Digital PD Meter and Pupillometer: What Actually Matters in the Dispensary

A digital PD meter — the instrument most opticians still call a pupillometer — does one small job that almost every other number in the dispensary depends on. It fixes where the optical centre of each finished lens will sit relative to the eye behind it. Get it right and a job is unremarkable. Get it wrong by a couple of millimetres and you have a remake, a non-adapt, or a progressive that never quite settles, and nothing on the order form will tell you why.

This guide is written for opticians, dispensing staff and practice owners who are weighing a digital pupillometer against the millimetre rule already in the drawer, or replacing an ageing unit. It is not a spec recital. It covers what the instrument actually measures, how much error the lens standards will absorb before a job goes out of tolerance, what the published comparison studies really found when they put rulers and pupillometers side by side, and the specific questions worth asking before a quote is signed.

What a PD meter measures — and why it is not the number a rule gives you

The two instruments are not measuring the same landmark, and this is the single most consequential fact in the category.

A dispensing pupillometer rests on the bridge of the nose and uses a coaxially mounted light source to sight the reflection of that light off the front of the cornea. Writing in 20/20 Magazine in January 2021, Barry Santini describes the mechanism plainly: these instruments "differ in their approach to the ruler/pupil center method by sighting the corneal reflection with the use of a coaxially-mounted light source—which eliminates the parallax error." Because the light and the observation path share an axis, there is no angular offset between where the operator is looking from and where the light is coming from.

The corneal reflex is not in the middle of the pupil. Santini puts it at "approximately 1 mm nasalward of pupil center at the spectacle plane," and notes it "has been considered the most accurate marker for the eye's visual axis." Jung and Chu, in Clinical Optometry in 2024, describe the same split from the other direction: the pupillometer "measures the visual axis, which is slightly nasal from the centre of the pupil," while "the PD ruler without penlight torch measures from the pupil centre, which is the pupillary axis."

The two-millimetre offset that is built into the method

Because the reflex sits about a millimetre nasal in each eye, the offset doubles across the pair. Santini gives the figure directly: a binocular PD taken from the corneal reflex is "typically near 2 mm narrower at the spectacle plane than one measured using pupil center."

That is not an error. Both landmarks are real and both are defensible. It is a convention, and it means two competent staff members using two correct techniques on the same wearer can legitimately write down numbers 2 mm apart. Whether that matters depends entirely on how much room the lens standards leave you — which is the next question, and the one almost no page in this category answers.

How much PD error the lens standards actually absorb

Centration error does not stay a millimetre problem. A lens whose optical centre is displaced from the visual axis behaves as a prism, and the standards are written in prism for that reason. Eyecare Business states the relationship the industry uses, Prentice's rule, as "P = hD" where "P = prism diopters of displacement, h = centimeters from the optical center, and D = diopters of power," and works it through: "P = (0.5) × (4.00)" — a 5 mm decentration on a +4.00 D lens produces 2.00 prism dioptres.

The tolerances below are as published by Eyecare Business from ANSI Z80.1-2010, corroborated independently by a second source that reports the same figures and adds that they are "consistent across the 2010, 2015 and 2020 editions" with the current edition being Z80.1-2025. We are reporting what those sources publish; we have not read the standard itself, and any specification that has to survive an audit should be checked against the current text.

Running the numbers backwards: how many millimetres is that?

Prism dioptres are the right unit for a lab and the wrong unit for a dispensing table, so invert Prentice's rule. A horizontal centration error of h millimetres across the pair produces an imbalance of h × D ÷ 10 prism dioptres, so the error that consumes the entire ±0.67 Δ allowance is 6.7 ÷ D millimetres. Note that this branch of the standard only governs up to ±2.75 D — above that the standard stops asking about prism and simply requires the finished PD to be within 2.5 mm of the one you specified:

Read the bottom row against the previous section. At the top of the prism-governed band, the ordinary two-millimetre difference between a corneal-reflex PD and a pupil-centre PD accounts for about 82% of the entire horizontal allowance the standard gives the finished job — before the lab has cut anything, and before any operator error at all.

There is a counter-intuitive consequence worth naming, because it runs against instinct. The millimetre allowance does not get tighter as lenses get stronger. Below ±2.75 D the prism rule bites, and the tolerable millimetres shrink as power rises. Above ±2.75 D the standard switches to a flat ±2.5 mm and stays there no matter how strong the lens gets. The band where centration is least forgiving in millimetre terms is therefore the top of the moderate range, not the high-power end — which is precisely the band most dispensaries handle most often.

