Seed Drill Calibration from a Catch Test

A drill's rate chart is a manufacturer's estimate made with a seed lot you do not own. Seed size, seed shape, treatment, humidity, how full the hopper is and how worn the metering flutes are all move the delivered rate, and the only way to know what the machine is doing is to catch what comes out of it and weigh it. The arithmetic afterwards is one division: weight caught, over the ground area it represents.

Take a catch from four openers on a 17.5 cm spacing, driven 100 m. The width represented is $4 \times 0.175 = 0.70$ m, the area is $100 \times 0.70 = 70$ m², and if the bags hold 1050 g the rate is $1050/70 = 15$ g/m². A gram per square metre is exactly ten kilograms per hectare, so the drill is sowing 150 kg/ha. In imperial the same test with 8 oz caught over 100 ft behind a 10 ft width covers 1000 ft²; an acre is 43.56 of those, so the rate is $0.5 \times 43.56 = 21.8$ lb/ac.

The error that dominates this calculation is the width. It is the width of the RUNS YOU CAUGHT FROM, not the width of the drill, and putting the machine's 6 m working width into a test that collected from four openers overstates the area by more than eight times and returns a rate low by the same factor. Bags fall off, too, and an opener that quietly delivered into the soil instead of into a bag has the same effect. Count the bags, count the openers, and multiply the row spacing by the number of runs actually collected.

Two further cautions make a catch test worth the hour it takes. Do it on the move rather than by jacking a wheel and turning it by hand — a static test misses the way ground drive slips, the way an air system's fan pressure changes seed velocity, and the way a drill meters differently when the frame is bouncing, and it typically reads 3 to 8% high. And do two catches, at the start of the run and again with the hopper down to a quarter, because most metering systems deliver slightly less as the head of seed above them falls. If the two agree, set the drill and go. If they do not, the average is not the answer; the reason for the difference is.