Debunking 106.1 mph (April Trains)

[oltmannd]

Semper Vaporo

 

But I would question if his value for Horsepower is the same as the 1903 value.  Horsepower has been re-defined at least once since then... 

Wut?  550 ft-lbs/sec.  Now and forever.

The AUTOMOTIVE industry changed how the RATE HP on automotive engines - particularly which auxiliary loads to count in or out  (water pump, oil pump, fuel pump, etc).  This had nothing to do with how and where you measure HP on a steam locomotive.  The two most common places to measure it are at the cylinder using an indicator (traces pressure vs. position - the area is the energy per stroke - times stroke rate give you HP).  The other is at the drawbar - force X speed  = power.  There was quite a bit of sophistication to these measurements - even way back when.

[daveklepper]

I think we can all agree that the article was overly dogmatic in its criticism, just as the original claims were dogmatic in their claims.   The exact truth will be never ascertained, because it would involve duplicating the experiment, and even if the track, locomotive, cars, etc. received exact duplication, there would remain the weather and the wind.

[Anonymous]

daveklepper

I think we can all agree that the article was overly dogmatic in its criticism, just as the original claims were dogmatic in their claims.  

If Hanke’s criticism was overly dogmatic, and thus perhaps inaccurate, why should we assume that the speed claim is exaggerated, false, or overly dogmatic? 

 

If Hanke is wrong, I see no reason to assume that the speed claim is wrong. 

 

There is no middle ground when it comes to debunking.  Either you debunk the speed claim or you don’t. 

[oltmannd]

Bucyrus

 

 

There is no middle ground when it comes to debunking.  Either you debunk the speed claim or you don’t. 

There really isn't an "is" or "is not" with this stuff.  It's just about trying to see how bright or fuzzy the lines are you can draw.  

Also, the burden of proof is generally on the "bunk"-er, not the debunker.  

"My Toyota Camry went 273 mph on my way to work today".  Can anyone prove that it didn't?

[Anonymous]

For the record gentlemen, having lived for many years in New Jersey, I can attest that the particular stretch of Jersey Central mainline in question has had a history of  very VERY fast trains.

To suggest out of hand that it is not possible is slightly more than disingenuous.

A little research on the Jersey Central on your part should be enlightening to you.

 

[Anonymous]

oltmannd

Bucyrus:

 

 

There is no middle ground when it comes to debunking.  Either you debunk the speed claim or you don’t. 

 

There really isn't an "is" or "is not" with this stuff.  It's just about trying to see how bright or fuzzy the lines are you can draw.  

Also, the burden of proof is generally on the "bunk"-er, not the debunker.  

"My Toyota Camry went 273 mph on my way to work today".  Can anyone prove that it didn't?

I understand your point.  The record claim itself is beyond the point of being proven or disproved.  So we are free to believe it or not.  But issue is that Hanke claims to have debunked the ATSF record.  That is an impossible burden, as you point out.

To me, the most interesting aspect of this controversy is why there should be such a desperate need to debunk the claimed record.  Would-be debunkers seem to be piling up large heaps of little uncertainties in the hope that a large enough pile will win their case.  So we end up with claims that watches did not have second hands in the 1905 era, or an inch was not as long as today's inch.   

[Stourbridge Lion]

[Anonymous]

The main point that I see is that 106 mph is not all that difficult to believe.  If the claim was 206 mph, debunkers would have a lot more to work with.  But splitting hairs over 106 mph more than a century latter seems like sour grapes.  It seems petty.   

 

Hanke hangs his debunking hat on the laws of physics, and concludes that a speed of 80-90 mph is all that was attainable.  However, on the previous page, the Professor of Applied Thermodynamics, Heat Transfer, and Applied Mathematics has set 100 mph as plausible.  But perhaps more importantly, he refutes the claim that the laws of physics can be directly applied to come up with a certain answer, as Hanke claims to have done.  

[h2fe9x2]

Dear Sir [Semper Vaporo], qualitatively speaking all arguments you invoke are valid and we could not agree more (with the exception of the so called horsepower standards). Yet one of the advantages of Physics is to allow us to quantify the arguments we use. In doing so one can discard or consider a certain physical dimension while modeling a phenomenon, and as consequence use a more expedite (and controllable) calculus procedure, and avoid unnecessary or redundant approaches.

So, commenting your argumentation, you failed to mention that the changing of standards (units of measure) with time was not a prerogative, or a consequence, of early 20th century science. It has occurred regularly since the advent of formal Science Academies, throughout the world, to the present day, and will continue to occur in the future. You should have questioned not the natural occurrence of standards adjustments but the magnitude of those changes and its influence on the matter discussed, and if you had quantified the matter properly the answer would be: to consider such tiny standards changes in this case would be absurd.

 

At the end of this post I present the references (sources) I used previously and also a few examples of the speed capability of some 19th century U.S. trains, as shown in several technical publications of the day (routine trains and not special “wild” runs).

 

But let us consider the influence of the following factors, in terms of equivalent rail horsepower gains:

(Equivalent to a “100 mph run” with the Scott Special on straight level track on calm day.) 

 

20 mph tailwind         -> +435 hp  (in accordance with Davis Jr. formula of 1926)

0.2% downgrade        -> +362 hp  (for a total train weight of 339 short tons)

1’’ wear of drivers      ->    -14 hp  (increase machine friction and back-pressure losses)

Sum                             -> +783 hp  (this is simply a scale value, or an estimate)

 

 

 

 

Yet the HP adjustment in a 1500 hp reading made in 1904, due to the changes of the U.S. Standards of weight and length made since then, would be +0.0032 hp today, or +2.4 W. The calculations and references supporting this value are given bellow. This is a relevance ratio of 0.0004% (percent), when compared with the sum of the 3 factors considered above!

