Belt weighing depends on the accurate measurement of weight and belt travel. This involves a weigh frame and a tachometer. The tachometer is a simple device, but can be surprisingly error prone. The tachometer can be a wheel or pulley running on the belt which is generating pulses via a proximity sensor or an encoder.
Typically the tachometer is calibrated by measuring a length of conveyor belt and then finding out how many tachometer pulses occur in that length. The result is the ‘tachometer calibration constant’ a number expressed in metres/pulse. If this calibration is only done occasionally it is important that this relationship is stable! But research has shown that errors of 1% to 2% are common, even on a presumed 0.25% belt weigher! Errors come from temperature effects, wheel/pulley contamination, belt stretch and another phenomenon known as the load sensitive tachometer.
The turning radius of a tachometer roller may not be stable
The simple expectation is that the pulley or roller will turn at its surface diameter and that the belt will roll over it tangentially. Therefore, it is expected that the turning radius is exactly at the surface of the pulley. Experience has shown that this is not the case.
Each roller and pulley actually turns at an unknown radius inside the belt. When the belt is wrapped around a roller, the bottom cover becomes thicker and the rubber in contact with the roller moves slower than the belt carcase. This is equivalent to the radius of rotation being inside the belt.
This effect does not matter if the radius of turning is stable. But for tacho pulleys which are under the carry side of the belt, loading conditions can affect the conformance of the belt to the roller and the radius can vary with loading, easily 1% to 2%.
Belt stretch, pulley wear and tachometry
Belt stretch is an issue for tail drum tachometers and so is the wear of tail pulley lagging. But even if the tachometer device is perfect, belt stretch due to belt loading and ambient temperature can introduce significant errors. Steel cord belts can easily stretch 0.5% between the feed point and the belt weigher, synthetic fibre belts can stretch up to 3% depending upon how far the belt weigher is from the feed point.
The answer might be to mount the tachometer under the carry side of the belt at the belt weigher location, but its important to take into account any load sensitivity of the tachometer device because big errors can go undetected.
It is an interesting state of affairs that belt weighers are calibrated when the belt is empty and are never checked while material is running, unless a live load test is conducted. If there are concerns about the calibration of a tachometer it should have its calibration checked while material is running.
Alternatively, a real time tachometer monitoring and calibration system such as the CST Automatic Tachometer Calibration System could be employed.
Automatic tachometer calibration system
CST has developed an ‘Automatic Tachometer Calibration System’ (ATCS) which is optionally available on all of its belt weighers. The ATCS system involves a fixed ‘reference distance’ along the conveyor structure between two ‘belt mark’ sensors and a marker in the belt. This is usually a pair of magnetic sensors about 100m apart and a magnet in the belt.
The ATCS uses software in the CST weighing Controller to monitor tachometer calibration real time and to adjust tachometer calibration to account for error sources. When the ATCS is implemented all the weaknesses of the tachometer system become obvious in a log file, it catches all the variations in tachometry. Real Time logging is the ultimate way to see if a system is working well or not, and it can sometimes be quite disappointing!
Belt stretch is one of the hardest errors to compensate. That’s because stretch is proportional to conveyor loading as well as temperature, the belt is continually stretching and relaxing depending upon the flow rate on the conveyor. We have seen 0.5% variation in tachometry, even on a low stretch steel cord belt.
The ultimate solution to this is to put a high frequency tachometer right at the weigh frame in contact with the carry side of the belt and to have a relatively short reference distance around the weigh frame. Then with multiple marker magnets in the belt, the tachometer can be tuned multiple times per belt revolution to respond to belt stretch. (and this capability is already built into the CST system).
Tachometer calibration solved
Using the CST ATCS provides a window into tachometer performance which has previously not been possible. It has been said that ‘We don’t know what we don’t know’ and this is very much the case with tachometer calibration and belt weighers in general. There is evidence to show that tachometer calibration changes of 1% to 2% can occur for a large number of reasons, on belt weighers that are meant to be 0.25% instruments. It is no longer good enough to simply calibrate a belt weigher tachometer and hope for the best. The ATCS reveals the performance of the tachometer and best of all, it makes continuous adjustments to reduce tachometer error practically to zero.
Written by Ian Burrell, managing director of Control Systems Technology.