19

2026

-

11

Why Your Rice Harvester Is Producing So Much Broken Rice

Author:


A mill buyer takes a handful of paddy from the truck, shakes it flat on his palm, and looks at it for maybe three seconds. If the broken kernels stand out, the price is settled before  the weighbridge. Nothing at the mill caused that. It was decided in the field, one setting at a time.

Broken rice almost never comes from a single failed part. It comes from a drum turning forty rpm faster than the crop needs, a concave still set for wheat, a tailings elevator quietly sending clean grain back for a second beating, or paddy cut three days past its best moisture. Each one looks minor on its own. Together they put five or six percent brokens in the tank and cost you a grade.

What follows is the order we work through in the field: where the damage is actually created inside a rice combine, how to tell from the grain tank sample which step is responsible, and what to change first. Most machines can be brought back under two percent in an afternoon. The ones that cannot usually have a worn part or a genuine capacity problem, and that is worth knowing before the season starts rather than halfway through it.

What a few points of broken grain actually cost

In milling, the number that decides the price of your paddy is head rice yield, the share that comes out as whole, unbroken milled grain. Brokens still have value, but they sell as brokens, at a discount that varies by market and by end use. A lot arriving with six or seven percent visible brokens does not simply lose those six points. It usually drops a grade, because buyers price on the sample in front of them, not on the field average.

There is a second layer that never shows up on the truck. Threshing does not only break grain outright; it puts fine fissures through kernels that survive the journey to the dryer. Those cracks open weeks later on the milling line, and the miller blames the mill. If you grow seed paddy, the tolerance is tighter still: a lot with one or two percent visible brokens is often rejected outright rather than discounted.

For a benchmark, China's technical guidance for mechanised rice harvesting caps broken grain at 1.5 percent for whole-feed combines and 0.5 percent for semi-feed machines, on a sheet that also limits total loss to 2.8 and 2.5 percent and impurity to 2.0 and 1.0 percent. Buyers elsewhere set their own numbers, usually in the same range. If your tank sample sits at five percent, the machine is doing damage that does not have to happen.

Prove the machine is guilty before you touch a setting

Start with two samples taken the same afternoon. One from the grain tank, spread flat on a tray in daylight. One from behind the machine: pick up a dozen panicles out of the straw and split them by hand.

If the grain in those panicles is already cracked, chalky, or split at the tip, the damage started before the header reached it. Three or four rain-and-dry cycles in the last fortnight, sucking insects, or a crop left standing too long will all do it. No drum setting brings that grain back. If the straw sample is clean and the tank is not, the machine is doing it, and the next question is which part of the machine.

One more free clue: where the brokens sit in the load. Grain that has been through the return loop two or three times tends to end up in the middle of the tank, with cleaner grain toward the edges. Damage spread evenly through the sample points upstream, at the drum. It is a rough test, but it saves you opening the wrong inspection cover.

Cause one: drum speed above the threshing threshold

Rpm on its own tells you almost nothing. What the grain actually feels is tip speed, how fast the rasp bar is moving as it passes the concave, in metres per second. A 600 mm drum at 800 rpm and a 900 mm drum at 530 rpm put roughly the same energy into the crop. That is why a number copied from a neighbour's machine, or from a manual written for a different drum, can be wrong in both directions.

Paddy comes off the panicle through a mix of impact and rubbing, and impact is what fractures kernels. Above the speed needed to strip the panicle clean, every extra metre per second goes into the grain and nowhere else. In a 2024 threshing study across three paddy varieties, raising cylinder tip speed from 17.3 to 24.6 m/s more than doubled the broken-grain percentage in the sample and pushed fissured grain from about 15 percent to 19 percent. Same crop, same machine, one variable.

How to find your number. Run at rated engine speed, not part throttle: drum rpm is calibrated against rated PTO speed, and the tacho stops telling the truth when you throttle back to save fuel. Start at the low end for rice, for most machines somewhere in the 600 to 900 rpm band depending on drum diameter, and take a sample after fifty metres. Drop 50 rpm and sample again. Keep dropping until unthreshed panicles start appearing in the tailings return or on the ground behind the machine. Then step back up one notch. That is your working speed for this crop, in this field, at this moisture.

A surprising number of combines are running 15 to 20 percent above that point, and have been for years, because the rasp bars were already rounded and speed was the cheap way to keep the tank looking clean.

Cause two: concave clearance, and its shape

Under the drum sits the concave. The gap between the two is the other half of the threshing pair, and for rice a common starting point is 20 to 30 mm at the inlet, closing progressively toward the outlet so the crop is rubbed as it travels rather than hit once and released. Too tight and the whole panicle gets squeezed; too open and intact panicles ride out the back.

The two damage signatures are worth learning, because they point in opposite directions. Fine powder, with straw beaten short and fluffy, means impact, and the drum is too fast. Split kernels mixed with unbroken husk fragments and pieces of rachis mean compression, and the gap is too tight.

