How a steam engine works

[SpaceMouse]

My ignorance knows no bounds it seems. 

I now find that it would be appropriate to learn how steam engines work, not just what fuels the nuclear reactors, but all the dohickies and gizmos on the engines that model railroaders call details.   

So I'm looking for a book or books or other reference materials that shows the history, development and implications of all said dohickies.

Know what I mean?

BTW: I know how to utilize inter-library loans.

["JaBear"]

Gidday Chip, here’s a couple of basic videos to wet your whistle, so to speak.

 

 

 

I’m not sure if there is a definitive book so to speak, but have been lucky enough to have been given several handbooks for the operation of several NZR steam locomotives which go into more specific details. I have also been given handbooks on the operation of the Walschaerts, and Baker valve gear.

 

I would presume that similar handbooks would have been issued to US railroad enginemen.

Cheers, the Bear.

[Overmod]

Keep in mind that there are at least two stages to this: initially learning the history of steam power, and then learning the 'fine points' of detail design.

Both of these are separate from learning the actual thermodynamics of how steam works, and then why thermodynamics alone can be a lousy guide for practical locomotive design, and then why the 'most efficient' locomotive designs might not be the ones practically best for actual railroad service...

This being separate from the history, and the design, and sometimes the corporate shenanigans, of the various 'doohickeys' (often patent, specialized, and 'invented here') over the years.  Much of this is imperfectly told and has to be put together from the literature rather than found in one or a couple of definitive books.

In my personal opinion you should start with a piece of overkill: get as many of the carefully labeled drawings that show 'parts of the steam locomotive' as you can.  Offhand I remember  seeing good ones for a NYC Hudson and a German locomotive as the later types, and one for a 'typical' American type from the late 19th Century (when what was 'good practice' could be wildly different from later developments!).  Then find some of the labeled backhead photos and keep them as reference on what to start looking up -- and unlabeled ones to compare.

The closest thing I can think of to a list of auxiliaries in one place is the Ransome-Wallis Encyclopedia of World Railway Locomotives.  That was published right at the effective end of big steam, at a time when I think most of the 'powers that be' expected steam in Britain to persist decades longer.  On the other hand, the actual listing of auxiliaries in the 'encyclopedia' part is wildly incomplete in many respects, while colossally overdetailed in others (British injectors in particular)

For what is actually involved in designing a working steam locomotive, I could point you to Ralph Johnson's book from the '40s (my copy came directly from Simmons-Boardman and was the 1983 edition!) but the better one now is Wardale's detailed FDCs (fundamental design calculations) done for the 5AT project but adaptable to other situations.  To these you need to add some backtranslated wisdom not found in books (like why air horns with proportional valves are better than whistles on big power using water treatment, and why air preheat and better water-leg circulation could pay outsize dividends) and make your own notebook of best design practice in things like low-water alarms -- where the 'mere function' of the device is not as important practically as some of the proprietary features of a given 'patented' type...)

[Overmod]

Many of the earlier references are, through the miracle of Google Books, free to download (as PDFs) or to print out.  The trick is knowing what they are, which is not in one place (as it probably should have been by now, and for all I know actually is).  There is a very good article from ~1849 on the development of early (!) valve gear which is fascinating both for what it says and for understanding what is practical about 'steam distribution' -- be prepared to break out the Anacin as you get into the subject of modern radial gear for piston valves, and then the merry world of poppet valves from Caprotti on.  Take copious notes on what you read ... and on what you don't (yet) understand ... then follow up on what you don't 'get' and repeat the exercise.

[Overmod]

I have not yet found a better 'on video' explanation than the one ATSF prepared, as I recall circa 1922, using the then relatively new technique of animation to show with technical correctness some of what was going on.  Fortunately at least some of this was preserved by Herron Rail Videos and a portion of it can be seen here:

https://m.youtube.com/watch?v=QYNRhL0ddDQ

Regrettably this was the 'silent film' era and the accompanying discussion, which would have contained additional important details such as why 2-cylinder double-acting locomotives need two cylinders in quarter, and why the 'compression' noted is important in operation, has apparently been lost.

Periodically as you learn more about steam technology you can come back to this illustration to clarify it.  For example the development of the 'combustion chamber' lengthens the distance to that 'rear tubesheet' to give increased radiant uptake; the precise motion of the piston valves involves longer travel than 'necessary' as well as long lap for good reasons; the behavior of the boiling water is very different from what is pictured, in significant ways; the plates and screens in a Master Mechanic front end are arranged differently from the illustration in important respects.

I do not see this in their current production list and you'll probably have to e-mail or contact them to find out how to get a full version:

http://www.herronrail.com/Title_List_12-09-19.pdf

["JaBear"]

I haven’t seen this manual, but it may be the go.

 

https://www.amazon.com/Locomotive-Enginemans-Manual-W-James/dp/1935327828

Cheers, the Bear.

[Overmod]

1. That manual was one of the better ones, but suffers from a fate more than usually common to steam-technology books: they were copied and reprinted in the era before Google Books and are therefore considered 'copyrighted' even with regard to the historic originals and hence not available as Google Books or archive.org scans even though the original editions are long out of copyright.  (Think of this as the print equivalent of why PRR 4700 is at MOT in St. Louis instead of the PRR collection at RRMPA...)

