Please continue reading in the other forum section:
viewtopic.php?f=6&t=48170

Note to most serious preservationists here: if you don't love minutiae of steam technology, this is REALLY going to be TL;DR material, so lasciate ogni speranza and all that if you continue reading.

I take up this discussion with a couple of basic considerations:

The purpose for which the locomotive is intended is very important. An eight-coupled locomotive no more capable than a GP40 is no longer particularly relevant, particularly if its cost and complexity far outweigh any putative saving of fuel and water. There is a long history of attempts at full or partial condensation, not one of which (including Holcroft-Anderson recompression) has worked in general railroad service "competitive with diesel-electric equivalent". Likewise, a single-unit locomotive of more than about 8000hp has historically had too ridiculous a water rate to prefer its operation to equivalent 4400hp commodity locomotives that can be easily MUed without the usual unattended-firing comedy.


Not a sensible requirement for a 5000hp locomotive that is intended to operate flexibly between random high and low power (which is something a typical 4400hp diesel-electric routinely handles nicely even if made by GE). There is a long history of unattended-firing design for the required combination of power and power density, none of which is particularly well-suited to solid fuel firing on railroad locomotives; we have see some frank disasters, for example in India and North America, when liquid-fuel firing is not properly appreciated. Even with automatic firing controls, you need the ability to provide a set of intelligent eyes on the unit at any given time, which at the outside involves two men per consist. Yes, I can program a Boston Dynamics Atlas to do some of it. That's not cost-effective yet.


This was one of the drivers of the abortive attempts in the Eighties to have 'native' alternatives to fossil oil locomotive fuel. Some interesting work, including the Porta revised 2-10-0 and 2-10-2 designs, came out of this, but nothing that would 'better' fuel synthesis for biofuel in diesel-electrics. (I am particularly sad that none of the GTCC approaches with SRC 'thrived' as it would have been fun to see how steam bottoming at appropriate locomotive scale worked out...)

I see you mention this and watertube locomotive boilers in the same post. Theoretically those don't really go together in the sense you're using, at least not for large locomotive practice on a typical PSR-addled operation. Where you want 'watertubes' is in the radiant section, as a defined-circulation waterwall; conventional HRSG tubes and flues do the convective job just fine in a way that is comparatively easy to maintain and service, and that accommodates the advantage of a Chapelon/Porta 'sectional feedwater heater' section with almost no overhead or maintenance downside.

We can take up the first cost and maintenance cost of various types of economizing and bottoming later. These do become more important with mandatory water treatment, and their practicability increases with modern materials and coatings.


The problem is that 'that ship has sailed' and no one cares about zero-net-carbon any more. The game is zero-carbon, and there is no cost-effective method, carrier or otherwise, to do this on a reciprocating steam locomotive. (I am interested to see if Steamology can manage it with modular steam generators and turbines using hydrogen, but I’m not too sure they are going to avoid things like long-term embrittlement during what will likely be frequent modulation)

I don’t think anyone serious would try running a locomotive on ammonia, but it would be very, very difficult to convince me it would be a good idea, and that would be before asking me if I’d invest in the development.


A great part of this is actually a non-issue, or rather is better solved with things other than choosing a boiler architecture that nominally survives large pressure and temperature swings in operation.

To me, the ‘proper’ solution to availability in service is to use an appropriate combination of ‘Direct Steam’ plant for bringing the boiler up to saturation temperature and pressure via staged steps (and dry-gas purging of oxygen content) and the bypass electric-heater approach that was used on converted Kriegslok 8055 (which used 35kWe or so to keep the boiler at 300psi net of auxiliary use during the time the locomotive was intended to stand idle without firing for a reasonable interval, i.e. a weekend). How you handle combustion-gas-side heating aside from good insulation of draft path is far less a complicated matter than keeping a complex watertube nest maintained clean and straight.


This is much, much less a concern in the modern age of PSR and high insurance rates for excursion or special passenger service in North America. Even in civilized places like Britain, where ‘historic preservation’ traffic is expected to be able to run fast enough to preserve complicated pathing, there isn’t yet a requirement for sustained speed over 100mph. And while I and others have had fun over the years figuring out how to get reciprocating locomotives to run at speeds up to 150mph, I certainly wouldn’t recommend actually going at very high speed with passengers aboard the train…

If you actually want something that looks like a legacy locomotive that does go 125+mph, the ‘proper’ solution is one of the evolutions of Turbomotive2, which was (erroneously, in my opinion) rolled into the 5AT effort to avoid any ‘distraction’.

