Balls or Rods: The Media Decide the Product, Not Just the Fineness

Spread a single layer of mixed stone on a flat plate — a good spread of sizes, nothing sorted. Lay a length of steel bar across it and press down evenly with both hands. Watch which stones crack. It will be the big ones, and only the big ones, because the bar is resting on them and cannot reach anything else. Now do it again with a ball bearing or a round pebble, pressed down at a point. Whatever is under the point breaks, and whether that is a large stone or a small one is entirely a matter of where you put your hand. That is the whole rung, done in two minutes with a bar and a ball. Everything that follows is putting numbers on it. Do the bar version twice: once on a single layer and once on a heap two or three deep. On the heap the selection gets weaker, because the bar is now resting on a lumpy surface rather than on individual stones. Rod mills are run with a thin charge for exactly that reason.

Three numbers describe a media charge and all three matter. **Fill.** The fraction of the mill's volume the charge occupies, usually a third or a little more. Too little and the charge has room to fall further than the toe, which hammers the liner instead of the material. Too much and there is no free space for the charge to cataract into, so it rolls instead, and the mill quietly turns into a cascading machine that grinds slowly and makes fines. **Grading.** A new charge is never one size. It is a distribution, because the mill needs the big media for the feed and the small media for the surface — the same argument as the tube mill's compartments, solved a different way. **Top-up.** Media wear away, and as they wear they get smaller, so a mill that is topped up only with the largest size drifts toward a coarse charge over months. Top up with a graded mix, and weigh what goes in. The commonest slow failure in a grinding circuit is a charge that has been maintained by eye. Measure the fill with the mill stopped: the charge's chord against the shell, taken with a tape, converts straight to a filled fraction. Record it every time. A drifting fill is visible in a log and invisible in a glance.

An open-circuit mill has to hold everything until the *worst* particle is fine enough, which means the average particle is ground far past the point of usefulness. That over-grinding is wasted power and unwanted fines, and it is entirely avoidable. A **closed circuit** puts a screen or a classifier on the mill's discharge and returns the coarse fraction to the feed. Each particle now leaves as soon as it is small enough. The coarse that comes back is the **circulating load**, and on a well-run circuit it is often larger than the fresh feed — two or three times, sometimes more — which surprises people the first time they see the figure. That number is not a sign of something wrong. A high circulating load means material is being removed promptly, which is exactly the point. What it costs is a bigger mill feed conveyor, a bigger classifier and more pumping, so there is an optimum rather than a maximum. The screening rung is what the classifier is doing, and its partition curve is why the circulating load can never be sharp: every pass returns some material that was already fine enough and passes some that was not.

**US 1,538,620, 'Ball and Rod Mill', Ferdinand Mora Canda of Chrome, New Jersey, application filed 19 February 1924, granted 19 May 1925.** The claims are about the mill's lining — 'a metal shell or casing, a circumferential series of brittle hard steel plates bolted to the shell, and a resilient shock absorbing medium interposed between the plates and the shell' — and the title covers both media in one breath, because the body of the machine does not care what is in it. That detail of the claim is worth noticing. By 1924 the mill itself was not the invention any more; what was being fought over was how to line it so the liners survived, which is the mark of a mature machine. Rod mills became the standard first stage of grinding in metalliferous plants for half a century, precisely because of the selection rule in step 4: gravity separation and flotation both work badly on slimes, and a rod mill delivered a narrow product without making them. They faded when autogenous and semi-autogenous mills took over the coarse grinding duty, not because the physics changed. **Honest limits.** A rod mill cannot be made long relative to its diameter — beyond about six to one the rods tangle, and a tangled charge is a day of work with a crane and a lot of swearing. Rods wear unevenly and have to be culled as well as topped up. A rod mill will not grind fine: the selection rule that protects the fines from being over-ground is the same rule that stops you reaching a fine product at all, so almost every rod mill in history had a ball mill downstream of it. And both are media mills, so both consume steel continuously — in a large plant the media bill can exceed the power bill.