PRR K4s 5399

2108 views
43 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    September 2003
  • 13,412 posts
Posted by Overmod on Thursday, September 24, 2020 8:18 AM

It was supposed to address some of the problems with 'steps' in RC poppet-valve gear, even present in the British Caprotti system on Duke of Gloucester, that do not allow the seamless continuous control of cutoff that good long-lap long-travel radial or link gears do.  (We would see this bug billed as a 'feature' in Franklin type D, but that's a different story with very different purpose in mind!)

The Franklin solution was of course the original "type C" with the continuous-contour shifting camshaft and the spherical roller followers, which was nifty as long as the gear wasn't operated in the 'best' notches for general running too long.  The approximation to line contact (really, point contact at tangency) caused distortion of even hard-plated surfaces over time; some but not all of this is relieved with either Wagner-throttle servo assist or "steam actuated" valves (as vertically in British Caprotti and somewhat dottily in Jones) which don't have necessarily heavy spring preload on the valves to "debounce" them in service, or require desmo.

I'd expect springs to be the excuse when the steam-valve maintenance gets troublesome, as it surely would in a post-'48 BR sort of shop environment, and even progressive springs effectively debouncing poppets at a meaningful rotational speed to justify poppets in the first place are going to play hob with the adjustable Reidinger nightmare box, sooner or later.  Of course the thermodynamic improvements weren't really there for cutoffs shorter than about 15%, nor I think were compression effects correctly considered in designing the exhaust cutoff duration and timing arrangements.  So it was an incompletely thought-through solution for a problem it created itself ... not an auspicious start to a revolution.

  • Member since
    January 2002
  • 4,087 posts
Posted by M636C on Thursday, September 24, 2020 7:48 PM

Reidinger was not really "infinitely" variable unless I misunderstand the purpose of the "dog clutches" shown in the patent. It did work well enought to keep it in service. The LNER rebuilt the six D49s fitted with Oscillating cam Lentz gear with piston valves (although it ahould be remembered that the centre valve was driven by Gresley Conjugating gear anyway....)

The timing of the Reidinger gear, with the patent application in 1938 in the UK, suggests that it was intended to address the problems of the P2 2-8-2 "Cock of the North" which was built with continuously variable cams but was rapidly converted to stepped cams due to the well described problems. Perhaps the LNER had already converted 2001 to piston valves by the time Reidinger was ready, but were happy to try it on a 4-4-0. The conversion of the five class 5 2-6-0s in 1954 suggests that BR were still willing to try an option for rotary cam drive.

Of course, as Vernon Smith points out, the gear on 5399 was built using Lentz patents. Smith was dismissive of the L.B. Jones poppet valves on PRR 5436, indicating that it had an irregular exhaust beat and unexpected loud noises were heard from the valves, suggesting that valves were sticking and releasing out of cycle. He says it was only ever used as second engine on double headed trains. I have seen a photo showing the cylinders and front end of 5436 and there is no indication of the drive.

Could this be the poppet gear drive of 5436?

Peter

  • Member since
    September 2003
  • 13,412 posts
Posted by Overmod on Thursday, September 24, 2020 8:54 PM

M636C
Reidinger was not really "infinitely" variable unless I misunderstand the purpose of the "dog clutches" shown in the patent

If I understand the design correctly, those are not Maybach clutches; they serve the same purpose offset keys do in setting position of a cam on a shaft, but do it with much more ling-term shear resistance.  One of those rings would be made up for each of the cams to set it precisely by degree so that it would never, ever slip or shift but be precisely adjustable in the shop.  See for example p.4 lines 31-35 in the patent.  This was arcanely but accurately referred to in the brief mention of this gear in the Encyclopedia of World Railway Locomotives, the place I came upon the promise of this gear as a kid.

Did we not decide that Spicer drive was an oil pump for the valve gear in 5399?

Incidentally, the Jones patent, 2,260,458A, can be viewed (and downloaded as PDF) here.  Who can be first to spot some of the issues this gear might have in practice? 