If you actually want steam that goes 186+mph, it would be GTCC+bottoming, or something like the Oxford Catalysts system, with electric transmission to more-or-less-conventional HSR trucks and suspension, in an appropriate carbody.

Not sure I’m convinced his approach scales to 5000-8000hp, which it would have to if you intend to make any economic sense vs. ‘commodity’ 4400hp diesel-electrics with MU…

To my knowledge, this would not convince any State agency concerned with licensing power boilers of the appropriate rating for anything more than a toy locomotive. I do not expect them to be impressed by the Mackwell design to issue waivers, either, although that certainly could be tried in particular states by groups with the right political savvy or connections.

The initial problem with this is that (again referring to the 8055 example) the actual cost of fuel represents an almost vanishingly small proportion of operating expense – on the order of 5% overall. If you add back added maintenance expenses, ash handling, or other concerns introduced by the wrong economy decisions about what ‘waste’ you’re going to fire (in particular, waste automotive lube oil, which will also be going away to a dramatic degree with supposed mandatory vehicle electrification, or solid-fuel firing with untorrefied renewable material… we already have seen WVO at locomotive scale become essentially unsustainable as ‘too cheap to meter’) you may in fact have to fall back on some kind of virtue signaling. As previously noted, no one in the AGW camp cares any more if your fuel is ‘carbon neutral’.


Steamology, if they are using what I suspect they are for steam-generator architecture, probably have a ‘structural’ module fire-up time no longer than that for a good Doble or Besler once-through generator – on the order of a minute or so. THAT is the criterion if you care about ‘rapid fireup independent of maintenance concerns’.

Staged direct steam with circulation, even with a conventional ‘Stephenson’ radiant-section construction, involves a controlled rise in the various temperatures of sections and components of the boiler structure in a way that avoids point stresses of the ‘usual suspects’ areas – if you then add control over the gas-side temperature up through initial firing to achievement of working saturation pressure, you solve most of the ‘other’ difficulties with primitive approaches to ‘cold’ firing. The necessary equipment can be vehicle-borne, probably using some variant of commercial swap-body mounting, and this of course suggests some independent development along the general lines Hulcher has come to offer for certain other types of economical support provision…


I have no intent, let alone any desire, to criticize Roger Waller and others, but I do not see an active expansion of ‘Plandampf’ over the past decade and a half, and I still am ‘nervous, like a Christian Scientist with appendicitis’, that the scheme to provide modern steam for Swiss commuter service went nowhere and has not been seriously resuscitated. In my opinion, a much more systems-oriented approach needs to be provided – just as was supposed to be done for those hydrogen battery trains – combined with commercial assurance that all the necessary supplies, maintenance, etc. will be available from multiple compatible sources over the anticipated lifetime of the equipment.


Here is where we get to the ‘catnip’ part, at least for me…


The actual limiting factor in the case of N&W 610 was valve tribology, and the concern was further expressed by David Wardale in some of his piston-valve detail design. I will take up the issue of inertia vs. opening and closing to steam in a bit, but there are in my opinion relatively few things that a good poppet-valve system (like that on BR 71000, British Caprotti) will do at high cyclic that a proper dual piston valve system (driven separately in duration and timing for inlet and exhaust) could not do without the need for complicated gearbox systems and convoluted tracts.


This is nonsense; the concern with inertia is not a major consideration in long-travel valves; only the weight of the valve spool is. Were you to extend the principle of the Willoteaux valve slightly, to comprise rotating the valve spool as well as reciprocating it, you can get relatively enormous port unshrouding rates and streamlined flow openings all the way through from the admission chest to the ends of the passages; if you then control your compression correctly (with perhaps a little Chapelon-style modulation of steam during the subsequent return) you can get admission pressure from passage to piston face consistent with chest-to-valve pressure, so there is no loss from larger or better streamlined passages as if they were ‘dead space’.