Here is a list of locomotive firing references found valuable by people at RyPN (some of which contain important details of locomotive or auxiliary-device design).

http://www.rypn.org/forums/viewtopic.php?f=1&t=30012&view=print

[gmpullman]

The PRR T1 Trust has made this excellent training book available as a .pdf file. It is quite worthwhile.

https://prrt1steamlocomotivetrust.org/bookclub/download.php

Good Luck, Ed

[Overmod]

gmpullman

The PRR T1 Trust has made this excellent training book available...

This reminds me that CSR/SRI went to some trouble a decade ago, when they were developing their 'Project 130', to prepare some white papers on modern steam technology, which can still be found with a little looking.  Here is an example:

https://static1.squarespace.com/static/55e5ef3fe4b0d3b9ddaa5954/t/55e6373fe4b04afd122b821d/1441150783767/%23+DOMS-1_Chapelon.pdf

[Overmod]

Lastspikemike

Again, for me,  the lightbulb moment was to realize that at full power the pistons are pushed directly by the full steam pressure developed by the boiler.

You almost couldn't be more wrong.  Among other things, you are neglecting the 'other half' of necessity, the getting rid of the steam mass once you're done using it for expansive thrust.

Interestingly, the ACTUAL cutoff that produces highest speed in well-designed reciprocating locomotives with piston valves lies in a surprisingly narrow band, between about 40 and 43%.   Far more 'interesting' are the effects of compression, both necessary and undesired, after the moment of exhaust cutoff back around to subsequent admission.

It is almost impossible to overstate the importance of the innovation of long-lap long-travel piston valves in true high-speed locomotives.  Some very late designs so valued high speed at the 'steam edge' that they used exaggerated valve travels at full -- see the nominal travel on de Caso's last 4-6-4s for example.  

Much more 'fun' can be had in looking at precisely why poppet-valve installations can progressively increase high-speed performance even though their nominal very short cutoffs become worthless in practice over as low a speed as 100mph.  The art involves getting the expansive steam mass into the cylinder very quickly (compared to the event length of a stroke at high cyclic) where it can then proceed to work its expansive magic proportional to the mechanical advantage of the 'crank' in the drive.  It also involves holding the exhaust open either to a larger net opening or a longer duration than that provided by 'symmetrical' radial valve gear, while still fine-tuning compression.

For true fun, look up the Franklin type D patent, where wire-drawing, that traditional ancient enemy of reciprocating-locomotive efficiency, becomes a necessary part of the control system!  Or consider what is needed to make a von Borries compound work at all, let alone be capable of achieving economical high speed... or the potential gains to be achieved by further reducing wall and nucleate condensation during expansion, the principal 'advantage' of locomotive superheat.

Theoretically, if you had no concerns about peak machine thrust or exhaust volume, you could rig something like a Stumpf uniflow engine with auxiliary exhaust porting  to work with full admission at high cyclic.  You'd need an enormous boiler, and either a great many auxiliary tenders or track pans for the ridiculous water rate, but you might develop comparable 'horsepower' to match starting tractive effort up to where the cyclic mass flow matched your steam-generation capacity.  Don't ask how you'd deal with augment or surge; they'd be hopeless without the equivalent of very substantial balance-shaft masses.  Note that even locomotives that could actually utilize that kind of mass flow (the B&O W-1 being a possible example, although its engines would have to be somewhat redesigned to handle the higher mass flow) would have traction issues making much of the increased power, well, intractable.  Even at 6400ihp the PRR T1 likely had great high-speed slipping difficulty with its 'rated' train at over 100mph, for reasons easily recognized (and, in some cases, relatively easily compensated for).

 

[gregc]

as an engineer interested in developing a software model of a steam locomotive, i've been disappointed by the books I've read: Semmens (2000), White (1968), Bruce (1952).   All provide a historical perspective and how designs evolved, but none provided a more thorough qualitative, much less quantitive description of steam locomotive design.

i also contacted someone invoked with the 5AT project which was investigating a modern design of a steam locomotive.    While some were looking at detailed fluid dynamic models of the flow of steam thru various piping of an engine, i never found a more general desrciption.

originally i failed to recognize the purpose of the reverser.    the reverser is a linkage in the valve gear that causes it to provide steam on the opposite side of the cylinder cycle so the engine can move in either direction.     but a significant secondary effect is to cut off the flow of steam to the cyclinder part way through it's cycle.  this has two advantages.   

the first is it extracts energy more efficiently from what steam is used by allowing the steam to expand in the cylinder rather than wasting the energy when exhausted.    The 2nd is to conserve steam which will be consumed more quickly at higher speeds.   i think setting the reverser is more critical than the throttle.

one mystery i was unable to resolve is the pressure in the cyclinder that accounts for the restricted flow of steam through the various piping at higher flow rates.   it is no doubt closer to boiler pressure when consumed at very low speed.   i don't know how to account for the pressure drop through all the piping.

another mystery for me is the properties of steam after superheating.

another intersting books is Firing the Steam Locomotive which describes the job of the fireman.   it makes you realize the both fireman and engineer need to plan ahead in order have the right amount of steam for the trackage ahead.    it's safe to refill the boiler which lowers it's temperature on the level track ahead of a grade when not much steam is needed and start building up the fire so that sufficient steam is available on the grade.

the above empasises that boiler pressure is hardly constant and operating a steam locomotive was quite an art.

i'm guessing that steam locomotive design ideas were held close to the vest by the various manufacturers.   I think the concepts were understood, but progress was made in small steps.   when considering both the complexity of the machine as well as the engineering concepts, there is little doubt why the railroads were eager to abandon steam engines.

