How to Read a Bearing Part Number: 5 Segments to Verify Before You Order

Split the designation into three blocks

Read a bearing number as three blocks, not one code. There is an optional prefix, then the basic number, then the suffix string^3,20,22,27.

The basic number carries the bearing type, the dimension series and the bore code. The suffix string carries sealing, clearance, cage and tolerance class^3,20,22.

Those two blocks are decoded by different rules, which is the whole reason the split matters. A buyer who treats the string as one code cannot tell whether a mismatch sits in the envelope — the part will not fit — or in the operating characteristics, where the part fits but behaves wrongly.

Take 6204-2RS/C3 (supplier-reported figure). The basic number 6204 fixes 20 mm bore, 47 mm outside diameter and 14 mm width¹⁹ (supplier-reported figure).

The suffix 2RS/C3 says rubber seals both sides and a C3 clearance class^3,26. One string, two independent readings.

JTEKT notes that standard bearing numbers correspond to JIS B 1512 boundary dimensions and are prescribed in JIS B 1513, and that manufacturers add supplementary codes beyond those the standard provides²⁰. That last point is the practical warning: the basic number is standardised, the suffix vocabulary is not fully.

Write the three blocks out separately on your RFQ line before you send it.

Decode the bore code before anything else

The bore code is a multiplication, not a size. For codes 04 and above, the last two digits times five gives the bore in millimetres, so 05 means 25 mm and 04 means 20 mm^3,22,26 (supplier-reported).

Below that, four special codes break the pattern: 00, 01, 02 and 03 stand for 10, 12, 15 and 17 mm, not 0, 1, 2 and 3 mm²². At 500 mm bore and above, the designation states the actual bore size directly rather than coding it²².

The mechanism is simple and unforgiving. A buyer who reads 05 as 5 mm orders a bearing that will not go onto the shaft at all, and the error survives every later check because the rest of the number looks right.

This is the first thing to rule out, because a wrong bore is not a specification disagreement — it is a part that cannot be mounted.

One source reports that 82% of cross-reference errors involve incorrect clearance specifications, which tells you clearance is the more common trap, but bore is the one that fails hardest²⁴. Check the bore code against the shaft diameter on the drawing before you compare anything else.

If the two do not agree, stop and re-read the code.

Use the series code to fix the envelope

The dimension series code sets the outside diameter and width for a given bore, so two bearings that both fit the shaft can still differ by 10 mm in OD^16,22,26. That is the second trap: a quote built on part-number similarity alone is not safe²².

Read the table by bore first, then by housing clearance (supplier-reported).

Read the seal suffix, not the fit

Seal and shield suffixes follow a letter-count rule. Z is a single-sided metal shield, ZZ is metal shields both sides, RS is a single-sided rubber seal and 2RS is rubber seals both sides³.

The count tells you how many sides are covered; the letter tells you the material. What the suffix does not do is change the installation dimensions.

Seal and shield variants of the same base bearing typically share the same bore, outside diameter and width, so the suffix decision is about grease retention and heat generation rather than whether the bearing will mount^18,19.

A 608 keeps its 8 mm bore, 22 mm OD and 7 mm width across the open, Z, ZZ, RS and 2RS variants¹⁸. The trade-off is thermal.

One source reports that metal shields stay stable above 120°C while rubber seals should not exceed 100°C, because rubber softens and loses sealing effectiveness as temperature rises³. Open bearings run cooler and last longer in high-speed, high-load duty, while shielded types retain grease but may run hotter⁴.

So the suffix is a temperature and contamination decision. If your application runs hot, ask the supplier which cover material they are quoting and at what temperature it is rated.

Confirm the clearance suffix explicitly

Clearance suffixes order from tightest to loosest: C2, then C0 (also written CN), then C3, C4 and C5³. The higher the number after C, the larger the internal clearance.

This is the segment most often lost in cross-referencing, and one source reports that 82% of cross-reference errors involve incorrect clearance specifications²⁴. The mechanism is thermal.

As a bearing runs hot, the inner ring, outer ring and rolling elements all expand, and the designed clearance shrinks³. Fit a C0 bearing into that condition and the working clearance can close to zero or go into preload, which raises friction heat and shortens life.

C3 exists so the working clearance after expansion lands back in the right band — not so the bearing stays loose forever³. One supplier recommends adding C3 clearance when a sealed or shielded bearing runs at high temperature and high speed with long life requirements¹⁸.

Note the counter-case too: a hot-running bearing is not automatically a C3 bearing, because too tight a fit, poor lubrication, overload and misalignment all raise temperature as well⁶. State the clearance code on the RFQ rather than accepting the supplier's default.