As for the dimensions of the cars of the Scott Special, they were no different from the cars depicted in Davis Jr. formula. This correlation, coefficients included, proposed in 1926, was based on controlled dynamometer tests made since 1906 with ordinary trains. Those tests were so detailed that still today we can know the train composition, the car type and weight, the trucks axel numbers, the train length, the wind speed, etc. The description of the cars used on the Scott Special is clear: searching on the NET one can conclude that they were similar in weight and dimensions to those included in Davis Jr. formula.

In regard to the lubricants used, that’s a more pertinent observation, so obvious that it was partially studied also at the St. Louis Test Plant in 1904. What’s more, and you probably know it better than me, the theory of lubrication we use today emerged in the late 19th century with the Tower’s studies and the Reynolds Theory. The fact is that the variety of lubricants used on car journals of those days cannot be compared with the variety of lubricants at our disposal today for a diversity of applications. So that uncertainty was not that important, since the Davis Jr. correlation was proposed discriminating rolling and journal friction, flange friction and air resistance. So the uncertainty due to lubrication is simply one the factors contributing to the observed experimental variance with this correlation. I recommend that you see it for yourself consulting the reports of the thorough studies about train resistance made by Edward Schmidt, of the Illinois University, between 1908 and 1916 [see references below]. The easiest way to do it would be simply to look at the graphics and compare the adjusted curves with the “cloud” of dispersed experimental points depicting accurately measured values of train resistance, corrected for gradient and acceleration. When I enlarged that probable maximum speed interval I was thinking precisely on the uncertainty affecting the coefficients of the Davis Jr. correlation, namely: the roadbed stiffness and rail weight, that would affect rolling friction; the influence of lubrication depicted as journal friction; track condition and alignment that would affect flange friction; and also on cars dimensions that would affect the air resistance coefficient. So the uncertainty analysis mentioned, but not performed in the previous post, was not intended as an exercise of cynicism, by saying that anything could be possible given the uncertainty of this and that…  When a problem is too complex we have establish plausible boundaries considering only the major factors, even if accuracy is lost, or else we will be lost in the midst of a major chaos.

The fact is that the adjustments of the standards you mention are of no consequence in this case and its consideration in an uncertainty analysis (a statistical study) would be absurd. Note that direct changes of the horsepower standard [hp] never occurred simply because the horsepower is a derived unit and not a fundamental unit, i.e., it results from a concept/definition based on fundamental standard units of length, weight (in a gravitational measuring system), and time, the definition being a work rate equivalent to lifting 550 lb at a speed of 1 ft/sec, already mentioned by someone in this forum. As to the horse that was doing the work that question was settled 230 years ago and not at the beginning of the 20th century!, when the scientific community was considering the questions posed by the emergent Quantic Mechanics and the Theory of Relativity ...

That HP definition has not changed since then (but this was already referred by someone). What has been changing are the norms, or procedures, used to rate the horsepower of an engine and not the physical unit of measure. For example, if one takes a given engine’s truck with 200 hp, SAE net HP, present day rules, and test it in accordance with the gross HP SAE rules of more than 40 years ago, this same engine would display probably more than 220 hp in accordance with that old norm. But not because today’s hp is greater than hp used then, no! But simply because today’s SAE net HP norm demands to test the engine in more realistic conditions, and this affect the HP displayed by the engine, such as the use of realistic exhaust manifolds, the use of the air filter and belt-driven auxiliary equipment, etc. In the 60s the engines would be stripped in the laboratory of all these components (except for those vital to engine function); doing the same today this same engine would be freer to run and would produce more torque and more brake HP, in the same measuring unit, the hp. Suppose that in the future a SAE norm demand to rate the HP of a car or truck in accordance with the measured wheel HP: then this same present day 200 hp, SAE net, will be probably less than 180 hp “SAE wheel HP” of the future (because the transmission losses would be considered in this case). Also the HP declared by internal combustion engine builders are conventional values correct for standardized atmospheric conditions (of pressure, temperature and moist), being perfectly normal to obtain during controlled tests horsepower readings within ±5% of the rated (expected) HP. For instance, if you take your car and test it over rollers on different places you will verify by yourself that although the engine is the same the HP is not; also you will obtain slightly different values from summer to winter even if one uses the same testing station. To that one can call informally uncertainty of the HP reading, and this is as true today as it was in those days, only the accuracy is better nowadays. But even so, one has to know what to compare: measurements made in a laboratorial environment, even if made in 1904, would be considerably more accurate than measurements made today in an automobile maintenance workshop. The point being, to believe that in late 19th century the mankind was in the stone age of science and engineering is nothing more than a sad display of ignorance. And that idea is implicit in Mr. Hankey argumentation when he suggests that they were unable to take proper time measures in those days…

 

Yet, these different norms, used to rate the HP of an internal combustion engine (SAE, DIN, ISO, EN), are not applicable to a steam locomotive because the physical dimension used to measured HP in such a machine already states the procedure used to obtain that power (in the same unit of measure, hp), since IHP means Indicated Horse-Power, or the mechanical power measured by an indicator in hp. And indicator is a device that allows the real-time recording of the pressure developed in the cylinders as a function of the piston position, allowing in this manner to compute the work developed by the steam against the pistons. So a 1600 IHP steam (piston) engine is a machine capable of a net piston horsepower of 1600 hp. Note that the IHP of gigantic marine slow-diesel engines is measured still today during engine testing, but such a thing is not practical with smaller and faster engines. Also worth noting that the first indicator tests performed with stationary steam engines were made more than 150 years ago, being an amply proven scientific procedure by the year 1904.

