Two tubes can magnify equally yet show different orientations
Imagine a distant sign with an upward arrow and a dot on its right. Through one lens tube the arrow looks larger and upright; through another it is equally enlarged but points down, with the dot now on the left. The second tube need not be broken: that can be how its optical layout works.
A Galilean layout combines a converging objective with a diverging eyepiece; a Keplerian layout combines a converging objective with a converging eyepiece. Changing the second element’s sign changes where the eyepiece intercepts the light. Orientation, the required element spacing and the field accessible to the eye follow from that change. Museo Galileo: the two layouts in a teaching model.
We will compare the same distant object at the same magnification magnitude, 10×. These are teaching calculations, not the description of a tested lens kit.
An upright view and a wider field answered different needs
Galileo’s surviving telescopes show early practice: separate objective and eyepiece cells, a wooden tube and a covering. He observed the sky with instruments of this kind from summer 1609. That does not mean he single-handedly invented the telescope. Museo Galileo: its instrument collection.
In Dioptrice, published in 1611, Kepler described another combination, using a positive eyepiece. Describing a layout should not become a claim that he personally built and tested such a telescope then. A wider practically accessible field became an important advantage, while inversion was acceptable for many astronomical observations. An upright view is more convenient on the ground; a Keplerian system can be supplemented with erecting optics. Museo Galileo: conceiving new telescopes.
This is an engineering choice rather than a division into “correct” and “incorrect” telescopes. Simple Galilean systems are used, for example, in opera glasses; more complex terrestrial instruments may contain additional prisms or lenses and no longer reduce to two elements. University of Arizona: Galilean telescope and comparison, slides 13-34–13-36.
The key difference is before or after the bundle crosses
A refractor gathers light with its main lens element, the objective. Light from one very distant point arrives as an almost parallel bundle. The objective makes it converge. Another point on the distant object supplies another bundle and, without further intervention, another image point. NASA: refracting telescopes.
In a Keplerian layout, the light has time to form a real inverted intermediate image. The eyepiece sits beyond it and, when adjusted for a relaxed eye, turns light from each image point into its own parallel output bundle. An accessible plane between the lenses can hold a reticle, for example.
In a Galilean layout, the negative eyepiece sits earlier. It weakens convergence just enough for the rays to emerge parallel, without crossing between the lenses. The location where the objective alone would focus them is beyond the eyepiece, but the installed eyepiece changes their subsequent path. Drawing a real intermediate image of the complete system there would be wrong.
Galilean layout: d=450 mm, M=+10
Keplerian layout: d=550 mm, M=-10
O: objective; E: eyepiece; F: objective focus; P: aperture image.
Thin lenses: fₒ=500 mm, fₑ=−50/+50 mm, D=50 mm, α=0.002 rad. Shared longitudinal scale; transverse scale enlarged; lens shapes symbolic. In the Galilean layout, grey dashes show where light would travel WITHOUT the eyepiece: these are not actual rays or an intermediate image. Dashed/solid P: virtual/real image of the objective as aperture stop, not of the star.
In both diagrams, each set of three rays represents one source point. “Parallel output” does not mean that all sky points merge: neighbouring points have different output angles. The eye perceives differences between those angles as a magnified object. University of Arizona: afocal systems, pages 4-1–4-4.
A shorter tube need not have less magnification
Our example uses the same objective in both systems. The diverging eyepiece goes before its focus; the converging one goes beyond. The thin elements are spaced 450 and 550 mm apart respectively. Both instruments magnify angular size tenfold: the different lengths do not imply different magnification magnitudes.
A Galilean layout preserves the orientation of a two-dimensional object. Without additional optics, a Keplerian layout rotates it by 180°: both up/down and left/right change. That is not merely a mirror reflection. Add a diagonal, prisms or other elements and the orientation must be traced again.
Branch · Spacing and magnificationOptional · safe to skip
One sign convention gives both layouts
Light travels left to right; f_(o) > 0 for a converging objective, f_(e) < 0 for a diverging eyepiece and f_(e) > 0 for a converging one. In the Galilean layout, the negative eyepiece focal length must have a smaller magnitude than the positive objective focal length, giving a positive separation. For distant objects and afocal output:
d = f_(o) + f_(e), M = (α′)/(α) = − (f_(o))/(f_(e)).
d is the separation between equivalent thin elements; α and α′ are small input and output angles in the same convention. With f_(o) = 500 mm:
- f_(e) = − 50 mm → d = 450 mm, M = + 10;
- f_(e) = + 50 mm → d = 550 mm, M = − 10.
Lengths add in mm and their ratio is dimensionless. A ray with slope α = 0.002 rad emerges at + 0.02 or − 0.02 rad. Positive preserves orientation; negative reverses it. This is a paraxial model for a relaxed eye, not a rule for measuring from the ends of arbitrary real cells. Thick multi-element groups have principal planes and their own mounting distances.
Why the edge of an upright view can be harder to find
The exit pupil is the image of the aperture stop formed by subsequent optics. In a model where the objective defines the working opening, the eyepiece images that opening. It is an image of the instrument’s entrance, not another picture of the star.
The positive Keplerian eyepiece forms a real image of this opening behind itself. The eye can be placed there to receive bundles from different directions. The negative Galilean eyepiece gives a virtual image of the opening inside the tube, where the eye cannot physically go. In real Galilean instruments, the eye pupil itself often becomes the system aperture stop, while the objective clips the field. Eye position therefore strongly affects the visible region. Greivenkamp: pupils and Galilean vignetting, slides 13-13–13-15 and 13-46–13-48.
The same asymmetric target
Galilean layout
Upright; M=+10. Part of the beam misses the pupil.