Match the precision class to the application

Precision classes come in three naming systems — ABEC, DIN and ISO — and they currently align well enough to compare across²³.

The class indicates tolerance strictness, and the selection rule the sources give is application-led: class 0 is adequate for general applications, while class 5 or higher is required for demanding applications and operating conditions⁹.

The mechanism is that a higher class tightens the tolerances the machine was designed around, so paying for a class the application does not need adds cost without adding benefit, while a general-purpose class in a demanding application leaves that tolerance unguaranteed.

One source reports that ABEC 5 dominates market revenue at 49.7%, but that is a stocking pattern rather than a selection rule — the same source says selection should start from the application, not a catalog default¹⁷.

So treat the share figure as a signal about what suppliers hold, not about what you should buy. State your speed and runout requirement on the RFQ and let the supplier propose the class, then check the proposal against the application rather than against what is on the shelf.

Run the pre-RFQ verification checks

A bearing number is not self-verifying, and part-number similarity is exactly how a wrong clearance or a wrong envelope gets quoted^17,22.

The checks below exist because the failure modes are silent: a 6004 and a 6204 share a bore and look alike in a quote line, and a C3 variant priced against a C0 requirement reads as a saving until the bearing runs hot (supplier-reported).

Write the dimensions and suffixes out in full so the supplier prices the part you specified, not the part the number resembles.

• ✓Write bore, outside diameter and width on the RFQ even when the part number is legible
• ✓Confirm the bore code against the shaft diameter, including the special codes 00-03
• ✓State the seal or shield suffix and the operating temperature it must survive
• ✓State the clearance code explicitly rather than accepting the supplier's default
• ✓State the precision class or the speed and runout requirement that fixes it
• ✓Ask the supplier to confirm the quoted designation segment by segment before you release the PO

Scale documentation to the consequence

Documentation should be matched to the purchase requirement, and the level scales with what a wrong bearing costs you²⁵. A general hardware order may need only a certificate of conformity.

A bearing OEM order may require material certificates, inspection reports, hardness records and control plans. Automotive customers may go further, into PPAP-related documents, process flow charts, FMEA and measurement system analysis, and medical or aerospace work may require additional validation²⁵.

The mechanism is that the buyer who does not set the level in the RFQ gets whatever the supplier sends by default, and a certificate that does not cover the actual manufacturing scope tells you very little.

When verifying certificates, check validity dates, issuing bodies, scope of certification, company name and factory address, and whether the certified scope actually includes the product you are buying — a trading company may present a certificate that belongs to a different factory²⁵.

That is not automatically dishonest, but you need to know who controls production and quality. Decide the documentation level before the RFQ goes out, not after the goods arrive.

Weigh MOQ and lead time before you commit

Two commercial factors sit outside the designation but decide whether the order is workable. Minimum order quantity varies by type, specification and whether customisation is needed, so there is no single figure to plan around⁶.

One supplier states that standard and custom parts carry different conditions and asks buyers to share the model, clearance code, quantity and application so it can quote and assess lead time⁶. Lead time behaves the same way.

One source reports an average lead time within 15 workdays, but that is a market average across peak and off-peak conditions, and the same source advises asking each supplier for their own committed date rather than planning against the average¹⁷.

The mechanism is that a buyer who assumes a standard MOQ over-commits volume, and a buyer who plans a shutdown around a market-average date is planning around a number no supplier has agreed to.

One supplier reports a cross-referenced 22330 CA/W33 arriving within 72 hours against a four-week original-manufacturer quote, but that is a single case, not a planning assumption²⁴. Ask each supplier for their own MOQ and committed date in writing.

Where the sources disagree

Four points in this article rest on sources that do not agree with each other. The table keeps both sides visible so you can see which figure is a claim and which is a rule, and the last column tells you what to do about it.

Where the two sides conflict, the safer reading is usually the one that makes you verify rather than assume (supplier-reported).

What the sources do not establish

• No official brand catalog page that decodes every segment for a single manufacturer
• No ISO/ABMA boundary dimension tables with actual OD and width values per bearing type
• No complete OD and width table for a given bore across all dimension series
• No speed rating, friction, temperature or ingress protection values for each seal type
• No numerical clearance ranges in micrometres for each bore size, or the effect of fit and temperature on them
• No tolerance values for bore, OD, runout and width per precision class
• No specification of what must appear on the bearing itself (brand, designation, country of origin, batch or date code)
• No specific MOQ numbers, price break tables, or standard versus non-standard suffix lead times