Multi-Lens System Calculator: Intermediate Images and Equivalent Focal Length
Chains the thin lens equation across two or three elements so you don't have to. Tracks intermediate image positions, signs of virtual objects, total magnification, and equivalent focal length for compound microscopes, refracting telescopes, and telephoto pairs.
Prerequisite: the single-lens thin lens equation
This page assumes you already know the single-element thin lens equation. If you don't, start with the Thin Lens Equation Calculator, which covers 1/f = 1/d_o + 1/d_i, the sign conventions for a single element, and how to read magnification from one lens or mirror. Everything below treats those fundamentals as given and builds on them to handle two-element and three-element systems.
The hard part of a multi-lens system isn't the equation. It's the bookkeeping. You apply the same thin lens relation to each element in turn, and the only thing that changes is what counts as the "object" for the next stage. Image 1 becomes the object for element 2, and the object distance for element 2 is measured from element 2's plane, not from element 1. Skip that translation and every downstream sign goes wrong. Magnification accumulates as a simple product, M_total = m_1 · m_2 · m_3 · …, with each negative sign flipping orientation once.
Wave Geometry and Phase: The Quantities That Define a Wave
Geometric optics models light as rays that travel in straight lines and bend at interfaces; wavelength doesn't appear anywhere in the thin-lens equation. The wavelength enters only when you ask about diffraction-limited spot size, chromatic effects (n depends on λ), or interference between rays that take different paths through the system. Multi-element setups are still ray-traced, but the wavelength-dependent corrections show up in two places: chromatic focal-length shift between blue and red, and the diffraction-limited point-spread function of the whole instrument.
Plane-wave description of light:
E(x, t) = E_0 cos(kx − ωt + φ)
k = 2π/λ, ω = 2πf, v = c/n in a medium of refractive index n. For visible light, λ runs from 380 nm (violet) to 750 nm (red). Each element along the optical axis is a thin slab that delays the central ray more than the edges (or the reverse for a diverging element), bending the wavefront on the way through.
The three principal rays you draw at each element are shorthand for three special wavefronts: a parallel ray represents a plane wavefront from infinity; a focal ray represents a wavefront that's diverging from the front focal point; a central ray represents an undeviated wavefront passing through the geometric centre. Phase only matters explicitly when the ray approximation breaks. The diffraction limit on the spot a multi-element system can form is roughly 1.22 λ f_eq / D, where D is the entrance-pupil diameter. For a 200 mm telephoto at f/4 imaging green light (λ ≈ 550 nm, D = 50 mm), that's about 2.7 μm. That sets the floor; aberrations and atmospheric seeing add to it.
In a multi-element system, the wavefront undergoes a sequence of curvatures. A converging element bends the wavefront to converge faster; a diverging element flattens it. The thin-lens calculation tracks the conjugate-distance bookkeeping of those curvatures element by element. The wave description and the ray description are equivalent in the paraxial limit, and the wave picture is what tells you when the ray picture breaks (very fast lenses, very small apertures, very wide fields).
Sign Conventions in Optics: What's Real and What's Virtual
The single-element conventions are covered in the thin lens primer. What multi-element work adds is a reliable rule for the object distance of element k+1, plus a sharper distinction between real and virtual objects than you usually need with a single element. For each element along the axis, you do three things. Solve the thin lens (or mirror) equation for the image distance s_i. Read off the local magnification m = -s_i / s_o. Then translate: the object distance for the next element is the separation minus the previous image distance, s_{o,2} = d - s_{i,1}. That last step is where most errors enter. It's the only non-obvious move in the whole procedure.
