commutator bars is the result, and if the number of commutator bars is arbitrarily increased in order to obtain low average voltage between bars, the armature reaction is correspondingly increased, and the distortion of the main flux becomes greater, resulting in a high peak voltage between adjacent commutator bars. As has been shown in the compensated machine, there is no distortion of the main pole flux, consequently with the same degree of safety, a higher average voltage between commutator bars is permissible. That is, with the same peak voltage between commutator bars, the compensated machine can have a higher average voltage between bars than the commutating-pole machine, hence, for very high speeds the compensated machine makes higher ratings possible. For the same reason, the compensated machine makes higher voltage machines for a given speed possible. The pole-face windings and the necessary connections for these introduce a certain amount of complication, which is inherent to the compensating windings. It is a fact that the compensated machine is not as easily dismantled and not as easily repaired as the commutating-pole machine, and for pure simplicity of construction and minimum number of parts we must admit that the old non-commutating pole machine is superior to either of the later types. Hence, the better performance is obtained at a sacrifice of simplicity, and a universal application of compensated machines would be as much of an error as would an attempt to apply non-commutating pole machines to all classes of service. Read Before the National Electric Light Association PUBLISHED BY ORDER OF THE NATIONAL ELECTRIC LIGHT ASSOCIATION NEW YORK CITY REPORT OF COMMITTEE ON UNDERGROUND CONSTRUCTION AND ELECTROLYSIS The report of the Committee on Underground Construction and Electrolysis for this year deals with the subject under the following heads: • Current Carrying Capacity of Cables Cable Specifications for Various Classes of Service Use of Split-Conductor Cable New Design of Three-Conductor Cable Special Devices and Methods As has been the practice in the past, the material for the report has been gathered from engineering literature and from the experience of member companies, and the Committee hereby makes acknowledgment of their assistance. Recently the American Committee on Electrolysis submitted to its principals a preliminary report presenting the subject of study of the several sub-committees. This report presents only such statements of facts as its members could at this time unanimously agree upon. President Wagner has referred the preliminary report of the American Committee on Electrolysis to the Technical Section for review by the Committee on Underground Construction and Electrolysis. The document is of importance, first, because it is an' aggregation of facts accepted by both the railway representatives and the representatives of the pipe and cable owning interests, and, secondly, because the material contained in it is intrinsically excellent. Much of the material includes first-hand information of the most illuminating and practical character for the solution of electrolysis problems'. The subject of electrolysis is of interest to practically all member companies and it is recommended that the report receive Manuscript of this report was received March sixth. careful study with a view to making comments and suggestions to guide the American Committee in formulating its final report. CURRENT CARRYING CAPACITY OF CABLES The question of the safe current carrying capacity of cables has been under consideration for some time past, and the need for definite information on the subject has been felt more acutely on account of conditions which have recently developed in the central station industry. The increase in price of copper, which has accompanied the existing condition of prosperity, and the increases in industrial load on the central stations, have made it imperative that cable systems be operated at the maximum load consistent with safety. Most operating companies have come to realize the importance of this subject and there are indications that a uniform system of rating will soon be developed. Some information on the practice of member companies has been obtained from answers to a Questionnaire recently sent out. A number of companies experienced trouble during the summer of 1916, due probably to dielectric losses occasioned by high temperature. The trouble appears to have been confined to sections of duct lines where local conditions impeded the dissipation of heat, and to cables operating at voltages above 10 000 volts. The overheating of duct lines may be remedied by supply-· ing water around the duct structure in locations where the soil characteristics are such as to interfere with the conduction or heat. Note The Committee on Underground Construction and Electrolysis has made a careful study of this report and will present its recommendations and criticisms to the Technical Section at the Annual Convention in 1917. It is a well known fact that the load factor is of importance in determining cable ratings, and experience has shown that approximately 40 000 kw-hr. per day is the maximum energy which may be safely transmitted over a 4/0, three conductor, 13 200volt cable. This corresponds to a load factor of 73 percent for a maximum current of 100 amperes, and 37 percent for a current of 200 amperes. As pointed out in previous reports, the carrying capacity of cables is affected by the temperature attained by the insulation. The effect varies with the impressed voltage, but as yet very little information has been obtained which will give an accurate idea of either the variation of the losses with voltages, or the actual magnitude of the losses themselves. Experience has shown however that the trouble is not serious below 10 000 volts, a fact which is borne out by an examination of the cable failures which may be attributed to excessive operating temperatures. Recent tests made by the Consolidated Gas Electric Light & Power Company of Baltimore, bring out some interesting facts. The tests which were conducted on different specimens of cable laid in a duct line constructed to approximate closely actual operating conditions, showed that dielectric losses must be taken into account in arriving at a rating for underground cables. The results appear to establish the following facts: I At low temperatures the type of insulating com- 2 At high temperatures the type of compound 4 The magnitude of the loss may be decreased by |