Written by: HSD Optical Engineering Team — 15+ years making rod lenses
Published: July 2026 | Read time: 7 min
Last month, a customer called us. Their endoscope shipment was rejected. Image edges were hazy. Resolution failed spec. They’d been chasing the problem for a week.
It was a 5-micron decentration on one rod lens. Five microns. On a 1 mm diameter blank. That’s all it took to kill the relay train.
They’d been buying blanks from a supplier that quoted “standard centration tolerance” and shipped a pass/fail slip. No measured values. No distribution data. And no liability when the yield tanked.
This happens more than you’d think. And it’s almost always avoidable.
The Part Nobody Wants to Talk About
Rod lens centration isn’t something you can fix downstream. You can’t polish a blank back into centration. You can’t coat it straight. You can’t align the error out in assembly — because when you stack 10 or 12 of these things in a relay train, every micron of decentration adds up.
A 3 µm offset on lens #3 becomes 6 µm by lens #6 and 12 µm by the distal end. Your MTF is gone before the light even reaches the coupler. In a high-resolution endoscope delivering 20,000 pixels or more, that’s the difference between a diagnostic-grade image and a rejected shipment.
The blank is where this is won or lost. Not at incoming inspection. Not at assembly. At the grinding spindle.
If your current supplier can’t give you a batch centration histogram, you’re gambling. Ask for one. If they push back — “that’s proprietary” or “we don’t normally provide that” — that’s your signal. Quality suppliers measure it anyway. The ones who won’t share the data are hiding something.
How Centration Is Actually Measured
There are two ways to measure rod lens centration, and they give you very different numbers.
Vertical-axis testing mounts the lens vertically and rotates it. For lenses above Φ2 mm, this works fine. But for micro rods — the Φ1.0 mm, Φ1.19 mm, Φ1.5 mm blanks that go into ureteroscopes and mini endoscopes — gravity flexes the lens in the chuck. The test reports a better centration number than the lens actually has. The error doesn’t show up until assembly, when it’s too late.
Horizontal-axis testing uses V-groove support so the lens rests on its cylindrical surface. Gravity acts perpendicular to the optical axis, not along it. This gives you the true centration value — the number your assembly line will actually see. It’s slower and requires more careful setup, but for anything under Φ2 mm, it’s the only measurement that means anything.
We use horizontal-axis testing exclusively for micro rods. Every blank under Φ2 mm gets measured this way, and the actual measured value goes on the COA per ISO 10110 requirements. Not a checkbox. A number.
What We Actually Do
We make uncoated rod lens blanks. Φ1.0 mm to Φ4 mm. Standard diameters and custom. Fifteen years, same factory, same engineers. We don’t run a marketplace where we flip parts from third-shift polishing houses. We make them.
A few things that matter if you’re buying:
- We don’t substitute glass. You spec Schott N-BK7, N-SK2, F2 — that’s what you get. No “equivalent” without your sign-off. Index mismatch in your relay design changes element spacing — we won’t make that call for you.
- Every blank is measured on a horizontal-axis centration setup with V-groove support. Vertical testers give you a gravity-flex number on micro rods, not a centration number. We don’t use them for anything under Φ2 mm.
- Every batch ships with a COA. Actual measured values — centration, diameter, length, surface quality per ISO 10110. Not a checkbox. Numbers.
Specs That Actually Matter When You’re Buying
| Spec | What to Demand |
|---|---|
| Diameter tolerance | ±0.005 mm or tighter. Loose diameter → tilt in the bore → another centration problem you didn’t need. |
| Centration | < 3 arcmin, with measured COA data per batch. Not “within spec.” The actual number. |
| Surface quality | 40-20 or better per ISO 10110. Cylinder scatter kills relay contrast. |
| Glass | Exact Schott or CDGM grade. Index mismatch changes relay spacing in your optical design — don’t let a supplier make that call for you. |
| Coating | Uncoated unless you need AR. Most coating shops want bare blanks for their own line. |
What This Means for OEM Assemblers
If you’re an OEM assembler — not a coating shop, but the company that specs, buys, and installs finished lenses into endoscope or ureteroscope relay trains — centration matters just as much, but your risk looks different.
You’re not grinding blanks. You’re buying fully coated, ready-to-assemble rod lenses and expecting them to drop into your relay train without re-measuring every piece. That only works if your supplier controlled centration at the blank stage — before coating, before cementing, before the cost of a failed assembly gets baked in.