Another aspect I would like to point out is that on my previous post I do not mention anything about Boiler Horsepower (BHP). This physical dimension was inappropriately used. Boiler Horsepower was (and still is) a thermal unit of power defined by ASME in 1884. The approximate equivalence to the kilowatt is: 1 BHP = 9.81 kW, depending on the standards used; by definition it means the energy transfer rate (to the water) needed to evaporate 34.5 lb of water at 212ºF (from 212ºF) in 1 hour (water boiling at the standard atmospheric pressure). Its use is older than 1884, although with a different definition, but similar in magnitude. The designation results from the fact that originally (mid 1870s) a boiler with X BHP associated with the non-condensing stationary steam engines of the day could also produce X IHP. Yet, as time passed the thermal efficiency of steam engine grew rapidly to the extent that in the 1940s a boiler with 1000 BHP (thermal, or 9810 kW) could allow a continuous cylinder HP in excess of 2000 IHP (mechanical, or 1491 kW), in the most economical steam rate of a locomotive. Precisely due to the ambiguity of this designation, it was never used outside the United States. So when consulting old test reports one has to be aware of these subtle things, and to try to contextualize the information seen. If an author presents a tractive power curve as a boiler horsepower curve, what he is actually saying is cylinder horsepower sustainable by boiler (without the water level dropping or the boiler pressure falling); if that value is simply a rating of boiler capacity or a boiler output measured in a plant test, then what we are seeing is a thermal power rating, and not mechanical HP. This can appear as a big confusion, but physics usually has that effect on people.

 

But returning to accuracy, the calibrations made with the speedometers used in a dynamometer car of the 1890s revealed uncertainties inferior to 1% of the readings; but in a test plant such an uncertainty would be irrelevant, because the speed was held constant by water brakes and the number of drivers rotations (wheel turns) would be counted in a 1 to 3 hours test, evaluated to the second; and the plant's dynamometer was so sensitive that the touch of a man’s finger in the traction bar would be registered; also a test plant indicator of those days could allow a 3% uncertainty; and the cylinder dimensions need to compute accurately the IHP were evaluated with an accuracy of one-thousand of an inch. In fact the measured cylinder diameter was the result of six measures, 3 sets of a horizontal and vertical diameter measurements, made at crank-end, at half-stroke and at the head-end of each cylinder. The piston-rods were also measured in a similar way to accurately compute the cylinder effective sectional area for each side of the piston. And in the test program I’ve mentioned, made in 1904, all instruments were recalibrated at the end of every set of 8 tests...

 

But such a thing as the irrelevance of the adjustment of Fundamental Standards, to the Industry in general, is quite obvious, no? I am not an American, but do you think by a single second that the U.S. of 1959 (the date of these major adjustments) would comply with an international convention, of the English-speaking nations, that could cause a massive negative impact upon the nation’s industrial capacity, if the customary standards were suddenly changed dramatically? Surely the U.S. would refuse such an agreement. These needed changes (by reasons of uniformity) affected only the national metrology laboratories in the English-speaking nations, but were of no consequence to fabrication procedures (in terms of tools, industrial measuring instruments, etc.). I recommend reading the inquiry made by NIST [National Institute of Standards and Technology] about the impact on US Industry of the reassignment of mass values and uncertainties to NIST working standards made in 1990 (yes, the standards are changing still today). I transcribe here a part of the conclusions of that inquiry that you can read in the Journal of Research of the National Institute of Standards and Technology, Vol. 95, Jan.-Feb., 1990, “New Assignment of Mass Values and Uncertainties to NIST Working Standards”:

 

«Effect on Industry and Technology

An Ad Hoc Committee of the National Conference of Standards Laboratories (NCSL) was formed in order to help assess industrial and technological implications of the actions contemplated for January 1, 1990 [New Assignment of Mass Values and Uncertainties to NIST Working Standards]. Members of the Committee include representatives from civilian and military standards laboratories, balance manufacturers, and weight manufacturers. All were asked to estimate the impact which a change of roughly 0.15 mg/kg would have on their programs. The members could not identify a single instance where such a change would affect a manufactured product or a critical measurement. Virtually all concerned, however, recognized that a change of this magnitude could be noticeable within their metrology laboratory. This is not surprising since typical NIST calibrations give an uncertainty of about 0.075 mg (3 standard deviations) for calibrations of 1-kg standards and users of these standards often have balances of comparable precision to our own.

In recent years, calibrations for primary national laboratories of other countries have been carried out using secondary standards CH-1 and D2 with assigned values based directly on measurements against K20 [build in 1889]. These measurements are not, therefore, in need of correction.»

 

That 0.15 mg/kg estimate change in the standard kg, considered as irrelevant by the U.S. Industry of 20 years ago, was a result of previous studies made to establish the U.S. mass standards time-stability. And yet this value was simply 4 times greater than the mass uncertainty certificate established with the technology of 1889 in regard to the Standard Kilogram supplied by the French government to the U.S.; the precision balance then used had a reading standard deviation of 10 mg (1/100’000’000 of 1 kg); the ones used in the study made above, initiated in 1985, had a standard deviation of 4.5 mg. Apparently the late 19th century technology was not that bad (at least reading those NITS reports full credit is placed upon the capacity to build, measure and replicate such precise standards in those days, and the mass recalibrations made today display quite similar uncertainties one century after). I will transcribe bellow the abstract of the paper associated with this NIST work, published in the Journal of Research of the National Bureau of Standards, Vol. 90, Jul.-Aug., 1985, “Recalibration of the U.S. National Prototype Kilogram”:

 

«The U.S. national prototype kilogram, K20, and its check standard, K4, were recalibrated at the Bureau International des Poids et Mesures (BIPM). Both these kilograms are made of platinum-iridium alloy. Two additional kilograms, made of different alloys of stainless steel, were also included in the calibrations. The mass of K20 in 1889 was certified as being 1 kg-0.039 mg. Prior to the work reported below, K20 was most recently recalibrated at the BIPM in 1948 and certified as having a mass of 1 kg-0.019 mg. K4 had never been recalibrated. Its initial certification in 1889 stated its mass as 1 kg-0.075 mg. The work reported below establishes the new mass value of K20 as 1 kg-0.022 mg and that of K4 as 1 kg-0.106 mg. The new results are discussed in detail and an attempt is made to assess the long-term stability of the standards involved with a view toward assigning a realistic uncertainty to the measurements.»

