2 Months to Go and We already failed one of our Goals :(

Hello All -

We had set a goal to have less than 1.8% of a certain type of internal rejection.

Despite our efforts, we have failed this goal. Even if we go the next two months without an internal rejection, we still fail.

What should we do next? I know a corrective action will be needed. But I want to attack this issue correctly.

Thoughts?

 

we produce plastic pellets for Medical companies.

Our goal of to reduce color variation internal rejections to less than 1.8%. We figured out that our blending of the colors process is not good enough. We have tried a few different blenders but nothing seems to work. We are again waiting for another blender to show up for us to see a demo.

 

I don't know, but it sounds to me you're on top of it. Trying different blenders is no easy task.

 

Is that "Another blender" the correct attack for this issue ??
What is different in that blender than the few different blenders that you have tested out ?
If nothing significant, than the challenge is elsewhere.
What is your current internal rejection percentage, which you want to hit to 1.8 %
Its your company and your objective. You are free to reset your objective, based on data you have.
If the target of 1.8 % an error, admit it, and justify the new objective target.

 

The color variation is within the batch. So a mixture of colorant is produced. It is fed into the extruder. The job may start off the correct color and very inconsistent - and then suddenly BOOM!! The feeder will feed in some of the mixture where there are clumps of colorant which were not blended evenly. The result is pellets that have a "salt & pepper" color - very light and dark.

Our customer's use very small extruders to process our material, so color consistency is a major key to them.

 

The 1.8% percentage was determined by reviewing the rejections from 2014.

During our Management Council Meeting we reviewed the metrics and all of management, including process and production, decided that this was the major cause of our down-time and honestly we are losing money having to correct each one of these rejections. The correction method is taking the material and reprocessing it so re-blend the color.

In the previous years - we did not have such a problem with color variation. This was due to our customers giving us more freedom on what colorants we could use. As the FDA cracks down harder and harder, we are limited on the colorants we can use and unfortunately, a lot of the ones we can use can't be blended that easily. Most of the time now, our customers tell us the colorants we MUST use.

 

It does seem to the blender in our eyes.

We have a larger blender that works great! But unfortunately, a lot of our feeder batches are too small for it to blend the mix correctly and evenly.

So we have been on the hunt for a smaller blender. The manufacturer of the large blender we have does not produce a smaller version.

 

We measure the color variation visually. It is extremely noticeable even from a distance that the pellets are different shades of color.

The whole lot has to be rejected as the color variation comes on suddenly and without notice. By the time it is noticed, we have salt and pepper pellets mixed in the with the "good" pellets.

The color variation occurs batch to batch - it depends on the formulation and the amount of colorants in each. We have process concentrating on reviewing the formulas before they go out on the floor. Process is included to try to adjust the process by slowing the line down to help incorporate the "clumped" pockets of color, but process doesn't always work either.

 

Nikki,
Beverly has published some Measurement Systems Analysis in the Resource section. Some of them are very practical to use. We have the same situation but ours is Bright and clear or Hazy. We use attribute MSA to test the operators. This could also be a nice little six sigma project for you and your team. The goal is to improve profitability by reducing defect. There are consultants in this Forum that can help you out. Hope it helps.

 

Is it possible to improve when it's noticed? I thought you might be able to learn more about the root cause if you could see it happen rather than not be able to until later.

 

Hi James,

Yes it is possible... As soon as it is noticed, correction is attempted - usually by process - once it is fixed, the containers "catching" the pellets are switched out.

Unfortunately, the color variation usually occurs at the end of the job - which forces us to correct the entire container of material.

 

Just a thought. Have you checked your mix ratios. Maybe you need a little more of one and less of another to get your current blenders to work adequately with the small batches.

 

FWIW -

"...Even if we go the next two months without an internal rejection, we still fail...."

Allow me just an elaboration on the unsaid side issue, as the experts here have given you plenty to work on for finding causes.

My first impression upon reading your post brought back ancient memories of when I worked in extrusion as well. Our company had quarterly goals for things like this including accidents, sales, output, etc. and if we met our goal there was a cash bonus at the end of the year.

One problem with this type of goal, when it is blown early, there is a terrible morale drop, especially if the goal is financially incentivized, which they sometimes are. For the remaining months a “what does it matter” attitude settles in, since nothing can repair the damage to a missed target, it is usually mathematically impossible.

While having measurable goals is great, they come with some negative effects. I have found it better in these cases to use the percentage number in a control chart, so instead of a term limited duration of time (quarterly), it is continuous data, and the number is simply taken, (like temperature), when needed.

This keeps everyone involved without feeling that “the next two months are wasted, we cannot hit the goal anyway”. Instead, using a moving number, improvement can immediately be applied and results seen without throwing out the next two months and then trying again.

Yes, I know there is no real difference quarterly-wise, the difference is in perception.

In the quarter, you have already lost, why improve? - In a “limitless” duration, the number can be affected immediately, so improvement is not “wasted”.

I know it is all semantics, but perception is everything, especially when more than a handful of people are involved.

 

Nikki,

I know it's not methodically sound to jump to "solutions" (before you know what the problem really is), but you already have a battery of analysis experts lined up here so I'll just add my 2 cents on the technological side.

Additive mixers tend to be a huge headache with small batch processes, especially with very potent colorants that are required in very low doses. One possible technological solution is to switch from pre-mixing to online dispensing. You need a very accurate dispenser for every additive, and they all feed directly into the extruder's hopper. It's an expensive solution but it provides great control. Might be worth it if you run lots of such small batches.

The fact that you get the variations near the end of the batch hints that this might have something to do with hopper dynamics. When the hopper is not full it changes and you might get segregation of the different ingredients due to their different physical properties. If this is the cause you just need to make some extra mixture so that you can have a full hopper all the way through.

One last thought is maybe you need a longer extruder, or another type of extruder / screw / barrel - one that mixes better and is thus less sensitive to the input material.

FWIW.

Ronen.