have to read 41.0 ÷ 1.047 = 39.2 lx. You need to recover 1.6 lx of the 3.4 lx you are down — a little under half the fall-off — which a deeper chamber or a second diffuser can plausibly do and a coat of paint cannot.

Question 2. Why is the chamber below the diffuser lined matt white when Part VI insisted that the inside of a camera be dead black?

• Because white card is cheaper than black card and the chamber is the larger area
• Because a ray thrown sideways in the chamber has not reached the film yet, so returning it to the diffuser turns it into illumination rather than flare
• Because a white chamber reflects less heat back onto the LED junction
• Because opal acrylic transmits more light when it is lit from a white surround

Show the answer and why

Answer: Because a ray thrown sideways in the chamber has not reached the film yet, so returning it to the diffuser turns it into illumination rather than flare

The two spaces have opposite jobs and the paint follows the job. Below the diffuser, no light has yet been graded by the wedge or aimed at the film, so a sideways ray returned by a white wall arrives at the panel as useful, undirected illumination — it raises the level and it fills in the fall-off at the ends. Above the diffuser, a ray that reaches the film off a wall has bypassed the wedge entirely and adds an equal veil to every step, which is flare in Part VI's exact sense: it lifts the toe and flattens the curve. Hence black above, white below. Option 3 is a real effect and is far too small to matter at a watt; option 4 confuses the panel with its surround.

Question 3. The wedge is loaded base-down instead of emulsion-down. What appears on the strip, and why?

• Nothing: the wedge attenuates the same amount either way round
• The whole strip is one step darker, because the base adds its own density
• Every step boundary softens, because light spreads inside the 0.18 mm of polyester before it reaches the emulsion
• The steps reverse order, because the wedge is now read from the other end

Show the answer and why

Answer: Every step boundary softens, because light spreads inside the 0.18 mm of polyester before it reaches the emulsion

The total attenuation really is the same either way, which is why option 1 is tempting and why the fault is so often missed. What changes is the geometry at the boundary between two steps: with the base between the two emulsions, light that has passed through step 8 has a couple of tenths of a millimetre in which to spread sideways before it lands, so it strays into the territory of step 9. Each step therefore has a graded edge rather than a sharp one, and a densitometer reading near an edge — or a step read at all, if the aperture is wide — no longer reports that step's density. It is a small distance and it matters because a step is only 6 mm long.

Question 4. The fog test strip shows a general veil. You re-run it with a strip of tape over the outside indicator and it comes back clean. What have you established, and what should go in the build record?

• That the box leaks at the lid, and the lid should be rebuilt
• That the indicator fogs film at the working distance, so either it is masked or the box is loaded before it is switched on — and either way the decision is recorded
• That the film was old, since the second strip was cut later
• Nothing, because a single pass and a single fail cannot separate the two

Show the answer and why

Answer: That the indicator fogs film at the working distance, so either it is masked or the box is loaded before it is switched on — and either way the decision is recorded

One variable changed between the two runs and the result followed it, which is exactly what a controlled test is for. What you have not established is a general fact about indicators: you have established a fact about this one, at this brightness, at this distance, with this film. That is why it goes in the build record with those particulars rather than into the course as a rule. Both responses are legitimate — mask it, or make it part of the operating procedure that the box is loaded and closed before the module is powered — and the honest instrument is the one whose certificate says which was chosen.

Question 5. A builder wants a shorter box and proposes 120 mm from emitter to film plane instead of 255 mm, arguing that the diffuser will sort it out. Working the bare-source fall-off at the end of a 127 mm wedge, what is the argument against?

• None: the diffuser makes the source distance irrelevant
• At 120 mm the bare fall-off is about 0.20 log H, more than a whole nominal wedge step, so the diffuser is being asked to remove ten times the error it is being asked to remove at 255 mm
• At 120 mm the exposure would have to be shortened below the reciprocity-valid region
• At 120 mm the wedge would no longer fit inside the box

Show the answer and why

Answer: At 120 mm the bare fall-off is about 0.20 log H, more than a whole nominal wedge step, so the diffuser is being asked to remove ten times the error it is being asked to remove at 255 mm

Half the wedge is r = 63.5 mm, so at d = 120 mm, tan θ = 0.529, θ = 27.9° and cos θ = 0.884. The fall-off is −4 log₁₀(0.884) = 0.214 log H, against 0.053 at 255 mm. Both need the diffuser's help, but the shorter box needs four times as much of it, and the flattening a panel can deliver is bounded by its own size relative to the distance — which the short box has also made worse, because the panel is now closer to the film and the film sees it over a wider angle. The right way to read the design is that the geometry gets you most of the way and the diffuser closes a gap, not that the diffuser is a substitute for the geometry.

Check my answersClear