William L. Gibson, Assistant Chief, Standardization and Quality Assurance Branch (Quality Assurance) , Memphis, TN
Since the enactment of the Smith-Doxey Amendment in 1937, which opened the door for the government classification of cotton samples for acceptance into the CCC loan program, the Cotton Division has maintained a Quality Assurance program to ensure accurate classification results and has actively worked to improve on the quality of services provided to our customers. The USDA, Cotton Division is responsible for providing cotton classification information to the cotton industry. Providing our customers with high quality classification data is a prime objective for the Cotton Division. Since 1980 when the Division began its HVI operations, we have gone from providing HVI quality data for 300,000 bales to over 18,000,000 in 1997. This large scale testing has resulted in various testing procedures being changed, omitted, incorporated or added. The Cotton Division is constantly implementing stricter and more rigid testing procedures to improve the accuracy of our classing information. Our objective is to continue to provide high quality fiber measurements and grade determinations; operate a cost efficient operation by making maximum use of available resources; and being that cotton is a commodity whose value can change rapidly over time, classify each sample as quickly as possible while still maintaining the quality of the classing service.
The Cotton Division currently operates 13 classing offices located throughout the Cotton Belt from California to South Carolina. At peak operation, approximately 240 HVI lines and 350 classers are utilized in these offices to provide efficient and accurate classification data. The primary responsibility of the Quality Assurance Unit (QA) of the Division is the technical supervision of our classing programs. In order for the classification data assigned to a given sample of cotton to be accepted by the entire cotton industry, be it growers, merchants, ginners, spinners or researchers, it is imperative that the data be at a uniform level based on the official standards.
Measures utilized by classing offices to ensure quality assurance include pre-season precision and accuracy testing for HVI's, environmental controls and conditioning, calibration and in-house checks and checklots. Pre-season precision and accuracy testing includes each HVI line in every office passing an evaluation. Cotton with known values is tested a multiple of times on each instrument. All instruments must demonstrate precision and accuracy of known value cotton. HVI Lines are not put into production until they successfully complete all preseason precision and accuracy test.
Environmental controls are necessary because some of the fiber determinations are moisture sensitive (primarily strength, mike, and length). All USDA laboratories and receiving rooms are maintained within the ASTM standards for temperature (70 degree F +/-1.0 degree) and humidity (65% relative humidity +/-2%). The cotton samples must fall within the range of 6.75% - 8.25% moisture prior to HVI classification.
Cotton samples received for classification are often outside this allowable moisture range and must be conditioned. In the past the samples had to undergo 48 hour conditioning to ensure that each had reached equilibrium within the prescribed moisture range. Today, in most Cotton Division facilities, Rapid Conditioning Units (RCU's) are used to quickly condition samples for testing. The RCU is a wire mesh conveyor connected to a plenum through which ambient conditioned air is drawn by a high-speed fan. Samples passed over this plenum can usually be conditioned within 10 minutes.
Calibration measures include using known value cottons to calibrate HVI lines for micronaire, length, strength and length uniformity. The colorimeters and trash meters are calibrated using tiles. Classers "calibrate" themselves by reviewing and referring to the Universal Standards, grade boxes prepared annually to represent all physical grades and all leaf grades.
Control charting and in-house checks are utilized in conjunction with calibration to verify such is maintained at a minimum of every two hours. In-house checks verify that the line can reproduce the test results of known value cotton. Process control charting (PCC) utilizes statistical sampling techniques to determine whether or not the process is in "control". If either in-house or PCC checks indicate that the line is not testing correctly, the entire process is checked for error, i.e. sample conditioning, operator error, HVI malfunction. If none is found, the line is re-calibrated.
In order to maintain a consistent level, approximately 1.0 percent of all cotton classified in each Classing Office is randomly taken for retest by Quality Assurance. This cotton what is known in the Cotton Division as a checklot sample. These checklot samples are samples randomly selected by a mini computer after classing information has been assigned for the sample. The classer is alerted to save the selected sample, which is then submitted, to QA. No quality data can be changed on the checklot samples after the mini-computer has selected it.
By re-classifying the exact sample used at the classing offices we are better able to determine where there may be problem areas and where to concentrate on the elimination of such. With the monitoring of all checklot samples and interpretation of standards being handled by the same group of HVI lines and permanently employed Cotton Division supervisors, we are better able to ensure that samples are classed uniformly throughout the entire cotton belt. Or in other words, instead of 13 different supervisors in 13 different offices determining what he or she interprets to be a color grade 31 for that given area, you have a group of 4 -5 supervisors looking at cotton harvested and ginned from all over the cotton belt ensuring that the samples are graded the same division-wide.
When QA notices that an office, a shift within an office, particular classer or instrument is having a reproducibility problem or a tendency to classify cotton too easily or hard, immediate interaction with the Classing Office Area Director takes place. In instances where a problem is considered of sufficient importance, staff members of Quality Assurance are sent to given offices to assist in its correction.
The Quality Assurance Unit not only keeps each office informed on their reproducibility results, but also maintains an ongoing comparison of classer reproducibility and HVI supervision results for the Division. This keeps all key personnel abreast of how well we are performing in comparison to the previous years as well as how we are performing in regards to our goals established prior to each season.
In our effort to ensure international uniform testing, the Division also conducts a HVI Check Test and an International Level Assessment Program. This allows the USDA to work cooperatively with Domestic and International Associations in pursuit of uniform HVI fiber quality measurements. These associations are provided a comparison of their HVI measurement results obtained from the same sample for the seven HVI quality factors. We have found that this results in more uniform testing and reduced HVI variability throughout the world cotton industry.
The Cotton Division's HVI check test is a monthly program where samples are sent each month to participating laboratories in the program. These laboratories test the samples for fiber property measurement by HVI and fax the testing results to the Branch. Once the testing has been completed, the participants receive monthly reports on the actual established values of the samples, the averages of results of all participants on each of the two samples and a comparison of their individual results to the established values. About 50 international laboratories and 40 United States laboratories currently participate. The participants calibrate the HVI instruments to ICCS (Micronaire Only) standards for Micronaire and to HVI calibration cottons for lengths, length uniformity and strength.
Under the International Level Assessment Program an organization will test and send a lot of cotton consisting of ten to twenty samples to QA. A report of the organization's test results compared to QA's test results, showing reproducibility and bias for each fiber measurement, is returned to the organization as soon as testing is completed. A cumulative report for the year is also returned to the organization. This program allows international organizations to test cotton samples at the same level as the USDA, much like our check lot program assures our classing offices test at the same level. There are currently four organizations participating in the program.
While these are not all the Quality Assurance measures employed to provide the highest quality testing and classing services strived for, they are some of the most important. With these measures the Cotton Division has taken throughout the years to maintain and improve quality assurance it is no surprise why we have been at the forefront of much of the modern technology and changes that are in use today in cotton classification. As we continue to enhance the quality of our services and cotton classing and HVI data become more repeatable and accepted, we are confident that the art of cotton classification will become the science of cotton classification. |
