More Aspheres in a Shorter Tip: What to Demand Before You Accept

Before you accept a tip design that leans on more aspheres than the original concept, ask for four things in writing. A side-by-side comparison of the original and revised prescriptions with the system benefit stated in a number, not an adjective. A tolerance and sensitivity study that ends in a predicted yield at your assembly capability. A metrology route that someone other than the maker can reproduce, with the uncertainty stated against the tolerance. And a list of everything else in the tip that the change touches, from the window and the seal to the illumination path and the working distance. Miss the third and the drawing becomes unenforceable. Miss the fourth and a shorter tip arrives with a longer problem list.

The Trade You Are Actually Being Offered

Adding aspheric surfaces buys design freedom. That freedom is real: extra variables let a designer shorten a system, hold a numerical aperture in a smaller envelope, or push aberration correction into fewer elements. The bill arrives somewhere else, in fabrication, in metrology, and above all in centering, because an asphere has an optical axis that must be found and a vertex that must be aligned, and a tip that is only a few millimetres across leaves very little room to be wrong. A scope tip is the hardest place to spend that freedom, because the assembly is miniature, the elements are close together, and the sensitivity of each surface to decenter and tilt lands on an assembly process with a fixed capability.

The question to put to the maker is not whether the aspheres help. It is which system requirement they buy, by how much, and what the buyer inherits when the design is frozen. A shorter tip is a legitimate answer. A shorter tip whose surfaces only the maker can measure, assembled to a centering tolerance your process cannot hold, is a design change that transfers risk to you without saying so.

Symptoms That Appear After a Design Leans on Aspheres

These are the complaints that surface months after acceptance, when the design is frozen and the tooling is paid for. Each one has a root cause that was visible in the design review, and each one has an evidence request that would have caught it.

Symptom in production	Root cause that was present at design review	What to require before accepting
Resolution passes at the centre, fails at the edge of field	Nominal performance was quoted at one field point, with no field-dependent tolerance study	Performance at named image heights, and the tolerance study run at those same heights
First article is excellent, the third lot is not	The surface sits near the edge of what the process can hold, so yield is being harvested by selection	The asphere’s departure, slope and form tolerance alongside the process capability the maker claims for it
Assembled MTF scatters unit to unit despite good incoming parts	Centering and clocking sensitivity of the added surfaces was never quantified for the assembly	A sensitivity table for decenter, tilt and clocking per surface, with the assembly compensation named
A barrel that worked in the prototype cannot be assembled in volume	The design assumed an alignment step or a manual tweak that production will not repeat	The alignment procedure in writing, with the adjustment range and the time allowance per unit
Nobody at the buyer can confirm the part meets drawing	The only metrology for the surface lives at the maker, with no transferable method	A measurement route a third party can reproduce, with full-aperture data and stated uncertainty

Centering and Datum: The Part That Bites Inside a Tip

Most of the review energy goes into surface form, because form is the number that appears on the drawing. In a tip, the assembly usually fails on the axis instead, and the axis is a weaker part of the documentation. Ask for the three items below separately, because makers often answer them as one.

Surface form is not the same as centering

A surface can meet its form tolerance and still be mounted off axis. Form describes the shape relative to a best-fit reference; centering describes where that shape sits relative to the mechanical datum the barrel will hold. Ask which datum the tolerance is referenced to, how it is established on the shop floor, and whether the vertex, the mechanical axis and the aspheric datum axis are declared to coincide. When they do not coincide, the design has to state the offset and the assembly has to control it, and that control has a cost that belongs in the quotation rather than in the ramp.

Clocking, wedge and the datum axis

An asphere with rotational asymmetry cannot be dropped into a barrel in any orientation. Clocking, the rotational position relative to a reference feature, becomes a controlled characteristic, and it needs a marking on the element and a feature in the barrel that makes the correct orientation unambiguous under a microscope. Wedge between surfaces compounds the problem. Ask for the rotational tolerance, the marking method, and the inspection step that verifies orientation after assembly. If the answer is a hand-written note on the drawing, it will not survive the third shift.

Assembly sensitivity at tip scale

Request a sensitivity table rather than a sentence. For each added surface, how much does system performance move per unit of decenter, tilt and spacing error, and what compensation is available at assembly. A design that is sensitive but compensable is a different proposition from one that is sensitive and rigid, and the difference shows up in yield. Where the tip splits into an objective and a relay, the interface between the halves deserves its own document, as set out in the interface sheet that lets each half stand alone, because a centering error at that joint is invisible to both suppliers separately.

