Build: The Contact Printing Frame and Its Light Source

Purpose

To build the two things a contact print needs and nothing else: a frame that presses negative and paper together hard enough and flat enough that the gap between them is not a variable, and a light source whose output across the printing area you have measured on paper, in density, at a fixed geometry you can return to next week.

It is the cheapest instrument in this part and the one that pays back fastest. It needs no enlarger, which is why it comes before one; every alternative process in Parts XXI to XXV prints through it; and it makes the proof sheet, which is the tool photographers actually edit with rather than a technical exercise.

Two claims govern the whole build, and both are geometry rather than opinion.

Contact means contact. Any gap between the negative’s emulsion and the paper’s lets the source’s angular size smear the edge of every detail. The size of that smear is arithmetic, and the arithmetic says the frame matters more than the lamp.

Even means measured on paper. A box that looks even can be half a stop down in a corner, because the eye adapts across a field and paper does not. The acceptance test here is a sheet of paper, developed, dried and read — not a meter reading and not a maker’s claim.

Learning objectives

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

• derive the penumbra a finite source casts across a gap, and use it to predict the softening from a curled negative, a dust speck or printing through the base;
• explain why a small distant source gives a crisper contact print than a large near one, and why that fights the requirement for an even field;
• choose glass for a printing frame on thickness, flatness and transmission, and say which of those matters for silver paper and which only for the ultraviolet processes;
• say when Newton’s rings form, what anti-Newton texture does about them, and what it costs;
• identify the emulsion side of a film and of a paper by three independent signs, and state what printing through the base costs;
• design an LED light box whose emitter spacing, diffuser distance and interior finish are chosen for a stated evenness target rather than for brightness;
• state the exposure by time at a fixed current, and give the reason dimming is the wrong variable;
• lay out and expose a proof sheet to the convention that the clear rebate just reaches maximum black, and say what that convention buys;
• map the uniformity of the finished box in density units and decide, from the shape of the map, which of three faults you have.

Prerequisites

Low-voltage electronics for the darkroom, in full. The constant-current driver, the LED as a diode, the heatsink as a photometric component, the current budget and the five datasheet numbers are established there and are used here without being taught again.

The safelight fog test, because the calibration prints on this page are made under a safelight with a measured working time. Without that number, a corner reading 0.05 high could be the box or could be the lamp, and you will not be able to tell which.

The step-wedge exposure series, which is where the step wedge comes from and where reading density off a print stops being mysterious.

Building the sensitometer’s enclosure is not required, but its uniformity map is the same procedure at a smaller scale, and this page’s version is deliberately its sibling: same grid, same deviation arithmetic, different detector.

Safety classification

Level A. The hazards are a workshop’s, not a laboratory’s: cut glass, a blade, a drill, and the ordinary print chemistry of Part IV at working strength. Nothing is heated, nothing concentrated is handled, and the electrical work is a certified enclosed low-voltage supply feeding an LED strip, at the same extra-low voltage Part XIV established and for HSE’s reason: keep the supply voltage down to the lowest that will do the job.

One step is declared higher. If you choose to cut your own glass, that step is a declared Level B operation under the mixed-level rule, and its controls are stated inline in stage 2: cut-resistant gloves, eye protection, a covered surface, and the edges arrised before the sheet is picked up by anything but its faces. It is avoidable, and the page prefers that you avoid it: a glazier will cut and polish a sheet to size, and the cost of that is in the parts table.

What is not a hazard here, and why. There is no ultraviolet in this build. A silver gelatin paper is printed by visible light and the emitters on this page are ordinary white LEDs at tens of milliamps per centimetre, which is a lower photobiological exposure than looking at a domestic ceiling light. There is no mains work, because nothing on this page is switched at mains potential and the supply arrives as a moulded lead from a certified enclosure. And there is no chemical hazard that Part IV has not already covered at Level A: the developer, stop and fixer used in Testing and calibration are at ordinary working strength, and the same gloves, tongs and eye protection apply. The one substance whose absence is worth naming is ultraviolet-blocking eyewear, which belongs to the UVA unit and would be theatre here.

Hazards

Hazard	Where it arises	Control
Cut from a glass edge	Handling a cut sheet before its edges are dressed; a sheet that breaks in the frame	Buy it cut and polished. Otherwise: cut-resistant gloves, eye protection, arris every edge with a stone under water before the sheet leaves the bench, and tape the perimeter as Kodak’s own home-made frame does
Glass breaking under spring pressure	Over-tightening a back onto a warped baseboard, or a point load from grit under the sheet	Foam pad, flat baseboard, and sweep the glass before every closing. Springs give pressure, not force: if you are pushing hard, something is not flat
Blade and saw	Cutting board, foam and battens	Cut away from the body on a mat, fresh blade, work clamped
Drill grabbing in thin sheet	Pilot holes for the hinge and the spring anchors	Step drill or a slow speed, workpiece clamped, eye protection
Print chemistry at working strength	The calibration prints	Part IV’s Level A controls: nitrile gloves, tongs, eye protection when pouring, dedicated vessels
Low-voltage electrical	The light box’s strip and supply	Certified enclosed supply with a moulded lead, fused on the low-voltage side, nothing at mains potential opened

Required PPE

Eye protection for every cutting, drilling and sawing operation, and when pouring the print chemistry. Cut-resistant gloves for glass handling before the edges are dressed — a nitrile glove is a chemical control and stops nothing sharp, so they are two different gloves for two different parts of the session. Nitrile gloves for the tray line, with tongs, on Part IV’s terms.

The control that is easy to skip is a clear bench: glass is dressed and taped on a surface with nothing else on it, because a sheet set down on a dropped screw is a sheet that breaks in your hands.

Ventilation

Ventilation is a control here for the adhesives and the print chemistry rather than for anything the build itself emits. Use a water-based PVA or a double-sided tape for the pad and the battens, and if you use a solvent adhesive or spray paint for the box interior, that step goes outdoors or at an open window with the room’s air moving, and it is done on a different day from any coated material in the room. The tray line runs with the darkroom’s ordinary air change, as Part II established.

Parts

Two assemblies. The frame is joinery; the box is a light source with a lid.