| Quantity | Positive | Negative |
|---|---|---|
| Object distance for element k+1, s_{o,k+1} = d - s_{i,k} | Real object: intermediate image lies between elements | Virtual object: rays still converging at the next element |
| Local magnification m_k | Upright wrt the local object for that element | Inverted wrt the local object for that element |
| Total magnification M_total = ∏ m_k | Even number of inverting stages | Odd number of inverting stages |
| Mixing lenses and mirrors in one system | Light propagates left-to-right after a mirror reflection (you've unfolded the path) | Folded path: extra care needed; convention varies by textbook |
When the bracketed quantity comes out negative, you have a virtual object. The light arriving at the second element is already converging toward a point past the element, so the "object" sits behind it. That's the case in essentially every telephoto design (the rear group acts on light that's already heading toward focus), and it's also what happens in beam expanders run backwards. Negative s_{o,2} isn't a glitch. It's the calculator catching a real geometric situation that the formula is happy to handle as long as you keep the signs.
The mirror-and-lens combination (Newtonian-style optics, or a Schmidt-Cassegrain) needs one extra discipline. Pick a sign convention before you start, write it on the page, and don't switch halfway through. The most common convention for mixed systems "unfolds" the path so that after each mirror, distances are measured along the new propagation direction with the same positive-toward-image rule. Mixing the lens convention (positive-for-real-image-on-the-far-side) with the unfolded mirror convention (positive-for-real-image-in-front-of-the-mirror) silently flips a sign and you won't catch it until the magnitudes disagree with the sketch. Drawing principal rays for each element in turn is slow but it catches gross sign mistakes that no amount of formula manipulation will. If your sketched intermediate image lands between the lenses, your s_{o,2} should come out positive. If your sketch shows the rays still converging when they hit lens 2, your s_{o,2} should be negative. Trust the sketch and look for the sign error.
Working through the intermediate image step by step
Concrete numbers help. Take a two-lens setup with f_1 = 10 cm, s_{o,1} = 20 cm, lens separation d = 15 cm, and f_2 = 5 cm. Here's the sequence the calculator runs internally; doing it by hand once is the fastest way to see why the "virtual object" case isn't a special rule but the same rule with a sign you accepted.
Step 1. Solve element 1.
1/s_{i,1} = 1/f_1 - 1/s_{o,1} = 1/10 - 1/20 = 1/20, so s_{i,1} = 20 cm. Local magnification m_1 = -s_{i,1} / s_{o,1} = -1. The intermediate image is real, inverted, same size as the object, and forms 20 cm to the right of lens 1.
Step 2. Translate to element 2.
s_{o,2} = d - s_{i,1} = 15 - 20 = -5 cm. Negative, so element 2 sees a virtual object 5 cm behind it. Geometrically, lens 1 wanted to form an image 20 cm to its right, but lens 2 is only 15 cm away, so the converging light hits lens 2 before it has a chance to come to a point.
Step 3. Solve element 2.
1/s_{i,2} = 1/f_2 - 1/s_{o,2} = 1/5 - 1/(-5) = 0.2 + 0.2 = 0.4, so s_{i,2} = 2.5 cm. Real image, 2.5 cm to the right of lens 2. m_2 = -s_{i,2} / s_{o,2} = -2.5 / (-5) = 0.5.
Step 4. Multiply.
M_total = m_1 · m_2 = (-1)(0.5) = -0.5. Final image is half the size of the original object and inverted with respect to it. One inversion came from lens 1; lens 2 didn't add a second because the virtual object kept the orientation.
Sound, Light, Mechanical Waves: Where the Equations Differ
All three obey v = fλ, but the lensing equations on this page apply only to light. Sound bends at boundaries between fluids of different speeds (Snell's law in the form sin θ_1 / v_1 = sin θ_2 / v_2), and you can build acoustic lenses (a CO₂-filled balloon focuses sound in air). The geometry mirrors optics, but the analogue of refractive index is c_ref / v_local, which inverts the role water plays for sound versus light. Mechanical waves on a string have no lensing analogue because the wave is one-dimensional; the analogue is impedance matching at junctions.