There is a third case that gets missed because it does not average out. A concave that is tight on one side and open on the other cracks grain on the tight side while throwing unthreshed heads out the open side, and no single adjustment can fix it. Before trusting any setting, measure the gap at three points across the width. On most machines it is a ten-minute check with a feeler gauge, and it explains a class of problems that otherwise look like bad luck.

One change at a time is not fussiness. Speed and clearance compensate for each other; move both and you lose track of which one helped. Set speed first, then open or close the gap to match.

Cause three: cutting paddy too dry

Crop moisture is the most influential variable in threshing damage, more influential than any setting on the machine. The same combine, untouched, will run clean at 24 percent and put three times as much broken grain in the tank at 16 percent. Grain that has lost its internal moisture is brittle, and the husk stops cushioning anything.

The working window for machine harvest is generally quoted at 20 to 25 percent moisture, at 90 to 95 percent maturity. Below about 18 percent you are harvesting glass. Above 28 percent the problem flips: incomplete threshing, blocked sieves, high impurity, and grain going out the back with the straw.

Weather is not yours to control, but the hour of the day is. If the crop has dried down hard, cut in the early morning while dew is still on the straw, or after dark. Leave the hottest, driest part of the afternoon to someone with a smaller machine and no quality contract. The moisture that comes back into the grain overnight is free insurance.

Damage that starts before harvest is the hardest to see and the least worth chasing. Paddy that has been through repeated wetting and drying in its final two weeks accumulates internal fissures while looking perfect on the truck. It will break in the mill whatever you do at the drum. If the season has gone that way, expect a lower head rice yield, price the crop accordingly, and do not spend a week hunting for a setting that cannot exist.

Cause four: feed rate that will not hold steady

Ground speed is a threshing setting. Push more crop into the drum than it can digest and the material packs into the concave instead of rolling through: the drum slows, threshing goes soft for a few seconds, the operator hears it and speeds up, and the next half hectare gets beaten. Neither the soft pass nor the hard pass is good for the grain.

The tell is in the engine note and in the sample. Uneven feeding gives you alternating clean and broken patches through the tank, and a drum that never settles. In rice, with heavy straw and often soft ground, the temptation to run hard at the start of a pass and crawl through the turn is strong. Hold a steady speed, keep the engine at rated rpm, and let the machine's capacity set the pace rather than the clock.

Lodged crop makes this worse, because the header has to come down and the reel has to move forward, and the feed turns lumpy. In lodged rice, slow down rather than compensate with drum speed. Speed is the wrong tool for a feeding problem.

Cause five: the return loop is double-threshing your grain

Every combine has a tailings elevator that sends unthreshed material back into the drum. That is normal and necessary. What is not normal is that elevator carrying mostly clean grain, which then takes a second and a third hit for no reason at all.

The check takes two minutes with the machine stopped and the key in your pocket. Open the tailings inspection port and look. You want to see unthreshed or half-threshed panicles, and nothing else. Whole, clean kernels in that stream mean the sieves are too tight or the fan is too low, and the drum is innocent.

This is one of the more common reasons a tank sample refuses to improve no matter how carefully the threshing section is tuned, and it is the cheapest of all the causes to fix.

Cause six: worn parts, and the speed added to cover for them

Rounded rasp bars, polished spike teeth, concave wires worn thin, one bent bar nobody has straightened. Worn threshing parts do not rub grain free as well, so the natural response is more speed or a tighter gap. Both buy threshing performance with grain damage, and neither works for long.

Open the inspection cover before each season and run a hand over the bars. Sharp edges rub grain off the panicle. Rounded, polished edges hit it. Replace in sets: one new rasp bar beside three worn ones simply moves the problem around the drum.

Cause seven: damage after the drum

Some of your brokens never went through the threshing section. Clean-grain elevators, tank augers and unloading augers all handle paddy, and paddy is abrasive. Worn flighting with sharp edges, a chain running too tight, grain dropped two metres into a truck, or a dryer pushed too hot will all break kernels that threshed perfectly.

The diagnostic is simple: if changing drum speed and concave clearance barely moves your broken percentage, stop tuning the threshing section and look downstream. Handling damage tends to hold a flat, steady percentage across every field and every day, whatever the crop is doing.

Cause eight: variety, and why last week’s settings fail this week

Two fields, one machine, one day, different result. Long, slender indica grain breaks more easily than short, round japonica; hybrids with heavy panicles and thick straw load the drum differently again. The same 2024 study found the long-slender variety produced the highest broken-grain percentage of the three tested at identical settings, and the medium-grain variety the lowest.

Which is why the notebook matters more than any chart, including the charts in this article. Crop, variety, moisture, drum rpm, concave gap, fan setting, ground speed, sample result. Two seasons of that and you have a table that actually fits your fields.

