Negative popping

Lifted from enlarger optics and design, which owns film popping as a movement fault belonging to the holder, derives the depth of focus this entry measures a bow against, and carries the honesty note about what the course could not source; from the LED enlarger head, whose rule that the heatsink fins go outside the chamber exists for this fault and no other; and from commissioning the enlarger, whose four-panel softness diagram makes this panel 4 and whose A3 reads a dry print with a loupe. Part XVII owns it twice: as one of the nine causes on the timer that lies, which is where the sharpness, not density discriminator comes from, and as the exposure record that makes any interim correction repeatable.

What you see

A print whose fine detail is all there, in the right place, and weak. Not blurred away — present, and low in local contrast, as though a thin veil had been laid over that part of the sheet. Hard edges carry a soft halo. Somewhere in the affected region a loupe finds the fine detail written twice, once sharply and once faintly beside it.

That is the whole appearance, and it is why the fault is so often filed under something else. It does not look like a focus error, because a focus error takes the fine detail away altogether. It looks like flare, or a dirty lens, or a negative that was never sharp.

Four observations settle it, and each costs nothing.

• The negative is innocent. Light box, loupe, both sides. The detail is sharp on the film. A camera-side softness travels with the negative; this one exists only in the projection.
• The softness is a band or a patch across the middle of the frame, and the corners are sharp. It is not arranged about the optical axis and it is not arranged about the diagonal. It runs across the free span of the carrier’s aperture, which is where a clamped sheet bows.
• It lands somewhere different on every print. Reseat the negative, print again, and the band moves, changes width, or does not appear. Nothing bolted to the enlarger behaves that way.
• It is worst on the first print after a break. Print the same negative three times in ten minutes and the first is the bad one; go and make a cup of tea and the fault comes back on the next sheet. A defect with that schedule is a defect about warming up.

Two more signs are worth knowing because they arrive before the print does. With a grain focuser on the projected image and the lamp left on, the grain visibly softens and sometimes jumps in the first minute or two, and on some glassless carriers there is a faint tick from the negative stage as it goes. That tick is the name of the fault.

What the dry print shows	What it is	Where it goes
Fine detail present but weak, haloed at edges, faintly doubled, in a band across the middle, in a different place on each sheet	Negative popping — this entry	The pre-warm test below
One corner soft, the diagonally opposite corner sharp, with a measurable keystone	Misalignment	one soft corner
Everything soft, centre included, evenly	Focus, or diffraction at an effective aperture of N(1 + m)	Refocus at full aperture
Closed, nested, irregularly spaced grey loops	Newton’s rings in a glass carrier	Newton’s rings
Flat grey shadows and a compressed tonal scale, coarse tones affected as well as fine	Veiling flare	flare veil
A denser centre, bowed straight lines, a centre very slightly sharper	A sheet not flat in a camera	buckled-sheet softness
A complete second image displaced in one direction, all across the frame	Something moved that was not the negative	motion smear
Prints getting darker or lighter through an evening, sharpness unaffected	The source or the developer	session exposure drift
Corners darker but not soft	Falloff, source, or a mechanical vignette	uneven enlargement

The doubling has a direction, and it is the direction that proves the case. The faint second image is displaced radially — outward from the centre of the print, or inward towards it — and by more the further from the centre you look. It is not displaced the same way everywhere, which is what camera shake, an easel that shifted or a head that was knocked would give. The reason is in the arithmetic below, and it is the one observation on this page that no other fault can imitate.

On an alternative-process sheet made by contact there is no focus to lose, so the fault shows differently: as a loss of contact in the middle of a frame that has been in a warm box for ten minutes, printing as a soft-edged patch, and — where two printings must register — as a negative that is no longer where it was for the first one.

Likely causes

Every one of them is a statement about heat arriving at a sheet that is not free to grow, and none of them is chemistry.