Where multi-element optics differs sharpest from acoustic or mechanical-wave problems is in dispersion. n(λ) for glass varies smoothly across the visible spectrum, and the variation is what creates chromatic aberration. A single converging element focuses blue closer than red, smearing white-light images. Achromatic doublets fix this by pairing a converging crown-glass element with a diverging flint-glass element, chosen so two wavelengths (typically the F and C lines, blue and red) come to a common focus while the net positive power survives. Apochromats stretch the correction across three wavelengths. None of this matters for a non-dispersive medium like air at audio frequencies, where a complex sound waveform travels intact.
Lensing analogues across domains
- Optical lens: n > 1 in glass, curved surfaces. Lensmaker's equation: 1/f = (n − 1)(1/R_1 − 1/R_2). Multi-element: ray-trace each surface, or use ABCD matrices.
- Acoustic lens: CO₂-filled balloon focuses sound (lower v, acts like a converging lens). Used for ultrasound focusing in medical imaging.
- Gravitational lens: mass curves spacetime, deflecting light by 4GM/(rc²). Galaxy clusters magnify background galaxies; Einstein rings form when the alignment is exact.
- Electron lens: magnetic fields focus electron beams in TEMs (transmission electron microscopes). Aperture, focal length, magnification all have direct optical analogues, with field strength replacing surface curvature.
Pierce's "Acoustics" covers the acoustic-lens analogue carefully and connects it to the more general impedance-mismatch picture that handles partial reflection at every boundary. The rule of thumb: where dispersion is small (sound in air, light in vacuum), the simple ray picture works to high accuracy; where dispersion is meaningful (light in glass, deep-water gravity waves, plasma waves), you have to track wavelength explicitly through the chain.
Multi-Element Setups (or Multi-Source Interference)
Sequential ray tracing handles any number of elements, with the bookkeeping local: element k depends on element k-1 only through the single number s_{i,k-1}. There's no global coupling, which is why the ray method scales easily from two to three to ten elements without fundamentally changing the algorithm. Three-element systems work the same way as two: solve element 1, translate, solve element 2, translate, solve element 3. The final image position can be reported relative to the last element (useful for placing a sensor) or relative to the first element (useful for sketching the layout to scale).
Equivalent focal length: two thin lenses in contact
When two thin lenses sit so close together that you can take d ≈ 0 (a cemented doublet, or two pieces of plastic stacked against each other), the system behaves as a single thin lens with focal length given by the simplest possible rule:
1/f_eq = 1/f_1 + 1/f_2
Powers add. That's easier to remember if you switch to diopters (P = 1/f with f in meters), since then it's literally P_eq = P_1 + P_2. Eyeglass prescriptions use this: a +2.0 D reading add over a -3.0 D distance correction gives a -1.0 D net power for near vision through the bifocal segment. Two converging lenses of f_1 = 10 cm and f_2 = 20 cm in contact behave as one converging lens of f_eq = 1 / (1/10 + 1/20) ≈ 6.67 cm. A converging f = 10 cm cemented with a diverging f = -25 cm gives f_eq = 1 / (1/10 - 1/25) = 1 / 0.06 ≈ 16.7 cm: still converging, but weaker than either component on its own. This is exactly how achromatic doublets work, with the two glass types chosen so chromatic dispersion cancels but power doesn't.
Equivalent focal length: lenses separated by distance d
Once the elements are separated, the "in contact" rule breaks. The correct expression for the equivalent focal length of two thin lenses separated by air gap d is:
1/f_eq = 1/f_1 + 1/f_2 − d / (f_1 · f_2)
The third term is what makes camera-lens design interesting. For two converging lenses, increasing d weakens the system (f_eq gets longer). For one converging and one diverging element with d chosen carefully, you can engineer a long effective focal length out of a short physical assembly, and that's the whole idea behind a telephoto. When d goes to zero, the third term vanishes and you recover the in-contact formula above. When d equals f_1 + f_2, the system is afocal: parallel rays in, parallel rays out, no finite focal point. That's the telescope condition.
