pressure engine, and the load to be carried throughout the test; then to add successively the main journals in the engine to be tested, the condenser and air pump, the connecting rod and crossheads, the valve and valve-rod, and finally the piston and piston

rod.

The plan was in great part successful, but could not be entirely carried out.*

Table VIII. shows the distribution of the friction in this series of trials.

TABLE VIII.

DISTRIBUTION OF FRICTION.

Condensing Engine.

21 x 20 inches built by Lansing Iron and Engine Works.

There were several conditions during this test that were not uniform throughout; nor were they normal. For instance, in cases C, D and E cards 7 to 15, the valve or piston of the engine or both was in motion, and it was impossible to lubricate the valve, although the piston was lubricated without difficulty. The results in both cases, C and D, can be considered as relative only, not

* A serious accident was barely escaped after the air pump had been attached and while the connecting rod was detached from the crank pin. The vacuum caused by the working of the air pump, drew the connecting rod toward the rapidly revolving crank; the meeting of the two was heralded by a sudden and violentounding. The rapidity of the motion prevented any serious damage, until the connecting rod could be safely put out of the way.

absolute. Again, in case E, it was found impossible to prevent the action of the air pump on the piston; the work done was however shown by the vacuum gauge, and was equivalent, respectively, to a pressure of 9.82, 10.3 and 10.3 pounds, in cards 13, 14 and 15, or a work equivalent to 69.26, 61.12 and 61.12, with an average of 63.8 horse-power, over and above that due to moving the engine. The average of a large number of previous trials gave 7.13 H. P. as the total friction of this engine together with air pump and condenser; and this fact must be noted in considering cases C and D.

Combining these various observations, we get the following results: Friction on main journals equals

Engine complete, air pump at work, case E, gives 113.88 H. P.; without piston, air pump and unlubricated valves as before, gives us case D, 48.60 H. P. The difference, 65.28 H. P., includes work of air pump and friction of piston and rod. The work of air pump is 63.8 H. P., leaving for friction of piston and rod 1.48 H. P.

The difference between the total friction, and that on the parts, already found, is the sum of the friction for cases C and D, which must be divided in the proportion indicated by the observations. Table IX. is a summary of the deductions from this set of trials. of a condensing engine.

Co-EFFICIENT OF FRICTION.-The co-efficient of friction can be deduced with certainty only for the main journals of the engine; since there is a variation in pressure of piston rings, stuffing boxes, and in angularity of the connecting rod, which is, at least to a great extent, unknown.

If we call the co-efficient of friction, p the pressure on the bearings in pounds for engines light, and plus mean pressure on piston for engines loaded, c the circumference of the bearings in feet, n the number of revolutions per minute, fpcn will thus equal the "lost work" of friction; which has been determined in the previous experiments, and is expressed as horse-power; this is transformed to foot pounds by multiplying by 33,000.

Table X. shows the value of this co-efficient for the several engines tested, and Table XI. is a summary of all results.

TABLE X.

CO-EFFICIENT OF FRICTION FOR THE MAIN BEARINGS OF STEAM ENGINES.)

0.85

1500 3 2600 5

500 23.

6" x 12" Straight Line... *12" x 18" Automatic (L. I. W.)...... 3.70 7" x 10" Traction (L. I. W.)....... 0.68 21" x 20" Condensing (L. I. W.). 3.30 4000 51

.09 .04 206

* The 12" x 18" automatic engine was new, and gave, throughout, an excessive amount of friction as compared with the older engines of the same class and make.

9

In each case the engine to be experimented with was first examined, by the process which has been so fully described in the preceding papers on this subject, to determine whether its friction under varying loads was actually constant, as in the engines previously tested. This was found to be practically the case in every instance, and even the compound engine, contrary to the expectation of the writer, exhibited substantially the same internal friction at all loads up to its full rated power, and with no load at all. The minimum friction was 13.5 H. P., the maximum, 17.5, varying irregularly, with the character of the lubrication, probably, and giving the higher or the lower figure indifferently whatever the work done, and however the power might be distributed between the two cylinders. As this engine was non-condensing, the problem still remains to be solved with respect to condensing engines, unless, indeed, the few experiments thus far reported may be taken, as is very probably the fact, as indicating the truth

of the general principle of constancy of internal friction for all engines, whether condensing or non-condensing.

These engines were also all tested to determine whether the previously reported increase of internal friction with speed were here to be accepted as correct. It was found that the several engines differed somewhat in this respect, but that the variation was in all cases slight, and in some instances insensible or even reversed, the friction decreasing in one case, observably, with increasing speed. It was sufficiently evident, for all the engines here considered, that this variation was so unimportant as to be negligible. The figures given in the several tables which have been presented in the preceding pages are therefore to be accepted as not only correct and reliable, but also as not likely to be affected by construction or method of operation of engine to such an extent as to be inapplicable to steam-engines generally. The writer, in the light of existing knowledge, would assume that it is the rule, with all the usual forms of engine, and under all common conditions of operation, that the internal friction of the machine is practically invariable with variation of useful work, and that it is very nearly independent of the speed of rotation and of piston, varying slightly, as a general rule, in the direction of increase with increase of speed. This latter principle leads to the conclusion that the friction co-efficient of the rubbing surfaces decreases with the load on the engine and with increase of pressure on them, a result confirmed by numberless experiments of the writer and others, independently. With good lubrication, the coefficient of friction rapidly decreases with intensifying pressures, and to such an extent as to make the actual resistance to movement very nearly constant. It is now possible to study the reported data intelligently, and to deduce useful and reliable. Conclusions relative to the effect of these new facts upon the theory and upon the principles of designing and constructing as well as operating steam machinery.

The last table presented, summarizing the work of the whole research, gives in most readable and intelligible form the data and the laws which it has revealed. The most important item of friction waste, in every instance, is that of lost energy at the main bearings. In every case it amounts to one-third or one-half of all the friction resistance of the engine, the higher figures being given by the condensing, the lower by the non-condensing engines, except that the first experiment, with the Straight Line Engine,