Document jmO3L2oV8mQmXwvggYYLYKnp5

ENERAL ELEC1 COMPANY JOHN STREET, HUDSON FALLS, NEW YORK 1283? . . . AREA CODE 518, TELEPHONE 747-3341 INDUSTRIAL AND POWER CAPACITOR DEPARTMENT February 8, 1973 Mr. J. R. Kuster Monsanto Chemical Co. Sauget, Illinois 62201 Dear Ron, Enclosed, please find control charts on water in tank car samples of Arochlor. You will note a fair agreement between our two labs and a low water content (averaging about 20 PPM) during the first few months of 1972. Start ing about in May, we began to read consistently higher than you and after midJune, we got no readings under 25 for 3 months and 5 of those readings were above the specification value of 35. At the time, I was uncertain as to the effect of our sampling so did not become too alarmed. However, since then matters have worsened. Since sometime in October, we really began to diverge. Your measure ments indicated decreasing water while ours continued a slow climb. Counting the 47 PPM on lot KC01-707, reported to you by phone, but not shown on the chart, four out of the last six readings have been above specification and all values have been rather high. Before you can take meaningful corrective action you obviously must be certain of your measurements. You have indicated a willingness to take appropriate action and I appreciate your prompt attention to this matter which at the moment does not appear to be serious. We will be happy to join in the inter-laboratory round robin with you to aid in an early solution to the problem. Our Engineering people have been alerted and will be kept appraised of the progress made. Very truly yours, GENERAL ELECTRIC COMPANY ep att. cc: RJ Boudreau LL Barker CR Zecchini R. H. Dewart, Manager M. & P. T. Laboratory r STLCOPCB4070182 DSW 310196 O* STLCOPCB4070183 Monsanto _ " J. R. Kuster - WGK FROM (NAME ft LOCATION) DATE . February 2, 1973 SUBJECT REFERENCE Aroclor 1016 - Water Content General Electric Co. Hudson Falls, New York H. J. Horner cc H. H. Munch - TIB P. G. Benignus - B2S J. R. Savage - B2SL G A. Hippe w. C. Engman J. K. Grant Bob Dewart of. GE at Hudson Palls, N. Y. called me early In January and said that he had found some of our 1016 cars out of specification on water. His samples were taken from the tank car on arrival at his plant. ^ Monsanto results are on samples taken from the tank cars upon loading prior to shipment from WGK. A tank car sample is forwarded to GE at the time of shipment. I asked Bob if he would check some of the advance samples to ascertain if we had an analytical problem or we were experiencing some-water pickup in the cars in transit. His attached letter indicates we have an analytical disagreement between our laboratories. , . GE uses a Photovolt Aquatest II in making their . determinations. WGK uses a Karl Fischer titration to an electrometric "Dead Stop" endpoint; which has been our method for many years. - We then ckecked some of our reserve samples using the same lots GE had checked on our advance samples.' Our . results using three different "Dead Stop" instruments checked our previous findings with answers ranging from 15 to 20 ppm, which are well within our specification range of 35 ppm maximum. We sent a portion of these samples to Bill Haynes at Corp. Res. who ran them on*an "Aquatest IIM. His results agreed Identically with those obtained by GE of 32 and 33 ppm. We did not have a sample of the Lot 7015 which was over specification limits. (See note on GE tabulation sheet.) A meeting was held on 1/31/73 to discuss this situation. Those attending were K. Horner, R. Kuster, J. Grant and B. Engman. It was decided to take samples and have them run by both methods at different locations to statistically determine if the two methods gave biased results. qSNN s| . REV. 1/72 STLCOPCB4070184 H. J. Horner Aroclor 1016 - Water Content Page 2 WGK 2/2/73 On 2/1/73 I called Bob Dewart at GE, Hudson Falls and Stewart Kernaghan of GE at Pittsfield, Mass, to see if they would participate in a round-robin testing procedure. They both agreed to help.. Jim Gloeckner, Lab Statistician, informed us that ten samples would give us enough data to provide a "t" study. ' The samples are currently being prepared and will be sent to the various locations by Monday, 2/5/73. Each laboratory will receive a sample taken from the same reserve sample bottle of a shipment made to GE at Hudson, Falls during the past ten months, making ten similar samples to each location. The WGK lab and the GE lab at Pittsfield, Mass, will make their determinations using the "Dead Stop" endpoint procedure. The GE lab at Hudson Falls and our Research Group (Bill Haynes) will use the "Aquatest II" for their measurements. - Vie will advise you of the results of this round-robin as soon as we obtain the information from the various locations. C7J. R. Kuster, dy Anal. Lab Supv DS* STLCOPCB4070185