Whether this lets you actually decrease the acceleration needed for the valve spool to get ‘acceptable’ mass flow for high speed is a question for detail design. But I suspect you’re going to run into guiding and critical-speed concerns before it starts to get excessive…


Mass balance in the valve drive was essentially solved by Cossart, who also had a reasonably low-mass alternative to pure poppet valves in his ‘drop valves’ (which have much the same advantage as piston valves in requiring no flow vector changes during unshrouding). One might argue that we should return to finding out exactly what mass distribution in ‘salmon rods’ actually produces the best overall augment reduction; we can also start determining whether variable mass in parts of the rods (done with fluid admission) is of practical benefit in overall balancing.


For those who are not crazed steam technologists like me, the ACE 3000 adopted William Withuhn’s ‘conjugated duplex’ system as described in an early-Seventies issue of Trains Magazine: it is essentially a double-Atlantic duplex with the two engines 180-degree antiphased, with a set of quartered inside rods holding them in that (anti)phase, and the cylinders as they are in the B&O and ATSF duplexes or the PRR Q1. The fact that the ACE 3000 was a compound requires that careful (but not difficult) accommodation to piston and rod masses be done – in my opinion, the system would have needed Chapelon-style IP injection to keep proper balance at very high speed, but the locomotive was never intended to be operated above ‘conventional’ SD40-2 speed.

What was NOT covered in the ACE 3000 patent, and in my opinion needed to be, was the proposed frame construction that would have been necessary. What is in the patent is laughable, on the order of a child’s drawing, and (just as an example) would almost certainly have needed tunnel-crank bearings on both inside mains even for 3000 equivalent horsepower.

I would argue that a much better solution is Deem-style conjugation (as designed in the ‘80s for a revived version of the Q2) which uses gears between the engines to maintain phasing. In my opinion, though, it would be a mistake not to include a considerable amount of ‘quill drive’ accommodation in the drivetrain; in fact, I advocate the provision of a Ferguson clutch (or its magnetorheological equivalent) between the engines, so that the engine can run, and be suspended and guided, as a regular duplex at any time conjugation isn’t required, and so that any differential wear between engine driver diameters doesn’t have to be frequently dressed out.

You might then use 135-degree phasing (as on some fairly successfully balanced designs) to get better distribution overlap of torque, with the effect of spreading any admission or exhaust issues over eight rather than four pulses per revolution with typical 2-cylinder DA engines. Nominally this would add back a small fraction of necessary overbalance with the engines kept conjugated, and this would likely be more than the ~80lb Voyce Glaze kept in the N&W J mains (it is supposed to be the vertical component of piston thrust with the locomotive at ‘dash’ speed with a typical consist). BUT with a locomotive the size of a double Atlantic, there is really no reason not to go far toward zero overbalance, with better lateral steering at the ends of the locomotive chassis handling the yaw moment, and better damping to prevent augmentation into hunting.

The major concern at that point is also partially ameliorated by a large heavy locomotive with a large heavy tender: surge. It turns out that there is a definitive solution to surge (although not, as its patent hinted, to overbalance-induced yaw or hunting) in the Langer balancer (patented by Westinghouse) which acts like a large inertial transverse balance shaft on the locomotive centerline.

There are a number of weird things so far. You would not use 80” spacing on a ‘converted’ T1 because there is no real advantage to maintaining the same rigid wheelbase with smaller drivers; you’d change the frame spacing. “Less valve gear parts” indicates some suicidal idea that you only need one set of valve gear drive on a conjugated engine – perhaps you could approximate this on a locomotive with cam drive of typical poppet valves, with shafts from one eccentric-crank-mounted drive ‘fore and aft’ to the camboxes in a Withuhn arrangement… but the precision of anti-backlash gears you need to achieve – let alone maintain – the necessary timing precision for high speed will eat you alive.

