Paint Color Standards

[Jessterish]

I'm still fairly new to the world of quality, but came across something the other day that I am slightly confused on, hopefully some of you nice people can help me out.

We have a number of colors of products (baked enamel on steel) that we produce. Our standard is LAB +/- 0.5 from standards. So far, apparently so good. However, our standards are coupons, that are then referenced back to another standard at the manufacturers and have deviations written on the back, which are apparently based off another "Master-Master." That is in turn not defined from another absolute value of some sort, but sent to purchasing on some basis that no one in my department is apparently actually able to tell me, and they approve (or reapprove) based on customer requirements and visual only comparisons. In fact our stock material has no individual final customer, so it's just approved by purchasing with no outside influence that anyone can tell me about.

This seems crazy to me? In doing comparisons, I can reliably pick out deviations of .7 between products (which would be in spec), and because we dispose of our out of date paint samples, we have no idea about drift across years. I mention because we have repairable products still in circulation after 10+ years. I haven't heard of a lot of issues with customers ordering things and complaining about it, but It certainly seems like it's just waiting to happen.

Am I being paranoid? I keep looking for answers, but the further I go with this, the more it looks like no one else cares about it at all.

 

[John Predmore]

What exactly is your question?

 

[ScottK]

I'm still fairly new to the world of quality, but came across something the other day that I am slightly confused on, hopefully some of you nice people can help me out.

We have a number of colors of products (baked enamel on steel) that we produce. Our standard is LAB +/- 0.5 from standards. So far, apparently so good. However, our standards are coupons, that are then referenced back to another standard at the manufacturers and have deviations written on the back, which are apparently based off another "Master-Master." That is in turn not defined from another absolute value of some sort, but sent to purchasing on some basis that no one in my department is apparently actually able to tell me, and they approve (or reapprove) based on customer requirements and visual only comparisons. In fact our stock material has no individual final customer, so it's just approved by purchasing with no outside influence that anyone can tell me about.

This seems crazy to me? In doing comparisons, I can reliably pick out deviations of .7 between products (which would be in spec), and because we dispose of our out of date paint samples, we have no idea about drift across years. I mention because we have repairable products still in circulation after 10+ years. I haven't heard of a lot of issues with customers ordering things and complaining about it, but It certainly seems like it's just waiting to happen.

Am I being paranoid? I keep looking for answers, but the further I go with this, the more it looks like no one else cares about it at all.

Color has been one of the most challenging things I've dealt with in my career in quality so I feel your pain.

In theory, whatever you are using to measure the Lab numbers is what you should be going by. Are you using a spectrophotometer?

I would look at how long you keep your comparison samples and see if that time frame is justified. And when it's time to make new samples what are you making them to - the Lab value graph or visual comparison to the old standard? And if you are doing visual do have light/target/dark coupons or just target?

 

[greif]

A very widely used color reference std is ceramic tiles made by Lucideon ( used to be called Ceram or sometime British Ceramic Tile Research Association (BCRA))
I purchased 3 oof these (white, mid gray, and deep gray) then had them calibrated by NCR in Canada. They provided k=2 uncertainty typically;
L* =0.18, a* = 0.04 b* = 0.04

On my xrite Exact I end up with total uncertainty and residual error of typ;
L*= 0.4 a*= 0.35 b* = 0.45

Xrite specs inter-instrument agreement at a max of 0.45 delta E*ab

So, I wonder about the accuracy you really have in reading 0.5 from standards, it you are just working from periodically replaced paint standards.

Maybe not important to you if you don't need to match up to the rest of the world.

 

[Jessterish]

I have several questions:
is this normal? not comparing across a length of time? is this a normal spec with the +/- .5 LaB range? (we do use a handheld spectrometer for that) Is it normal for purchasing to OK the color samples? Is it good to be comparing to a series of physical standards deviated instead of an absolute number set? It just seems imprecise and like we are just waiting for a major issue to arise. Frankly, I am surprised it hasn't yet.

 

[Jessterish]

Also, I don't know that anyone in my department is actually privy to how they do new paint samples. It seems that purchasing just does a visual comparison to previous sample, but I am not positive on that.