[BATMAN]

gregc

originally i failed to recognize the purpose of the reverser.    the reverser is a linkage in the valve gear that causes it to provide steam on the opposite side of the cylinder cycle so the engine can move in either direction.     but a significant secondary effect is to cut off the flow of steam to the cyclinder part way through it's cycle.  this has two advantages.    the first is it extracts energy more efficiently from what steam is used by allowing the steam to expand in the cylinder rather than wasting the energy when exhausted.    The 2nd is to conserve steam which will be consumed more quickly at higher speeds.   i think setting the reverser is more critical than the throttle.

I always wondered why the engineer was always playing with the reverser in all the Royal Hudson videos I watch. Well I learned my one thing for today and I am still on my first cup of coffee.

[SpaceMouse]

Yeah, but I was hoping for enough to get started. 

[SpaceMouse]

Holy crap guys, this is great. 

[Overmod]

gregc

I also contacted someone invoked with the 5AT project which was investigating a modern design of a steam locomotive.    While some were looking at detailed fluid dynamic models of the flow of steam thru various piping of an engine, i never found a more general description...

Part of the problem is that the 5AT people intentionally deprecated theoretical design in favor of 'practical' computations that even people in primitive parts of the world, or after a catastrophic EMP strike on the West, could use to design locomotives effectively.  (This is some of the same principle that governed the reprinting of the Red Devil book -- something you should have in your library right next to Johnson if you're serious about steam design)

You need to invest in the full version of the FDCs (you may have to sweet-talk Chris Newman at the successor to the 5AT consultancy to arrange for a full set) and learn how to abridge and then run them for the design and scale of a 'project' locomotive.  As I've previously noted, Fry's analysis in the early 1920s was reasonable, but involved far too many unexplained empirical constants that were derived from specific norms of locomotive design at that time, norms that do not apply to most modern power or circumstances.  This for example renders the lolog approach to heat transfer in the tubes almost functionally useless, especially if you have an erroneous conception of how Besler tubes and the like work.

Originally i failed to recognize the purpose of the reverser.

Note that the original name for the device was 'cutoff' (when it was a Johnson-bar type device) and the reverse involved using a different set of hooks in the gear.  We use the term 'power reverse' because it makes 'lock-to-lock' adjustments of radial gear like Walschaerts (where the link serves for both forward and reverse motion as well as cutoff adjustment) easier under power -- as in switching or backing down on a train -- but it would indeed be more correct to call it a reversing cutoff.  Very often this is plain in the design of the 'reverser' in the cab, particularly the Franklin Precision wheel type which involves a great deal of cranking across mid to actually get the engine 'reversed'...

... a significant secondary effect is to cut off the flow of steam to the cylinder part way through [the piston stroke]...  

Think of this as the primary effect, with selecting reverse across mid secondary.  

Incidentally the correct approach to starting a locomotive in slippery conditions does not involve the 'canonical' getting the throttle wide open and then cranking up the reverser.  It involves using intentionally longer cutoff and 'driving on the throttle' up to what might be 30-35mph ... the lower MEP but longer effective pressure and lower pressure/torque 'peakiness'  combining with the quick 'exhaustion' of available pressure in a developing spin to make recovery quick and simple, often just with quick followed by fine adjustment of the reverser instead of trying to hog a throttle to fine position, a largely impossible task without a precision air throttle.

The 2nd [reason to use cutoff] is to conserve steam which will be consumed more quickly at higher speeds.

This is reasonable, but don't forget that you have to accelerate engine and train to those higher speeds before sustaining them, and for that you need more, rather than less, cutoff at wide-open throttle.  I don't recall this being 'compensated for' in the detail design of the Valve Pilot or similar devices (which match cutoff with presumably steady-state road speed using a 'needle matching' interface) but it wouldn't be that difficult to make an analog device that would give you rate-of-acceleration vs. additional steam consumption for different operating conditions.  I know of no railroad that thought such a thing was important enough to 'design for' but that does not mean it wouldn't be highly useful in high-speed operation, particularly passenger traffic over 110mph on less than a highly straight and level profile.

I've concluded setting the reverser is more critical than the throttle.

Much more critical: in a modern superheated engine with front-end throttle the 'accepted wisdom' is to get the throttle open as soon as possible and control speed (and acceleration) only with the reverse from that point on.  As I have said, that method was running into trouble with the development of unconjugated duplex locomotives (where you needed both fast-acting and repeatable drop in admission pressure to prevent spin propagation) but even there the 'throttle' modulation would be via fast-acting Wagner throttles (see his patent circa 1912, and it isn't the same 'Wagner' that did the bypass valves on locomotives like 3751 and 2926) on a separate proportional control from the reverse.

One mystery i was unable to resolve is the pressure in the cylinder that accounts for the restricted flow of steam through the various piping at higher flow rates.

