1. Was this conversion by Lima in 1939 the only one? 2. Were others modernized by PRR itself. 3. How does a front-end throttle improve performance?
One advantage of a front-end throttle is that it reduces the response time of the locomotive to throttle changes.
No other K4 got rebuilt the way 5399 did. One or two others got poppet valves.
What Type popet-valves were used?
daveklepper What Type poppet-valves were used?
What Type poppet-valves were used?
5399 was fitted with Franklin Type A valve gear.
This conversion is dealt with in detail in Vernon L Smith's "One Man's Locomotives".
The conversion to a front end throttle was part of a conversion to increase the steam flow area right through from the boiler to match the capability of the poppet valves. This involved changing the dry pipe, the throttle and the superheater. 5399 originally ran with the poppet valves and the original boiler and throttle, but was modified to increase its power. It could haul a train normally requiring two standard K4 locomotives.
Peter
daveklepperWhat Type poppet-valves were used?
On 5399, the Franklin System type A with oscillating-cam drive was used: this has two small admission valves and two larger exhaust valves per cylinder end (this is the layout on the 'heraldic shield' in contemporary Franklin advertising!) This is the system adopted for the T1 construction, and in the rotary-valve T1 conversion (which used bridges to activate the paired valves from a single follower).
Oscillating cam poppet valves are something of a scam, because when operated from traditional valve gear the valve lift can be proportional to cam excursion, which is NOT productive of clean limited cutoff. (This is not necessarily a 'bad' effect when running at lower speed, for the same reasons the 'type D' in the Vulcan conversions for Army 2-8-0s were). It is interesting that the poppet-valve conversion was done before the large sine-wave superheater was put in.
3847, which was converted in 1945, uses a rotary-cam system, with the older very heavy Franklin outside drive frame similar to that on the C&O L-2 Hudsons with RC. That engine's appearance can be studied here:
http://www.northeast.railfan.net/images/prr3847.jpg
In some ways this is a superior method, but only if fitted with continuous-contour cams (which are expensive to machine and difficult to maintain). Most of the RC systems with discrete cam profiles (e.g. British Caprotti and some Reidinger) tended to have poor performance off the regime for which the cams were machined -- which was fairly often. It might be interesting to see if modern hard coatings can solve some of the issues with spherical followers running on continuous-contour surfaces at the speeds and pressures involved with the heavier progressive spring pressures in practical 'debounced' Franklin-style poppet valves.
The Hagley Museum and Trust has several PRR negatives of the 5399 in their collection.
From the captions it would appear the testing was done in Fort Wayne.
Good Luck, Ed
Overmod daveklepper What Type poppet-valves were used? On 5399, the Franklin System type A with oscillating-cam drive was used: this has two small admission valves and two larger exhaust valves per cylinder end (this is the layout on the 'heraldic shield' in contemporary Franklin advertising!) This is the system adopted for the T1 construction, and in the rotary-valve T1 conversion (which used bridges to activate the paired valves from a single follower). Oscillating cam poppet valves are something of a scam, because when operated from traditional valve gear the valve lift can be proportional to cam excursion, which is NOT productive of clean limited cutoff. (This is not necessarily a 'bad' effect when running at lower speed, for the same reasons the 'type D' in the Vulcan conversions for Army 2-8-0s were). It is interesting that the poppet-valve conversion was done before the large sine-wave superheater was put in. 3847, which was converted in 1945, uses a rotary-cam system, with the older very heavy Franklin outside drive frame similar to that on the C&O L-2 Hudsons with RC. That engine's appearance can be studied here: http://www.northeast.railfan.net/images/prr3847.jpg In some ways this is a superior method, but only if fitted with continuous-contour cams (which are expensive to machine and difficult to maintain). Most of the RC systems with discrete cam profiles (e.g. British Caprotti and some Reidinger) tended to have poor performance off the regime for which the cams were machined -- which was fairly often. It might be interesting to see if modern hard coatings can solve some of the issues with spherical followers running on continuous-contour surfaces at the speeds and pressures involved with the heavier progressive spring pressures in practical 'debounced' Franklin-style poppet valves.
5399 had the cams driven from the crosshead, the Walschearts valve gear having been removed.