The ‘magic’ driver diameter if you’re staying up in the 70”-plus range is 76” (or its nearest metric equivalent) – that being the dimension that ‘fits’ under the double-Belpaire chamber at the boiler dimensions consonant with traditional AAR loading gage with lowest running center of mass. If you want short rigid wheelbase and overall length reduction, which in both cases are advantages for most practical locomotives in “modern” service, you’d go straight to the one-piece-cast 58” wheels in the ACE 3000 proposal… or somewhat smaller if you think you could make a Roosen-style motor locomotive with detail improvements. (The Roosen engine apparently worked with a single, albeit complicated, reverse gear per side, something I suspect would be done differently on a modern version…)

Since I personally have little faith that the Besler constant-torque locomotive would have worked out in actual practice, even if there were no comparison with early E-unit diesel-electrics as an alternative, I won’t say anything other than ‘look elsewhere for technology’. I could probably be convinced out of that if you’re good enough with detail design, but it would have to be very good in an awful lot of places that the Beslers didn’t attend to…


Oh, hell-to-the-no NO to that idea. The Russians tried it and it predictably ‘failed to thrive’. Even if you tried necking the cylinder support to get the rigid wheelbase anywhere remotely sane, you’ll run into rod angularity issues really, really quick… unless you even further extend the rigid wheelbase with three sets of rod-connected driver axles per ‘side’. Then you can start explaining how you modulate steam flow to the outer ends of the cylinder vs. the much larger ‘opposed-piston’ space between pistons. While this would have comic value around April 1st, it is unlikely to be something that safely and cost-effectively satisfies the necessity that pure electronic timing and control be used on a large reciprocating locomotive.

Antiphase as on the Withuhn plan does just as good a job at ‘horizontal balancing’ as central cylinders does; in fact you will note that all the four-cylinder early duplexes put the cylinders where Withuhn does for highly advisable reasons. Conjugation with Q2-style rear cylinder mounting does most of the job a
center cylinder block would, without all the weird tinkering, although we’d still need to remember that the Q2 benefits from having the three-axle rear engine to enable the overall wheelbase reduction at acceptable rod angularity…

[
Why did they have to be ‘balanced’ in pairs at any sane achievable cyclic?

Which any double-acting engine could have with nothing more complicated than Chapelon-style Willoteaux valves (which further were made of thin material fabricated by welding, something greatly facilitated by modern machining techniques)


Until you have very careful compression control, “larger sectional area” in your passages is going to have no better result than was observed on the Q2 during testing – where they remarked that the observed ‘dead space’ was excessive. Think for a moment about how you have to compensate for that…

I’m surprised no one has mentioned that great innovation by Bulleid, the sleeve-valve system essayed on the Leader-class engines. If you want the shortest possible passages, the straightest flow from ‘chest’ to cylinder, and the easiest (at least nominally) way to reciprocate and revolve the valve for fastest port opening, you need look no further than what Bulleid did (admittedly you will need a somewhat more sophisticated drive for the fore-and-aft sleeve motion than he used, but it shouldn’t be rocket science to design). The actual required power to drive these ought to be less than that for poppets of typical Lentz style with suitable spring restoring force for high cyclic (which is higher than expected given the necessity of debouncing valve closure without going to desmo cams) and might in fact be less than that for a proper steam-closed British Caprotti setup.

This is a fun discussion as far as I’m concerned, although I admit its connection with strict historic preservationism is more than a bit tenuous even with exhaustive research of historical sources.

The German translation for "inertia" means both rotational ans translational masses. Of course I meant the mass effects of linear movement when I wrote "inertia". And this also meant the rods and levers moving in about the same directions.

Of course I should have said 'mass' instead of 'weight' in English.

The inertial force on a traditional piston-valve spool, even at less than 'de Caso' travel and its associated peak acceleration and implied 'machinery speed' for valve lubrication, can be significant. For example, valve-stem stresses for an early-twentieth-century 16" piston valve (massing 244lb) are already at 5825lb at 60mph (with 80" drivers) whereas a 12" valve under similar conditions involves only (!) 4600lb. I confess that I am not in possession of the peak force in de Caso's U class that had (as I recall) about 15.9" valve travel, but as this was on a nominally-high-speed design in the French context (which may be artificially constrained by their 120km/h peak speed limit) I can't help but conclude he thought the peak stresses in the valve-gear components were acceptable even at that extreme peak machinery speed with careful detail design.

To this consideration, as noted, we have to add the balance forces associated with the valve drive, which includes the combination of revolving and reciprocating force in the eccentric rod (which we can approximate via center-of-percussion to a reasonable approximation, and then partially compensate out via assembled dynamic balancing).