Where the standard's two branches meet, and where they do not

The tolerance table switches units at a power threshold, which invites an obvious arithmetic check: at the crossover, do the prism rule and the millimetre rule give the same answer? Running the cited figures through the cited formula:

• Vertical. 1.0 mm is 0.1 cm. 0.1 × 3.375 D = 0.3375 Δ, which is the published ±0.33 Δ. The two branches meet essentially exactly — the crossover power is not arbitrary, it is the power at which 1.0 mm and 0.33 Δ describe the same displacement. (One source prints the threshold as ±3.37 D and the other as ±3.375 D; the arithmetic shows 3.375 is the exact value and 3.37 a truncation of it.)
• Horizontal. 2.5 mm is 0.25 cm. 0.25 × 2.75 D = 0.6875 Δ, against a published limit of ±0.67 Δ. These do not reconcile exactly; the power at which they would is about 2.68 D. The practical consequence is small but real — right at the ±2.75 D threshold the millimetre branch is marginally more generous than the prism branch it replaces.

The reason to do this arithmetic is not pedantry. It is that a specification you can check is a specification you can defend, and running the conversion catches transcription errors for free.

The number that matters most: progressives get one millimetre

Notice what the table does for progressive lenses. There is no power banding. The horizontal allowance is ±0.67 Δ and ±1.0 mm total, at every power, and the vertical carries a ±1.0 mm difference at every power. A second source reports that the international standard, ISO 21987, takes the same view from the other side, asking "for the fitting cross to be within 1 mm of the specified monocular PD regardless of power" on progressives.

Two standards, arrived at independently, landing on the same millimetre. That is the number to hold in your head, because it reframes everything above: the two-millimetre difference between measurement conventions is twice the entire centration allowance a progressive is given. On a progressive, the choice of method is not a refinement. It is the budget.

What the published comparisons actually found

Two peer-reviewed studies have put a millimetre rule and a digital pupillometer on the same wearers. They are worth reading together, because their headline numbers disagree by roughly a factor of two — and the disagreement is more useful than either number alone.

Why the two disagree — and why that is the finding

The temptation is to pick the better study and quote one number. That throws away the useful part. The two studies used different ruler techniques, and Jung and Chu name the distinction explicitly: their ruler was used "without penlight torch," so it read the pupil centre. A ruler technique that sights a penlight reflex and a ruler technique that sights the pupil centre are separated by exactly the systematic offset Santini quantifies at roughly 2 mm binocularly — which is the right order of magnitude to account for a mean difference of 1.25 mm in one study and 0.54 mm in the other.

So the honest reading is not "rulers are off by about a millimetre." It is: a ruler and a pupillometer differ by an operator-noise component plus a convention component, and the convention component depends on a technique detail that is rarely written down anywhere. Jung and Chu also adopted the pupillometer as their reference instrument on the strength of its repeatability — r = 0.98 against r = 0.96 for both the ruler and a smartphone app — and cautioned that apps "warrant caution for complex prescriptions" even while working as screening tools.

"Not clinically significant" is not the question a lab asks

Gantz and colleagues concluded that binocular and monocular manual and automatic measurements "were significantly different statistically, but not clinically." That conclusion is sound, and it is answering a question about vision care.

It is not answering the question a dispensary has. The same study's 95% limits of agreement on distance binocular PD run from −1.25 mm to +3.76 mm — a spread of five millimetres. The upper limit alone, +3.76 mm, is nearly four times the entire ±1.0 mm centration allowance a progressive lens is given. One dataset, two questions, opposite answers, and the difference between them is simply whether you are asking about the eye or about the finished lens.

This is worth being blunt about because it is where the category's marketing usually goes vague. A pupillometer does not make anyone see better. It narrows the distribution of a number that a manufacturing standard polices to one millimetre.

Monocular PD, and why it stopped being optional

A single binocular PD assumes the face is symmetric about the bridge. Faces are not, and a progressive corridor placed on the assumption that they are starts every wearer off-axis on one side.

Santini's recommendation is to take "the PD monocularly, using the occlusion bar and noting any measurement differences indicated," specifically because a wearer's phorias affect what a binocular measurement records. The occlusion bar matters mechanically: it lets each eye be measured while the other is dissociated, so a latent deviation does not quietly bias the reading.

Two practical consequences for a purchase decision:

• Confirm the instrument reports right and left separately, not just a sum. A unit that only outputs a binocular figure cannot support a progressive order that is specified per eye.
• Confirm it has an occlusion or dissociation function. Monocular numbers derived by halving a binocular reading are not monocular measurements, and they carry none of the asymmetry information that made you want them.

By ruler, the same measurement is harder still: Santini notes monocular measurements taken with a rule are "far more influenced by both operator and patient parallax."

Near PD and the convergence problem

Near PD is not distance PD scaled down; it depends on the wearer actually converging to a working distance, and that is a behavioural requirement rather than an optical one. Santini identifies "unwanted convergence" as "the most insidious contributor" to error, "as it is almost impossible to get an untrained consumer to provide a distance fixation while looking at a proximal mirror or camera."