You will have to rephrase this, or say more about it, as I can't figure out what you mean.  Chapelon, among others, was instrumental in determining the need to reduce pressure losses up to port admission to the cylinder.  There is a necessary tradeoff between 'dead space' in the passage from port into cylinder volume and expansion volume, but this is solved only with better heat retention/jacketing and reversible compression control for high speeds -- the Q2 testing proved there was no way to optimize 'both' in the absence of those methods; other testing at that period (and after) indicated that higher superheat was if anything counterproductive.

don't know how to account for the pressure drop through all the piping.

The easiest way is probably to cheat: design for least flow resistance and consistent practical sizing down to the ports, then measure the instantaneous pressure with a sensible transducer setup that can graph instantaneous fluctuation and also calculate/integrate averages.  In practice some of the manometers used for engine indication could actually do this at test cyclic (up to about 10rps); there is much more advanced modern instrumentation that can actually do this at points 'out in the mass flow' and measure temperature there as well.

Another mystery for me is the properties of steam after superheating.

This is actually covered (to the extent you need to care about it) in the Rubber Bible (for millennials and you other kids, that's the CRC [Chemical Rubber Company] Handbook of Physical Constants); it is also much more carefully determinable with thermodynamic calculations using the 'right' variables for steam (which is far from an ideal gas even when superheated).  An important message is that each additional degree of superheat adds very little to 'additional pressure' in the cylinder; this is one reason why saturated engines appear to have good power at low speed compared to superheated ones ... and the main point of superheat is NOT to increase the piston thrust.

Consider what happens in an actual cylinder where the steam is doing work on a piston that moves and increases the contained volume.  We accept that all the 'work' is generated by available heat energy, but remember that it is coming out of the heat energy in the available mass of steam, as pressure, and it essentially comes out of the whole mass of steam nearly simultaneously (as the 'vibration' between molecules transfers relatively quickly at these pressures).  If steam were a gas, like that in an IC engine, this loss would be incidental, but steam isn't a gas; it has a phase boundary right at the temperature and pressure involved here, and as the work is done, saturated steam "microcondenses" around convenient nuclei (perhaps dust, or derived from evaporated or carbonized lubricant) in the volume, contracting more than a thousand times as it does so.  Unlike wall condensation, where the steam goes to liquid phase upon contacting the relatively cold inside wall metal surface of the cylinder, or the piston or piston rod, this is a necessary consequence of steam doing expansive work -- and the key is to have the steam so much 'hotter' than saturated that even at the end of its expansion it still has adequate internal heat energy to remain in the gas phase.  And that in a nutshell is the 'miracle' of the Schmidt superheater ... and the degree to which practical superheat has to 'make it down the tract' and into the cylinder to work.  (Remember in this discussion that compression heating also 'superheats' the contained volume, and that is part of the essential thermodynamics of locomotive design too... although you won't find that discussed much outside of SACA)

Most of the 'improvements' pending for steam locomotives during the relatively brief part of the Golden Age where they 'mattered' became unimportant over the same half-decade that more conventional steam did: 1946 to 1951.  And the principal reason was difficulty in finding and keeping the combination of skilled and low-paid workers that was a precondition for steam as railroad motive power at large scale.  One visible sign of the change was that by the time the Roosen motor locomotive was brought to the United States and 'made available for research' no one cared about that sort of improvement, and by the time the Army realized this there was no one back in Germany that cared about it either.  And this was a design more valuable than turbines for typical American railroad work...

 

[Overmod]

BATMAN

I always wondered why the engineer was always playing with the reverser in all the Royal Hudson videos I watch.

He is actually 'driving' the engine doing that; he probably had the throttle full open by the time he got above 20mph and he will keep it open as long as he can, right up until he hits a downgrade and wants the engine to stop making power or 'coast' (and there is a whole master class of fun information about how he will do that).

More fun is video in which you see someone starting a large modern locomotive with a considerable train -- see if you can find the very early sound film of an engineer working one of the NYC Hudsons along the Hudson River.  You will notice just the fine coordination between 'taps' on the throttle and spins of the reverser wheel to get effective acceleration out of a locomotive with which he may have had only a few trips' familiarity ... and, as with most classes of steam locomotive, had very different throttle response or behavior from her 'sisters' and needed to be felt out in the first few miles.  It's in here somewhere:

https://www.youtube.com/playlist?list=PLqKGkdgzE1eMUUM-H1f-WYLXUvE6Ayeje

but I can't get video to run on my current Mac system...

For fun listen to this, and hear the effect of fine changes in the throttle and reverse:

http://youtu.be/X5MtamH9ju8

[SeeYou190]

Hey Chip, I did not read all the responses, so sorry if this is repetitive.

When I was in High School I joined the Florida Live Steamers, a 1/8" (7 1/2" gauge) group that was very active down here. Over a few years I actually helped build a 4-4-2 Atlantic and a 2-6-0 Mogul.

This taught me more about how a steam locomotive works than I could have predicted. Actually building one, you cannot help but learn a lot.

I currently have "part of" a 2-8-2 in my backyard being neglected.

My suggestion would be, if at all possible, to look into live steam locomotive contsruction manuals and maybe find a group. These are sort of a "real thing", and these guys love to chat about their builds.

I would probably be very ignorant of how steam works had I never joined that group.

-Kevin

[Overmod]

SeeYou190

My suggestion would be, if at all possible, to look into live steam locomotive construction manuals and maybe find a group. These are sort of a "real thing", and these guys love to chat about their builds.