The frame

Part	Quantity	What it does	Substitutes and notes
Glass, 3 mm float or picture glass, cut and edges polished, 25 mm larger each way than your paper	1	The pressure plate and the flat reference	Have it cut. 3 mm is the sweet spot for an 8 × 10 in frame: stiff enough not to bow under the springs, light enough to lift with one hand. 2 mm bows; 4 mm is heavy and no flatter
Baseboard: 12 mm MDF or void-free plywood, same outside size as the glass	1	The flat the whole sandwich is pressed against	Flatness is the specification, not the material. Check it against a straightedge before you cut anything else; a bowed board defeats every spring you fit
Split back: two pieces of 9 mm ply, together the size of the glass, with a 6 mm gap at the join	1 set	The pressure member, in two halves so one can be lifted	The gap at the join is deliberate: it is where the hinge tape or the piano hinge lives
Closed-cell foam, 6 to 10 mm, one piece the size of the printing area	1	Equalises pressure across the sandwich	Closed cell, not open. Open-cell foam takes a set, sheds and holds moisture. Camping mat or the packing foam from an instrument case both work
Leaf springs or over-centre latches, 4	4	Pressure, applied and released repeatably	Two along each long edge for 8 × 10 in. Bought printing-frame springs are the neat answer; steel strip bent over a former is the cheap one
Piano hinge, 150 mm, or cloth hinge tape	1	Joins the two halves of the split back	Kodak’s own home-made frame uses wide adhesive tape as the hinge, and it works
Registration battens: 6 mm strip, two lengths	2	Locate the paper in the same place every time	Glued inside the baseboard rebate at one corner, so paper and negative go back where they were
Perimeter tape, cloth or gaffer, black	1 roll	Dresses the glass edge and blacks the frame	Kodak tape the remaining edges of a home-made frame explicitly so you do not cut yourself. Do it even on polished glass: it stops chipping

The light box

Part	Quantity	What it does	Substitutes and notes
Neutral white LED strip, 12 V, in a length that gives the row spacing stage 5 calculates	1	The source	Constant-voltage strip with its own series resistors is the simple route. High CRI is irrelevant here — this is monochrome paper — but a stable strip matters: buy one part number and buy enough of it, because a second reel from a different bin will not match
Certified, enclosed 12 V supply with a moulded lead, rated at least 1.5 times the strip’s draw	1	Power	Check the plug and fuse markings as the safelight build sets out. Oversize it: a supply run at its limit warms, and a warm supply sags
Fuse holder and fuse, low-voltage side	1	Fault protection at the battery end of the circuit	Sized to the strip, not to the supply
Opal acrylic sheet, 3 mm, the size of the printing area plus 20 mm	1	The diffuser	Opal, not frosted-one-side and not clear. Two thinner sheets spaced apart diffuse better than one thick one and cost more light
Box: 12 mm MDF or ply, internal depth set by stage 5	1	The integrating chamber	The depth is the design variable. Build it so the diffuser can be moved on ledges before it is fixed
Matt white paint, interior	~250 mL	Makes the chamber integrate rather than image	Matt white inside the box and matt black on everything outside the diffuser aperture. This is the opposite of the safelight housing and for the opposite reason: there we wanted no reflections, here we want nothing but reflections
Toggle switch and a socket for the timer	1 each	Manual on for setting up; timed exposure for printing	The timed channel is Part XVII’s low-voltage output, or a bought timer switching the same 12 V
Frame stops: three battens on the box lid	3	Fix the frame’s position over the box	Two edges and one corner. The whole calibration is void if the frame can sit anywhere

🛠Three things this page will not buy for you

A strip part number. No specific LED strip available in the United Kingdom was verified against its own datasheet while this part was written, so none is named. What is specified is the geometry, which outlives any part number, and the acceptance test, which does not care what you bought.

A ready-made evenness figure. The emitter spacing and diffuser distance that meet the target on your box are established by measurement in Testing and calibration. This page gives the physics that says which way to move and the map that says whether you have arrived. It publishes no number of its own, because a number measured on somebody else’s box is a decoration.

An account of anti-Newton glass from its makers. No manufacturer’s or conservator’s description of how the texture is produced, or of what it costs in resolution, was read while this page was written. T2 therefore gives the optics of why the texture works and asks you to observe the cost on your own sheet, and gives no figure for it.

A low-iron glass transmission curve. Ordinary window glass and acrylic are covered by a source below. Low-iron glass is not: no manufacturer’s transmission data for it was read while this page was written, so the page says what low-iron glass is for and gives no numbers.

Tools

Tool	For	Notes
Straightedge, 600 mm, and feeler gauges or a torch	Checking the baseboard for flat	The torch is the poor man’s feeler gauge: light under a straightedge is a gap
Saw or track saw, drill, countersink	Board and battens
Craft knife, steel rule, cutting mat	Foam and tape
Carborundum stone or wet-and-dry, 240 grit	Arrising a cut glass edge, if you cut your own	Under running water. Dry glass dust is not something to make
Multimeter	Strip current, once, at the fuse	Written on the box’s label afterwards
Loupe, 8× or 10×	Reading the proof sheet, and the Newton’s-ring check	The single most useful tool in this list
Tape measure and a set square	Diffuser distance, frame position, box square	Every number in the calibration is relative to these
A densitometer, or the Part XV head	Reading the uniformity map	See Testing and calibration for the visual fallback and what it costs you

Estimated cost

Cost band ££, and it splits unevenly: the frame is cheap and the box is where the money goes.

The planner prices a bought contact printing frame at £74.80 to £77.99 — a Paterson 35 mm/8 × 10 in frame and a Clearfile proofer, read on 5 September 2026. That is the honest benchmark and this build should come in under it, with a larger printing area and a light source the bought frame does not include at all.

The planner prices none of this build’s own materials — no glass cut to size, no LED strip, no low-voltage supply, no opal acrylic, no springs, hinge or foam. Its own gap register names float glass cut and dressed; this page adds LED strip, opal acrylic and frame hardware to that list.

Estimated consumables cost

What one run of this page uses up. The frame, the box, the glass, the strip and the supply are capital and are excluded: they survive the session and every session after it. So is the loupe, the densitometer and the timer.