Reading the tank sample

Table 1  What the grain tank is telling you

What you see in the tank

What it usually means

What to change first

Fine powder, straw beaten short and fluffy

Drum or rotor speed too high

Drop 50 rpm and resample

Split kernels with husk fragments and crushed rachis

Concave too tight, or tight on one side

Open a quarter turn; measure the gap across the full width

Whole panicles in the tailings or on the ground behind

Threshing too light

Close the concave slightly, or raise speed one notch

Grain and chaff blown out the back

Fan too high, sieve too open

Lower the fan until grain starts to carry over, then back up

Brokens concentrated in the centre of the tank

Return loop carrying clean grain

Check the tailings port; reset sieve and fan

Steady broken percentage that ignores every threshing change

Damage after threshing, in handling

Inspect elevators, tank auger, unloading auger

Sample sticky, drum note laboured, sieves blocking

Harvesting too wet

Let the dew lift, slow down, clean the sieves

Grain in the straw behind the machine already cracked

Field damage, before harvest

Not a machine problem; adjust expectations

Starting points by crop condition

Table 2  Rough starting points for paddy, by moisture

Parameter

Dry and brittle (below 18%)

Normal (20 to 25%)

Wet (above 26%)

Drum speed (600 mm drum)

600 to 700 rpm

700 to 800 rpm

800 to 900 rpm

Concave gap at inlet

28 to 30 mm

22 to 25 mm

20 mm

Fan

Lower

Mid

Higher

Ground speed

Steady, moderate

Normal for capacity

Slower

Watch for

Fissures, powder in the sample

Sample every 50 m

Clogging, unthreshed heads

These are starting points, not targets. Scale drum rpm to your drum diameter and check the operator’s manual for your machine before anything else.

A twenty-minute sequence that converges

  1. Prove the machine is doing the damage: split a dozen panicles from the straw behind the machine.
  2. Check the tailings port. Clean grain there means the shoe is wrong, not the drum.
  3. Set the engine to rated rpm, then reduce drum speed in 50 rpm steps, sampling every 50 metres.
  4. Only once speed is settled, adjust the concave a quarter turn at a time.
  5. Set the fan: lower it until grain starts to carry over with the chaff, then bring it back up just enough.
  6. If nothing responds, stop tuning and inspect the wear parts.
  7. Slow the ground speed last, and only as far as the crop demands.

Change one thing, run fifty metres, take a sample. It is slower for the first hour and much faster for the rest of the season.

What to ask before you buy the machine

Settings can only take a machine so far. A few design choices decide how easy broken grain is to avoid, and they are worth asking about before money changes hands.

  • Drum diameter and length. A larger drum reaches the same tip speed at lower rpm, which is gentler on grain and easier on the drive train. A longer threshing arc gives the crop more chances to separate at low speed instead of taking one hard hit.
  • Concave adjustability. Being able to change the gap from the seat, in small increments, is what makes the 50-rpm method practical. Machines that need a spanner and twenty minutes get set once a season and never touched again.
  • Sieve type. Adjustable-angle fish-scale sieves let the cleaning section be matched to moisture without flooding the returns.
  • The return system. How much the tailings elevator can carry, sized against the sieve area, decides whether clean grain ends up going round twice.
  • Tracks and ground clearance for paddy. On paper this is not a grain-quality specification, but a machine that bogs down gets driven too fast in the dry patches and too slow in the wet ones, and feed rate suffers at both ends.

Golddafeng's 4LZ-7 fits that brief: a crawler whole-feed rice combine with a 2.0 m axial-flow threshing drum, 110 HP turbocharged engine, 7 kg/s feed rate, 2.4 m header, 550 mm tracks and 340 mm of ground clearance for wet fields. The long drum is the part that matters in this article, because it lets the crop separate over a longer arc at lower drum speed, which is the direction you want to move when broken grain is the problem.

Questions we get asked

Is three percent broken rice normal?

It depends who is buying. For milling into domestic table rice, three percent will usually be accepted and discounted. For seed paddy, or for a contract with a tight specification, it is a rejection. The more useful question is what your machine is capable of: if a well-set combine can hold 1.5 percent in the same crop, every point above that is money you chose not to collect.

Will slowing the drum leave grain in the field?

Only if you go below the threshing threshold, which is why you test downward until unthreshed panicles appear and then step back up one notch. At the bottom of the working band you lose nothing. Above it, you lose quality instead.

Does a bigger machine break more grain?

Capacity is not the problem; overloading it is. A 7 kg/s machine running at 60 percent of its feed rate is gentler than a 3 kg/s machine running flat out. Broken grain tracks uneven feeding and running past capacity, not size on its own.

Why is the sample worse in the afternoon?

Because the crop dried. Same field, same settings, four hours apart. Check the moisture before blaming the machine, and if the answer is 16 percent, cut earlier tomorrow.

The short version

Broken rice is rarely one large mistake. It is four small ones left alone for a season: a drum set fast because the bars are worn, a concave still set for the last crop, a return loop nobody has opened, and paddy cut in the driest hour of the driest week. Fix the first three and the fourth gets easier.

Sample the tank every fifty metres for one afternoon and write down what you change. That is most of the job.