1. A glassless carrier at a high magnification. The commonest by a long way. The film is held by its rebate at two or four edges and the middle looks after itself, which is fine until the middle gets longer.
2. A cold negative, focused at once. The fault is not that the negative is warm. It is that the negative changes while the shutter is open. A negative that has come to the lamphouse’s temperature before you focused it does not pop, at that magnification, again this evening.
3. A tungsten head. A filament radiates far more infrared than visible light and sends nearly all of it down the same path. This is the historical reason for heat-absorbing glass in a condenser stack and for the large-format printer’s habit of letting a negative sit in the beam before final focusing.
4. An LED head that dumps its waste heat inside the chamber. The emitter radiates almost no infrared, so the radiative route is largely closed — and its own dissipation still has to go somewhere. A chamber warmer than the room warms the negative by convection instead, which is the same fault arriving by a different road. The build page puts the fins outside the chamber for exactly this reason.
5. A long exposure, or a long burn. ILFORD’s beginner sheet aims at about ten seconds. A ten-second exposure may be over before the negative has moved; a ninety-second burn at a small aperture is a different proposition, and so is a sequence of exposures with the lamp on between them.
6. Thin acetate rather than thick polyester, and a big span rather than a small one. ILFORD publish FP4 Plus on 0.125 mm acetate in 35 mm, 0.110 mm acetate in roll film and 0.180 mm polyester in sheet film. The sheet-film base is both a different polymer and 64 per cent thicker than the roll film, and it spans a much larger hole. Both of those matter, in opposite directions.
7. A carrier that grips unevenly. Two edges clamped hard and two free gives the strain exactly one place to go. A rebate with grit under it does the same locally.
8. A contact frame or UV unit that has been running for ten minutes. No focus is lost, but contact is, and registration with it.

Not this entry at all: anything you can find on the negative under a loupe, and anything that appears in the same place print after print. Both are excluded by the test below, on one sheet.

What is happening: the chemistry and physics

Nothing chemical happens at all, and that is worth saying because the appearance — a veil, a loss of local contrast, a stain-like softness — sends people to the developer. No developer, fixer or reducer touches a fault whose entire cause is that a piece of plastic got longer.

The negative is a laminate, and its flatness is a balance

The Image Permanence Institute’s account of the plastic-base negative describes cellulose acetate film as four layers: the gelatin emulsion on top, a cellulose nitrate subbing beneath it, the cellulose acetate support, and a bottom gelatin anti-curl layer. That last layer is not packaging. IPI say what it is for in as many words: it is always present on the base side of the film to counteract the tension exerted by the gelatin emulsion and provide dimensional stability, and the practice dates from 1903, when a gelatin layer began to be applied to the back of the support to compensate for the tension exerted by the binder layer. On roll film it also carries the anti-halation dye, which clears in processing.

So a flat negative is not a flat object. It is a sandwich held flat by two gelatin layers pulling against each other across a plastic core, and gelatin is hygroscopic. Anything that changes one layer more than the other — warmth, drying, a few minutes under a lamp — unbalances the pair and the sheet curls. That is a second mechanism running alongside the thermal one, in the same direction, on the same timescale.

📓Two mechanisms, one sourced coefficient between them: none

The film gets longer because it is warmer, and it curls because its two gelatin layers dry unequally. This course has no sourced coefficient of thermal expansion for the bases it uses, no sourced moisture coefficient for their gelatin layers, and no measured split between the two. The optics page already records the same gap and this entry does not fill it.

What the page does instead is work in strain — fractional change in length — rather than in temperature, and show that the strain needed is so small that arguing about which mechanism supplies it is beside the point. If you meet a coefficient elsewhere, check where it came from before you plan around it.

Heat gets in whether or not the lamp looks hot

Wall’s 1912 dictionary is a good witness that the enlarging lantern was always a heat problem. Its Condenser entry compares the condenser to a burning-glass concentrating “the light (and also the heat) of the sun”, and gives an instruction that is really an engineering principle: all condensers should be so loosely mounted in thin cells that they can be turned round, otherwise the expansion by the heat may cause them to crack. The glass in the lantern is hot enough to break itself if it is not left free to grow. The negative sits below that glass, is far thinner, and is clamped.

The same dictionary’s Calorific Rays entry notes that glass is highly adiathermanous, so photographic lenses pass few heat rays. That is a statement about the lens, below the negative, and it is no comfort at all: the negative is on the source side of the lens, where the heat is.

An LED head closes the radiative route and opens a convective one. Cree publish, for the emitter family this course works from, a thermal resistance from junction to solder point of 5.8 °C/W for white and a maximum junction temperature of 150 °C; the build page’s worked case, one white emitter at 700 mA, dissipates about 2.09 W and puts the junction about 12 °C above its solder point. Two watts is not much until it is inside a closed white box with the negative under its exit port and nowhere else for the heat to go.

Buckling: how a strain of one part in ten thousand becomes a tenth of a millimetre

Take a strip of film clamped at two opposite edges, a free span L apart, and let it grow by a fractional strain ε. The extra length has nowhere to go along the plane, so it goes out of it. Model the bow as a single half-wave, y = h sin(πx/L), and set its excess arc length equal to the extra length εL:

h = (2 L ÷ π) √ε

Bow height from a constrained in-plane strain

That square root is the whole story of this fault. The in-plane extension is εL; the out-of-plane displacement is larger than it by a factor 2 ÷ (π√ε), which grows as the strain gets smaller. At ε = 10⁻⁴ the factor is about 64.