Consumed	This session	Sourced price	Cost this session
RC paper, 8 × 10 in, variable contrast	6 sheets — one uniformity map, one Newton check, one step-wedge contact, three proof sheets	£33.50 for 25 sheets (£1.34 a sheet) to £96.18 for 100 (£0.96)	£5.76 to £8.04
Paper developer concentrate, at 1+9	50 mL, making 500 mL of working solution	£10.52 for 500 mL	£1.05
Stop bath concentrate, at 1+19	25 mL, making 500 mL	£10.66 for 500 mL	£0.53
Rapid fixer concentrate, at 1+9 for RC paper	50 mL, making 500 mL	£21.05 for 1 L	£1.05
Nitrile gloves	1 pair	£6.64 for 50 (£0.13 a pair)	£0.13
Matt white interior paint	about 150 mL	None. The planner carries no line for it	—
Perimeter and hinge tape	about 2 m	None. The planner carries no line for it	—
Adhesive for the foam and battens	about 20 mL	£11.99 for 5 L of PVA, which the planner itself calls a bad buy at that size	£0.05
Abrasive stone or wet-and-dry, if you cut your own glass	one sheet	None. The planner carries no line for it	—

Subtotal: £8.57 to £10.85, and it is a floor rather than a total. Three rows carry no dated price — paint, tape and abrasive — and none of the three is free. The paper figure assumes six sheets and no mistakes; budget ten. The one line worth arguing with is the paper: at £1.34 a sheet, buying the 100-sheet box before this session saves more than the tape and paint cost together, and the proof sheets you make here are a record you keep, not scrap.

Alternative route

No darkroom at all. The frame is built in daylight. What needs a dark room is loading it and exposing it, and a changing bag does the loading: stage 4’s opaque back makes a closed frame light-tight, so it can be loaded in the bag and carried to wherever you can darken a room for two minutes. Night working under the room page’s ordinary blackout is the other route. Neither changes any test here.

No light box. Use the enlarger, as Kodak’s own instructions do: empty carrier, f/11, head raised until the light covers an area just larger than the paper, about eight seconds. It is how most proof sheets in the world have been made. What you give up is repeatability between sessions — head height, aperture, lamp age and voltage all move and none is written down — which is the thing the fixed box exists to buy. Take the route knowingly and write the four numbers down every time.

No frame. Clean glass and a weight, which Parts IV, V and VII already use and which is ILFORD’s own instruction for making a paper negative. It gives real contact in the middle and less at the edges — which stage 1 lets you predict rather than discover.

Cannot lift or clamp. Build at 5 × 7 in rather than 8 × 10 in: the glass weighs about a third as much and four small over-centre latches replace the springs, each worked with one hand. It is the cheaper route on paper too — the planner prices 5 × 7 RC at £0.45 to £0.64 a sheet against £0.96 to £1.34 for 8 × 10.

Cannot fix anything to a wall. Nothing here is fixed to a building. The box is a box, the frame lives on it, and the whole instrument stores flat under a bed.

Build

Stage 1 — The arithmetic that decides everything else (20 minutes)

Before any cutting, work out how much gap your frame is allowed. It takes ten minutes and it settles the glass thickness, the foam, the spring count and the box depth at once.

A source is not a point. It has a width, and from the negative’s position it subtends an angle. Across a gap between negative and paper, that angle spreads the edge of every shadow into a penumbra.

Why contact means contact: the penumbra a finite source casts across a gap

1. 1Source of width S — the diffuser, the bulb, or the sun; what matters is its width as seen from the negative
2. 2Distance D, source to negative — the only lever that costs nothing but space
3. 3The negative, with an opaque edge — every detail in the picture is one of these edges
4. 4Gap g between the two emulsions — zero if you get orientation and pressure right; 0.110 or 0.180 mm if you print through the base
5. 5Penumbra b on the paper — b = S x g / D. This is the whole argument for a sprung frame

Drawn readable rather than to scale: the gap is exaggerated enormously, because at true scale it would be invisible, which is rather the point.

b = S × g ÷ D

Geometric penumbra across a gap

b is the width of the blurred band on the paper; S the width of the source as seen from the negative; g the gap between the two emulsions; D the distance from source to negative — all in millimetres. It is two similar triangles and nothing more: the source subtends an angle of roughly S ÷ D at the negative, and over a distance g that angle opens into a band of width g × S ÷ D.

🧮Three cases, with ILFORD's own base thicknesses

ILFORD publish the base thickness of FP4 Plus: 0.110 mm (4-mil) clear acetate for roll film and 0.180 mm (7-mil) polyester for sheet film. Those are the gaps you get for free if you lay the negative down the wrong way up.

Emulsion to emulsion, in a sprung frame. g is a few micrometres — dust and the roughness of the two gelatin surfaces. b is a few micrometres too, whatever the source. This is why the frame matters more than the lamp.

Through the base of a 120 negative, small distant source. S = 25 mm (a bare bulb), D = 600 mm (Kodak’s 2 ft), g = 0.110 mm:

b = 25 × 0.110 ÷ 600 = 0.0046 mm. Invisible under a loupe.

Through the base of the same negative, large near source. S = 300 mm (a diffuser the size of the printing area), D = 250 mm (a shallow box), g = 0.110 mm:

b = 300 × 0.110 ÷ 250 = 0.132 mm. That is a soft edge you can see at arm’s length, and it is twenty-six times the previous case for the same negative and the same gap.

The lesson is not “use a small source”. It is that the gap and the source’s angular size multiply. A large near source is a perfectly good contact printer provided the gap is zero, and it is a catastrophe the moment the gap is not. Everything from here on is about making g zero and keeping it there.

⚗The same relation, arrived at independently, with numbers

Ware sets out exactly this geometry for the alternative processes and writes it in a more useful form. He defines the aspect value of a source as its largest linear dimension divided by its distance from the print — w ÷ d, which is S ÷ D above — and then states, by the geometry of similar triangles, that blur = gap × aspect value. It is the same equation, and having it from a specialist monograph rather than from this page’s own derivation is worth more than the derivation is.

He also tabulates it. For gaps of 0.1 to 0.5 mm:

Source	Aspect value w ÷ d	Blur, mm
The sun	0.01	0.001 to 0.005
A NuArc 2125 exposure unit	0.1	0.01 to 0.05
A “light bed” or the open sky	4	0.4 to 2.0

And he gives a criterion this page had no other way of sourcing. The eye resolves about 0.1 mm at its near point of 250 mm, while 0.25 mm is what is conventionally taken as acceptably sharp in calculating depth of field — so Ware takes 0.3 mm of blur as the onset of a “fuzzy” or “soft” image, and adds that the blur will be most conspicuous where local contrast is high.