 

As to the uncertainty analysis I’ve mentioned in the previous post, as I said earlier, it was not intended as an exercise of cynicism. The expression “uncertainty analysis” is not mine. It is simply the branch of the Mathematics/Statistics that studies the error propagation throughout an experimental procedure and related physical model, due to the uncertainty about the exactness of the measures taken. Also it is not a new thing as you will see reading the NIST reports I’ve mentioned, and many others.  My intention was to point out that a 100 mph run is in fact so close to a plausible speed with that locomotive, running hard on level track with such a small train (looking not to physics but to other known and documented similar performances), that the influence of a moderate 10 to 20 mph helping wind, or a slight downgrade, cannot be disregarded in any serious scientific study of that claim. And I was not suggesting that the track had sunk to give a convenient 0.2% downgrade; what I was trying to point out was that Mr. Hankey did not care to provide the reader with a detailed gradient profile, because in the States the use of condensed gradient profiles is so widespread that is not uncommon in those charts to discard as near level such a slight gradient, especially in a run of 2200 miles long through mountain territory.

What’s more, it would be a question of purism to think that those alleged [by me] helping circumstances would diminish in some way the significance of the claimed performance. Why? Because the majority of the documented railway speed records we know were also achieved in favorable circumstances, natural or not (as to gradient, wind, or mechanical alteration of the standard equipment used). For example, the present day conventional train speed record of 357.2 mph [574.8 km/h], made by a French AGV set in 2007,  was nothing more than a manmade congregation of “favorable circumstances”, in the sense that nothing was left to chance and that nothing could be farther from the normal operating conditions: the train was reduced to the minimum, the gear ratio was changed, the electrical power equipment was reinforced, the protection systems against overload and over-speed were disengaged, the mechanical and electric tension of the aerial power cables was increased, the traction motors used were forced to produce 2 times the maximum rated HP of a standard AGV set, the entire train was instrumented and monitored (as a moving laboratory) in the course of that trial, a speed-distance curve was calculated in advance with the intention of combining the best track alignment with a down gradient, the atmospheric conditions were taken in consideration, the railway traffic was completely interrupted, and last but not least the trial we know was the culmination of a vast number of speed trials not divulged to the general public…  The result: that awesome speed!

Do you know any speed record made by a steam train involving such a vast planning? So if the driver knew that in a certain portion of the track the trains usually had an easier rolling (due to a very slight down gradient, or because the train was usually exposed to favorable winds at that spot) would it be diminishing to the engineer to use that intuitive knowledge to try to obtain an unusually high speed (given the opportunity to try it)? To me it would be nothing more than a display of wisdom. Claims of speed such as this cannot be viewed as a maximum speed by catalog, as declared by a car manufacturer, one that obeys to definite rules needed to make comparable competitive products.

 

So I choose to believe that a truly high speed was achieved that day, probably close to 100 mph, based on the knowledge I have of similar situations (and not on a Wikipedia research) and on my calculations (*),  until some other historian, or an expert and reliable person, bring more light into the subject. If that claim is such an absurd, as to deserved that kind of public debunking, as Mr. Hankey believed adequate, then he should have done the homework properly and investigate if there are verifiable accounts of bad riding qualities with locomotives of that class, or if there were “favorable circumstances” to consider, and a detailed gradient profile of that road is probably at the reach of Mr. Hankey and TRAINS resources. And the gradient is not actually a random “favorable circumstance”: it was there or it was not! He should have specified explicitly that information. Every rail fan in Europe knows that the U.K. Mallard’s speed record of 125 mph, made July 3, 1938, was achieved on a downgrade, at the end of a 0.42% grade.

 

(*) And as for the calculations I’ve made, actually only the results are my responsibility, which were made explicit to allow verification, because the expressions and the calculus sequence I used are not mine and do not reflect present day standards (criterions) but the standards of those days. The most recent of these expressions is the Davis Jr. train resistance formula, first made public in October 1926. It is an empirical expression base on controlled dynamometer tests made with ordinary vestibule trains since the beginning of 20th century. The others are simply conceptual physical laws (and as such not subject of discussion) that can be encountered in engineering books and test reports from more than 100 years ago till nowadays.

What you Sir should have asked is this: why are these (old) formulas still used, after a century, in present day performance calculators used by railroads and why are these formulas still referred in engineering college books or in Handbooks for Mechanical Engineers? William Hay, an authority in railway engineering in the 1980s (and respected as so today), still thought as pertinent to present and explain the meaning of the different coefficients in this old Davis formula.   

Why? Because those studies, in which Davis based is correlation, are still valued and recognized as reference studies by the present day railroad and academic communities. There are today other Davis formulas, the Adjusted or Modified Davis, and others, to contemplate today’s better equipment (track and cars), but the original one dates to 86 years ago and reflects early 20th century railroad equipment; and the cars and locomotive used in the Scott Special run were modern equipment in 1905, so the formulae I used in the previous post is perfectly adequate. Only the use I’ve made of my scientific calculator can be incorrect.

Obviously, anyone is free to disagree: the results presented previously are only valid in the assumption that the William Davis formula depicts adequately the train behavior at that place and instant, and in the circumstances I had to guess, given the absence of proper information (but apparently that did not constrain Mr. Hankey). Yet if one wants to use the physics to debunk some speed claim, then that person as an “obligation” to propose, or point, a better and more adequate calculus procedure, or else forget any pretense physics argumentation. And obviously the laws of physics are nothing more than human concepts and derived assumptions made in accordance with observation; as such they do not restrain nature behavior: a rock tossed in the air will not ask permission to Newtonian laws to fall…

These expressions and the statements I made here, and in the previous post, can be encountered in the references given bellow. Many of these are redundant, yet the majority is available freely on-line, the reason why I´ve chosen these and not others (that I have been collecting and studying for the past 20 years).

