The condenser is first thoroughly dried out, particular care being given to the outside of the tubes and the bottom tube-plate P. Water is then circulated through the tubes and chambers for an hour or two, after which the pumps are stopped, all water is allowed to drain out and a careful examination is made inside. Any water leaking from the tubes above the bottom baffle-plate will ultimately be deposited upon that plate. It is essential to stop this leakage if there be any, otherwise the condensed steam measured during the consumption test will be increased to the extent of the leakage. A slight leakage in a large condenser will obviously not affect the results to any serious extent. The safest course to adopt when a leak is discovered and it is found inopportune to effect immediate repair is to measure the actual volume of leakage over a specified period, and the quantity then being known it can be subtracted from the volume of the condensed steam at the end of the consumption test.
It is equally essential that no leakage shall occur between the bottom tube-plate P and the tube ends. The soundness of the tube joints, and the joint at the periphery of the tube-plate can be tested by well covering the plate with water, the water chamber W and cooling chamber having been previously emptied, and observing the under side of the plate. It must be admitted that the practice of measuring the extent of a water leak over a period, and afterward with this knowledge adjusting the obtained quantities, is not always satisfactory. On no account should any test be made with considerable water leakage inside the condenser. The above method, however, is perhaps the most reliable to be followed, if during its conduct the conditions of temperature in the condenser are made as near to the normal test temperature as possible. There are many condensers using salt water in their tubes, and in these cases it would seem natural to turn to some analytical method of detecting the amount of saline and foreign matter leaking into the condensed steam. Unless, however, only approximate results are required, such methods are not advocated. There are many reasons why they cannot be relied upon for accurate results, among these being the variation in the percentage of saline matter in the sea-water, the varying temperature of the condenser tubes through which the water flows, and the uncertainty of such analysis, especially where the percentage leakage of pure saline matter is comparatively small.