Read the last row of that table against the third worked example above. A light bed — which is precisely what the light box in stage 5 is — has an aspect value near 1 and can reach 4, and at a 0.5 mm gap it is producing blur of up to 2 mm. That is not a marginal softening; it is a different photograph. The whole case for the sprung frame is in that row.

Set your own limit now and write it in the notebook. Ware’s 0.3 mm is the onset of visible softness; this course works to 0.02 mm, a fifteenth of it. The margin is there because the gap is the quantity you cannot measure directly — you never see the dust speck, and you find out about the curl on the print — so the budget must absorb the errors you will not detect. The 0.02 mm is the course’s choice; the 0.3 mm it is measured against is Ware’s.

Rearranged, your limit tells you what gap the box you are about to build can tolerate:

g(max) = b(max) × D ÷ S

The gap your geometry allows

For b(max) = 0.02 mm with a 300 mm diffuser 250 mm away: g(max) = 0.02 × 250 ÷ 300 = 0.017 mm. Seventeen micrometres. That is not a gap you can eyeball. It is a gap you get by pressing two flat things together with foam behind them, which is the next four stages.

Stage 2 — The glass (30 minutes)

The glass has three jobs and only three: be flat, be stiff enough not to bow between the springs, and pass the light. Everything else said about printing-frame glass is a consequence of one of those.

Thickness. Ware is specific for the plate-and-clips route the alternative processes use: 4 mm plate glass, and not 2 mm picture glass, which may bend or crack under pressure. That is the figure to follow when the only thing pressing is a row of clips around the edge.

In a sprung frame, where four springs and a foam pad distribute the load rather than pinching the perimeter, 3 mm is the working answer for 8 × 10 in and it is the one this build uses. Ware’s lower bound still binds: at 2 mm a sheet that size bows visibly under four springs and the bow lifts the centre of the sandwich, which is precisely the place a proof sheet is least likely to show you the problem. Scale thickness with area, not with ambition — and if you are going to clip rather than spring, take Ware’s 4 mm.

Which glass. Ordinary float glass is flat to a standard no home workshop can check, because it was formed on molten tin. That is the reason to buy rather than salvage: a sheet from an old picture frame may be float and may be drawn glass with a slow wave in it, and you will read the wave as a soft corner for years.

Low-iron glass is float glass with the iron content reduced. Iron is what makes an ordinary sheet look green edge-on, and reducing it raises transmission a little across the visible band and rather more at the blue end. For silver paper it is a refinement you will struggle to measure. For the ultraviolet processes it is sold as a benefit. This course publishes no transmission figures for low-iron glass, because none was read from a manufacturer while this page was written, and the honest consequence is that the case for buying it is not made here.

Acrylic is lighter, does not shatter, and is the right answer for a large frame or a reader who cannot safely handle a big sheet of glass. It is also softer — it scuffs, and a scuffed sheet prints its scuffs — and it takes a static charge that attracts exactly the dust you are trying to exclude.

!The transmission fact that matters, and where it matters

ICNIRP’s guidance for workers records that window glass transmits some radiation down to about 310 nm, which is inside the UVB band, while most plastics such as polymethyl methacrylate — Perspex or Lucite — and polycarbonate normally do not transmit below about 370 nm.

Read that carefully, because it is the opposite of the folklore.

For this frame, printing silver paper by visible light, it does not matter at all. Both materials pass everything the paper can see.

For the UVA unit it matters enormously, and it is the glass that wins. Ordinary glass passes UVA freely. Acrylic cuts off at about 370 nm, which sits inside the UVA band and above the 365 nm peak of the shorter of the two common UV LED wavelengths. An acrylic-glazed frame under a 365 nm array is a frame with a filter in it.

So: if you will ever print an alternative process in this frame, glaze it in glass. If it will only ever see silver paper, acrylic is a legitimate choice made for weight and safety. One frame or two is a decision this page asks you to make now rather than discover later, and the cheap answer is one glass frame that serves both.

Edges. Buy the sheet cut and polished and the Level B step disappears from your session. If you cut it, the declared controls are cut-resistant gloves, eye protection, a covered bench with nothing else on it, and the edge arrised with a stone under running water before the sheet is lifted by anything but its two faces. Tape the perimeter either way: Kodak tape the remaining edges of their own home-made frame explicitly so you will not cut yourself, and tape also stops the chip that becomes a crack.

Stage 3 — Flatness, foam and pressure (30 minutes)

Three parts do one job: bring g to zero and keep it there across the whole area.

The baseboard is checked, not assumed. Lay a 600 mm straightedge across it in five directions and look for light underneath: a visible gap is 0.1 mm or more, five times your whole penumbra budget. MDF is flatter off the shelf and heavier, void-free birch ply flat enough and lighter; seal both faces either way against darkroom moisture.

The foam equalises rather than pushes. A 6 to 10 mm closed-cell pad turns four point loads into something like uniform pressure and accommodates the thickness step where a strip of 35 mm film ends and bare paper begins. Cut it 5 mm smaller than the printing area all round, so it cannot creep out and hold the back open at an edge.

The springs give pressure, and pressure is not force. Four leaf springs on an 8 × 10 frame is right; two is a frame that touches only in the middle. If you are pressing hard to close the latches, stop: something is not flat, and forcing it is how glass breaks under a spring.

✕The four ways a frame that looks closed still has a gap

Curl. Film off a roll curls across the strip, and curl lifts the ends of a strip off the paper by more than your entire budget. Flatten negatives under weight in their sleeves for a day before a proof session.

Grit. One grain of dust between glass and negative props the sandwich open in a tent-shaped region tens of millimetres across. Sweep the glass and close the frame the moment it is loaded.

A warped back. Two pieces of thin ply bow in the middle of each half. 9 mm is the minimum, and a batten glued across the outside of each is cheaper than thicker ply.

Foam that has taken a set. Open-cell foam compresses permanently where the springs press hardest, and after a year it holds the sandwich open exactly where it used to hold it shut. Closed cell, and replace it when a fingernail no longer springs back.

Stage 4 — The split back, and why it is split (30 minutes)

The back is in two halves, hinged at the join, with the springs arranged so that one half can be released and lifted while the other stays clamped. That is the entire purpose: you can look at the print without losing registration, and put the half back down and carry on exposing.

For silver paper that is a convenience. For the printing-out processes it is the method. Ware’s instructions for the simple cyanotype say it plainly: because the process prints out, a traditional hinged-back frame lets you inspect the result, and you continue the exposure until the high values are light green, the mid-tones firm blue and the deepest shadows reversed to a pale blue-grey. You cannot judge that from a clock. The frame is the instrument that lets you judge it from the image.