  • Member since
    January 2002
  • 4,087 posts
Posted by M636C on Friday, September 25, 2020 3:03 AM

Did we not decide that Spicer drive was an oil pump for the valve gear in 5399?

I certainly suggested it but that was before I saw this drawing of 5436...

 

Peter

  • Member since
    September 2003
  • 13,412 posts
Posted by Overmod on Friday, September 25, 2020 7:10 AM

Certainly the right size and configuration, but the drive in the patent is presumably a multiple-thread worm on a sleeve over the axle driving the timing shaft, while the arrangement pictured would not work that way, perhaps involving spiral bevel gears.  Note that this is the opposite of an automotive rear end where the action optimized here would be 'backdriving'; in the Jones gear the shaft only rotates the 'distributor' either way (presumably with angular precision) the cutoff being determined from 'full gear' through a neutral to full reverse by sliding part of the driven distributor mechanism axially.  I think the presumption was that shaft torquing or cumulative play wouldn't be a problem because it's only turning a fluidic distributing valve, all the sealing being in shear...

I did think it was too stout for a lube pump, but I thought the same thing of reverse gearing and planetaries for transverse direct-drive steam turbines...

Some valve gears -- the one on Bulleid's Leader, for example -- used one method to drive the gear and a separate one to time it.  That allows the valve motion to be quick and positive, but its timing and duration at high speed precise.

In Bulleid's case, he had 'a plethora of rings' in a Meehanite sleeve that had to be kept consistently lubricated well enough not to break.  To facilitate this he had a drive arrangement that gently rolled the sleeve to and fro while the main gear urged it 'to and fro' past the port openings (which were all the way around the cylinder at each end, the effective dead space limited to the aggregate volume of the ports through the sleeve).  A chain drive from a sprocket on the leading bogie axle -- correctly deployed running in one direction instead of periodically reversing -- ran the stiction-precluding, oil-spreading twist.  The combined motion was said to be fascinating to watch...

Something to remember in Jones gear is that it is mandatory full servo: there is no physical connection to the valves at all so that if any part of their construction should stick or bind there is nothing but fluidics (or assisting return springs) to 'urge' them.  I have always thought of this principle as a grave mistake on an operating high-speed reciprocating locomotive.

Part of the issue is to be read 'behind the scenes'.  A popular pastime of poppet people was to make the use of higher superheat look more practical by eliminating lubrication (see Wardale's discussion of this in the 5AT correspondence over adopting Brutish Caprotti) both in the valve motion and by reducing coking by providing steam displacement of air, or steam-displaced bypass as in Nicolas or Trofimov valves.  A problem is that there is already the possibility of ridiculously high superheat in locomotives eith front-end throttles operated at high speed (I think it was Ross Rowland who used the expression 'crazy high') and those of us who remember Stumpf and his barrel-shaped valves look very carefully at the tribology of the guides and pistons in the driving end of even a 'multiplicity' of small valves running in such steam for an extended period (that could he considered a 'soak' period).  To me this implies a need both for Wardale-style cooling of the 'back ends' of the valves eith exhaust steam and for some combination of high-volume circulation (independent of fluidic use) of the oil in the valve actuation circuits combined with oil-temperature tempering (again perhaps with an exhaust-steam or 'boiler water' intercooler but the risks of 'water in the oil' are disturbingly high (for anyone who has owned a Ford 6.0L diesel!!)

  • Member since
    January 2002
  • 4,087 posts
Posted by M636C on Friday, September 25, 2020 7:23 PM

Certainly the right size and configuration, but the drive in the patent is presumably a multiple-thread worm on a sleeve over the axle driving the timing shaft, while the arrangement pictured would not work that way, perhaps involving spiral bevel gears.

Having enlarged the drawing it does indeed look like a large diameter worm gear driving the shaft, but the size and shape of the casing is a very close match to that in the photograph. Certainly some of my HO models have a large diameter worm that also served as a flywheel. There is no indication of an offset that would be expected  with bevel gears. The casing over the axle would allow a helical gear as expected by a worm drive but not a bevel gear...