This makes sense entirely even outside a 'prototype' discussion -- and if you want to see a great piece of modeling, see if the detailed construction discussion of Ed Woodings' PRR T1 model is still online somewhere.  Every piece of that locomotive was built to scale, including the poppet valve gear...

[gregc]

Lastspikemike

Imagine my disappointment to learn that the driving cylinders are never opened to full boiler pressure and that pressure then allowed to  expand, and that the cut off is always used to prevent this from happening.

i hope it's obvious that the pressure differential between two ends of a pipe are what cause a fluid to flow in the pipe (if the pressures were the same, nothing would flow).    so the pressure at the cylinder intake is only every close to the boiler pressure when there is very little flow at very slow speeds.

early cutoff, "cuts off" the flow of steam into the cylinder, so that it will expand (not prevent it) as the cylinder moves expanding the volume of the steam.    there is probably minimal cutoff (late in the cycle) when starting to maximize the pressure in the cylinder for as much of its' cycle as possible.

of course yet another advantage of cutoff is to allow pressure to build in the steam chest outside the cylinder while the intake valve is closed, maximizing the volume of steam that enters the cylinder during the next cycle.     

as an EE I see analogies to electron flow: pressure is voltage, pipe diameter and length is resistance and steam chest volume is capacitance.

[ATLANTIC CENTRAL]

Chip, you really did it this time, I hope you are having fun and following this because you will learn a lot if you do.

I will give you the really simple practial description, that skips a lot of details.

The engineer has a throttle, which controls the volume of steam, and he has control of the cutoff, which is similar to valves/cam timing in your automobile engine.

By leaning how to use these two adjustments in concert with each other, the engineer does what would otherwise be considered magic, he moves thousands of tons with hot water vapor.

And then he has brakes when he needs to stop, do you know how the Westinghouse air brake works?

That will need to be another thread.....

There is plenty of expert info just from Overmod alone, I will not muddy the water any unless I see a good point to be made.

Everyone in this hobby should understand at least the basics of these things, just my old fashioned opinion.

Sheldon 

[SpaceMouse]

Lastspikemike

Incomprehensible prolix rebuttals will be included in the   model railroading chapter of my new book: "Semantics for Pedantics" quite possibly following my chapter on judicial writing ( not to be confused with judicious writing for which there will be no room in my book of course).

[gregc]

Lastspikemike

I see the valve opening to the cylinder at full steam pressure (assuming the throttle is  fully open) and the steam expands to fill the space inside the cylinder.

yes, but the steam chest pressure outside the cylinder is not necessarily at boiler pressure when at higher speeds.    there's the pressure drop between the boiler and steam chest and the ability to produce more steam.

And yes, while there continues to be flow into the steam chest, it may not equal the volume entering the cylinder at that time, resulting in the cylinder pressure being less than the steam chest pressure.

(i wish i could estimate both those pressures.   I have tried to estimate the pressure on the output side of a throttle based on the area of the opening).

Lastspikemike

If a cut off stroke is used for the valve movement then at some point the full steam pressure is cut off and the steam remaining trapped by the closed valve and the piston face expands by a completely different ratio.

again yes.   after the intake valve closes at some fraction of the cylinder cycle, as the cylinder moves, its volume increases resulting in a reduction in pressure (~~PV/T), but significant work continues to be performed with no additional steam.   this results in the more efficient use of steam, but not maximal possible work output (if steam pressure can be sustained).

the art is recognizing that without cutoff, allowing as much steam into the cylinder as possible each cycle, results in a lower sustainable pressure and sub-optimal work from the cylinder.   at "optimal" cutoff, maximal work output results from the cylinder.   

i think(?) optimal obtains the maximum average cylinder pressure.

imagine trying to control a model that models this behavior, trying to find that cutoff setting while at the same time keeping track of boiler water level, fire and anticpating future needs

 

[gregc]

Lastspikemike

prolix

enjoy looking words up and expanding my vocabulary.

[Overmod]

Lastspikemike

I look at cutaway drawings of the valve chest  and cylinders. I see the valve opening to the cylinder at full steam pressure (assuming the throttle is  fully open) and the steam expands to fill the space inside the cylinder.

Ignore previous 'explanation' in this space.

Keep in mind that a steam locomotive is a complex machine, not just a constrained projectile piston in a cylinder.  Many of the apparent complexities in cutoff stem from this.

First, there is no 'expansion' in lastspikemike's given example -- the drop in pressure is trivial, corresponding to that in a 'fireless cooker' locomotive after the first 'chuff'.  This is due to phase-change expansion in the supercritical boiler water mass, which responds to attempted drop in pressure with massive expansion -- the mechanism behind the rocket effect in boiler explosions, but I digress.  So it is reasonably safe to assume that admission pressure -- whatever that happens to be at a given time -- is reasonably maintained at the end volume as it is at the moment admission pressure is re-established in the cylinder via the ports and dead space -- a matter of milliseconds in a modern locomotive.

It then follows that the thrust of the piston is as lastspikemike suspects it to be, substituting only measured admission pressure for the approximation percentage in the canonical PLAN formula.  And that this produces the highest thrust per stroke that this expander will produce.

