[Dividing Line Image]



by Dom Ruggeri

December  2001:

Metalworking Fluid Formulation 10 Setting Specifications, or Confessions of a Chemist by Dom Ruggeri
Some of the best R & D Chemists I know cut their teeth in Quality Control.  It will never stop puzzling me why every company I know places their recent grad hires into QC with little to no training.

As I once said, my first industrial position was in QC, and I admit I made my share of blunders.  It was a Saturday afternoon I had been on the job 2 weeks.  The plant had to work overtime so someone was needed to QC the batch samples.  Things were going well, every sample thus far had passed muster.  About 12:30 PM the Master-Compounder brought in a sample of Honing Oil.  This particular product always needed a viscosity adjustment and today was no exception.  I never understood why we couldn't change the specifications to reflect what the plant was capable of producing, or modify the product and incorporate the viscosity adjustment into the base formulation.  Either solution made perfect sense to me because these constant adjustments wasted time and more importantly money.  I tested the sample and sent the viscosity adjustment out to the plant floor.

 After lunch the sample came back into the lab I was shocked.  The bright red color was now amber and the viscosity was out of specification.  I didn't know what to do.  I always double checked my calculations and recorded them in the adjustment log.  How could this happen?   I had only one option.  I called the chief chemist at home and explained my situation.  He said he would be there in 10 minutes.  The longest 10 minutes of my life passed before he finally arrived.  We went through my reasoning process and my calculations.  Everything checked out until we looked at the batch sheet.  The adjustment should have been 105.5 pounds.  I had written 1055 pounds.  I was sick.  That one forgotten decimal point was going to cost me my job.  I looked at the chief Chemist and said, "I guess I’m done." His answer was not what I expected.  He asked me if the compounder had questioned my adjustment.  He didn't.  “Well then it is not entirely your fault.” He replied.  He also told me that I was lucky since this product was a good seller and would not be around long.

But now I am more confused.  He explained that my adjustment was bigger then the batch size.  I had to agree since it was in my handwriting.  It was my haste to get the adjustment to the plant that caused my error.  His voice was calm as he explained always remember to not only check what you calculated, but make sure you check your writing.

The Chief Chemist called the compounder to his office and asked him why he made an adjustment that was larger then the batch size.  His answer was I do what the card says.  We doubled the batch size and that corrected my error.  What saved my tail was the compounder never questioned me.  He had over 10 years experience and should have known better.  I got a minor reprimand and for the next week the chief chemist checked my adjustments.  As far as I am concerned, I got off easy.

We have all seen this phrase on technical data sheets: "These are typical values and do not constitute a specification." Ok, so what does constitute a specification?

Those of you who have been following this series know we have formulated a masterpiece, passed the field trial, and are ready for a full production run.  The rub, as always, is what QC tests will determine that our product was made correctly.  Further, what specifications should we place on those tests; as of now we only have laboratory and pilot plant production batches?   Since these batches were made using very accurate weighing tools, the QC values should agree within operator error.

Now from direct percent by weight we are using gallons per inch (GPI), weight by difference, and pumping drums placed on a plant scale that is constantly moved around.  In short, a calibration nightmare.  After all this, we get a call telling us out masterpiece is out of specification.  But is it?

We could have the greatest product ever conceived but if our factory cannot produce it, we have nothing.  The specifications are the only way for QC to know the material is made correctly.  A good rule of thumb is to take your laboratory and pilot plant values, average them and set initial specifications to plus/minus 10%.  For example:

Average Acid Value (3 laboratory batches, 2 pilot plant runs) = 2.3 meq/gm

Initial Specification:  2.07 - 2.53 meq/gm

You probably think this is too wide.  I agree, but it is a start.  Reality is that the specifications are a statistical tool that is always being measured and always being refined.  Manufacturing specifications should not be set until you have 5 production runs completed.  After 5 more runs, you have 10 data points.  These specifications should be reevaluated using statistical methods.  Use these same methods again after 15 production runs now sets the final manufacturing specifications.  After 15 production runs you should have a very good idea of what your plant can produce.  Still, the specifications should be reviewed every 6 months.  Time consuming, yes but it is time well spent.

To all of our readers:
I wish you and those you hold dear a happy and peaceful holiday season.

Good Luck Dom