Build the back opaque and light-tight at the join: a 6 mm gap bridged by fabric hinge tape with a weave in it is a slot source aimed straight at your paper. Piano hinge on the outside face, or doubled tape with an internal felt light-trap strip.

Section through the closed frame, and the half that lifts

1. 1Glass, 3 mm, cut and polished — flat and stiff; taped at the perimeter
2. 2Negative: base up to the light, emulsion down onto the paper — the gap between the two emulsions is the quantity stage 1 budgets
3. 3Paper, emulsion up — shiny side up for ILFORD MULTIGRADE RC
4. 4Closed-cell foam, 6 to 10 mm — turns four spring loads into even pressure
5. 5Split back, two halves of 9 mm ply, hinged over a 6 mm gap — light-trap the join; a fabric hinge over an open slot is a light leak
6. 6Leaf springs, two per half — so that one half can be released while the other still clamps
7. 7Registration battens in one corner of the baseboard — the reason the sandwich goes back where it was
8. 8One half lifted, mid-exposure — inspect a printing-out image and close it again without losing register

Layers drawn readable rather than to scale; the emulsion-to-emulsion gap is drawn as nothing because that is the specification.

Stage 5 — The light box, and the design that decides its depth (45 minutes)

The box is an integrating chamber with a diffuser on top and the frame sitting on the diffuser. Three decisions make it, and they are coupled.

Interior finish: matt white, everywhere. The opposite instruction from the safelight housing, for the opposite reason: there every reflection was a leak, here every reflection is a contribution. A matt white chamber scatters light many times before it leaves, which is what turns a row of discrete emitters into an area source. Gloss white is worse, because it reflects the emitters as images and an image of an emitter is a hot spot.

Emitter spacing against diffuser distance. The ripple above a row of emitters dies away as the diffuser moves further from them, and faster the closer together the emitters are — so the two levers fight each other on cost: closer emitters mean more strip and more current, a deeper box means less light. The rule this page uses is a starting point rather than a law — diffuser distance at least the emitter spacing, preferably twice it — and T3 decides whether your box has arrived.

The falloff to the corners is real and is not the diffuser’s fault. A corner is further from most of the emitters than the centre and sees them more obliquely. The cosine-fourth law is the extreme case of that geometry — a point source on a plane falling off as the fourth power of the cosine of the off-axis angle — and a broad diffuse panel does considerably better, because every point of the panel is at a different angle. The page gives you no number for your box, only the direction: more depth and more overhang both flatten the corners, and both cost light.

The light box in section: spacing, depth, overhang and the frame's fixed position

1. 1LED rows at pitch p, on the matt white base — one part number, one reel; a second bin will not match
2. 2Matt white interior on every surface — gloss white images the emitters; matt white scatters them
3. 3Opal diffuser at height h on ledges — start at h = 2p, and move it before you fix it
4. 4Overhang: emitters and diffuser both extend past the printing area — this is what flattens the corners
5. 5Three battens fix the frame position — two edges and one corner; the calibration is void if the frame can wander
6. 6Grommeted cable entry and a fuse on the low-voltage side — certified enclosed supply, moulded lead, nothing at mains potential opened

Depth is the design variable. Build the ledges so the diffuser can be moved by 20 mm either way before anything is glued.

Wire it, fuse it and measure the current once. The strip is a constant-voltage part with its own series resistors: a stable 12 V and a fuse on the low-voltage side sized to the strip. Put the meter in series once at the fuse and write the figure on the label.

One box or two? Two. The two instruments want opposite things: this chamber is matt white and relies on a diffuser, while the UVA unit is matt black and cannot use a diffuser at all, most acrylic stopping below about 370 nm. A shared box would be a compromised version of both, and it would put a UVA array in the enclosure you open casually.

Stage 6 — Exposure by time, at a fixed current (10 minutes)

Fit a toggle switch for setting up and a socket for a timed channel — Part XVII’s low-voltage output, or a bought timer switching the same 12 V. Do not fit a dimmer.

That is a real instruction, not an aesthetic one. Time is the right exposure variable and drive current is the wrong one, for three reasons that stack:

1. An LED’s output is not proportional to its current, and the deviation is not something you can predict from a knob position. Two clicks of a dimmer is not two-thirds of an exposure.
2. Current changes the die temperature, and temperature changes the output on a curve that also depends on how long the box has been on. A dimmed box has a warm-up you have not characterised.
3. A dimmer setting is not a record. A time is. “Twelve seconds” survives a house move, a knocked knob and whoever used the box last; “about a third up” does not.

Fixed current, fixed geometry, and the whole of your exposure control is a number of seconds. If the box is too bright — and a well-built one usually is — the fixes are a second diffuser 20 mm below the first, a neutral density sheet, or fewer strip rows: all permanent, measurable and on the label.

Stage 7 — Registration and the label (10 minutes)

Fit three battens to the top of the box: two along adjacent edges and one at the corner between. The frame goes against all three, every time. This costs nothing and it is what makes an exposure recorded in March mean anything in September.

Then the label, on the box, in the same form as the safelight’s:

Contact printer. Diffuser height ____ mm. Strip current ____ mA at 12 V. Printing area ____ × ____ mm. Uniformity: worst corner ____ density from centre, measured on ____ (paper), ____ (date). Proof exposure: ____ s at this geometry for ____ (paper).

Four blanks stay empty until Testing and calibration fills them. Until then this is a box with a lamp in it.

Stage 8 — The proof sheet as an editing tool

The instrument exists to make these, so build the first one into the session.

What a proof sheet is for. Kodak put it in one sentence: the images are the same size as the negatives, they help you choose which to enlarge, and they make a record to file with the negatives. The second half is the half people skip and the half that compounds — a shoebox of proof sheets in date order is a body of work you can see.

Laying out the film. A 36-exposure 35 mm film cuts into six strips of six and fits an 8 × 10 sheet with a margin; a 120 film gives twelve 6 × 6 or ten 6 × 7 frames in three or four strips. Cut in the rebate, never through a frame line, face down on clean paper. Sleeve the strips first and proof through the sleeve if it is clear enough: you lose a little sharpness to the extra gap, which stage 1 lets you calculate exactly, and you gain never handling the negatives again.