The measured data reflect this. In Jung and Chu the near ruler-versus-pupillometer difference (1.08 ± 0.99 mm, limits to +3.03 mm) is twice the distance difference; in Gantz the near difference is smaller than the distance one. The two studies do not agree on the direction of the effect, which is itself informative — near PD variability is dominated by how the fixation target was managed on the day, not by the instrument class.

What to look for: a selectable working distance rather than a single fixed near setting, and a fixation target the wearer can genuinely hold at infinity for the distance reading. If your dispensary does much near-specific or occupational work, this is the feature that earns its keep.

Vertex distance: the second measurement some units make

Above moderate powers, where the lens sits relative to the eye changes its effective power. Writing in 20/20 Magazine in December 2025, Sam Winnegrad notes that "once prescriptions go beyond ±4.00 diopters, sensitivity to this distance increases dramatically," and gives the compensation formula as "Fe = F / (1 - dF), where F is the lens' nominal power, and d is the change in vertex distance in meters."

That formula can be checked against the article's own worked example. A −10.00 D lens moved from 12 mm to 8 mm is a 4 mm change, so d = 0.004 m:

Fe = −10.00 ÷ (1 − 0.004 × −10.00) = −10.00 ÷ 1.04 = −9.615 D

The article reports "approximately -9.62 diopters" for that move. The formula and the example reconcile, which is the useful property: this is a number your staff can reproduce rather than one they have to accept.

Relevance to the purchase: some PD meters read vertex distance as well as PD, and some do not. Of the units US Ophthalmic carries, the product page for the Ezer EPD-2600 states that it "features everything your practice needs—including vertex distance measurement capabilities," while the EPD-1800 page describes PD measurement and does not mention vertex. If you dispense many high powers, that difference is the one that separates the two Ezer units in practice.

What to check before you sign a quote

Because this is a low-drama instrument, it tends to get bought on a one-line description. These are the questions that actually change the outcome:

1. Which landmark does it sight — corneal reflex or pupil centre? Everything downstream, including whether your new readings are comparable with your old ones, follows from this.
2. Does it output monocular right and left values independently? Required for progressives specified per eye.
3. Is there a real occlusion or dissociation function? Not a halved binocular number.
4. What working distances are selectable for near, and is the distance target genuinely at infinity?
5. Does it read vertex distance? Only matters above roughly ±4.00 D, but above that it matters a lot.
6. How is it powered, and how is it cleaned between wearers? A forehead-and-nose-contact instrument gets handled all day.
7. What is the calibration and verification path, and who performs it? An instrument whose readings you cannot verify is an instrument you cannot defend when a lab queries an order.
8. Who services it, and where? US Ophthalmic keeps an in-house technical and spare-parts department, so support on a carried unit stays domestic rather than routing overseas.

Question 1 also implies a housekeeping job that costs nothing: write the convention on the record. If your practice standardises on corneal-reflex monocular PDs and says so on the order, a lab query five weeks later takes one minute instead of a remake. Mixing conventions across staff, across a ruler and an instrument, or across an old file and a new measurement, is worth up to about 2 mm systematically — against a progressive budget of 1 mm.

The pupillometers US Ophthalmic carries

The pupillometers collection holds four units, all currently available. It is worth naming a distinction the collection title flattens: three are digital instruments, and one is a different class of tool entirely.

The PM-120 is the odd one out and deliberately so: a knob-adjusted PD-and-pupil-height tool serves the fitting and marking side of the bench rather than replacing a digital reading. If you are shortlisting on the strength of the collection title alone, that is a distinction worth making before the quote, not after delivery.

One thing we will not do is invent numbers. We read all four product pages while writing this guide and none of them publishes a specification table — no measurement range, no resolution, no power source, no dimensions. Any figure you see quoted for these units elsewhere did not come from our catalogue. If you need the measurement range, the near-distance settings or the resolution in writing before you commit, ask us and we will send the manufacturer figures rather than a paraphrase. That is also a fair question to put to any supplier in this category.

Where a PD meter sits next to your lensmeter

The two instruments look adjacent and are not interchangeable. A PD meter measures a wearer before the job exists. A lensmeter measures a lens — including, on units with the feature, the distance between the optical centres of a finished or marked pair, which is how you confirm the lab honoured the number you sent.

They are the two ends of the same loop, and owning both is what lets you tell a measurement problem from a manufacturing one. Our write-up of the Ezer ELM-BH-PD auto lensmeter covers the verification end in detail, including what the PD function on a lensmeter does and does not tell you, and the ELM-BH-PD product page lists the model itself. For a wider view of the category there is our comparison of auto lensmeters and digital lensometers for 2026. If you are equipping a lane and a dispensary from scratch, the new practice equipment checklist places both instruments in the wider build.