Basically, the patent drawing and the photo match pretty well in my opinion....

Peter

  • Member since
    September 2003
  • 13,412 posts
Posted by Overmod on Friday, September 25, 2020 9:52 PM

In the longitudinal section the gearbox is illustrated with the driven-gear casing directly above the axle centerline, much as it would be in a typical automobile speedometer drive.  In the picture, there is nothing in that location; the casing over the axle and the domed forward face of the 'nose' are both visibly forward of the center line, which is what led me to think there is hypoid or spiral-bevel gearing in there.  There is vanishingly little issue with backdriving in the design as the shaft is only driving a fluid distributor, and that fluid is not 'steam' but pressure oil at that point.

I confess the more I look at the principle of the Jones gear, the more I appreciate what Tom Daniels tried to do with British Caprotti, using the steam only in lieu of a debounced counterspring that causes so many issues in Franklin-system Lentz.  If you want fluidic amplification of valve events, do it in the follower system on top of what the mechanical modified-trap cam does.

The question with the British Caprotti valve arrangement then becomes how do you regulate the throttle to be sure the valves get enough pressure to seat -- as otherwise you get the same sort of BLOW as with 2926 before the Wagner drifting valve seats...

  • Member since
    September 2003
  • 13,412 posts
Posted by Overmod on Friday, September 25, 2020 9:55 PM

In the longitudinal section the gearbox is illustrated with the driven-gear casing directly above the axle centerline, much as it would be in a typical automobile speedometer drive.  In the picture, there is nothing in that location; the casing over the axle and the domed forward face of the 'nose' are both visibly forward of the center line, which is what led me to think there is hypoid or spiral-bevel gearing in there.  But I don't think there is much doubt that this is what the Jones patent would call for. in being an enclosed drive in a maintenance-difficult environment going to a single complex fluidic distributor that does all the functions for the four 'ends' through the range of forward to reverse.

There is vanishingly little issue with backdriving in the design as the shaft is only driving a fluid distributor, and that fluid is not 'steam' but pressure oil at that point.

I confess the more I look at the principle of the Jones gear, the more I appreciate what Tom Daniels tried to do with British Caprotti, using the steam only in lieu of a debounced counterspring that causes so many issues in Franklin-system Lentz.  If you want fluidic amplification of valve events, do it in the follower system on top of what the mechanical modified-trap cam does.

The question with the British Caprotti valve arrangement then becomes how do you regulate the throttle to be sure the valves get enough pressure to seat -- as otherwise you get the same sort of BLOW as with 2926 before the Wagner drifting valve seats...

  • Member since
    September 2003
  • 13,412 posts
Posted by Overmod on Friday, September 25, 2020 9:56 PM

In the longitudinal section the gearbox is illustrated with the driven-gear casing directly above the axle centerline, much as it would be in a typical automobile speedometer drive.  In the picture, there is nothing in that location; the casing over the axle and the domed forward face of the 'nose' are both visibly forward of the center line, which is what led me to think there is hypoid or spiral-bevel gearing in there.  But I don't think there is much doubt that this is what the Jones patent would call for. in being an enclosed drive in a maintenance-difficult environment going to a single complex fluidic distributor that does all the functions for the four 'ends' through the range of forward to reverse.

There is vanishingly little issue with backdriving in the design as the shaft is only driving a fluid distributor, and that fluid is not 'steam' but pressure oil at that point.

The thing that bothers me now is what that disc at the end of the driven shaft does.  It does not look like part of any torque coupling, which it almost has to be.

I confess the more I look at the principle of the Jones gear, the more I appreciate what Tom Daniels tried to do with British Caprotti, using the steam only in lieu of a debounced counterspring that causes so many issues in Franklin-system Lentz.  If you want fluidic amplification of valve events, do it in the follower system on top of what the mechanical modified-trap cam does.

The question with the British Caprotti valve arrangement then becomes how do you regulate the throttle to be sure the valves get enough pressure to seat -- as otherwise you get the same sort of BLOW as with 2926 before the Wagner drifting valve seats...