Orientation, three independent signs. Get this wrong and the whole sheet is mirrored and soft.

• The notch, on sheet film. Foma state it exactly: the notch is on the right upper corner of the short side when the emulsion faces you. Bergger give the same rule with the symmetry made explicit — the emulsion faces you when the notch is in the upper right or lower left corner, which is the same sheet turned through 180°. ILFORD say it a third way for FP4 Plus sheet film: the emulsion faces the user when the sheet is held in the position their diagram shows. Three makers, one convention.
• The curl. A strip of roll film laid on the bench makes a shallow trough, and the emulsion is on the inside of it. This one is practice rather than a manufacturer’s statement: the course found no sheet giving the direction of curl, and some films carry a layer that reduces it — Bergger’s PANCRO 400 has an anti-curling layer in both its 120 and sheet forms. Use the curl to check the other two signs, never as the only sign.
• The sheen. Under a raking safelight the base is specular and the emulsion slightly matt. For paper ILFORD remove all doubt: MULTIGRADE goes into the easel shiny side up.

Then Kodak’s rule for the sandwich: the negatives’ emulsion faces the paper’s emulsion, and the negatives face the light source. Base up, towards the glass; emulsion down, onto the paper.

The exposure convention: the clear rebate just reaches maximum black. Expose the sheet so that the film’s unexposed edge — the rebate, which no image light ever reached — prints at the deepest tone the paper will give, and not one step more.

This is a printing convention rather than a measured standard, and the course states it as a convention with its reason and cites no source, because none in this corpus establishes it. The reason is that it makes the sheet comparable. The rebate is the one part of every negative that is identical — clear base plus fog, the same on every frame of that film type — so pinning it to the same tone on every sheet means the differences you see between frames are differences in the negatives rather than in your printing. A sheet exposed by eye tells you which frames you like today; a sheet exposed to the rebate tells you which negatives are thin, in the same language as the sheet you made last year.

Kodak’s own starting points are worth having: a bare 7 W bulb 2 ft above the frame for about ten seconds, or an enlarger with an empty carrier at f/11 for about eight — with the instruction to double the time if the sheet is light and halve it if it is dark. Your box will differ from both; T5 finds its number.

Reading it. A loupe, and a soft pencil on the back. Mark the frames worth printing, mark the crops, and — this is what makes the sheet an editing tool rather than an index — mark the ones that nearly work and say what is wrong with them. Six months of that in a box is the most useful teaching material you will ever own, and it is about your own pictures.

Testing and calibration

Nothing here is optional and the order matters: mechanical first, then optical, then photographic. Each test writes into the label.

T1 — The pressure check

Load the frame with a sheet of ordinary paper and a strip of clear film base, close it, and slide a strip of thin paper — a till receipt, about 0.05 mm — in from each edge and at the middle of each side.

Pass: the strip does not go in anywhere, or goes in less than 5 mm.

Fail at one edge: that spring is weak or its latch is not seating. Fail in the middle: glass or back is bowing — thicker glass, a batten, or a fifth spring. Fail everywhere: the foam has set, or the baseboard is cupped.

T2 — The Newton’s-ring check

Newton’s rings are thin-film interference: where the film base and the glass are separated by a gap comparable with the wavelength of light — a fraction of a micrometre, varying smoothly across the region of near-contact — the light reflected from the two surfaces interferes, constructively at some separations and destructively at others, and the result is a set of concentric coloured or grey fringes mapping the shape of the gap. They appear precisely where contact is almost perfect, which is why they are a sign of a good frame rather than a bad one, and they are worse with smooth, glossy surfaces than with textured ones.

Put a strip of clear, developed film base in the frame, close it, and look at it under a bright oblique light.

• No rings anywhere: either your contact is genuinely imperfect everywhere, which T1 should have caught, or the base has enough texture to prevent them. Check T1 again before you celebrate.
• Rings in a few small regions: normal, and usually invisible on the print at contact scale.
• Rings over large areas of every negative: they will print. The fix is anti-Newton glass, which carries a very fine texture on one face; the texture holds the two surfaces apart by more than a wavelength so the fringes cannot form.

Which way the texture faces, and what it costs. The textured face must go against the film, because that is the interface that is trying to form fringes. But the texture is then in contact with your negative and in the light path, and at contact-printing scale it prints: a fine grain-like structure laid over the whole image. That is the trade, and it is why anti-Newton glass is a fix for a problem you have rather than a default. The course could not source a manufacturer’s or conservator’s account of how anti-Newton glass is made or of how much resolution its texture costs, so the mechanism above is given as optics and the cost is given as an observation to make on your own sheet rather than as a figure. Ware’s 0.3 mm is the yardstick to make that observation against: if the texture is visible at all on a contact print read at a normal viewing distance, it is contributing something in the same range as the fringes you removed.

T3 — The uniformity map, in density on paper

This is the acceptance test for the box, and it is deliberately the same procedure as the sensitometer’s with one change: the detector is a sheet of paper, because paper is what the instrument is for and because a lux meter’s photopic response is not the paper’s response.

1. Switch the box on and leave it on for your established warm-up, or ten minutes if you have none. Do not switch it off between steps.
2. Put the empty frame in its battens with the glass in place and no negative, and expose a whole sheet of RC paper to a mid-grey — the exposure that gives roughly a middle tone, found with a quick test strip. Mid-grey is chosen because the paper’s curve is steepest there, which makes it the most sensitive detector you have.
3. Process it with the rest of the session’s prints, in the same developer 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 dry-down is not uniform across a sheet that has been squeegeed.
4. Read density at fifteen positions: five across the long axis and three across the short, with the centre reading as the reference. Express every other point as a difference from the centre.

ΔD = D(point) − D(centre)

Deviation of one grid point, in density on paper

The acceptance criterion this course uses is a worst-point deviation of 0.04 in density, and the page is explicit about where that number comes from: it is borrowed from ILFORD’s safelight sheet, which gives a change of about 0.04 in density as the change that means safelight conditions are inadequate. It is therefore a published indication of what detectable on paper means, pressed into service as a uniformity limit. No source in this course’s corpus establishes a uniformity tolerance for a contact printer, and rather than invent one this page reuses the only density-detectability figure it has and says so.

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

A symmetrical map, dark at all four corners and lighter in the middle, is geometry. The box is too shallow or the emitter field does not overhang far enough. The fixes are depth and overhang, and both cost output.