 

Historic note about the standards adjustments

Although the U.S. had a Standard Yard supplied by the British Government in 1856, the metric units were recognized as a legal system in 1866 by the U.S. Congress, making this use permissible in the States to the point that by a formal order of the Secretary of the Treasury, April 5, 1893, the Standard Meter and Standard Kilogram, became the legal standards in U.S., being denominated the "Fundamental Standards”: against which all copies would be compared (and still are today, in the case of mass standards).

In accordance with Avallone and Baumeister III in “Standard Handbook for Mechanical Engineers”, 1997, those standards became then:

 

1 Yard = 3,600/3,937 m; 1 lb = 0.453 592 43 kg

 

These Standards were real bodies, and although build with the greatest care they would vary slightly from nation to nation depending on the capacity to replicate the original ones. That care can be best appreciated by saying that in 1893 Brown & Sharpe Mfg. Co., the U.S. manufacturer contracted by the U.S. Office of Weights and Measures, was able of replicate the Standard Yard with a deviation of only 0.00002 inch! (for a 36 inches length). Although an excellent precision work, residual differences would always subside between Standard Units nominally equal. For that reason in 1959 the national laboratories of the English-speaking nations agreed to the following standards:

 

1 Yard = 0.9144 m (exactly); 1 lb = 0.453 592 37 kg

 

So, remembering the horsepower definition given above, these differences in standards would result in nominal horsepower values given by: 1 hp_1893,U.S. = 1.000 002 132 hp_today!

Well, on the laboratory tests I’ve mentioned in the previous post, made in 1904 with a locomotive similar to the one used by Scott Special on that 100 mph run, the time-average horsepower measured at the cylinders by indicator was 1621.5 hp for 2 hours, or, if we perform the HP correction above, 1621.503 hp using today’s Standards! But the usual stated uncertainty for an indicator reading was 3% at least, with very short connecting pipes, possible with a stationary engine, the case here. So to contemplate a HP correction due to the adjustment of Standards would be an absurd exercise: the power difference is less than 3 W, similar to the HP capability of a hamster! If this measured HP had been made in U.K., in those days, the correction would be minus 13 W, an irrelevant difference even if the locomotive had been tested with today’s best instrumentation (obviously for a machine that big).

 

 

Examples of high speeds with nineteenth century U.S. ordinary trains

Controlled run (by Baldwin) made in July 1, 1898, with the Vauclain compound Atlantic No. 1028 of the Philadelphia & Reading Railroad, with the ordinary train No.25, 5 cars and 202 tons weight, between Camden and Atlantic City, also a “level” and straight road, although with only 55½ miles (start to stop).

 

Engine No. 1028, Train No. 25, July 1, 1898
position	time			MP ave. speed
mile-post [MP]	[hh:mm:ss]			[mph]
Camden	3:	50:	00	-
MP 55	3:	52:	29	City limits
MP 54	3:	53:	44	48,0
MP 53	3:	54:	43	61,0
MP 52	3:	55:	39	64,3
MP 51	3:	56:	34	65,5
MP 50	3:	57:	28	66,7
MP 49	3:	58:	18	72,0
MP 48	3:	59:	08	72,0
MP 47	3:	59:	55	76,6
MP 46	4:	00:	42	76,6
MP 45	4:	01:	28	78,3
MP 44	4:	02:	11	83,7
MP 43	4:	02:	56	80,0
MP 42	4:	03:	43	76,6
MP 41	4:	04:	32	73,5
MP 40	4:	05:	21	73,5
MP 39	4:	06:	10	73,5
MP 38	4:	06:	56	78,3
MP 37	4:	07:	41	80,0
MP 36	4:	08:	24	83,7
MP 35	4:	09:	08	81,8
MP 34	4:	09:	52	81,8
MP 33	4:	10:	37	80,0
MP 32	4:	11:	20	83,7
MP 31	4:	12:	03	83,7
MP 30	4:	12:	47	81,8
MP 29	4:	13:	30	83,7
MP 28	4:	14:	13	83,7
MP 27	4:	14:	55	85,7
MP 26	4:	15:	37	85,7
MP 25	4:	16:	20	83,7
MP 24	4:	17:	03	83,7
MP 23	4:	17:	46	83,7
MP 22	4:	18:	30	81,8
MP 21	4:	19:	15	80,0
MP 20	4:	19:	59	81,8
MP 19	4:	20:	43	81,8
MP 18	4:	21:	25	85,7
MP 17	4:	22:	08	83,7
MP 16	4:	22:	51	83,7
MP 15	4:	23:	34	83,7
MP 14	4:	24:	18	81,8
MP 13	4:	25:	03	80,0
MP 12	4:	25:	48	80,0
MP 11	4:	26:	32	81,8
MP 10	4:	27:	15	83,7
MP 9	4:	27:	59	81,8
MP 8	4:	28:	42	83,7
MP 7	4:	29:	25	83,7
MP 6	4:	30:	09	81,8
MP 5	4:	30:	51	85,7
MP 4	4:	31:	32	87,8
MP 3	4:	32:	16	81,8
MP 2	4:	33:	07	70,6
MP 1	4:	34:	15	52,9
Atlantic City	4:	35:	17	-
Actual time: 45 min 17 sec.
Despatcher time: 45¼ min

 