Two warnings on this formula. First, f_eq alone doesn't tell you where the image forms. The system has principal planes that are shifted from the physical lens positions, and the image distance has to be measured from the correct principal plane, not from the rear lens vertex. The calculator on this page does the right thing because it ray-traces through both elements explicitly. Hand calculations using f_eq alone often miss this and place the image at the wrong location. Second, when 1/f_eq evaluates to zero (afocal system), f_eq is undefined and so is the "image distance for an object at infinity." That isn't a numerical bug. It's telling you the system images parallel to parallel and you should be asking for angular magnification (telescope magnification) instead of linear magnification.
Magnification cascade and three-element systems
Each element contributes a multiplicative factor m_k = -s_i,k / s_o,k, and the total is the product M_total = m_1 · m_2 · m_3 · …. Sign tracking is the whole game. A negative factor flips orientation once. Two negatives flip twice and put the image back upright. A microscope objective at 40× and an eyepiece at 10× would individually be inverting, so the product is a positive 400× (upright relative to the original specimen, which is why microscope vendors quote unsigned magnifications and call it a day). A simple refracting telescope, where both elements are converging in series, ends up inverting once overall because the eyepiece works in a different conjugate regime. That's why astronomical refractors show inverted images and terrestrial spotting scopes add an erecting prism.
The two-element separation formula 1/f_eq = 1/f_1 + 1/f_2 − d/(f_1·f_2) doesn't generalise cleanly to three or more elements. You can derive an analogue, but it's long, and once you're past two elements there's essentially no benefit over just ray tracing each stage. Sequential application of the thin lens equation handles any number of elements in the same time it takes to handle two, because the bookkeeping is local. A worked three-element pass: f_1 = +50 mm, d_12 = 30 mm, f_2 = -20 mm, d_23 = 40 mm, f_3 = +30 mm, with an object 75 mm in front of element 1. Solve element 1: 1/s_{i,1} = 1/50 - 1/75 = 1/150, so s_{i,1} = 150 mm, m_1 = -2. Translate: s_{o,2} = 30 - 150 = -120 mm (virtual object). Solve element 2: 1/s_{i,2} = 1/(-20) - 1/(-120) = -1/24, so s_{i,2} = -24 mm, m_2 = -24/(-120) = 0.2. Translate: s_{o,3} = 40 - (-24) = 64 mm. Solve element 3: 1/s_{i,3} = 1/30 - 1/64 ≈ 1/56.5, so s_{i,3} ≈ 56.5 mm. m_3 ≈ -56.5/64 ≈ -0.88. Total: M_total ≈ (-2)(0.2)(-0.88) ≈ 0.35. Inverted twice, ends up upright, about a third of original size.
Once you're routinely doing three or four elements, switch to ABCD matrix optics. Each element gets a 2×2 matrix, each air gap gets a 2×2 matrix, you multiply them in order, and the resulting system matrix tells you f_eq, principal planes, and image positions in one pass. Hecht chapter 6 and Smith chapter 2 both walk through this. The thin-lens chain is fine for two elements and tolerable for three; past that, matrices are easier and less error-prone.
Interference, Beats, and Phase-Shift Diagnostics
Geometric ray tracing predicts an infinitely sharp point image, which can't be right because diffraction sets a finite spot size. The Airy disk has a radius of about 1.22 λ f_eq / D where D is the entrance-pupil diameter of the whole system. For a 200 mm telephoto at f/4 imaging green light (λ ≈ 550 nm, D = 50 mm), the Airy radius is about 2.7 μm. That's the diffraction-limited resolution; aberrations add to it, and atmospheric seeing adds more for ground-based astronomy. The Hubble Space Telescope hits its diffraction limit because there's no atmosphere; ground-based telescopes don't, except with adaptive optics that flatten the wavefront in real time.