Moreover, at the moment of early cutoff the 'equalized' admission-pressure equality is indeed broken, and the subsequent 'expansion' is proportional to the mass of steam contained in the cylinder at the moment of cutoff, expanding against the piston and falling in effective pressure both due to expansion in volume and work done against the piston to move it.  

However, the piston is not driving a linear load (as in a Westinghouse air compressor); it is linked via a connecting (main) rod and a third-class lever arrangement to produce torque at the driver wheelrim.  We conveniently measure this resultant as it gives us an exact force match per driver to use against train resistance (e.g. the result of a Davis-formula calculation).

At or close to FDC (the point we begin this thought experiment) although the effective pressure is very high, the torque produced is very close to zero, not enough in fact to move the engine off center.  That is the reason for quartering a two-cylinder double-acting engine: at this point the expansion in the 'other' cylinder is at half stroke (not quite the point of maximum mechanical advantage on the main pin, but close) and it is doing essentially all the propulsive work at this point.  I invite readers to duplicate the graph in Wardale's Red Devil book to see the torque effect (he plots the torque resultant assuming there is no twist in the axle or play at the rod bearings, every 15 degrees over a full driver rotation) because there are aspects that will be important in a bit.

We assume full-stroke admission, as in the very early English high-pressure engines, and at the end of the piston stroke we will have the mass of steam that roughly corresponds to slightly more than the swept volume of admission-pressure steam -- nominally still with at least slight additional superheat over equilibrium at admission pressure so there is no substantial condensation yet.  All this steam has the ability to do considerable expansive work, but we have no place to send it; the entire water mass represented in the charge, plus the heat involved in vaporizing and superheating it to admission pressure, must now not only be thrown away, but fully expressed from the cylinder on the return stroke.  (Those familiar with Stirling engines will see a different possibility, but that's a different kind of expansion engine).

The catch here is that as the exhaust valves open, the admission-pressure steam starts to expand into the exhaust tract, increasing in volume as it loses pressure and choking the tract as far as exhaust mass flow goes.  This requires some percentage of return thrust to expel.

Now, torque is not the same thing as horsepower, as anyone actually familiar with internal-combustion engines knows.  It is well-established that most good modern reciprocating engines reach their maximum developed horsepower roughly over 35mph (which is why the later Franklin high-speed booster is balanced to spin nearly that high) but although this number would be remarkably higher in magnitude if achieved at full admission pressure mep, the corresponding mass flow can be calculated (from the number of swept volumes, four for a simple DA two-cylinder) times the revolutions per second at peak horsepower ... times 3600 to get the volume exhausted per hour ... and this is the water mass the boiler has to evaporate per hour (we'll assume better Rankine heat recovery from the hot exhaust, but even so...)

Stop and look at this number a moment.

Consider how much tender capacity this represents, and if you expect to operate the locomotive in service, how much tender capacity to do the trick between division points or crew changes.

Then consider how the boiler produces this mass flow, and how large a grate, firebox, and chamber area is needed to do it without losing pressure.  Then consider the effects on the boiler structure if the pressure is allowed to fall substantially, let alone cyclically.

Locomotives that required a large mass flow of steam to produce torque at low speed were indeed built -- Henderson's triplexes, quads and quints were premised on the idea.  Note that even with relatively high fixed cutoff these 'ran out of steam' remarkably quickly ... this being no fault of the induced draft, which was ample, but simply a failure of the boiler to source the required mass flow even at the conservative pressures in the early 20th Century.

Now turn to matters of practical adhesion.  Look at your derived torque curve, and overlay it with a line corresponding to the maximum safe adhesion (from the FA and the locomotive's weight on drivers, to a good first approximation).  Everywhere the graph lies above this adhesion line, your 'torque' will be wasted in wheelspin; the degree to which it varies up and down will likely appear in your trailing consist as surge.  Neither of these is good; you want to keep torque as constant as possible over a revolution.  Proper cutoff in admission helps with this; in fact you can get smoothness comparable to the Swiss-drive mechanics of a good three-cylinder simple.  

[pav]

Another good book that is easy to understand is 'Model Railroader Cyclopedia Volume 1: Steam Locomotives'.

Eric

 

[ndbprr]

I suspect that most of this is far beyond basic concepts. Think of a steam engine as a teakettle. Heat is applied, water boils and produces steam. When enough steam is produced it builds pressure and the tea kettle whistles to say it is hot enough. Plug the opening and eventually the steam pressure will build to wear it blows up the kettle.  Forget who it was in England who made a steam driven pump to remove water from the coal mines which started the mechanical revolution resulting in primitive locomotives that replaced horse drawn rail carts.  Everything since then is ways to produce higher efficiencies just like cars today compared to model T Fords

[Overmod]

Now look at what happens if you try full-stroke admission with the engine at any particular rotational speed -- in a low-drivered engine this might be no more than the 10 to 15 miles an hour where its boiler starts running out of volumetric effectiveness.  You have full thrust on the piston all the way to BDC ... as the rod starts to move around to the back of the crank and the lubrication starts to suffer ... then nearly instantly you relieve the pressure to exhaust and open the other end to steam.  Do not expect your rod bearings to survive this treatment very long, or your axle bearings or wedges or pedestals either.  