Free span L	ε = 10⁻⁵	ε = 10⁻⁴	ε = 10⁻³
24 mm — the short side of a 35 mm frame	0.05 mm	0.15 mm	0.48 mm
56 mm — a 6 × 6 frame	0.11 mm	0.36 mm	1.13 mm
102 mm — the short side of a 4 × 5 sheet	0.21 mm	0.65 mm	2.05 mm

A 24 mm span that grows by one part in ten thousand grows by 2.4 micrometres, which is nothing, and bows by 153 micrometres, which is not. That is the mechanism in one line: a film cannot resist a small strain by being stiff, only by being long enough to go somewhere else.

Two riders, both honest. The strains in that table are chosen to bracket the answer, not measured — a tenth of a per cent is the top of the range and a thousandth of a per cent the bottom, and the reader should read the table as a shape rather than as a specification. And the model assumes the excess goes into one smooth arch. A real negative held on four sides makes a shallower, messier surface with more than one hump, which is why the band of softness on a print is rarely a tidy stripe.

What the darkroom can afford: depth of focus at the negative

The optics page derives the tolerance and this entry uses it rather than re-deriving it. Depth of focus at the easel is T = 2cN(1 + m), where c is the largest blur you will accept on the print; longitudinal distances scale as m² between object and image space, so the corresponding tolerance at the negative is that divided by m²:

δu(max) = 2 c N (1 + m) ÷ m²

How far the negative may move before the print shows it

The course uses c = 0.05 mm, which is about the finest detail a good eye resolves on a print held at arm’s length, and says so rather than pretending the number is standard.

	f/2.8	f/5.6	f/11
2×	0.21 mm	0.42 mm	0.83 mm
4×	0.09 mm	0.18 mm	0.35 mm
8×	0.04 mm	0.08 mm	0.15 mm

Now put the two tables beside each other. A 0.15 mm bow — one part in ten thousand, on a 35 mm frame — is comfortably inside the tolerance at 2× and any aperture, is marginal at 4×, and is twice the tolerance at 8× and f/5.6. This is why the fault has the reputation it has: the same negative in the same carrier is innocent at postcard size and ruins an exhibition print, and nothing about the enlarger changed.

🧮What the blur actually measures on the print

Turn the tolerance round. A negative displaced by δu puts the image plane m²δu away from the paper, and the converging cone has effective f-number N(1 + m), so the blur circle on the print is

b = m² δu ÷ [N (1 + m)]

At m = 8, N = 5.6 and δu = 0.15 mm: b = 64 × 0.15 ÷ (5.6 × 9) = 0.19 mm, nearly four times the 0.05 mm this course is prepared to see.

At m = 2, N = 11 and the same 0.15 mm bow: b = 4 × 0.15 ÷ (11 × 3) = 0.018 mm, which is a third of the limit and invisible.

One movement, two verdicts, and the only things that changed were the magnification and the aperture. This is the course’s own arithmetic from its own derivation, not a published figure.

Why the print is not simply out of focus

Here is the part that explains the appearance, and it follows from one fact: the pop happens part way through. Call f the fraction of the exposure delivered before the negative moved. The paper does not average the two states — it adds them, in exposure:

E(x) = f · E(sharp) + (1 − f) · E(defocused)

What the paper receives when the negative moves mid-exposure

Defocusing does not destroy light, it redistributes it locally. So for any detail coarser than the blur circle, the two terms are the same and nothing at all has happened. For any detail finer than the blur circle, the second term contributes only its local average, and the modulation surviving in the sum is f times what a clean exposure would have given.

A pop halfway through therefore halves the contrast of the fine detail and leaves the tonal masses exactly as they were. The detail is still in the right place, at the right size, at reduced amplitude, sitting on a veil made of the light that was spread out of it. That is why the print reads as flared or dirty rather than as unsharp — and it is the discriminator against real veiling flare, which adds a roughly uniform term and therefore compresses the whole tonal scale, coarse tones included. Popping eats fine detail and nothing else.

It also explains why the fault is worse when the timing is unlucky. A pop in the last tenth of the exposure costs a tenth of the fine-detail contrast and is nearly invisible. A pop at the halfway mark costs half of it.

And it explains the row on Part XVII’s table: this is a sharpness fault and not a density fault. The total light through the negative did not change, only its distribution, so a densitometer laid on a large uniform patch finds nothing while a loupe finds everything.