Multi-element systems suffer chromatic aberration when the elements aren't corrected. Each element has a wavelength-dependent focal length f(λ) because n(λ) is dispersive. Sequential ray tracing at one wavelength is what this calculator does; at a second wavelength the focal lengths shift slightly and the final image position shifts with them. The longitudinal chromatic shift Δs_i ≈ −s_i² × Δ(1/f) gives a quick estimate. For a single converging element with f = 100 mm and a glass dispersion such that f shifts by 0.5% between blue and red, Δs_i is a few hundred μm at standard conjugates, which is enormous compared to a sensor pixel. That's why every camera lens past a kit zoom is a multi-element design with at least one achromatic pair.
Phase-shift diagnostics in multi-element systems
- Anti-reflection coatings on each surface use destructive interference. A λ/4 layer of MgF₂ (n ≈ 1.38) cancels reflection at one wavelength; multi-layer broadband coatings stretch the cancellation across the visible. A 6-element lens has 12 surfaces, each losing about 0.5% with coating versus 4% without. Without coatings, transmission would be 0.96^12 ≈ 61%; with single-layer coatings, 0.995^12 ≈ 94%.
- Newton's rings show up between element surfaces during alignment. Optically contacted surfaces (no air gap) show no fringes; air-spaced doublets show a few rings whose count tells you the gap thickness to within λ/4.
- Airy-pattern asymmetry in the final image diagnoses miscentred elements. A perfectly centred system produces a circularly symmetric Airy disk; a decentred element introduces coma, which shows up as a comet-shaped spread on one side of the image. That's how astronomical optics get aligned to the arc-second.
When the calculator returns a magnification that disagrees with what you sketched, the issue is almost always one of the intermediate magnifications, not the formula. Check m_1 against its principal-ray sketch. If lens 1 produced a real, inverted intermediate image, m_1 has to be negative. If you wrote it as positive in your scratch work, every later sign will be wrong.
Where the paraxial model stops being honest
Every formula in this section assumes paraxial rays: light close to the axis at small angles where sin θ ≈ θ. Push past that regime (fast lenses, wide field of view, or heavy off-axis points) and aberrations show up. They aren't errors in the formula; they're higher-order terms the paraxial expansion threw away.
- Spherical aberration. Marginal rays focus closer to the lens than paraxial rays. Worst at large aperture, zero at small aperture. The reason your f/2 prime gets sharper when you stop down to f/8.
- Coma. Off-axis points form comet-shaped blobs instead of round spots. Distinctively asymmetric, and hard to confuse with anything else once you've seen it.
- Chromatic aberration. Refractive index depends on wavelength, so different colors focus at different distances. Cured (in part) by achromatic doublets that combine crown and flint glass.
- Astigmatism, field curvature, distortion. Tangential and sagittal rays from off-axis points don't focus at the same place; the image surface bends; straight lines bow. Each gets corrected with extra elements, often aspheric.
For homework, lecture demos, and first-cut conceptual designs, the paraxial model is honest and adequate. For real lens design you switch to ray-tracing software (Zemax OpticStudio, Code V, or the open-source RayOptics Python library) that traces thousands of skew rays through real surfaces and computes spot sizes, MTF, and chromatic focal shifts. The calculator on this page is paraxial only, and that's a feature: it gives you the right answer for the regime where the formula is exact, instead of a wrong answer dressed up as a precise one.
Worked Example: Microscope, Telescope, Telephoto
Compound microscope (objective + eyepiece)
Configuration: f_obj = 4 mm, f_eye = 25 mm, tube length L = 160 mm, specimen placed 4.1 mm from the objective.
Step 1. Specimen at s_{o,1} = 4.1 mm, just past the objective focal point. Solve: 1/s_{i,1} = 1/4 - 1/4.1, giving s_{i,1} ≈ 164 mm. The objective forms a strongly magnified, real, inverted intermediate image roughly at the standard tube length, near the field stop of the eyepiece.
