Ask how a bearing is designed and most people start at the outside: bore, outside diameter, width. Those are the constraints you are given, not the design. The design happens in the middle of the part, and it runs in an order that surprises people the first time they see it.
This article walks the chain on a real bearing — a 5203-2RS double row angular contact sample we measured, dismantled and reverse-engineered in our own shop. Every number below is from that sample.
The design order
The instinct is to ask "what is the inner ring outside diameter?" That is the last number in the chain, not the first. The actual order is:
Ball diameter + contact angle + groove radius + clearance target
↓
Groove centre positions (inner and outer, separately)
↓
Inner groove bottom diameter / outer groove diameter
↓
Groove depth
↓
Inner ring shoulder OD / outer ring shoulder bore
Everything visible from outside the raceway is a consequence. The shoulder diameters exist to retain balls whose position was already decided three steps earlier.
The measured sample
5203-2RS, measured 2026-06-18. Sample data from one batch, not a specification.
| Category | Item | Value |
|---|---|---|
| Envelope | Bore d | 17 mm |
| Envelope | Outside diameter D | 40 mm |
| Envelope | Width B | 17.47 mm measured (nominal approx. 17.5) |
| Ball | Diameter Dw | 6.35 mm (1/4″) |
| Ball | Count | 8 per row, 16 total |
| Inner ring | Shoulder outside diameter | 25.00 mm |
| Inner ring | Groove bottom diameter | 22.60 mm |
| Inner ring | Groove depth | 1.20 mm |
| Inner ring | Groove centre distance Mi | 6.5497 mm |
| Outer ring | Shoulder bore | 33.05 mm |
| Outer ring | Groove diameter | 35.38 mm |
| Outer ring | Groove depth | 1.165 mm |
| Outer ring | Groove centre distance Mo | 6.77 mm |
| Raceway | Groove radius, two rows | 3.2880 / 3.2853 mm |
| Raceway | Average groove radius | 3.2867 mm |
| Raceway | Curvature ratio R/Dw | 0.5176 |
| Geometry | Mo − Mi | 0.2203 mm |
| Geometry | Axial offset per row | 0.11015 mm |
Reading the numbers
Groove depth is arithmetic, not a design input
Inner ring: (25.00 − 22.60) / 2 = 1.20 mm. Outer ring: (35.38 − 33.05) / 2 = 1.165 mm. Both fall out of the groove diameter and the shoulder diameter — you do not choose depth directly, you choose what sits either side of it.
Curvature ratio R/Dw = 0.5176
This is the single most informative number on the sheet. It says the groove radius is 51.76% of the ball diameter — the groove is slightly larger than the ball, which is what allows the contact to spread into an ellipse under load rather than a point.
Common practice for ball bearings puts this ratio in the region of 0.515 to 0.530, so 0.5176 sits in the normal band, toward the tight end. Tighter conformity means a larger contact ellipse, lower contact stress and higher stiffness — but more spin friction and heat. Looser means the opposite. It is one of the genuine trade-off dials in bearing design, and it is invisible from outside the part.
The groove centre distances differ on purpose
Mi = 6.5497 mm on the inner ring. Mo = 6.77 mm on the outer ring. The difference, 0.2203 mm, gives an axial offset of about 0.11 mm per row.
That offset is the contact angle. Because the two rows' groove centres are staggered between inner and outer ring, the line joining the contact points runs at an angle to the radial plane instead of square to it — which is precisely what makes the bearing angular contact rather than a double row deep groove.
The honest limit of a reverse-engineered angle
From the groove centre distances, the groove radii and the ball diameter, the contact angle on this sample works out to roughly 29–30°.
That number deserves a warning label, and we are going to give it one rather than quote it as a finding.
It is a geometric back-calculation from a free, unloaded, dismantled sample. The real operating contact angle depends on the actual radial and axial clearance, on the fits after mounting, and on load. The angle a bearing works at in service is not the angle you compute from a part sitting on a bench. Before any such figure is used to freeze a design, it has to be confirmed against measured clearance, working clearance after mounting, and the contact track on the raceway.
This distinction matters commercially, not just academically. A supplier quoting you a precise contact angle for a bearing they reverse-engineered — with no clearance measurement and no contact-track evidence behind it — is quoting a calculation, not a property. It is a reasonable starting point for a sample plan. It is not a specification.
What this means if you are sourcing a copy of an existing bearing
A common request is "here is a bearing, make me the same thing". The chain above shows why the useful conversation is not about the outline:
- Envelope dimensions are the easy part. Bore, OD and width can be measured with hand tools and matched by anyone.
- The raceway decides the behaviour. Groove radius, curvature ratio, groove centre distances and the resulting contact angle determine load capacity, stiffness, noise and life. None of these is visible without a profilometer and a dismantled sample.
- Clearance is a separate decision. Two bearings with identical raceway geometry behave very differently at different axial clearances.
So when a copy "fits but does not last", the outline was matched and the raceway was not.
Working with us
If you have a sample and no drawing, that is a normal starting point for us rather than a problem. We dismantle it, take profilometer readings on both rings, record groove radii, groove centre distances and derived geometry into a measurement file, and identify which figures are measured, which are calculated, and which still need confirmation before anything is frozen. From there we propose a sample plan against your application conditions.
FAQ
What is the groove curvature ratio and why does it matter?
It is groove radius divided by ball diameter — 0.5176 on this sample. It sets how closely the groove wraps the ball, which drives contact stress, stiffness and friction. Typical ball bearing practice is roughly 0.515–0.530.
How do you get the contact angle from a dismantled bearing?
By back-calculation from the inner and outer groove centre distances, the groove radii and the ball diameter. It is an estimate from free geometry, and must be confirmed against measured clearance and contact track before use.
Why are the inner and outer groove centre distances different?
Deliberately — that difference creates the axial offset that gives the bearing its contact angle. On this sample the difference is 0.2203 mm, about 0.11 mm per row.
Can you copy a bearing from a sample?
Yes, starting from a dismantling and measurement pass. Matching the outline is straightforward; matching the raceway geometry is the actual work.
Is groove depth a design input?
No. It falls out of the groove diameter and the shoulder diameter, both of which are set earlier in the chain.