Those Baldwin Vauclain compounds revealed the ability to sustain speeds in excess of 80 mph “routinely”, while pulling train consists of 5 to 7 cars and 200 to 300 tons weight (cars only), much longer than the Scott Special. What’s more, the aggregate start-to-stop average speed achieved with engine No.1027, for the months July and August of 97, was an astonishing 69 mph, supplanted the following year by engine No.1028 with this same train, namely, an aggregate start-to-stop average speed of 70.5 mph for 53 trips (or 2941 miles), the best run being made 5 August 1898 in 44¾ minutes, that’s an average speed of 74.4 mph start to stop, the cruising speed being in excess of 80 mph by train Dispatcher’s sheet (written to the nearest ¼ minute). But on July first that year, the train was timed to the nearest second at which successive milepost (MP) and between MP48 and MP3 the average speed was 81.5 mph, and those 45 miles included the line heaviest grades, 5 miles at 0.6% made at a minimum speed of 73 mph! As to maximum speed, an 87.8 mph was achieved between MP 5 and 4 (this mile being made in 41 sec, the preceding one taking 42 seconds), this with the help of short 0.67% downgrade one mile long, near Pleasantville. Certainly that the random uncertainty is 1 second, but as the train was accelerating fast the maximum speed can be assumed safely as being 88 mph. The train was one made of 5 cars, with 4 wheels trucks, 202 tons trailing weight, and 311 tons total train weight (the locomotive alone weighted 71 tons). The maximum speed achieved on level track was 84 mph, between MP 18 and 6, representing a cylinder horsepower of 1480 hp if computed in accordance with the Davis Jr. Formulae presented in the previous post. Yet Baldwin maximum declared IHP was 1450 hp at 70 mph. What’s more, the run made in 5 August was slightly faster and the train heavier, 6 cars instead of 5, so the IHP was undoubtedly still larger… Or else, running in a good track, the train resistance was actually less than the computed value (Davis formulae). These locomotives were pure Vauclain compounds with two pairs of 2 cylinders, one HP and one LP, connected to the same crosshead; although not Balanced locomotives, they were able to make 90 mph, or close to, in “normal” service with trains heavier than Scott Special (and the locomotive weight was 26 tons less than AT&ST No.510).

Also in the Proceedings of the Forty-Sixth Annual Convention of the American Railways Master Mechanics’ Association, 1913, we can see speed-distance charts depicting the speeds practiced on this same road by engine No. 303, class P6-a, a 3-cylinder simple expansion Atlantic, non-superheated; the runs were made in August 1910 with train No. 17. The engine had entered service in June 15, 1909. The trains pulled were 5 to 7 cars long, up to 300 tons weight (cars only), and cruising speeds up to 90 mph on level track were sustained. A curve representing the average speeds made on 27 trips is shown there.

 

Engine No. 303, Train No. 17, August 1910 (ordinary runs)
position	speed  (*)
mile-post	fastest run	average of 27 runs
[MP]	[mph]	[mph]
Camden	-	-
MP 55	City limits	City limits
MP 54	47,0	43,5
MP 53	54,0	50,5
MP 52	56,0	53,5
MP 51	56,5	53,5
MP 50	61,0	60,0
MP 49	62,0	60,0
MP 48	68,0	64,5
MP 47	70,5	66,5
MP 46	70,0	67,5
MP 45	75,0	72,0
MP 44	80,0	76,0
MP 43	74,5	71,5
MP 42	71,5	68,5
MP 41	67,5	63,5
MP 40	67,5	63,5
MP 39	71,0	67,0
MP 38	73,0	70,5
MP 37	80,0	75,5
MP 36	81,5	78,5
MP 35	81,0	78,0
MP 34	80,0	77,0
MP 33	82,0	79,5
MP 32	84,5	81,0
MP 31	84,5	80,0
MP 30	83,0	79,0
MP 29	85,0	81,0
MP 28	86,0	82,0
MP 27	91,0	85,0
MP 26	90,0	85,0
MP 25	89,5	85,0
MP 24	89,0	85,0
MP 23	90,0	84,5
MP 22	86,5	82,5
MP 21	86,5	83,0
MP 20	86,5	83,0
MP 19	88,0	84,5
MP 18	91,0	87,5
MP 17	90,5	85,5
MP 16	90,0	85,0
MP 15	88,5	84,0
MP 14	90,0	83,0
MP 13	86,5	82,0
MP 12	86,5	82,0
MP 11	86,0	82,0
MP 10	90,5	85,0
MP 9	90,5	85,0
MP 8	90,0	84,5
MP 7	89,5	84,5
MP 6	89,5	84,5
MP 5	93,5	88,5
MP 4	91,0	85,0
MP 3	88,5	82,0
MP 2	81,5	76,0
MP 1	45,5	30,0
Atlantic City	-	-
Start to Stop average speed:
average for 27 runs  ->  69 mph		(for 1499 miles)
fastest run  ->  73 mph		(for 55½ miles)
M48 to MP3 average speed:
average for 27 runs  ->  77 mph		(for 1215 miles)
fastest run  ->  83 mph		(for 45 miles)
(*) Speeds rounded to the nearest ½ mph (by graph)

 

Yet the simple expansion Atlantics of the class 340-349 made considerably faster runs in 1906 and 1907. In the Report of the Proceedings of the Thirty-Ninth Annual Convention of the American Railway Master Mechanics’ Association”, p. 426-427, Atlantic City, June, 1906, we can read the following communication:

               

«(…)

                The President: The secretary has a few notes to read.

                The Secretary: It will probably be interesting to the members to hear about the run made to Philadelphia yesterday by the Reading engine [June 19, 1906]. The train was hauled by Engine No. 343, one of the new Atlantic type locomotives designed and built by Mr. Taylor of the Reading Road. Six cars were hauled, three eight-wheel day coaches and three twelve-wheel Pullmans. The weight of the cars [tare] was about 246 tons.

                The 55½ miles from Atlantic City to Camden were run in exactly 45 minutes, which represents an average speed of 74 miles per hour, for each mile [start to stop]. On the return journey, with the same engine and train, the 55½ miles were run in 43 minutes and 25 seconds, which gives an average speed of 76.7 miles per hour for each mile [start to stop]. Taking out the first 1½ mile from Camden, and the last mile into Atlantic City, the remaining 53 miles were run at an average of 44.15 seconds for each mile, or 81.54 miles per hour. From mile post 29 down to mile post 7 these 22 consecutive miles were each made in 40.33 seconds, or at a speed of 89.46 miles per hour.

                The fastest mile was run in 38 seconds, or at a speed of 94.7 miles per hour. On both runs Mr. Taylor had issued orders that the train time should not be more than 50 minutes and not less than 45 minutes.»