Why the second image is displaced radially

The paper stays where it is, so what fixes the position of a point’s blur on the print is the ray through the centre of the lens. That ray puts an object point of height y at −y·v/u on the paper, with v fixed. Move the negative by δu and the whole affected region is re-scaled about the axis by δu/u, so a point at radius R on the print shifts by

δR = R · δu ÷ u

Lateral shift of the second image

with u = f(1 + 1/m) the negative-to-lens distance. The shift is zero on the axis and grows outwards. For a 50 mm lens at 8×, u = 56.25 mm; at a point 50 mm out from the print centre, where a 0.15 mm bow is still about 0.10 mm deep, δR = 50 × 0.10 ÷ 56.25 = 0.09 mm — nearly twice the acceptable circle, and plainly a doubled edge under a loupe.

Two consequences the reader can check on a single print. The doubling is radial, so its direction is different in different parts of the sheet, which no external movement can imitate. And the two symptoms peak in different places: the defocus is worst at the middle of the bowed span, on or near the axis, where the lateral shift is zero, while the doubling is worst part way out, where the bow is still deep and the radius is already large. A popped print therefore looks softly veiled in the middle and doubled around it, and looks messy rather than uniformly bad.

This derivation is the course’s own paraxial arithmetic from the geometry the optics page establishes. No published photographic account of the radial signature was found in this course’s corpus, and it is offered because it predicts something checkable rather than because somebody else wrote it down.

Why polyester sheet film behaves differently from acetate roll film

Two published facts and one piece of reasoning.

IPI describe how polyester base is made: cast from a melt, biaxially oriented by stretching in both directions, then — while still mechanically restrained — heated considerably above the stretching temperature, which grows crystallites and locks the polymer chains together so that the sheet is dimensionally stable up to much higher temperatures. They add that PET has the highest strength, toughness, stiffness and tear resistance of any commercial film base at normal temperatures, and that it was adopted first for negatives that required high dimensional stability. ILFORD, for their part, put sheet film on 0.180 mm polyester and roll film on 0.110 mm acetate.

The course’s own inference, stated as one: a polyester sheet should both grow less for a given warming and resist the bow more, being 64 per cent thicker than roll film in a material of higher stiffness — and it also spans four times the distance in a 4 × 5 carrier, which the bow relation punishes directly. The two effects pull opposite ways and this course has not measured either, so it does not claim a winner. What it will say is that the span term is the one with no square root in it: at a fixed strain, h is proportional to L, and format is the variable the printer chose.

!Under glass, the same heat makes a different fault

A glass carrier removes the bow by force: the film cannot leave the plane because there is a sheet of glass on each side of it. The heat does not go away, and neither does the expansion — the film grows against friction between two flat surfaces, so the fault it produces is not defocus but movement of the gap between base and glass. That is exactly the mechanism the Newton’s rings entry lists last: a negative warming in the beam, so the fringe pattern moves part way through the exposure and smears.

One fault is traded for another, deliberately. The trade is the optics page’s to make, and it is made with knowledge of the third term — four more surfaces to keep clean.

Diagnostic questions

• Is the detail on the negative? Light box, loupe, both sides. If the softness is on the film, this page has nothing for you: go to the camera-side entries.
• Is the fine detail missing, or present and weak? Missing is a focus error. Present, weak and haloed is this fault, and the distinction is made with a loupe on a dry print and nothing else.
• Are the tonal masses affected? If the shadows have gone flat and the whole scale is compressed, you are looking at flare, not at popping.
• Where is the band? Across the middle of the frame with sharp corners is a bowed sheet. At one corner with the opposite corner sharp is alignment. Everywhere including the centre is focus.
• Does it land in the same place on the next print? A fault that moves when you reseat the negative is not bolted to the machine.
• Is the first print of a session the bad one? A fault that improves through a run and returns after a break is a fault about warming up. This is the strongest single indicator on the list.
• Under a loupe, which way is the doubling? Radial, and different in different parts of the sheet, is this entry. One direction everywhere is something that moved.
• How long was the exposure, and what was the magnification? Ninety seconds at 8× is the fault’s natural habitat; ten seconds at 2× is not.
• Glassless carrier? Then this fault is available to you. Glass carrier, and you should be hunting rings instead.

!The confirming test: change the negative's temperature and nothing else

Print the same negative twice. Between the two prints change exactly one thing — whether the negative was at the lamphouse’s temperature when the exposure began.

1. Print one, cold. Take the negative from its sleeve, put it in the carrier, focus at full aperture on the grain with a focus finder as ILFORD’s sheet directs, stop down, and expose immediately.
2. Then leave it there with the lamp on and the lens capped. The cap stops the paper being exposed and stops nothing reaching the negative, which is the point: the negative receives the whole beam and the easel receives nothing. Two to three minutes.
3. Refocus — it will have moved, and watching how much it moves under the grain focuser is half the diagnosis on its own — then uncap and print two, same paper, same aperture, same time, same developer.

