TECHNOLOGY FOCUS

CENTRAL SYSTEMS -  PART 6

POUND OF CURE v. OUNCE OF PREVENTION

by Dom Ruggeri

NOVEMBER 2002:

I was working for a small formulating house back in the mid nineties.  This house specialized in coolants for the silicon slicing industry.  Now this cousin to the Metalworking Industry has two distinct areas of operation; the clean side where there are clean rooms, people in white coveralls using very high tech tools, and a dirty side, where the ingots of silicon are sliced into wafers.  We sold to the dirty side.  This particular operation was a chemical manufacturer’s dream.  The fluid was diluted, used, and shot down the tubes.  Ten to Twenty drums of neat coolant was considered normal usage for these people.

The products ran trouble free.  One pass and away; how much trouble can that cause?  We had just finished field-testing our flagship product at a very large wafering account.  All went well and we got our first 20-drum order.  We were quite pleased with ourselves.  The account finished their stock and switched over to our material and the trouble began.  Saw blades began to blow apart.  Wafer surface flatness had gone to pot causing more lapping time to achieve the desired flatness.  Everything that could go wrong did.  As was usual, the powers that be at my end said we were sabotaged.

Now we all have heard the stories of how sabotage was always happening in the old days, but that was past.  I found it hard to believe that it could be happening in the 1990’s.  The national sales manager and I flew out to the account.  By the time we got there they had switched back to the competitive product.  I considered this reasonable as production had to continue, but the sales manager was livid.  He could not understand how our material ran so great in trial but fell flat on it’s face in production.  When we spoke to the production engineers, they were less then cooperative.  Out and out hostile was more like it.  Basically they told us to get our coolant and get out.  I was not convinced the problem was entirely our fault.

I managed to convince the production engineers to work with me to resolve this problem.  We began by reviewing how the coolant was controlled during the trial:

1) Isolated five ID saws from main system
2) Set up a mix tank and a feed tank to supply the above saws with fluid
3) Charged the coolant by hand using graduated cylinders
4) Added the coolant directly to the mixing tank from there to saws

Now when they changed over to our product, they took a wand and went from the competitor's drum right into ours.  The problems began about 2 hours after they began to use our stuff.  I found out later that it would take 2 hours for the competitive product to completely clear the lines.  I checked out the proportioning unit.  What I found was a maze of pipes leading from the drum to a proportioning unit, this bit of technology did everything but change the drums.  It was a masterpiece of technology designed to control concentration at any temperature without human intervention.  Yeah, right!

We changed from the competitive product to ours once again.  This time I took a sample of our material.  I went to their chemistry lab and prepared a concentration control graph (I offered them one during the trial, however, they had faith in their equipment).  When two hours had passed, we began to see minor problems.  I checked the color of our fluid as it sprayed on the blades, and it looked weak.  I took a sample and checked the concentration by titration using the graph I generated moments ago.  The concentration should have been 500 ppm or about 0.05% it was in reality 250 ppm or about 0.025%.  I presented these data to the engineers, who immediately rejected it; their proportioner would never be that far off.  We investigated the problem and they were right.  However, this marvel of technology did not take the fluid viscosity into account, our fluid was about twice as viscous as the competitor’s product.  Fortunately this was a simple matter to correct with the appropriate human intervention.

We now have the perfect coolant in our system; it is running as the salesman said it would.  So how do we keep it perfect?  System control is the answer.  Work closely with your coolant supplier to ascertain just how to control your particular coolant system.  As I always say, they are the experts where their coolant is concerned.  Let them guide you on how to best keep the coolant running well.

Most shops I have visited have the QC group maintaining the coolant as well as doing the necessary testing on their finished goods.  So what tests and what records should the end user keep on their coolant?

Step 1:
Since this is a central system, you only need to keep records for each system.  These records would entail concentration by various methods and any other tests you are able to run as recommended by the coolant supplier.

Step 2-Testing:
There are many concentration control tests; some very sophisticated, others no so sophisticated.  For me personally, the easier the better.  The following are a few simple tests that have never failed me:

1.  Refractometer:
Once a shift is generally enough, this test is a quick measurement of coolant concentration.  Your coolant supplier should have concentration control graphs for your coolant.  If not, make up a fresh solution at the concentration you’re shooting for.  Measure that, and check the coolant in the system.  Adjust accordingly.  Remember, tramp oil will cause inaccurate readings

2.  Titration:
If possible run this once a day, however at least once a week will suffice.  This test measures the water-soluble components of the coolant.  The concentration is a bit more accurate than the refractometer, however if someone dumps anything that has excessive alkalinity into the system it will interfere with this test.  Your coolant supplier should have the concentration factors for this test.

3.  Acid Split:
Assuming the coolant contains oil, another test is the acid split.  This measures the lubrication content of the coolant.  Once again, your coolant supplier should have the necessary charts to run the test.  This test should be run at least once a week to be sure the lubrication of the coolant is not compromised.

A simplistic broad brush approach of course.  However, every end user should partner with their coolant supplier.  This partnership will yield handsome dividends down the road.  Until next month thanks for reading and if you have any questions please feel free to e-mail me at the magazine.

Good Luck
Dom