 

This is a quite respectable performance for an early 20th century “ordinary” steam train, no? To make 22 miles of “straight level track” at almost 90 mph with 6 cars weighting 260 tons at least, considering the weight of the passengers carried… And the well-known run made in May 14, 1905, between Atlantic City and Camden, in 42 min 05 sec, start to stop, was not the fastest ever made! In a visit to the States made by Gérard Vuillet in 1926, a man of influence (in the financial world) and a railway expert, he was authorized to search the files of the Reading Company’s Mechanical Department at Reading, Penn.: in those files he encountered a run made in June 14, 1907, also by the engine No. 343, from Camden to Atlantic City, in 41 min exactly (by Train Dispatcher’s time, presumably given to the nearest ¼ minute); the fastest mile was made in 36 seconds, near Pleasantville, or at a speed of 100 miles per hour, this with a train of 260 tons (cars only).

Well, seeing the speed charts mentioned above and the way those times were recorded to the second at every mile post, I’m convinced that in those days they manage quite well making careful time measures (to the second and in sequence), because although the runs were made with different locomotives (not of the same class) the times taken, and correspondent speeds, are consistent in terms of the ratio of the average speeds practiced to the maximum speeds obtained. 

The Atlantics referred above had grates of 76 (the compound) to 95 sq.ft, much bigger than the AT&SF locomotive, but they burned anthracite so a direct comparison cannot be made in these terms. Actually the boilers of these locomotives were equal or smaller than the one used on No. 510.

 

 

Conclusions

If speeds of 90 miles per hour could be sustained on a level track with ordinary trains, regularly, although on special circumstances resulting from a question of prestige and from the need to compete for the same traffic with a rival road, the PRR, this with heavier, longer and older trains than the Scott Special, then why is it so difficult to believe that a 100 mph burst of speed could be achieved with such a small train on a special run? One made with a single paying lunatic like Dead Valley Scotty, and with a clear road ahead of the locomotive, and not an ordinary run with hundreds of “innocent” passengers! And this is not physics talk. This is only to put things in perspective, considering what the technology of the day could actually do. If this 100 mph run was an impossible accomplishment, one that cannot be easily refuted by the laws of physics (in the absence of detailed information), then believing in this can only result from a preconceived idea; and probably it is just that: a question of belief (and a pointless one).

Well, apparently the one person that does not like numbers is Mr. Hankey himself, because this information (and a lot more) was made public at the time in the technical (and reliable) publications of the day. One has only to research a little and study properly what we encounter.

 

 

References (supporting the information posted here and on the previous post)

 

-          Maclean, Magnus, “Physical Units”, London, Biggs and Co., 1896

-          Cotterill, J. H., “The Steam Engine”, Third Edition, Spon & Chamberlain, 1896

-          Smart, R. A., “Performance of a Four-Cylinder Compound Locomotive”, Purdue University, paper presented before the St. Louis Railway Club, February 11, 1898

-          “Vauclain System of Compound Locomotives: Description, Method of Operation and Maintenance”, Baldwin Locomotive Works, Burnham, Williams & Co., 1900

-          “The Pennsylvania System at the Louisiana Purchase Exposition: Locomotive Tests and Exhibits”, the Pennsylvania Railroad Company, 1905

-          “Report of the Proceedings of the Thirty-Ninth Annual Convention of the American Railway Master Mechanics’ Association”, Atlantic City, NJ, June, 18-20, 1906

-          “Measuring Tools”, Third Edition, Machinery’s Reference Book No. 21, The Industrial Press, 1910

-          Schmidt, E. C., “Freight Train Resistance: Its Relation to Average Car Weight”, University of Illinois Engineering Experiment Station, Bulletin No. 43, May 1910

-          Heck, Robert, “The Steam Engine and Turbine”, D. Van Nostrand Company, 1911

-          Clayton, J. P., “The Steam Consumption of Locomotive Engines from the Indicator Diagram”, University of Illinois Engineering Experiment Station, Bulletin No. 65, January 1913

-          “Report of the Proceedings of the Forty-Sixth Annual Convention of the American Railway Master Mechanics’ Association”, p. 282-285, Atlantic City, NJ, June, 11-13, 1913

-          Wood, A. J., “Principles of Locomotive and Train Control”, McGraw-Hill Book Company, 1915

-          Tuttle, Lucius, “The Theory of Measurements”, Jefferson Laboratory of Physics, Philadelphia, 1916

-          Cole, F. J., Chief Consulting Engineer of the American Locomotive Company, Train Resistance in “Locomotive Hand-book”, American Locomotive Company, 1917

-          Schmidt, E. C., and Dunn, H. H., “Passenger Train Resistance”, University of Illinois Engineering Experiment Station, Bulletin No. 110, December 1918

-          Shealy, E. M., “Steam Engines”, McGraw-Hill Book Company, 1919

-          Davis, W. J., Jr., “Tractive Resistance of Electric Locomotive and Cars”, General Electric Review, Vol. 29, pp. 685-708, October 1926

-          Johnson, R. P., Chief Engineer of the Baldwin Locomotive Works, “The Steam Locomotive”, Simmons-Boardman Publishing Company, 1942

-          Kalmbach, A. C., “Railroad Panorama”, Kalmbach Publishing Company, 1944

-          Chapelon, André, “La Locomotive a Vapeur”, English translation by George W. Carpenter, Camden Miniature Steam Services, 2000, reprint of the  Second French Edition, Paris, 1952

-          Vuillet, Gérard, «Railway Reminiscences of Three Continents», Thomas Nelson Ltd, London, 1968

-          Hay, W. W., “Railroad Engineering”, Second Edition, John Wiley & Sons, 1982

-          Davis, R. S., “Recalibration of the U.S. National Prototype Kilogram”, Journal of Research of the National Bureau of Standards, Vol. 90, Number 4, July-August 1985