Step 2. m_1 = -s_{i,1}/s_{o,1} ≈ -164/4.1 ≈ -40. That's the "40×" on the objective barrel, give or take. Sign is negative: real inverted image.
Step 3. The intermediate image sits just inside the front focal point of the eyepiece, so the eyepiece works as a simple magnifier. Object distance s_{o,2} ≈ 25 mm (chosen so the eyepiece sends rays out parallel for the relaxed eye). Image goes to infinity: s_{i,2} → ∞, and the eyepiece's contribution is best expressed as an angular magnification m_eye ≈ 25 cm / f_eye = 250/25 = 10×.
Result. M_total ≈ 40 × 10 = 400×. The two-stage architecture is what lets a microscope reach 400× without needing a single absurdly short focal length lens. Any one element with f ≈ 0.6 mm would have impossible tolerances; splitting the magnification across two stages relaxes them to manageable values.
One subtlety. Microscope vendors quote "total magnification" as the product of objective and eyepiece factors, treating both as positive numbers, and that's the convention you'll see on a lab slip. The calculator on this page tracks signs explicitly, so a real-conjugate microscope setup may show a negative M_total. Same physics, different sign convention. Both are correct; just be aware of which one you're reading.
Brief: refracting telescope (long objective + short eyepiece)
Configuration: f_obj = 900 mm, f_eye = 20 mm, separation d = f_obj + f_eye = 920 mm, distant astronomical object.
Step 1. Object at infinity. The objective forms its image at its back focal point: s_{i,1} = f_obj = 900 mm. That intermediate image is real, inverted, and very small (because the object is essentially infinitely far and only the angular size matters).
Step 2. s_{o,2} = d - s_{i,1} = 920 - 900 = 20 mm = f_eye. The intermediate image sits exactly at the front focal point of the eyepiece. That's the afocal condition, and it's on purpose: it's how you build a telescope.
Step 3. With the object at the eyepiece's focal point, the equation gives s_{i,2} → ∞. Linear magnification is undefined. The relevant quantity is the angular magnification:
M_angular = -f_obj / f_eye = -900 / 20 = -45×.
Result. The telescope magnifies angles by 45×. The negative sign means the image is inverted (and stays inverted; refractor users get used to this, or add a star-diagonal prism that introduces an extra reflection). The afocal system has 1/f_eq = 1/900 + 1/20 − 920/(900·20) = 0 to working precision, so f_eq goes to infinity. The right output is angular magnification, and M_angular = -f_obj / f_eye is what every astronomy text uses.
Brief: telephoto pair (positive + negative thin lenses)
Configuration: f_1 = +100 mm (front group, converging), f_2 = -25 mm (rear group, diverging), separation d = 80 mm, object at infinity.
Step 1. Parallel rays from infinity. Front group focuses them at s_{i,1} = f_1 = 100 mm behind the front element.
Step 2. s_{o,2} = d - s_{i,1} = 80 - 100 = -20 mm. The rear group sees a virtual object 20 mm behind it: light converging toward a point that's past the second element when it arrives.
Step 3. Apply the equation: 1/s_{i,2} = 1/f_2 - 1/s_{o,2} = 1/(-25) - 1/(-20) = -0.04 + 0.05 = 0.01, so s_{i,2} = 100 mm. Real image, 100 mm behind the rear element.
Effective focal length. Apply the separation formula: 1/f_eq = 1/100 + 1/(-25) − 80/(100·(-25)) = 0.01 − 0.04 + 0.032 = 0.002, so f_eq = 500 mm. The system images like a 500 mm lens. Physical length from front element to final image is 80 + 100 = 180 mm. The system is roughly 500/180 ≈ 2.8 times shorter than a single lens of the same focal length would be.