  • Member since
    January 2002
  • 4,087 posts
Posted by M636C on Saturday, September 26, 2020 1:34 AM

In the longitudinal section the gearbox is illustrated with the driven-gear casing directly above the axle centerline, much as it would be in a typical automobile speedometer drive.  In the picture, there is nothing in that location; 

That is because the photo has been retouched to remove the background...   we can't say whether there is a casing there or not. On the other hand, the casing around the axle does not look deep enough to cover a bevel gear....

Peter

  • Member since
    September 2003
  • 13,412 posts
Posted by Overmod on Saturday, September 26, 2020 5:21 PM

M636C
That is because the photo has been retouched to remove the background...   we can't say whether there is a casing there or not.

Not only do I think you are correct about the retouching, but on reflection I think I see the outline of a rear cover and at least two bolts in the retouching.  If you hadn't suggested the possibility I would not have looked.

  • Member since
    June 2002
  • 16,983 posts
Posted by daveklepper on Sunday, September 27, 2020 7:33 AM

If one were to build as modern and up-to-date rod steam locomotive, twocylinders, probably a 4-8-4 or 2-8-4, dual-service, what valve-gear would you choose?

  • Member since
    September 2003
  • 13,412 posts
Posted by Overmod on Sunday, September 27, 2020 10:24 AM

daveklepper
If one were to build as modern and up-to-date rod steam locomotive, twocylinders, probably a 4-8-4 or 2-8-4, dual-service, what valve-gear would you choose?

Since very high speed is no longer appropriate for this kind of reciprocating locomotive, the 'best' choice will be some flavor of Baker, constructed entirely with needle or similar bearings on the pin joints.  In addition to very long life with minimal maintenance concern, this has extremely good precision and little change of events over time.

I am partial to David Wardale's idea of separating admission from exhaust timing and duration by using two parallel piston valves (of his general construction, articulated in three pieces with suitable rings and defined steam edge, steam or overcritical-water cooling of the admission-valve liner, etc) and this would require some modification to the Baker gear somewhat along the lines of 'sociable' bicycle construction; you could also probably drive the exhaust valve via derived drive but packaging it becomes a bit of an issue.

Long-lap though perhaps not the long-travel of the late French (de Caso) practice; jacket the hell out of the cylinders as well as insulating them; capable 'enough' reversible compression control (which may prove to involve multiple reversible relief valves with insulated reservoirs) -- this also helping to avoid the 'unwinding' of Baker gear sometimes notable on accounts of N&W testing... if it can happen there it can happen anywhere, and needs to be guarded against...

  • Member since
    January 2002
  • 4,087 posts
Posted by M636C on Sunday, September 27, 2020 8:31 PM

This thread has caused me to look up a number of things that were just names. Reidinger valve gear was a prime example, paticularly since there was no visible evidence on the locomotive to differentiate between rotary cam Lentz gear (which all Reidinger locomotives had when built).

It is clear that as late as 1954, enough senior engineers in British Railways were concerned about the long term reliability of varible cutoff British Caprotti gear to be testing a fairly complex alternative on locomotives in regular service.

Baker valve gear was little used in Australia, maybe eleven locomotives. The New South Wales Railways had 120 locomotives fitted with Southern valve gear, which in theory required even less attention than Baker gear, and notably eliminated the combination lever. It was said that the Southern gear worked so well and required so little attention that the first indication of something being wrong was a major component falling off the locomotive. There were 470 similar locomotives with inside Allan straight link motion, a simplified version of Stephenson gear with, as the name suggests  a straight rather than curved reversing link, which presumably absorbed much more time in inspection and repair. Queensland Railways had one locomotive with Southern valve gear and there were tales of the gear swinging from side to side to the extent that the valve events were affected. Perhaps Southern gear ended up being too complicated for the theoretical benefits.

Peter

SUBSCRIBER & MEMBER LOGIN

Login, or register today to interact in our online community, comment on articles, receive our newsletter, manage your account online and more!

FREE NEWSLETTER SIGNUP

Get the Classic Trains twice-monthly newsletter