-          Davis, R. S., “New Assignment of Mass Values and Uncertainties to NIST Working Standards”, Journal of Research of the National Institute of Standards and Technology, Vol. 95, Number 1, January-February 1990

-          Avallone and Baumeister III (editors), “Standard Handbook for Mechanical Engineers”, Tenth Edition, McGraw-Hill International Editions, 1997

-          Hugh, W. C., and Steele, W. G., “Experimentation, Validation, and Uncertainty Analysis for Engineers”, Third Edition, John Wiley & Sons, 2009

 

[Anonymous]

h2fe9x2

 

Conclusions

If speeds up to 90 miles per hour could be practiced on a level road with an ordinary train, although on special circumstances resulting from a question of prestige and from the need to compete for the same traffic with a rival road, the PRR, this with heavier, longer and older trains than the Scott Special, then why is so difficult to believe that a 100 mph run could be achieved with a lighter train on a special run? One made with a single paying lunatic like Dead Valley Scotty and with a clear road ahead of the locomotive, not an ordinary run with hundreds of “innocent” passengers! And this is not physics talk. This is only to put things in perspective, considering what the technology of the day could actually do. If this 100 mph run was an impossible accomplishment, one that cannot be easily refuted by the laws of physics, then believing in this can only result from a preconceived idea, and probably it is just that: a question of belief (and a pointless one).

Well, apparently the one person that does not like numbers is Mr. Hankey itself, because this information (and a lot more) was made public at the time in technical (and reliable) publications of the day. One has only to research a little and study properly what we encounter.

 

 

References (supporting the information posted here and on the previous post)

 

-          Maclean, Magnus, “Physical Units”, London, Biggs and Co., 1896

-          Cotterill, J. H., “The Steam Engine”, Third Edition, Spon & Chamberlain, 1896

-          Smart, R. A., “Performance of a Four-Cylinder Compound Locomotive”, Purdue University, paper presented before the St. Louis Railway Club, February 11, 1898

-          “Vauclain System of Compound Locomotives: Description, Method of Operation and Maintenance”, Baldwin Locomotive Works, Burnham, Williams & Co., 1900

-          “The Pennsylvania System at the Louisiana Purchase Exposition: Locomotive Tests and Exhibits”, the Pennsylvania Railroad Company, 1905

-          “Report of the Proceedings of the Thirty-Ninth Annual Convention of the American Railway Mechanical Association”, Atlantic City, NJ, June 1906

-          “Measuring Tools”, Third Edition, Machinery’s Reference Book No. 21, The Industrial Press, 1910

-          Heck, Robert, “The Steam Engine and Turbine”, D. Van Nostrand Company, 1911

-          Clayton, J. P., “The Steam Consumption of Locomotive Engines from the Indicator Diagram”, University of Illinois Engineering Experiment Station, Bulletin No. 65, January 1913

-          “Report of the Proceedings of the Forty-Sixth Annual Convention of the American Railway Master Mechanical Association”, p. 282-285, Atlantic City, NJ, June, 11-13, 1913

-          Wood, A. J., “Principles of Locomotive and Train Control”, McGraw-Hill Book Company, 1915

-          Tuttle, Lucius, “The Theory of Measurements”, Jefferson Laboratory of Physics, Philadelphia, 1916

-          Cole, F. J., Chief Consulting Engineer of the American Locomotive Company, Train Resistance in “Locomotive Hand-book”, American Locomotive Company, 1917

-          Schmidt, E. C., and Dunn, H. H., “Passenger Train Resistance”, University of Illinois Engineering Experiment Station, Bulletin No. 110, December 1918

-          Shealy, E. M., “Steam Engines”, McGraw-Hill Book Company, 1919

-          Davis, W. J., Jr., “Tractive Resistance of Electric Locomotive and Cars”, General Electric Review, Vol. 29, pp. 685-708, October 1926

-          Johnson, R. P., Chief Engineer of the Baldwin Locomotive Works, “The Steam Locomotive”, Simmons-Boardman Publishing Company, 1942

-          Kalmbach, A. C., “Railroad Panorama”, Kalmbach Publishing Company, 1944

-          Hay, W. W., “Railroad Engineering”, Second Edition, John Wiley & Sons, 1982

-          Davis, R. S., “Recalibration of the U.S. National Prototype Kilogram”, Journal of Research of the National Bureau of Standards, Vol. 90, Number 4, July-August 1985

-          Davis, R. S., “New Assignment of Mass Values and Uncertainties to NIST Working Standards”, Journal of Research of the National Institute of Standards and Technology, Vol. 95, Number 1, January-February 1990

-          Avallone and Baumeister III (editors), “Standard Handbook for Mechanical Engineers”, Tenth Edition, McGraw-Hill International Editions, 1997

-          Hugh, W. C., and Steele, W. G., “Experimentation, Validation, and Uncertainty Analysis for Engineers”, Third Edition, John Wiley & Sons, 2009

h2fe9x2,

 

In your quote above, you ask why the AT&SF speed record is so difficult to believe.  I don’t think that it is difficult to believe so much as there is a refusal to believe it.  There is a widespread bias that nothing great that happened until the modern era of railroading.  These early speed records are a serious threat to that school of thought.  Obviously, Mr. Hanke began with the pre-existing belief that the AT&SF speed record was false before he set out to “prove” that was the case.  

 

Mr. Hanke says there are three reasons not to believe the claimed speed record:

 

1)      It fairly reeks of wishful thinking and corporate spin.

 

2)      There is no credible scenario in which the railroad (or anyone on the train) could have accurately timed that feat.

 

3)      It was physically impossible. 

 

 

People defend Hanke by saying that the record cannot be proven one way or the other.  That is true, but that is not what Hanke is saying. 

 

So, thanks for your effort in making a technical review of the physics of the debunking.  I would say that you have debunked author Hanke’s invocation of the laws of physics, and restored the AT&SF speed claim as standing free of bunk.