A lopsided map, falling steadily from one side to the other, is not geometry. It is a diffuser not sitting flat on its ledges, a box out of square, or one end of the strip run at a lower voltage than the other because the supply feeds it from one end. Feed a long strip from both ends and re-map.

A map with a stripe in it, running parallel to the strip rows, is the emitter pitch printing through. The diffuser is too close. Raise it 20 mm and re-map; that is why the ledges were left adjustable in stage 5.

A map with one bright patch is a gloss spot inside the box, a gap at the diffuser’s edge, or a brighter bin of strip spliced into the middle of the run.

With no densitometer, the fallback is honest and coarse: cut the sheet into its fifteen squares, shuffle them, and try to sort them 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 result, and a weaker one than 0.04. Write down which test you did.

T4 — The step-wedge contact print

The uniformity map says the field is even. This says it transfers a tonal scale.

Contact-print a 21-step transmission step wedge at your mid-grey exposure, process it with everything else, and read the steps. You are not looking for a particular curve — that is Part XIII’s subject — but for two failures the frame can cause and the uniformity map cannot see:

• Compressed shadows with a soft edge to every step: a gap, with the stage 1 penumbra smearing each step’s edge into its neighbour. Check T1.
• A different number of steps at one end of the wedge than the other: the uniformity map again, in the units that matter.

Record the exposure and the step count beside the map.

T5 — The proof exposure, established and written down

Finally, the number the box exists to give you. Lay out a real film, expose a proof sheet, and adjust by Kodak’s rule — double it if light, halve it if dark — until the clear rebate just reaches maximum black. Two or three sheets will find it.

Write it on the label: the seconds, the paper, the developer and its dilution, and the date. The number is only true for that combination and the label says so — but when the paper changes, finding it again costs one sheet rather than an afternoon, because everything else is nailed down.

Troubleshooting

Symptom	Likely cause	Test that distinguishes it
Every frame on the proof sheet is soft, evenly	Printing through the base, or a systematic gap	Look at the sheet: if it is also mirrored, the negatives went in the wrong way up. If not, run T1
Some frames sharp, some soft, on the same sheet	Curl in individual strips, or grit under one strip	Flatten negatives under weight overnight and re-proof. If the soft ones are the outer strips, it is spring pressure at the edges
A soft, tent-shaped region around a point	One grain of grit propping the sandwich open	Brush the glass; the region is centred on the speck
Coloured or grey concentric fringes over the image	Newton’s rings	T2
Corners lighter than the centre, symmetrically	Box too shallow or overhang too small	T3, and read the map’s shape
One side consistently lighter	Diffuser not flat on its ledges, or the strip fed from one end only	T3; feed the strip from both ends
A regular stripe pattern	Emitter pitch printing through a diffuser that is too close	Raise the diffuser 20 mm and re-map
The exposure that worked last month is wrong	The geometry moved, the paper changed, or the strip aged	Check the frame is against all three battens; check the current at the fuse against the label
Sheet fogged at one edge, no image structure	Light leaking from the box past the frame, or from the safelight	Re-run the box’s own leak check with the frame in place and the room dark, then re-read the safelight’s working time
Rebate never reaches maximum black however long the exposure	Exhausted developer, or fixer in the developer	This is a chemistry fault, not a frame fault; see Part IV

Questions

1. Your diffuser is 320 mm wide and sits 180 mm below the negative. You print a 5 × 4 sheet negative the wrong way up, through its 0.180 mm polyester base. Calculate the penumbra. Then calculate how far away a 30 mm source would have to be to give the same blur, and say what that tells you about which of the two errors matters more.
2. A reader argues that because their light box is “perfectly even to the eye”, the uniformity map is a waste of a sheet of paper. Give two independent reasons the eye is the wrong detector here, and name the one circumstance in which their claim would be defensible.
3. Anti-Newton glass fixes the fringes and adds a texture to the image. Under what circumstances is that trade worth making, and what would you measure to decide?
4. Explain why the course specifies a fixed drive current and a timed exposure rather than a dimmer, and then give the one situation in which a dimmer would be the right answer for a light box of this kind.
5. You are given a bought contact printing frame at £74.80 and told it is better than a built one. List what it does and does not give you against the specification on this page, and say which of the tests in Testing and calibration it exempts you from.
6. The proof-sheet convention pins the clear rebate to maximum black. Suppose you instead exposed each sheet so that the best frame on it looked right. Describe two decisions you would then be unable to make from a stack of proof sheets, and say why.
7. Your uniformity map shows the four corners 0.03 down, the centre flat and one edge 0.07 down. Which fault is which, and in what order would you address them?

Further experiments

Measure the penumbra rather than calculating it. Contact-print a razor blade’s edge with a 0.5 mm shim under one end, so the gap varies smoothly from zero to 0.5 mm along it. Under a loupe the edge softens as the gap grows; mark where you first see it. Divided by your aspect value, that is your own threshold in the only units that matter — a better number than either Ware’s 0.3 mm or this page’s 0.02 mm.

Map the same box twice, once on paper and once with a meter. The two maps will not agree, and the disagreement is informative: it is the difference between the paper’s spectral response and the meter’s photopic one, plus the paper’s own curve steepening the map where the meter is linear. Whichever map is prettier, the paper is the one that decides.

Take the depth trade seriously. Map the field at three diffuser heights — h = p, h = 2p and h = 3p — reading a mid-grey sheet each time and also recording the exposure needed to reach that mid-grey. You will have measured, on your own box, the exact rate at which evenness is bought with light, and you will be able to choose a point on that curve instead of guessing at one.

Print the same negative in the frame and under a sheet of glass and a weight. Same paper, exposure and processing; compare the corners under a loupe. That is the alternative route’s cost, measured — and it is usually smaller than expected for a flat sheet negative and larger for a curled 35 mm strip.

Recap

A contact printer is two instruments with one specification between them. The frame’s job is to make the gap between the two emulsions zero, because blur is that gap multiplied by the source’s aspect value — a relation that punishes a large near source and forgives a small distant one. Flat baseboard, stiff glass, closed-cell foam and enough springs deliver it; curl, grit, a bowed back and tired foam defeat it.

Glass is chosen for flatness and stiffness, and float glass bought cut and polished answers both. Its transmission is irrelevant for silver paper and decisive for the alternative processes, where ordinary glass passes UVA and acrylic cuts off at about 370 nm — so a frame that will ever see a UV unit is glazed in glass. The back is split so a printing-out image can be inspected without losing registration: a convenience for silver, the method for cyanotype.