Meanwhile you had to make your rods very stout to absorb the monster thrust.  These have appreciable inertial mass, and momentum, and this appears in part as reciprocating augment (the thing overbalance is for).  These forces go up as the square of the peak speed, and while you can compensate for this somewhat by making the rods heavier still, you then have the overbalance guiding problems, including severe nosing and hunting, and very few cost-effective ways to deal with them.  

This method completely falls apart if you expect to run heavy trains at diameter speed or above, where any real interest in steam power starts  Here the physical time available to get steam mass into and out of the cylinders effectively starts to matter; it would be physically impossible to move valves quickly enough to get long expansion (and in fact when N&W tried this with some of the best-designed Baker gear in the industry, it 'unraveled' with somewhat dismal predictability from inertial forces alone when let out to corresponding valve opening)  This is where the fast porting/unporting of poppet valves, and separation of admission and exhaust duration, come into their own, and while cutoff much shorter than 24% had little historical value, improved approaches can use shorter overall admission to good result in faster rotation (and hence higher horsepower per unit time).

I will take up effective compounding in the next post, as most people who haven't studied steam design don't really understand how it works or what to do to improve it.

[Overmod]

Lastspikemike

Hornby tried an electrically heated OO scale Pacific live steamer. I neariy bought one (sale price  CAD$1,000.00. !?!@#?) it was so pretty. Quite the technological tour de force. 

I did buy one of these, precisely because it was a tour-de-force probably never to be repeated, like those expanding tables on yachts from the same general era, or twelve-cylinder production cars from multiple makers.

The problem I had with it was that, like many live-steam models, what it actually did with the steam is not the same as in a 'real' locomotive.  The "throttle" is somewhat similar to how you do steam control in a once-through flash boiler: the more you pump in, the more goes out to the cylinders.  There is no adjustable valve gear at all, as on a donkey engine, and no care for balance at all.  In other words, cool as hell to do, but no more interesting that a typical putt-putt steam launch in complexity.  (And you can teach a kid how to set up, run, and put away one...)

The real problem with it was that, at the time, it was too expensive and too operationally limited* for 'serious modelers' -- it fell into a hole between geek tribes, as it were.  That is something I frankly didn't see coming, and it's a little sad there will likely never be a 're-run' in properly-improved form.

(* and then, there was the oil issue.  If you don't like smoke on a layout, you won't like oil...)

On the other hand, I've been crying for scale operating rodwork for many years.  That is not just scale rods and pin detail instead of soft-metal stampings and slotted screw heads; it's proportional valve gear actuation from reverse position right through to proper Walschaerts and Baker operation servo-driven proportional either to road speed and 'load' (for conventional model throttles controlling DC motors to get speed) or to an actual model power-reverse control with calibratable scale (and perhaps a simulated version of Valve Pilot) for the true tech nerds (and perhaps some geeks as well! )  Note how very simple this would be to implement in its coarsest form (with gear moving either side of mid depending on direction, and spring-centering with power off).  But not one person here, including quite a number of model-railroading gods, sees even much need to replicate prototype valve-gear detail, let alone the finer points of its proportioning.  Some argue that robustness rather than fidelity is the mark of a good model, and I would be among the last to suggest that they, especially by their chosen standards, are wrong even if it produces (to me) toy-train-like appearance. 

And yes, I rack that up in no small part to 'steam technology' being boring to most model-railroad fans.  Perhaps this thread can start to change some hearts and minds...  

[gregc]

ndbprr

Everything since then is ways to produce higher efficiencies just like cars today compared to model T Fords

not sure what you mean by efficiency.    it typically means extracting the usefullness from something with minimal waste.

i've described how allowing steam to expand within a cylinder improves efficiency -- more work from the same volume of steam.  one way of evaluating this efficiency is the energy in the coal consumed compared to the work produced by the engine

automobile technology has certainly improved since the model-T.   internal combustion engine efficiency has certainly improved, but is still only 20%

i don't think it's accurate to say "Everything since then is ways to produce higher efficiencies".   there are different goals.

a major drawback of steam engines is the time required to obtain full power.   other technologies such as diesel or turbine engines minimize this time.   not sure the benefit is efficiency or less complexity and easier maintenance.

not sure how to evaluate the efficiency and benefits of electric vehicles.  efficiency may depend on the source of power: coal, nuclear, water, ...   Benefits may be zero pollution from the vehicle, zero startup time, higher torque, ...

while efficiency or fuel cost may be important, performance and other contraints may be even more important.   electric locomotives to go underground have a unique purpose

[Overmod]

Lastspikemike

That control conserves steam and gives finer control over cylinder pressure and hence torque at the driver than the throttle can. That results from switching over from "boiler pressure"  control to effective control of the actual expansion ratio inside each driving cylinder.

But don't make the mistake that this 'fine-tunes torque' exactly; it does adjust it, but makes it peakier at the same time, which contributes to both a propensity to stall and to spin at times ... not a fun place to be!

Better to think of the controls of a steam locomotive as double-salient in a sense: the throttle controls the admission effective pressure, and the cutoff controls the mass flow available for expansive working starting at that pressure.  There will be times -- plenty of times, on excursions! -- where you don't need more than 150psi or so throttle pressure, but you need to start a train without slipping on uncertain track and then operate it efficiently.  You do this by coordinating the throttle and reverse appropriately.  For fast or 'most efficient' running, you want minimum actual impediment in the steam flow ... and the poppets in a multiple front-end throttle are an impediment.  Then you crack it open and, once you're at a speed where surge and low-speed augment are no longer problematic, drive on the reverse ... until you get into the range of 'high-speed slipping'.  At that point things start to get interesting...

