Advanced Colour Grading for Bread Flour and White Maize
It is a long time (1959) since the introduction of the Kent-Jones Flour Color Grader and while the equipment is no longer available, and the units in service hardly reliable, the system still persists today.
Based on fundamental CIE norms of the day, the KJ system offered a breakthrough in optimising milling properties with a quick procedure, rather than relying on long laboratory routines and the human senses. The single numerical number result was also easy to understand and to communicate.
By 2006 we started receiving inquiries if there was a better newer way to measure the same parameters. It would appear that as Colour Measurement had reached very high standards by this time, as well as a rising need for customers to communicate the values quickly and easily within a group of mills or customers, anomalies in the KJ system were being noticed.
When working in a single site environment, it is quite simple to compare values from one batch to the next, as more or less the same conditions hold. This is particularly true if a “Target / Sample” approach is used, where every time a measurement is taken, it is compared to a “Target” or already approved sample, under the same conditions. Unfortunately, the KJ system lent itself to reading the absolute numbers, regardless of the environment. By 2006 if not earlier, it was clear that the same colour was returning different values. This was due to varying environment, as well as an inadequate inter-instrument agreement. (IIA). In other words, when a known colour was measured at different times on the same instrument, or at the same time on a number of instruments, different values were returned.
Introducing the Solution:
To optimise this situation, an independent body, in the local scene the Southern African Grain Laboratory (SAGL), was requested to “Ring test” submissions from the various millers, and arbitrate or moderate these to each other, by an adjustment to the SAGL values, considered the norm.
This system worked quite well, but as millers looked to optimise production quality as well as measurement results, waiting for a ring test result was no longer acceptable.
At this time, we together with Konica Minolta looked to the similarly configured Tristimulus filter based Chroma Meter 400 series for a solution. While the KJ system used Hunter values to result a single value, based on the calculations.
- Whiteness = G-A+B
- Yellowness = (A-B)/G
Konica Minolta favoured an actual CIE Norm approved L*a*b colour difference result.
Trials were carried out locally, in the USA and Australia, and in regions trading with those countries, with quite good results, the outcome being that in the USA especially, the Konica Minolta CR-410 Instrument became quite popular. It had a much higher accuracy and repeatability, and although more expensive, more reliable results were obtained. Locally however, millers were reluctant to change. One of their valid objections is that while initially the IIA was excellent, being filter based, over time these the IIA values would decline.
In 2009 Konica Minolta launched a new Bench-Top Instrument, the CM-5 Spectrophotometer, designed especially for food processing laboratories.
With special features like freedom from a PC, local result printing (Also possible with the CR-410) and added accuracy, impeccable IIA and long term reproducibility, data export ability and user control functions, we decided early in 2010 to look for a customer in the Flour Colour Grading business to test it out.
Fortunately, we were partnered by an “Early Adopter” customer willing to pioneer a new system for White Maize and Flour Colour grading who agreed to test out our “Target / Sample” theory.
To this end, we arranged to loan them our CM-5 Demo Instrument, and patiently over nearly a year, this customer was able to create a workable colour grading scale, based on L*a*b values, ∆ a*b values (Difference to target) as well as investigate new Whiteness Indices approved by the CIE.
During this time, we agreed that ash cannot be accurately graded as a component of the flour. The instruments act like an eye, and what it sees is colour. The source of the colour is not discernible by the eye unless say red sweets and green sweets are clearly visible in a bowl. In a powder blend however, the colour is the colour but the ash contribution is unknown. We also had a closer look at the “Wet” test carried out by KJ.
While it is clear that dry flour and wet flour would differ in colour, we could not really find a reliable way to have a “Stable” wet sample to measure. Its contribution to the colour grading process was also dubious.
In 2011, Konica Minolta supplemented the CM-5 range, with a CR-5 Colorimeter. The CR-5 is somewhere between the CR-400 Series, but with the IIA an accuracy of the CM-5. Again, our pioneering customer decided to try a CR-5, as its lower cost and simplified single button measurement procedures would be suitable for production sites. This would take results straight from production to management, with the option of verifying top the master results of the CM-5. A six month trial was carried out to examine the IIA and Intra Instrument Agreement of the CR-5, and a 99.9% result was resulted from all the trials. The identical design and operation procedures of the CM-5 and CR-5 are the perfect Development Laboratory and Production Quality Assurance partners.
The customers completed scale was shared with SAGL, to be utilised by the milling industry as a whole, and to date the majority of the millers use this system to a greater or lesser extent, each focussing on their own IT. The procedure is also documented on the SAGL website.
We invite customers not yet using this system to contact us for assistance and support based on our long experience, and great success of this new process.
To date, we have never had a miss-identified colour grade reading, unless an error was present. The instruments themselves are largely self-diagnostic, but there is always room for Human Error.