Lens Barrel Machining: Bore Roundness, Thread Concentricity and Black Anodize
A lens barrel holds a stack of glass elements at fixed distances and on a common axis. The glass is made to micron-level centering by the optics supplier; the barrel's job is not to throw that away. Every micron of bore clearance, every degree of thread tilt, and every 10 µm of anodize on the wrong surface turns into decentering, tilt, or a reflection that shows up as flare in the corner of an image.
Lens barrel machining is turning work, mostly, with a boring bar and a thread mill and a very specific measurement routine. The dimensions are not exotic. What makes it difficult is that the functional surfaces are usually the ones coated last.
What the barrel actually controls
Four things, in order of sensitivity: the centering of each element relative to the optical axis, the axial spacing between elements, the tilt of each element's seat, and the internal scatter that reaches the sensor.
Centering depends on bore diameter and roundness. Axial spacing depends on shoulder position and face squareness. Tilt depends on the seat face being perpendicular to the bore within a few microns over the seat width. Scatter depends on the internal finish and coating, and on whether a machined edge catches light at a grazing angle.
A 50 mm focal length lens with 0.01 mm of decentration has about 40 arcseconds of pointing error, enough to visibly soften a corner on a 5 MP sensor with 2 µm pixels.
Bore roundness and what clearance costs
The bore is usually H6 or H7 with a roundness callout of 0.005 mm and cylindricity of 0.008 mm. Those are not arbitrary. Clearance between the bore and the element's ground OD is what allows the element to sit off-axis, and the element sits wherever gravity and the retaining force put it.
| Bore tolerance on Ø25 | Worst-case diametral clearance | Max decentering | Angular error at f = 50 mm |
|---|---|---|---|
| H6 (+0.013/0) | 0.026 mm | 0.013 mm | 54 arcsec |
| H7 (+0.021/0) | 0.034 mm | 0.017 mm | 70 arcsec |
| H8 (+0.033/0) | 0.046 mm | 0.023 mm | 95 arcsec |
| +0.05/−0 | 0.063 mm | 0.032 mm | 130 arcsec |
The table assumes a lens OD ground to h6 and no centering adjustment. In practice, most designs either select-fit the element to the bore in 0.005 mm steps, or use a tapered seat plus a retainer that pushes the element into contact on one side. Either approach works; what does not work is a nominal H7 bore and a hope.
Roundness is the harder callout to hold because boring a thin wall releases stress. A barrel with a 3 mm wall and a 40 mm bore will go out-of-round by 0.005 to 0.010 mm after the finishing pass if the blank was not stress-relieved. Rough, relieve at 160 °C for 4 hours, semi-finish, then finish with a sharp boring bar at 0.05 mm depth and 0.03 mm/rev.
See also: Drone Frame Machining: Billet Center Plates, Ribs and Motor Patterns and Robot End Effector Machining: Where Gripper Accuracy Actually Lives.
Threads and the seat behind them
Retaining ring threads are usually M42 × 0.75 or a similar fine pitch, and the callout that matters is concentricity to the bore, not thread class. A 6g thread that is concentric within 0.02 mm TIR will hold a ring that pushes evenly on the element. A thread that is 0.05 mm off-axis will cant the ring, and the ring will tilt the element no matter how flat the seat is.
Cut these threads with a thread mill on a turning center with live tooling, or single-point them with a full-profile insert. Taps are a poor choice in thin-wall aluminum barrels because the tap lifts material and can distort the bore 0.01 mm at the thread root.
The seat face behind the thread needs squareness of 0.005 mm to the bore over the seat width, and Ra 1.6 µm or better. A seat that is square but rough will let the element rock by the height of the surface peaks under clamping force, which is 0.002 to 0.004 mm for a turned finish at Ra 3.2 µm.
Black anodize and stray light
Type II black anodize at 10 to 20 µm is the standard internal coating for aluminum barrels. It is not decorative. A black, matte, mildly porous surface absorbs light that would otherwise bounce off the barrel wall and land on the sensor as flare.
Two machining consequences. Anodize grows about 50 percent inward and 50 percent outward, so a Ø25.000 mm bore becomes Ø24.980 mm after a 20 µm coating, which is outside H7. Either mask the bore and the thread, or leave 0.02 mm on the diameter and finish-machine after anodize. Masking is cheaper for small quantities; post-machining is more reliable for anything with a tight bore.
The second consequence is edge quality. Anodize does not cover a sharp edge uniformly — the coating thins at a 90° corner and can craze. Break every internal edge to 0.2 mm minimum, and radius the aperture edges that face the optical path so no bright ring reflects at a grazing angle.
Some designs blacken the aluminum behind the last element with dye or flat black paint. If you do, specify the thickness: 30 µm of paint on a seat face changes the axial spacing.
Thermal focus shift
Aluminum at 23.6 µm/m·K moves a 50 mm barrel by 0.035 mm over a 30 °C rise. Glass moves far less — N-BK7 at 7.1 µm/m·K, F2 at 8.2 — so the barrel grows away from the glass and the focus shifts toward infinity. For a fixed-focus lens, that is a blur of about 0.04 mm at the focal plane, which is visible at f/2.
There are three fixes, and they all change the machining. Use a low-expansion barrel material such as Invar 36 at 1.2 µm/m·K, which is expensive, hard to machine, and usually reserved for athermal instrument optics. Use a hybrid mount with a titanium or stainless section that compensates the aluminum. Or design a cam slot so a focus group moves to compensate, which turns the barrel into a multi-axis part with a helical slot held to ±0.02 mm.
If you stay with aluminum, put the thermal expansion where it does least harm: keep the distance from the last element to the sensor flange short, and put the long spacing between elements that are less sensitive to defocus.
Machining sequence around anodize
The sequence that avoids rework: rough turn everything leaving 0.5 mm, stress-relieve, semi-finish leaving 0.15 mm on bores and threads, finish all non-cosmetic features, anodize, then finish-bore and finish-thread with diamond or uncoated carbide. Mark the datum face with a laser before anodize so the post-coating setup has a reference that survives the process.
Barrels are usually turned from bar or tube on a CNC turning center with a bar capacity up to Ø80 mm, then finished with live-tool milling for slots and set screws. Coordinated surface finishing matters as much as the turning, because the anodize line is where a good barrel becomes a bad one.
Frequently asked questions
What bore tolerance does a lens cell need?
H6 with 0.005 mm roundness for a select-fit element, or H7 if the design uses a tapered seat and a centering retainer. Anything looser than H8 adds more than 90 arcseconds of pointing error at f = 50 mm.
Does anodize change the bore diameter?
Yes. Type II at 20 µm closes a bore by about 0.02 mm on diameter, because roughly half the coating grows inward. Mask the bore, or leave 0.02 mm and finish-machine after coating.
Why is thread concentricity specified instead of thread class?
Because a 6g thread that is 0.05 mm off-axis will cant the retaining ring and tilt the element. Concentricity to the bore within 0.02 mm TIR controls the tilt; thread class only controls fit.
How do you reduce thermal focus shift in an aluminum barrel?
Shorten the distance from the last element to the sensor flange, move long spacings to elements that tolerate defocus, or switch to a low-expansion material such as Invar 36 at 1.2 µm/m·K. A 30 °C rise moves a 50 mm aluminum barrel by 0.035 mm.
Related reading
Send a drawing for a quote and DFM review.
STEP, IGES, X_T, PDF or DWG up to 50 MB. Tell us quantity and finish, and the quote comes back with tolerance and process notes, not just a price.