Frequently asked questions

What is a PD meter, and is it the same thing as a pupillometer?

In optical dispensing they refer to the same instrument: a device that measures interpupillary distance so lenses can be centred correctly in a frame. Note that the word "pupillometer" is also used for a different class of instrument used in neurological monitoring, which is unrelated to spectacle dispensing. If you are searching for equipment, "PD meter" is also used for positive-displacement flow meters in industry — worth knowing when you are comparing supplier listings.

How much more accurate is a digital PD meter than a millimetre rule?

Two published comparisons put the difference between a ruler and a pupillometer at 0.54 ± 0.74 mm (Jung & Chu, 2024) and 1.25 ± 1.27 mm (Gantz and colleagues, 2021) for distance binocular PD. The pupillometer showed the higher test–retest correlation in Jung and Chu, r = 0.98 against r = 0.96. Part of the difference between the two studies is operator variability and part is a measurement convention — see the section above on the corneal reflex.

Why does a pupillometer give a narrower PD than a rule?

Because it sights the corneal light reflex rather than the geometric centre of the pupil. The reflex sits about 1 mm nasal to pupil centre in each eye, so a binocular reading taken from the reflex runs roughly 2 mm narrower than one taken from pupil centres. Neither landmark is wrong; what matters is that you do not mix them within a single job or across a wearer's file.

How accurate does a PD have to be?

It depends on the lens. For single vision and multifocals the published ANSI Z80.1 horizontal tolerance is ±0.67 prism dioptres up to ±2.75 D, switching to ±2.5 mm from the specified PD above that. For progressives there is no power banding: ±0.67 Δ and ±1.0 mm total at every power, and the international standard ISO 21987 similarly asks for the fitting cross within 1 mm of the specified monocular PD. In short, progressives allow about one millimetre.

Do I need monocular PD, or is a single binocular number enough?

For progressives and for any asymmetric face, monocular values are what the order needs. Halving a binocular reading is not a monocular measurement — it reproduces the symmetry assumption you were trying to avoid. Look for an instrument that reports right and left independently and has a genuine occlusion function.

Are smartphone PD apps good enough for dispensing?

In Jung and Chu's 2024 comparison a smartphone app differed from the pupillometer by −0.59 ± 1.17 mm at distance, with limits of agreement from −2.89 mm to +1.70 mm, and matched the ruler's test–retest correlation of r = 0.96. The authors' own conclusion was that apps work as screening tools but warrant caution for complex prescriptions. Set that spread against a progressive's one-millimetre centration allowance and the answer for a dispensary is straightforward.

Does a PD meter measure vertex distance too?

Some do. Of the units we carry, the EPD-2600 product page states vertex distance measurement capability; the EPD-1800 page does not mention it. Vertex matters once powers go beyond about ±4.00 D, where the effective power of the lens shifts measurably with how far it sits from the eye.

How do I get pricing on a pupillometer?

Equipment in this category is quoted rather than listed. Tell us how your dispensary works — whether you fit many progressives, whether you dispense high powers, whether you need near working distances — and our specialists will come back with the units that fit, plus availability and lead time. Request a quote and we will match the instrument to the bench. US Ophthalmic is a direct provider to eye care practices nationwide, backed by an in-house technical and spare-parts department and full US warranty and service — reachable through customer service for calibration, verification and parts questions after the sale.

Sources

• Santini B. "Meet Your Pupilometer." 20/20 Magazine, January 2021.
• Winnegrad S. "Vertex Distance: Understanding its Role in Accurate Vision Correction." 20/20 Magazine, December 2025.
• "Understanding Prism Part 2: Verifying Prescribed and Unwanted Prism." Eyecare Business, December 2012 (Prentice's rule; ANSI Z80.1-2010 tolerance table).
• Gantz L, Shneor E, Doron R. "Agreement and inter-session repeatability of manual and automatic interpupillary distance measurements." Journal of Optometry. 2021;14(4):299–314.
• Jung YR, Chu BS. "A Comparative Analysis of Interpupillary Distance Measurement Techniques Evaluation in Modern Times: From Rulers to Apps." Clinical Optometry. 2024;16:309–316.
• Secondary corroboration of the ANSI Z80.1 tolerance figures and the ISO 21987 progressive fitting-cross requirement, reporting the figures as consistent across the 2010, 2015 and 2020 editions.
• US Ophthalmic product pages for the EPD-2600, EPD-1800, PM-100 and PM-120, and the pupillometers collection listing, all read on 26 August 2026.

The tolerance figures above are reported as published by the cited trade sources, which cite ANSI Z80.1-2010 and note consistency through the 2020 edition. They are given here to explain why centration precision matters, not as a compliance reference; verify against the current edition of the standard before relying on them in a specification or a dispute.