The light box is a matt white integrating chamber whose depth is the real design variable, traded against evenness and against output. Its acceptance test is a mid-grey sheet of paper read at fifteen points against the centre, with a 0.04 density limit borrowed openly from ILFORD’s safelight sheet because the course has no uniformity figure of its own. Exposure is by time at a fixed current, because a time is a record and a knob position is not.

And the thing it makes is a proof sheet exposed so the clear rebate just reaches maximum black — a convention whose whole value is that it makes every sheet you will ever make comparable with every other one.

Check your understanding

Question 1. A negative is printed base-up through its 0.110 mm acetate base, under a 300 mm diffuser 250 mm away. Roughly what penumbra does each edge acquire, and what is the single most effective correction?

• About 0.005 mm; the correction is a thicker glass
• About 0.13 mm; the correction is to turn the negative over so the two emulsions touch
• About 0.13 mm; the correction is to raise the diffuser further from the negative
• About 1.3 mm; the correction is anti-Newton glass

Show the answer and why

Answer: About 0.13 mm; the correction is to turn the negative over so the two emulsions touch

b = S x g / D = 300 x 0.110 / 250 = 0.132 mm. Raising the diffuser does help, and it is the third option for a reason - but to get the same improvement geometrically you would have to move it to more than 6 m, which is not a box. Turning the negative over takes g from 0.110 mm to a few micrometres and takes b with it, because the two quantities multiply. It is the cheapest correction in the whole build and it is free.

Question 2. Why does this page insist that the uniformity map be read on a mid-grey sheet of paper rather than with a lux meter?

• Because a lux meter is expensive and most readers do not own one
• Because the paper is the actual detector the instrument serves, and its steepest response is at mid-grey, whereas a lux meter reads a photopic weighting the paper does not share
• Because a lux meter cannot resolve differences as small as 0.04
• Because paper integrates over the exposure and a meter does not

Show the answer and why

Answer: Because the paper is the actual detector the instrument serves, and its steepest response is at mid-grey, whereas a lux meter reads a photopic weighting the paper does not share

Two separate things are wrong with the meter and only one of them is about precision. A photopic meter weights the spectrum the way the eye does, and a bromide or variable-contrast paper does not respond that way, so a spectral difference across the field can be invisible to one and not the other. And the paper is steepest in the middle of its curve, so a mid-grey sheet is the most sensitive detector you have for a small difference in exposure. The Further experiments section asks you to make both maps precisely because the disagreement between them is informative.

Question 3. Your uniformity map shows a regular stripe pattern running parallel to the LED rows. What has gone wrong and what is the fix?

• The strip has a faulty section; replace the reel
• The diffuser is too close to the emitters, so the emitter pitch is printing through; raise it and re-map
• The interior is gloss rather than matt white; repaint it
• The frame is not against its battens; reposition it and re-map

Show the answer and why

Answer: The diffuser is too close to the emitters, so the emitter pitch is printing through; raise it and re-map

A stripe whose spacing matches the emitter pitch is the source structure surviving to the diffuser. The ripple above a row of emitters dies away as the diffuser is moved further from them, which is why stage 5 asks for a diffuser height of at least the emitter pitch and preferably twice it, and why the ledges are built adjustable before anything is glued. A gloss interior gives a different signature - discrete bright patches where emitters are imaged, not a regular stripe - and a misplaced frame gives no pattern at all, only a shifted field.

Question 4. A reader plans one frame for both silver paper and cyanotype, and proposes acrylic glazing because it is lighter and will not cut anyone. What is the objection?

• Acrylic is not flat enough for contact printing at any size
• Acrylic scratches, and the scratches print
• Most acrylic does not transmit below about 370 nm, which sits inside the UVA band and above the 365 nm peak of the shorter common UV LED, so it would act as a filter in the UV unit
• Acrylic takes a static charge and attracts dust

Show the answer and why

Answer: Most acrylic does not transmit below about 370 nm, which sits inside the UVA band and above the 365 nm peak of the shorter common UV LED, so it would act as a filter in the UV unit

All four statements are true and only one of them is disqualifying. ICNIRP record that window glass transmits down to about 310 nm while polymethyl methacrylate and polycarbonate normally do not transmit below about 370 nm - so for visible-light silver printing acrylic is a legitimate choice made for weight and safety, and for a 365 nm UV array it is a filter. Scratching and static are reasons to be careful with acrylic; the cut-off is a reason not to use it for one of the two jobs at all.

Question 5. What does exposing every proof sheet so the clear rebate just reaches maximum black actually buy you?

• The deepest possible blacks in the final enlargements
• A sheet whose exposure is standardised on the one part of every negative that is identical, so differences between frames and between sheets are differences in the negatives rather than in the printing
• Protection against safelight fog, because maximum black hides it
• A guaranteed match to the enlarger exposure for the same negative

Show the answer and why

Answer: A sheet whose exposure is standardised on the one part of every negative that is identical, so differences between frames and between sheets are differences in the negatives rather than in the printing

The rebate is clear base plus fog: the same on every frame of every film of that type, and the only part of a negative that carries no picture. Pinning it to the same tone makes the sheet a measuring instrument rather than a set of small pictures - a thin negative reads thin against the same reference on every sheet you have ever made. The course states this as a printing convention with its reasoning and cites no source for it, because none in its corpus establishes one.

Question 6. Why does the build specify a toggle switch and a timer socket, and explicitly no dimmer?

• Because a dimmer would introduce flicker that could beat with the exposure
• Because an LED's output is not proportional to current, current shifts the die temperature and hence the output, and a knob position is not a record that survives a knock or another user
• Because dimming a constant-voltage strip damages it
• Because the timer cannot drive a dimmed load

Show the answer and why

Answer: Because an LED's output is not proportional to current, current shifts the die temperature and hence the output, and a knob position is not a record that survives a knock or another user

Three effects stack, and the third is the one that bites in practice. Output against current is a curve rather than a line; changing the current changes the junction temperature, which changes the output again on a time constant you have not characterised; and a dimmer setting cannot be written down in a way that another person, or you in a year, can reproduce. A time can. If the box is too bright, the page's fixes are all permanent and measurable: a second diffuser, a neutral density sheet, or fewer rows.

Check my answersClear