• Print two is sharp and print one is not. The diagnosis is made. Nothing changed but the negative’s temperature at the moment of exposure.
• Both prints are equally soft. Not this fault. Go to alignment, to focus, or to the lens.
• Both are sharp. You did not reproduce it. Check the magnification and the exposure length against the print that failed, because this fault is conditional on both.

The test that costs no paper at all is to watch the grain focuser for two minutes with the lamp on before you print anything. If the grain drifts out of focus while you look at it, the negative is moving, and you have the answer before a sheet is committed.

No published account of either test was found in this course’s corpus. Both are offered as reasoning from the mechanism: temperature is the only variable that changes the bow without changing anything in the light path.

Corrective action

Do not refocus and reprint at the same setting and hope. The negative will be at a different temperature next time and the fault will be in a different place. And do not take it to a bath: there is no chemistry in it.

Work in this order.

1. Pre-warm the negative and focus after it has settled. Free, immediate, and it is the fix Part XVII’s table gives. Lamp on, negative in the carrier, lens capped, two to three minutes; then focus, then print. This is the whole repair for a print that must go out tonight, and it is a good habit permanently.
2. Fit a glass carrier where the enlarger allows one and the format needs it. It makes the negative flat by force and it is the other fix on Part XVII’s row. Read the trade before you buy: four more surfaces to keep clean, the possibility of Newton’s rings, and anti-Newton glass with its own cost if the rings actually print. The optics page owns that decision.
3. Get the heat out of the lamphouse rather than out of the negative. If you are building the head, the fins go outside the chamber; the build page’s stage 3 makes that a rule and this is the fault it exists to prevent. If the head is a donor tungsten lamphouse, its heat-absorbing glass — if it has any — is a component and not an optional extra.
4. Shorten the time the lamp is on. Not only the exposure: the focusing, the burn, the dodge and the gaps between test-strip bands are all time with the beam on the negative. Working at a wider aperture shortens the exposure and narrows the depth of focus, so it trades one term for the other; working at a smaller one does the reverse, up to the diffraction limit Rodenstock warn about, which bites at the effective aperture N(1 + m) rather than at the number on the barrel.
5. Split a long burn. A ninety-second exposure given as three thirty-second exposures with the lamp off in between is not the same thermal history, and where a print needs a long total it is worth trying. Record it on the exposure record if it works, using the f-stop timing convention, because an unrecorded change is one you will make differently next week.
6. Check the carrier grips evenly. Grit under a rebate, a bent frame or a mask fitted to the wrong format all clamp unevenly, and an unevenly clamped sheet pops at less provocation than a fairly clamped one.
7. In a UV unit or contact frame, this is a scheduling problem rather than an optical one. Load the frame, close the box, and give the negative the same settling time before the exposure starts that you gave the lamps; the unit’s own temperature log tells you what the box is doing.

What you cannot do is burn or dodge it away. Dodging changes density; the loss here is a loss of fine-detail contrast at unchanged density, and adding or subtracting exposure over the region gives a lighter or darker region with the same weak detail. If a print with a known popped band leaves the darkroom, write it on the back with the magnification, the aperture and the exposure length, so the next print of that negative starts from information.

Prevention

Make pre-warming part of focusing, not a remedy for a fault. Lamp on, negative in, lens capped, two minutes, then focus. It costs two minutes per negative and removes the commonest sharpness fault in enlarging from the list of things you have to diagnose. This is the same discipline as ILFORD’s focus at full aperture then stop down: do the demanding thing first, so that everything afterwards is inside the setting you found.

Focus at full aperture, and know that you are focusing at the worst case. Depth of focus is narrowest wide open, which makes best focus easiest to find and also means the setting you found is the one least tolerant of a negative that then moves. Stopping down widens the tolerance around a setting that was already right.

Record the carrier on the enlarger’s certificate, as commissioning requires: glass or glassless, anti-Newton or plain, and the date. A machine whose configuration is written down is a machine whose faults can be attributed a year later, and the choice between the bow and the rings is the single line on that certificate that decides which of these two entries you will be reading.

Design the heat out at build time. Fins outside the chamber; a baffled vent that is a thermal control; a temperature log taken with the lid shut. Every one of those is on the build pages already for other reasons, and this fault is one more.

And keep the pairing straight. A glassless carrier can bow and cannot ring; a glass carrier can ring and cannot bow. They are the two ends of one decision, they cannot both be cured by the same adjustment, and a printer who knows that will change the carrier rather than chase the developer.