Enforceability: Who Can Actually Measure the Asphere?

A tolerance that only the supplier can verify is a promise, not a specification. This is the point that separates a design review from a supply decision. Aspheric surfaces are commonly measured by interferometry with a computer-generated hologram, by profilometry, or by a dedicated asphere measuring instrument, and each route has a different usable aperture, a different sensitivity to alignment during measurement, and a different uncertainty. Ask which route the maker will use, what the uncertainty is relative to the tolerance, whether you will receive full-aperture maps rather than a single RMS figure, and whether the reported error is referenced to the nominal prescription or to a best-fit sphere. Facilities that publish their asphere fabrication and metrology routes, such as Optimax, show how much of the answer is process description rather than a single number.

Then answer the buyer’s half of the question: what will you check at goods-in. You will not duplicate a CGH interferometer, and you should not try. What you can do is measure the assembled tip against the parameters that predict station performance, resolution and distortion at the image heights you use, back focal length, and field coverage on the actual sensor. Ask the maker to show how those assembled measurements correlate with the surface data, so a goods-in result can be read back to a surface cause. Where the correlation is not offered, treat the surface tolerance as an internal manufacturing control and buy on assembled performance instead. Metrology houses that supply both the instruments and the acceptance practice, such as Trioptics, are a reasonable place to check what a transferable measurement for centering and focus looks like.

What Else in the Tip Must Re-Qualify

A shorter tip is a mechanical change as much as an optical one, and the optical review will not catch the rest. Walk this list with the design owner before the change is frozen, and give each line a named owner and a completion date.

• Window and seal. A shorter barrel moves the front window closer to the object and changes the seal geometry. Hermetic packaging and any IP67 or IP68 duty claimed for the assembly must be re-verified on the new geometry, not assumed from the old one.
• Illumination path. Light guide or fibre clearance, exit angle and falloff at the new working distance. A shorter tip changes the illumination geometry even when the optics are untouched.
• Working distance and depth of field. The reduced length usually comes with a changed front focal distance. Re-state the working distance and the depth of field at the same aperture, and confirm the station’s fixturing still reaches focus.
• Sterilization and cleaning. Any cycle the device sees, temperature, chemistry and number of repetitions, must be re-run on the new assembly if the materials or bonds changed.
• Sensor pairing. Confirm the revised design still matches the sensor it will ship with, including chief ray angle and cover glass, rather than assuming the original pairing survives a prescription change.
• Assembly time and tooling. A design that needs a new alignment fixture or a longer takt time is a cost change, and it belongs in the quotation before the design freeze, not in the first invoice after it.

The Evidence Sheet

Reduce the acceptance decision to one page that the maker signs and you file with the drawing revision. The rows below are the ones that decide whether the change is safe to accept; the exact limits come from your application, not from a template.

Item	What to state	Why it decides the acceptance
System benefit	The requirement the shorter tip satisfies, and the measured or modelled gain, at named field points	Without it there is no case for added complexity, only a preference for a smaller package
Prescription delta	Original versus revised surfaces, glasses, thicknesses and clear apertures, with the drawing revision number	Fixes what is being approved, so a later change cannot hide inside the same drawing
Surface specification	Departure, slope, form tolerance, and the datum each is referenced to, per aspheric surface	Makes the surface testable and separates shape from position
Sensitivity and yield	Decenter, tilt, clocking and spacing sensitivity per surface, with predicted yield at the agreed assembly capability	Shows whether the design survives your process rather than the maker’s best bench
Metrology route	Measurement method, uncertainty against tolerance, full-aperture data, and who else can repeat it	Decides whether the tolerance is enforceable after the order is placed
Re-qualification scope	Seal, window, illumination, working distance, sterilization and sensor pairing, each with an owner	Catches the failures that the optical review is not looking for

Conclusion

Accepting a design that leans on more aspheres is a fair decision when the evidence matches the added complexity. Ask for the system benefit in a number, the tolerance study that ends in yield, the centering and clocking data tied to a declared datum, and a metrology route that a third party can reproduce. Then walk the rest of the tip, because a shorter barrel re-opens questions about seals, illumination, working distance and sterilization that have nothing to do with surface form. Put the whole answer on one signed page and file it with the drawing revision. A maker who can supply that page is telling you the change is genuinely manufacturable, and a maker who cannot has just shown you where the risk would have landed. If the program needs a custom objective built to your drawing rather than a modified catalogue design, the route described in custom optics from drawing to mass production covers the same acceptance sequence, and the endoscope family pages show how sensor pairing is normally fixed, at endoscope lenses.