To note that these issues are significant to high-speed steam design would be an understatement -- one that figures significantly in the history of the Franklin System of Steam Generation, for instance, and somewhat differently in the history of Cossart's approach (cf. the Algerian Garratts). Ed King makes reference to the tendency of some N&W installations of Baker gear to (in the charming term they used) "unravel" at excessively long cutoff at speed, and I believe issues with valve tribology not independent of valve-spool mass were the cause of failure in the PRR tests of J 610.

In the Franklin system, the mass is reduced to the minimum, unshrouding is accomplished at two sets of sealing surfaces, and the drive is supplied via cams (which can be ground with a modified-trapezoid profile if jerk should become an issue in high-speed actuation). The cam drive, in turn, can be taken off an external return crank, as with 'usual' radial valve drives like Walschaerts or Baker, but with the initial drive gearbox driven directly off the arm of that crank -- this reducing most of the lateral 'hammer blow' contribution to augment to the lever arm from the center of mass of the crank and associated assemblies carried by the main-driver suspension across to the contact patch of the other driver in the pair. Note that this mass can be effectively balanced even though the driveshaft is 'hinged' at its forward end as it is nearly entirely rotary; there is a little length change during suspension movement, but its effect is essentially trivial.

Cossart's system is a bit more interesting. He installs the reciprocating valve motion antiphased with the cylinder drive, so that its inertial forces at least partially 'cancel out' those in both the vertical and longitudinal planes, making the forward end of the eccentric rod more heavily massed (the typical English translation of these being 'salmon rods'. Interestingly, French practice appears to be to progressively lighten these rods, first with small holes and then much larger ones, which might indicate that lightweight mass might be more significant than compensating inertial forces in the reciprocating masses.

As with the Berry accelerator, Cossart also has his valve drive considerably 'out of phase' with how his valves will be moving, so he uses a jackshaft at the forward end of the eccentric rods, which then drives a camshaft arrangement. This could probably be arranged to drive Lentz/Franklin style poppets, but the required shafting or linkage to get a proper 'straight-line' steam path might be... let's say 'somewhat involved'. What he has instead are what (again in English translation) are called 'drop valves' -- these are like a 'best combination' of piston and poppet valves, and as with British Caprotti, they are held seated by steam pressure (which can be facilitated using the design of Porta's beloved Wagner (he spells it 'Waggoner' but don't be fooled) proportional 'servo' actuation) and have ring seal for positive location at the top, and an extended poppet-style seat at the bottom. This is a much more attractive arrangement than Franklin type A in a number of respects, although I have no equivalents of indicator diagrams or test-plant results that show the steam mass flow at high speed and load.

Certainly Chapelon's design of Willoteaux exhaust valve qualifies as an attempt to produce minimum mass in a piston-valve spool -- it is fabricated out of welded plate!

I've seen news of plenty of new build locomotives completed lately in the UK, including a chance to see one in person on the Ffestiniog this year. I think financial, business, and use case planning coupled with fund raising were as if not more essential than the engineering concerns. None of them had all the bells and whistles a modern disciple of Porta may think of, but all are beautiful machines that are functional this very minute.

So what is better, to postulate the thermodynamic and mechanically perfect steam locomotive that somehow saves the environment? Or build a locomotive that while archaic matches a historic preservation need? Turns out the school of the archaic replica machines has yielded more real world fruit than the theoretically perfect path.

This of course is much more a relevant formulation of the 'original posed question' for a historic-preservation "Interchange" discussion!

Porta himself could be a proponent of 'appropriate technology' -- the 0-6-2 prominently if a bit misleadingly marked 'biomasa' being an example that jumps immediately to mind. We have had multiple discussions here over the years about incremental improvements to 'historic' steam through the use of modern materials, and the periodic pitfalls that lurk for innovators (Teflon seals in valve gear, anyone?)

Much of the last few posts in this thread 'presume' one of the conditions that the OP posed as important for a modern design: its ability to reach high road speed. This is in some respects more of a 'European' (and particularly British) consideration, as both high peak speed and reasonably assured reliability reaching and sustaining that speed are essential to their mainline operation.

Historically, the most 'important' design factor is to assure economical 'automatic action' across the usual range of loads and road speeds without complicated 'mecanicien' adjustments on the locomotive or things like jumper caps in the front end. One could argue that that is something significant in 'compromise' design for a fairly wide range of potential 'new steam' opportunities that have to appeal to old steam aficionados...