Keplerian layout
Rotated 180°; M=−10. The beam fits.
Dashed: eye pupil; colour: beam
Orientation symbols are separate and do not scale magnification. Circles are sections at the eye pupil, not photographs of the field. Teaching example: α=0.004 rad, eye diameter=5 mm, beam diameter=5 mm; eye 5/55 mm behind eyepiece. Beam centres +2/0 mm. The 20 mm eyepiece does not clip these bundles. This is one vignetting case, not a universal field in degrees or a retina model.
This explains the Keplerian layout’s typical field advantage without making it unlimited. Eyepiece edges, field stops, internal baffles and eye dimensions can still clip bundles. Vignetting means losing some required light through limited openings; it differs from losing sharpness through aberrations.
Branch · The pupilOptional · safe to skip
Where exactly is the objective’s image?
The objective acting as stop is an object for the eyepiece at u_(p) = d > 0. Using the same thin-lens convention, with positive image distance for a real image behind the lens:
v_(p) = (1)/(1/f_(e) − 1/d), m_(p) = − (v_(p))/(d), D_(p) = ∣m_(p)∣D.
Our two layouts give v_(p) = − 45 and + 55 mm from the eyepiece. With D = 50 mm, both give D_(p) = 5 mm. Negative indicates a virtual image on the objective side; positive indicates a real one on the eye side. Pupil diameter is always positive, unlike signed pupil magnification m_(p).
Do not confuse the sign and location of the stop image with those of the intermediate object image. If the eye actually limits the system, its system entrance and exit pupils are defined from that stop instead. These numbers describe the image of the specified objective, not every possible real configuration.
Branch · Field geometryOptional · safe to skip
One numerical example of a clipped bundle
For the figure, choose another direction, α = 0.004 rad, and a hypothetical 5 mm eye pupil. Place it 5 mm behind the negative eyepiece or 55 mm behind the positive one. Thin-lens calculation gives equal 5 mm bundles at the eye plane, but their centres are +2 and 0 mm from the axis.
The first disk partly misses the eye opening; the second fits. Reversing the input angle makes the first displacement −2 mm. On axis, both bundles fit in this configuration. This checks geometry rather than measuring a light percentage or full field. Different diameters and eye distances give different results; the 20 mm eyepiece is not the limiting opening in this example.
Two lenses still need to share an axis
The objective faces the object and the eyepiece faces the eye. A cell holds each element, a tube connects them, and a focuser changes axial spacing to suit object distance and the observer’s eye. An aperture stop defines the working opening; correctly placed internal baffles intercept unwanted light while preserving the required bundles.
- Aperture stop: defines the working opening
- Objective and cell: hold the element without tilt
- Tube and internal baffles: maintain the axis and reject unwanted light
- Focuser: changes position along the axis
- Eyepiece in its cell: negative or positive to suit the layout
Assembly gaps, thicknesses and diameters are arbitrary; this is not a dimensioned drawing or a verified kit. The ± symbol means a choice of eyepiece sign. Baffles must not arbitrarily block the required field; check their positions against the bundles.
Check an educational assembly on paper first: lens signs and properties → required spacing and focus travel → cells and alignment → baffles → image check. Two lenses are not suitable merely because they fit inside a tube. Even prepared educational kits have separate holders, spacers and moving tube parts. Taking a Closer Look archive, parts list on page 4.
The word “converging” alone does not specify a real surface’s orientation. Museum telescope 2428 has a plano-convex objective with its convex side outward; its surviving biconcave eyepiece is a nineteenth-century replacement. A modern group may contain several lenses and require its own orientation prescription. Our thin symbols are not such a prescription. Museo Galileo: detailed description of 2428.
Three similar complaints can have different causes
Coloured fringes can arise from dispersion: a simple lens does not focus all wavelengths identically. Turning the focus knob selects a compromise without removing the colour dependence itself. The next article examines achromats and apochromats; “Keplerian” and “Galilean” describe arrangement and signs rather than the degree of colour correction. University of Arizona: Chromatic Aberration, page 4-5.
A ghost or veiling glare may come from unwanted reflections and light travelling along unintended paths. Coatings, internal surfaces and baffles have jobs different from changing eyepiece sign. Arbitrarily narrowing every opening risks sacrificing useful aperture and field.
Asymmetric blur or field movement during focusing calls for checking centring, tilt and play in cells or the focuser. The ideal layout assumes a common axis; misalignment violates that assumption. Real components also have thickness, curved surfaces and tolerances that this thin-lens comparison does not calculate. Edmund Optics: first-order limits and real elements.
An upright view, a wide field and photography need different solutions
For the same focal-length magnitudes, a Galilean layout gives an upright view and shorter element spacing. Its limitations concern the eye’s access to bundles and a narrow practical field. A Keplerian layout offers a real intermediate plane and an accessible external stop image, but needs additional optics to erect a terrestrial view. Neither pair of simple lenses promises freedom from aberrations.
A camera without its own lens needs a different arrangement: put its sensor in the objective’s real image plane. That differs from attaching it in place of the eye at a parallel afocal output. A camera with its own lens can receive that output in a suitable configuration; that kit is neither designed nor tested here.
Try answering these questions: where would the Galilean objective’s light meet without its eyepiece, and why does it not meet there with the eyepiece installed? How can 450 and 550 mm both give 10×? Why does an upright image not guarantee a wide field? These answers connect geometry with what the observer sees.
Revisit the common principle in “How does a telescope work?”. The course next turns to colour correction and a documented educational assembly using prepared lenses. No physical assembly or testing of the examples described here has been performed.