Aspheric Design Guide For Manufacturability

size  |  shape  |  surface definition  |  tolerances  |  general guidelines  |  testing methods  |  final note

Size:                                                                                                                                                                                                                                   [ back to top ]
• 15 to 120mm diameter preferred
•  5 to 200mm with limitations

(If an aspheric shape is such that it can be processed out to a 15mm diameter, it may then be edged down to the desired final diameter)

• Minimum 10mm local radius preferred
• Concave:  Minimum 40mm local radius preferred.

• Maximum sag height (either convex or concave) of 12mm preferred
• Monotically convex surfaces are easiest to process.

Shapes going through an inflection point make form correction more difficult.

The processing diameter of an aspheric lens can be 4 – 10mm larger than the finished lens; therefore, it is desirable for the aspheric shape to be “well behaved” in this regard.

A down-turned edge can cause minor difficulty. 

An up-turned edge can significantly interfere with processing.

Flange interfering with a convex surface is bad.

Flange cut into a surface is okay (but adds a manufacturing step).

Annular flats on concave surfaces are routine.

The tolerances to follow are only guidelines because many factors come into play to determine the cost and difficulty of making any given part.  What may be a “slam dunk” for one part can be a “half court shot” for another, depending on size, shape, glass type, etc.

Avoid multiple tight tolerances, they compound difficulty.  For example, requiring a 1/10-wave accuracy, center thickness of ± .025mm, surface quality of 10-5 scratch-dig and a difficult glass type may each individually cause only a modest price increase but requiring them all on the same part will make for a very expensive optic.

Tolerances tend to be independent from each other in their effect on system performance but have a compounding effect on manufacturability.  DON’T PICK ALL HIGH PRECISION TOLERANCES IF THEY DON’T ALL NEED TO BE!

Base radius is a separate tolerance from form error since it can be focused out in most systems.  Avoid specifying a tighter radius tolerance than needed.

Give a larger share of your “error budget” to the aspheric surface.  For example, when designing a lens system with 7 spherical and 1 aspheric surface, callling for ¼-wave surface irregularities, instead of specifying ¼-wave on all 8 surfaces, specify .5 microns on the asphere and 1/8-wave on the spherical surfaces.  The net tolerance is tighter and the system is more manufacturable.

Try to use a mainstream glass (BK-7 if possible) for the aspheric element(s).

In general, not just for aspheres, avoid letting the computer pick the glass. Check on glass availability during the design process; many “new” glasses (Ohara ”S” type and Schott “N” type) are not readily available.

** Centration vs. wedge.  For spherical lenses, centration and wedge are the same thing.  This is not true for aspheres. The process we use at Kreischer Optics insures the aspheric surface is VERY WELL CENTERED relative to the outside diameter.

Form Talysur F:
Precision profilometry (contact measurement of sag height across a diameter which is compared to the ideal Form) is the industry standard in spheric metrology. Profilometry is versitle in both ground and polished surfaces may be trested adn degree of depart use from best fit sphere is not a limitation. Taylor Hobson with their Form Talysurf instruments in the industry leader in precision prfilimetry. Kreischer Optics has two From Talysurf, the S5, with 120mm diameter and 12mm say height capacity, and the PGI 1240 which is Taylor Hobson's premier metrology system for aspheric Form measurment . The PGI 1240 has a diameter capacity or 200mm with say height measurement ranges of 12.5, 25, or 38mm (dependent on length of stylus arm) with a measurement resolution of 0.8nm and overall measurement accuracy of under 0.1 microns.

Zygo GPI:
Very mild aspheres (up to about 8 waves departure) may be tested without null optics, otherwise, a null lens (designed in house at KOL) or a computer generated hologram will be needed.  This has the advantage of [potentially] full aperture testing to 1/8-wave accuracy or better, but not all aspheric shapes can be tested this way.  Null lenses can add $1k - $7k of non-recurring cost and may add to lead-time.  CGH nulls will cost $7k and up and take 6 – 8 weeks.  With either null lenses or CGH nulls, only polished surfaces may be tested.