That ratio (effective focal length divided by physical length) is called the telephoto ratio, and it's the whole reason camera telephotos use a positive-then-negative pair instead of a single long lens. A 500 mm prime lens that's actually 500 mm long is unwieldy. A 500 mm telephoto that's 180 mm long fits in a backpack. The price you pay: aberrations are harder to control because the rear element is working with already-converging light, and the entrance pupil is smaller than the front element for a given f-number, which limits how fast the lens can be made. Real telephoto designs add more elements to fix this; the two-element pair is the conceptual core.
References
- Hecht, E. (2017). Optics (5th ed.). Pearson. Chapter 5 walks through multi-element systems and matrix optics. Chapter 6 is the aberration material if you want to see where the paraxial model breaks.
- Born, M. & Wolf, E. (1999). Principles of Optics (7th ed.). Cambridge University Press. The authoritative reference for diffraction theory, aberration analysis, and the link between geometric and wave optics. Heavier reading; the treatment is full-strength.
- Smith, W. J. (2007). Modern Optical Engineering (4th ed.). McGraw-Hill. Practical lens design with worked compound microscope, telescope, and camera-lens calculations, including ABCD matrix techniques.
- Pedrotti, Pedrotti, & Pedrotti (2017). Introduction to Optics (3rd ed.). Cambridge. Cleanest treatment of paraxial sign conventions in the undergraduate texts; the multi-element worked examples are close to what this calculator does internally.
- Kingslake & Johnson (2009). Lens Design Fundamentals (2nd ed.). Academic Press. Older book, still the best on the geometric reasoning behind telephoto and retrofocus designs.
What this calculator doesn't model
- •Thin-element idealization. Each element is treated as zero thickness. Real lenses have principal planes separated from the physical surfaces, which shifts effective positions. For thick or compound elements use ABCD matrix methods.
- •Paraxial only. No spherical aberration, coma, astigmatism, or distortion. Wide-aperture or off-axis rays don't obey 1/f = 1/s_o + 1/s_i exactly, and the error grows with the cube of the angle.
- •Single wavelength. No chromatic effects. Real glass disperses, so the focal length you compute here is the focal length at one (unspecified) wavelength.
- •Perfect alignment. No tilt, decenter, spacing tolerance, or surface figure error. Real systems require tolerance analysis before they're built.
The right use case is homework, concept exploration, and getting a first-order layout sketched out before you commit to detailed ray tracing. For an actual instrument, hand the layout to optical-design software and check tolerances against manufacturing capability.
Debugging Multi-Element Optics Calculations
Real questions from students stuck on sequential ray tracing, sign convention errors, and interpreting virtual vs real images.
How do you solve a multi-lens optical system?
Solve a multi-lens system one element at a time. Apply the thin lens equation 1/f = 1/u + 1/v to the first lens, find the image. Then treat that image as the object for the second lens. The catch is the object distance for lens 2. Measure it from lens 2, not from lens 1. If the lenses are separated by distance d and the first image forms at v₁ on the far side of lens 1, then u₂ = d − v₁. If v₁ > d, the first image lies past lens 2. That gives a virtual object: u₂ comes out negative, and you keep the sign in the second lens equation. Total magnification multiplies: m_total = m₁ · m₂ = (−v₁/u₁) · (−v₂/u₂). Three lenses, multiply three. Inversions cancel in pairs. Worked example: lens 1 (f = 10 cm) with an object at u₁ = 15 cm gives v₁ = 30 cm. If lens 2 (f = 20 cm) sits 50 cm from lens 1, then u₂ = 50 − 30 = 20 cm. That's at the focal point of lens 2, so the final image forms at infinity (1/v₂ = 0). A telescope works exactly this way: the objective forms its image at the focal point of the eyepiece, producing parallel rays for the relaxed eye. This sequential approach handles thin lens systems with any number of elements, including mirrors and mixed configurations.
My telephoto setup gives a positive total magnification, but the image looks inverted when I look through it—why?
