It depends on the rod mill work index and on the desired product size. It will also depend on if you are grinding open or closed circuit and wet or dry. And if you have any stages of grinding, crushing, or classification before it! It will generally be between 10 mm and 50 mm.

But is 50 mm suitable feed size for rod mill?

Depends on what you are feeding, the desired product, the feed rate, the connected motor size, are you grinding open or closed circuit, etc. A blanket statement cannot be given it must be calculated from the operating parameters.

A first approximation for optimum feed size is Fo=16,000(13/Wi)^.5. Coarser (and finer) feeds are possible. 50 mm is approaching the top end. What is the context of the question? Are you optimizing an operating plant or designing a new installation? You would operate a rod mill in closed circuit?

Not normally, but it can be done.

19 mm for Taconite with 100 mm diameter rods! 75mm for Dolomite and Limestone mix with 100 mm diameter rods. Your experience will be somewhere in between.

This is based on local experience with over 70 rod mills ranging from 800 to 1650 kW. All are grinding wet in open circuit.

I think we should forget about Rod Mill and concentrate on Ball Mills in the primary grinding stage especially we have a close-circuit grinding circuit.

I am not sure if I can validate above statement (?) as the breakage mechanism is not necessarily the same and this depends on a lot of parameters: rocks hardness, rocks type, composition, product size required, rods filling degree, design of the process.

Bond would say ideal feed is 80% -16mm for a 13 W.I> material. If the Material has a higher W.I. there is a Bond formula to calculate the ideal feed size for a normal 1.5 to 1.8 L/D ratio rod mill.

Dependent upon the Material and the Process! I seen feed as large as 100mm on soft material. There are specialty rod mills where you target a low ratio of reduction such as end and center peripheral discharge than feed need to be targeted to product size required based upon the calculate reduction ratio in the rod mill.

Rod mills behave differently from ball mills or SAG / AG mills.

Ball mills need very large balls to begin to break down a hard crushed particle like taconite. They tend to not be very efficient and produce a lot of very fine product, especially if the balls aren't big enough to smash the largest particles.

Rod mills excel in making the largest particles smaller without generating a lot of fines. For coarse separation of banded magnetite like taconite they are the most energy efficient tumbling mill for taking fine crushed product and reducing it to about 80% passing 4-5 mm.

The University of Minnesota has done studies on converting the many rod mills around here into ball mills with large balls. In our particular situation where we are trying to specifically make particles around 5 mm for the first stage of magnetic separation, the economics didn't work out at all, too much power, too little capacity.

Rod mills have become niche comminution devices since about 30 years ago. Because of mechanical limitations on rod length and the need to have an aspect ratio for the mill that avoids tangling (a potentially deadly failure for those tasked with rectifying the problem) rod mills have a maximum installed power of about 2.5 MW. With SAG mills now at 32 MW and ball mills at 20 MW the rod mill is now left to lower throughput applications where it has demonstrable advantages over other technologies. Specific advantages include fines minimization, problem topsize elimination, avoiding an extra crushing stage, high power in a small footprint, and a few others. As others have stated, the maximum topsize is entirely material dependent. However, it is not enough to know the average hardness. An understanding of hardness (Bond RWI) variability is essential when making such decisions.

Rod Mills are used to prevent over grinding. If the grinding mill is close circuit with a good classification system, why should we use Rod Mills especially there are so many inherent problems with the management of grinding media?

Why not crushing another stage down to 5 mm with cone crusher or a vertical shaft mill?

Again, VSI will create too much fines which is detrimental for subsequent stages of processing. With a good sizing machine like Derrick screen, sizing can go down to as fine as 75 microns!

19-25mm