For assemblers, we offer broadband AR-coated rod lenses (400–700 nm) across 245 standard sizes, all built from the same centration-controlled blanks. The same COA data follows the part through coating. You get measured centration values on the finished lens, not just the blank. That’s the data your incoming inspection team needs to skip 100% part-by-part measurement and move to batch audit against the COA — which is exactly what our coating-lab customers do.
Whether you’re a coating house buying uncoated blanks or an assembler buying finished lenses, the centration spec is the same: ≤ 3 arcmin, measured, documented, on every COA.
What Happened to That Customer
They switched to us. Three changes:
- Incoming inspection went from 100% part-by-part to batch auditing against our COA. Throughput doubled.
- Assembly first-pass yield rose over 20%. Fewer relay train teardowns. Fewer Friday-night panic calls.
- They built a centration database from our COA data — now they run SPC on incoming blanks instead of treating every shipment like a mystery box.
For coating shops, it’s even simpler. You coat. You assemble. It works. For OEM assemblers buying finished lenses, the same centration data means your incoming inspection can trust the COA instead of re-measuring every lens on the line.
Ready to Test Us?
Send us your spec — diameter, length, glass type, coating preference, rough volume. We’ll send back a datasheet, a sample COA format, and a lead time. Usually within 24 hours.
No pitch deck. No “solutions provider” language. Just the numbers.
If you want an NDA first, that’s standard. Say so and we’ll get one in place before the technical discussion.
Send your spec. Get a quote with real COA data.
NDA-protected. Response within 24 hours.
Send Us Your Spec — Get a Quote → Browse Rod Lens Product LineFAQ — Rod Lens Centration Questions, Answered
What is centration in rod lenses?
Centration measures how well the optical axis of a rod lens aligns with its mechanical axis. In practice, it’s the tilt or offset between the polished cylindrical surface and the center of curvature of the end faces. For medical endoscope relay trains, even 3–5 µm of decentration compounds across 10+ lenses and destroys image quality.
What centration spec should I require from a rod lens blank supplier?
Demand ≤ 3 arcmin with measured COA data per batch — not a pass/fail checkbox. The actual number matters because it feeds into your assembly tolerance stack. If you’re building high-resolution endoscopes or ureteroscopes, ask for ≤ 2 arcmin.
Do you provide batch centration histograms?
Yes. Every shipment from HSD Optical includes a Certificate of Analysis with actual measured centration values for that batch. No averages. No “within spec.” Raw numbers you can feed into your SPC system.
What’s the difference between horizontal and vertical centration testing?
Horizontal-axis testing (V-groove support) gives the true centration value. Vertical testing introduces gravity-induced flex on micro rods (Φ2 mm and under), producing an artificially better number. We use horizontal testing exclusively for anything under Φ2 mm — because that’s the number your assembly line actually sees.
What glass types do you stock for rod lens blanks?
N-BK7, N-SK2, F2, and other Schott / CDGM equivalents. We don’t substitute without written approval. Index mismatch in your relay design changes element spacing — we won’t make that call for you.
What’s your MOQ for custom rod lens blanks?
Typically 100 pcs for standard diameters. Custom sizes and tight tolerances may require a slightly higher MOQ depending on tooling. Contact us with your spec and we’ll confirm within 24 hours.
How can I verify your quality before placing a large order?
Two ways:
- Request a sample COA — we’ll send you a real batch COA so you can see the format and data granularity.
- Order a trial batch — 50–100 pcs, tested to your spec. If it meets your standard, you scale. If not, you walk away with zero commitment.
Do you offer NDA protection?
Yes. NDA is standard before any technical discussion. Mention it in your inquiry and we’ll get one in place first.
About HSD Optical
15 years manufacturing precision rod lenses. We supply uncoated blanks to endoscope OEMs, coating shops, and integrators across Europe, North America, and Asia. ISO 9001. Full melt-to-shipment traceability. Every batch: measured, documented, COA included. NDA as standard.
HSD Optical — Precision Rod Lens & Micro Lens Supplier for Endoscope & Ureteroscope Tier 1 Assemblers
Specifications subject to confirmation at inquiry. Customer details anonymized. HSD Optical — ISO 9001:2015.