A positive magnification M_total means the image is upright relative to the object only if both individual magnifications have the same sign. For two converging lenses, M₁ = −s₁′/s₁ is usually negative (inverted), and M₂ = −s₂′/s₂ can be positive or negative depending on object placement. If M_total > 0 but the image appears inverted, double-check that your sign for one of the intermediate magnifications isn't flipped.
I traced rays through a two-lens system but my final image is on the wrong side—where did I mess up?
Check whether you used the correct object distance for the second lens. The image from lens 1 becomes the object for lens 2, but the object distance s₂ is measured from lens 2, not from lens 1. If the intermediate image lands to the right of lens 2, it's a virtual object (use negative s₂). Most errors come from forgetting to subtract the lens separation d: s₂ = d − s₁′.
I got an 'undefined' equivalent focal length—did I break something or is that normal?
It's normal for afocal systems. When the system matrix element C equals zero, f_eq = −1/C becomes undefined. This happens intentionally in telescopes and beam expanders where parallel input rays exit parallel. It's not an error; it means your system doesn't converge parallel light to a single focal point.
Can I model a compound microscope with this? The magnification seems way off from my textbook.
Yes, but you need the right configuration. The objective lens creates a real intermediate image at distance s′_obj (typically 16–18 cm from the objective). The eyepiece then magnifies this. Textbook formulas often use M = (L/f_obj)(25 cm/f_eye) where L is the tube length. This calculator gives the exact ray-trace result, which may differ slightly if your object distance doesn't match the standard working distance.
The calculator says 'real image' but I can't see anything on my screen during the lab—what's happening?
Three common issues: (1) The image may be too dim—check that your light source is bright enough and apertures aren't blocking rays. (2) The image distance might be longer than your setup allows—verify s′ and ensure your screen is at that exact position. (3) Aberrations can blur the image beyond recognition if you're using large apertures or cheap lenses. Try a smaller aperture to improve sharpness.
My object is exactly at the focal point of the first lens and the calculator shows infinity—is that a bug?
No, that's physically correct. When s₁ = f₁, the thin lens equation gives 1/s₁′ = 1/f₁ − 1/f₁ = 0, meaning s₁′ → ∞. Rays leave the first lens parallel. If there's a second element, those parallel rays will converge at its focal point. This is actually how telescopes and collimators work by design.
I'm getting confused about when to use positive vs negative focal lengths for my concave mirror—help?
For mirrors, use the convention: concave mirror = positive f (it converges parallel light), convex mirror = negative f (it diverges light). The focal length is f = R/2, where R is the radius of curvature. R is positive if the center of curvature is in front of the mirror (concave), negative if behind (convex). This matches the Cartesian sign convention used in most physics courses.
The paraxial approximation is mentioned everywhere—when does it actually break down in practice?
The paraxial (small-angle) approximation assumes sin θ ≈ θ, which is accurate to <1% error for angles under about 5°. It breaks down when: (1) your aperture is large relative to focal length (fast lenses, f/2 or faster), (2) objects or images are far off-axis, or (3) you're using spherical mirrors/lenses at wide angles. Real optical design software uses exact ray tracing and accounts for spherical aberration, coma, and other errors the paraxial model ignores.
How do I know if my intermediate image is real or virtual? The sign conventions are killing me.
After applying the thin lens equation for element 1: if s₁′ > 0, the intermediate image is real (forms on the opposite side of the lens from the object). If s₁′ < 0, it's virtual (on the same side as the object). For mirrors, positive s′ means the image is in front of the mirror (real), negative means behind (virtual). The key is consistency—pick one sign convention and stick with it for the entire problem.
My professor's formula for telescope magnification is M = −f_objective/f_eyepiece, but the calculator gives a different number—why?
That formula applies only to afocal telescopes at infinite conjugates (infinitely distant object, relaxed eye viewing). If your object is at a finite distance or the eye isn't relaxed, the actual magnification differs. Also, the negative sign indicates an inverted image for astronomical telescopes. The calculator computes the exact magnification for your specific object distance, which may not match the idealized textbook formula.