... some turbocharged engines have a driver controlled boost pressure control which has a similar function.

Be careful with analogies to internal-combustion engines.  The point of turbocharger boost is to increase the amount of fuel that can be burned during a stroke, and there will be diminishing returns at some point where the complications of higher peak firing pressure, peakier torque, and exhaust back-pressure outweigh the added boost.  That is not a complication with an external-combustion expander, which is not volumetrically limited in its ability to utilize pressure expansion.

It will not help when some astute readers find out about John Sharpe, Gotaverken, and 'steam turbocompounding'.  It does NOT help to recompress steam with exhaust-steam energy to get more thrust out of it.  The recompression is to get a phase change to lose some of the 'excessive' heat in exhaust steam so it can be run back through the boiler at 'liquid' density without losing the latent heat of vaporization ... something that would induce MEGO syndrome in most everyone here to explain   (See also the glory that was Holcroft-Anderson 'recompression'...)

While we're on the subject of turbocharging, it is fun to consider the Velox boiler, a fascinating dead end in sophisticated design.  It starts with recognition that external combustion works ever so much better if the fire can be turbocharged -- pressurized up to 30psi as in some (ultimately misguided) electrical generation power boilers.  The Swiss neatly used a gas turbine(!) to perform this in a small boiler to generate steam; I can't do nearly the justice to it that Duffy did for the Newcomen Society -- or that Douglas Self did in a somewhat less enthusiastic way on his delightful site.

There is another perhaps interesting story about MEP in steam engines.  A man named Tuplin, in England, was something of a bore on the subject of high boiler pressure and superheat -- he thought them unjustifiable in terms of increased maintenance costs for the benefits gained.  He was fortunate to observe the crew of a NYC Niagara, probably doing a break-in run on local freight out of Harmon, performing the work of a relatively small 2-8-0 on the 2-8-0's consumption of coal and water.  That was done by carrying fire on only some of the grates, with careful firing attention, and sliding-pressure firing the boiler only up to what I recall being a peak of about 180psi.  Now this was something of a waste of the capital involved in a 6000hp high-speed passenger locomotive, but it demonstrates a different way of working a locomotive efficiently to task.

 

The gas pedal in a steam locomotive is a combination of the cut off ("reversing lever" ) controlled by the engineer and the combination lever which is automatic and fixed in its pattern of operation. The really fascinating part about all this monkey motion is the skill of the erecting engineers in matching the combination lever geometry to the reversing arm. The combination lever moves the actuation point of the  radius rod (the rod controlled by the engineer by actusting the reversing lever) on valve motion to keep in sync with the linear stroke of the crosshead by matching that motion to the eccentric motion of the valve gear crank which in turn is offset approximately 90 degrees off the  crank pin orientation. The length ratio of the combination lever is critical to the fine control of the reversing lever by the engineer.

I am not sure if you understand what the 'combination' lever in Walschaerts is actually there to combine.  It synchronizes some of the crosshead motion with the 'commanded' valve excursion through the reverse to get effective valve motion; it is only a baseline setting and doesn't affect 'fine control' over the valve motion at all, really only its phase relative to mid.  (The existence of the combination lever is the explanation for 'nightwalking' of locomotives left in mid position on the reverser but with slightly leaky throttles...)  It may be easier to see the 'result' on locomotives like the D&RGW M-64 4-8-4s which had variable lead adjustment.

Perhaps the best way to see what goes on is to download and run Charlie Dockstader's old DOS valve-gear modeling programs, which let you play with the dimensions and see the effect on the valve motion and timing in realtime: a picture is worth a thousand prolix words.  They were updated for Windows XP and have been verified to run on VMs available 'free' in the cloud for those whose systems won't allow legacy software from that era to run.

Something that ought to be of interest is the device in the cab that controls the reverser position.  Even early versions of power reverse could have interesting precision combined with speed, an early example of high-power servomechanism without 'force feedback' much like the controls in a Hulett unloader (which you have to see operated to believe!)  The version on UP 3985 could be slewed from full forward to full reverse in less than a second, with locating precision 'locked' at any point in that travel when achieved (needless to say if you did that at speed you wouldn't enjoy the brief time for observing the results!) and there were several different approaches to getting the fine precision control over the reverse that could be implemented for control.  

There were two styles of Franklin Precision reverse -- one was a wheel type, as fitted to locomotives like NYC Hudsons; the other was a lever type that could be moved more radically in a short time, as on Nickel Plate 765.  The 'catch' is that it's difficult to do fine positioning of the lever type, or quick movement and recovery of the wheel type.  [Incidentally the reverse problems on Blue Peter are not to be found on any American locomotive with wheel reverser...]  

The British had a power reverse of equal precision, for locomotives in their practice that used vacuum brakes and had no 'cheap' source of power air.  This was the Hadfield, circa 1950, which used steam as the working fluid (much as steam was used for debounced valve return in British Caprotti poppets)  To my knowledge they did not evolve a better fine-or-fast control than Americans did.