1883: Fischer and the Ball Grinding Machine

Before bearings could be mass-produced, someone had to solve a problem that looks trivial and is not: making steel balls that are round to within a micrometre, one after another, in large numbers. In 1883 Friedrich Fischer patented a grinding machine at Schweinfurt in Germany that turned ball making from a craft into a process. That patent is the quiet precondition for the entire modern bearing industry.

This article explains why precision balls were the bottleneck before 1883, how the grinding machine removed it, how Schweinfurt grew into a bearing town, and how the workshop that Fischer built became the FAG company in 1905. It closes with the part of the story that is still engineering, not history: ball grade, roundness and the fact that bearings depend today on the same property the 1883 machine first made repeatable.

The bottleneck before 1883

A ball bearing is a set of very precise rolling elements running in very precise raceways. The rings can be turned and ground to a specification, but if the balls themselves are not uniform, the whole assembly is compromised. Before a mechanical grinding process existed, balls were finished by hand, and hand finishing cannot hold a diameter to a few micrometres across a batch of thousands.

The consequences were immediate and mechanical. A ball that is slightly oversized changes the internal clearance of the bearing it sits in. A ball that is slightly out of round does not roll smoothly; it lifts and drops the ring once per revolution, which is vibration, noise and heat. Multiply that by the number of balls in a bearing and by the number of bearings in a machine, and the whole system runs rougher and hotter than any drawing predicts.

Hand-finished balls

Hand finishing meant lapping individual balls between plates and checking them by gauge. The work was skilled, slow and, crucially, variable. Sorting into size bands could tighten the fit, but sorting cannot make a ball round, and it wastes the balls that fall outside the bands.

Why inconsistent diameter caused vibration

In a bearing, all the balls in a row are meant to share the load. If their diameters differ, the largest balls take most of the load and the smallest take little, so a few balls do most of the work and fail early. Variable roundness adds a once-per-revolution disturbance, which is exactly the frequency that turns into audible noise and into bearing damage.

Fischer's grinding machine, patented in 1883

Fischer's insight was geometric. Two grinding discs, set at an angle to each other, will drive a ball between them in a compound motion, abrading it from every direction at once and driving it towards a sphere. By adjusting the disc angle and pressure, the process could be tuned to produce balls of consistent size and roundness, and, once set up, it could do so continuously rather than one ball at a time.

How the grinding worked

The ball is held between a rotating grinding wheel and a regulating disc. The contact geometry rotates the ball so that material is removed evenly over its whole surface, and the gap between the discs sets the finished diameter. The key property is reproducibility: the machine holds a dimension rather than an operator holding it by feel.

From balls to complete bearings

Once balls could be made uniformly, the remaining problem was the rings. A complete bearing needs an inner and an outer raceway ground to a matching curvature, hardened and finished. The firms that mastered ball making were one step away from mastering complete bearings, because a ball plant already had the grinding culture, the gauging and the hardened-steel experience.

The route from component to bearing

Fischer's workshop supplied balls to other makers before it supplied complete bearings. That is the classic path: start with the hardest single component, prove the process, then integrate forwards into the assembly as the ring-grinding capability matures.

The sequence also explains why ball makers and bearing makers were, for a time, separate businesses. A workshop that ground balls well could sell them to several bearing assemblers, each of which ground its own rings. Integration forwards happened only when a ball maker could also guarantee the raceway curvature that matched its balls, because the relationship between ball diameter and raceway groove radius is what keeps the contact stress within design limits. Get that match wrong and the bearing carries a fraction of its expected load before the raceway fatigues.

Schweinfurt grows as a bearing town

A successful process attracts the suppliers it needs. Ball grinding calls for grinding wheels, hardened steel, gauges, precision machine tools and skilled operators, and once those exist in one town the next bearing business starts there rather than elsewhere. Schweinfurt became exactly that: a concentrated bearing cluster whose advantage lay in shared capability, not in any single patent.

• Grinding wheels and abrasives. The consumable that every ball and raceway grinding line depends on.
• Hardened and clean steel. Bearing steel of a cleanliness a general steelworks did not supply.
• Gauges and metrology. The instruments that let a shop hold micrometre dimensions in daily production.
• Skilled grinders and gaugers. A trade passed between generations within the same town.

The cluster effect also worked on the skills side. Ball and raceway grinding is a trade, and a town with generations of grinders and gaugers produces new plants faster than a location with none. That is why a nineteenth-century process decision still explains the geography of the European bearing industry.

The FAG name and the 1905 company

The workshop that Fischer founded became, in 1905, the company known by the FAG name - an abbreviation of the German for "ball bearing factory". The formation date of 1905 is the documented anchor for the company, five years and more after the 1883 patent that made its product possible, and it marks the point at which ball making turned into complete bearing manufacture on an industrial scale.

The two dates together tell the story cleanly. 1883 is the process patent; 1905 is the company. Between them lies the transition from a component workshop to a bearing manufacturer, and after them lies the long international expansion that put the name into catalogues worldwide. You can read the modern range on our FAG page and its later sibling on the INA page.

Ball grade, and why roundness still rules

Modern balls are classified by grade, which is a statement of permissible deviation. The ABMA grade system, and the equivalent ISO 492 style tolerance thinking applied to bearings, expresses the point simply: a lower grade number means a tighter tolerance, so a grade 10 ball is more precise than a grade 25 ball.

Reading a ball grade

The grade number is essentially the ball diameter tolerance in parts of a micrometre. A grade 10 ball holds its diameter to roughly 0.25 micrometre, while a grade 100 ball is an order of magnitude looser. The grade also constrains sphericity and surface roughness, both of which feed directly into noise and fatigue life, which is why a bearing's quiet running can usually be traced back to the balls it was built with.

Indicative grades and tolerances. Confirm the exact bands and the sphericity and roughness limits against the ball supplier's specification before specifying on grade alone.

Why Ball Uniformity Still Sets the Quality Bar

The 1883 machine solved a process problem, and the same problem still governs quality: uniformity of the rolling elements. The bearing families that depend most on that uniformity, the deep groove and self-aligning types, are also the most common. Our deep groove ball bearing range and the deep groove deep dive show how ball quality propagates through the assembly.

Use a short step-by-step check when you specify or accept a ball bearing.

1. Confirm the envelope against ISO 15 so the fit and the shoulder dimensions are correct before anything else.
2. Identify the ball grade that matters for the duty; a quiet motor needs a tighter grade than a slow, coarse-duty position.
3. Check the ring tolerances under ISO 492, because ball quality and ring quality are separate specifications and both feed vibration.
4. Match clearance to the operating temperature using the standard range, not by habit.
5. Look at the closure. Seals and shields control contamination, which is the main way a good ball set is destroyed in service; see the bearing accessories range for the components around the assembly.
6. Confirm the calculated life using ISO 281 ratings rather than comparing only a nominal load figure.

For context on how the designation system encodes these choices, see the designation guide; for the other major family that grew from the same grinding culture, see the cylindrical roller range. The company history continues on the INA and Schaeffler story, and the interchange lookup and brand directory place the names in the wider market.