Even before the boiler improvements, 5399 could haul a 1000 ton train at 94.7 mph under AAR test conditions. I seem to recall that this was about the same as the original UP 800 when tested.
It was allocated to heavy trains which would normally have required two K4s. On 5 Nov 1939, it worked The General weighing 940 tons, saving 21 minutes on the schedule at an average overall speed of 65 mph.
After the boiler improvements, it was tested against a standard K4 5341.
At 40 mph, it was 16.2% more powerful in indicated horsepower.
At 60 mph,17.1%
At 80 mph, 22.9%
At 100mph, 46.8% (4070 ihp against 2770 ihp)
Of course, if 5341 had had the boiler improvements and had the piston valves been enlarged, the difference might not have been as great.
In France, although many locomotives were rebuilt up to the standard of the Paris Orleans 3700, it was found that dramatic increases in power and economy could be obtained simply by applying improved Low Pressure piston valves, along with improving the steam flow from the boiler to the HP cylinders and between HP and LP cylinders. This was much cheaper than the full Chapelon treatment.
M636C5399 had the cams driven from the crosshead, the Walschaerts valve gear having been removed.
RC gear, of course, has neither a substantial shrouding of valve openings or a periodic reversal of momentum of the cam to accommodate.
Why were not more K4s improved? It would seem like reduction in crew costs would have made the investment worthwhile. Or were savings eaten up in additional maintenance costs?
How did these improved K4s compare with the two K5s?
What you forget is that the crosshead drive did not move the cams directly: there are two additional pieces of equipment in there.
I didn't forget it. I checked the drive from one of the photos in the Hagley group linked earlier.
This is in principle the same drive as the Young gear used on the UP 7000 4-8-2, which was replaced by Walschearts gear, most memorably on the streamlined 7002. Whatever the theoretical disadvantages, 5399 actually worked rather well. It had the advantage that the gear was readily accessible in that application. Smith commented that the subsequent application on the T1s resulted in poor accessibility, not only in the rear cylinders, but due to the streamlined boiler casing extending over the forward gear.
As Dave no doubt recalls, one of the K5s had Caprotti gear which did not show the advantages that 5399 demonstrated.
This goes to show the evils of assumption, where references talk about the crosshead drive being similar, and I look at photos just enough to conclude the crosshead was driving mechanical lubrication. Upon review I believe this is the system shown in Woodard's patent 2234614 of March 11, 1941.
This raises something else interesting: the substantial cylindrical device on the lower rear cylinder head visible in the picture. I don't know if this is a typical compression-relief device but it is considerably longer than any I (admittedly, again, not very carefully) have seen in late practice. Good secondary compression relief (and I agree with Jay Carter that having it reversible, like an extension of 'cushioning' compression volume in the endspace up to the valves, would be a great advantage.
Interestingly, the nightmare-box part of the T1 gear (the thing that was on end and crammed between cylinders for the rear engine) was apparently not the thing that needed serious maintenance attention -- it could be sealed and given oil changes and according to a couple of notes in the T1 Trust engineering committee discussions very few of them seem to have required working on. It was the other piece, the little radial gear in a box, that seems to have been involved in many problems. I don't remember any definitive answer on how much of this was related to increasing the spring pressure to 'debounce' the poppet valves. (Better control of the valves was one of the projects for 1948 that were not 'proceeded with'.)
The Caprotti gear was on there largely because Baldwin made an agreement in the late '20s, thinking Caprotti poppet-valve gear was going to be the wave of the future, and succeeded in flogging it to a large number of railroads. They even featured logging and narrow-gauge export locomotives with it. To my knowledge it was successful nowhere Baldwin tried it, and eventually had to be removed; to my knowledge none of them were subsequently converted to Franklin System so the issues were more than in the cambox and general frailty of the actuation system.
The K5s were unfortunately timed: they came just after the massive orders for K4s in the late '20s, and had to sit out the capital dedication to mainline electrification, by which point the frenzy over duplex augment reduction on large passenger power (and perhaps the bad taste of the ACL R1 fiasco) was in full cry. By the time the results of Timken lightweight rods on eight-coupled heavy engines were becoming evident PRR had already committed to something larger and smaller at the same time!
Since the K5s were orphans they weren't used as testbeds for much of anything, being stuck on the Northern Central where high speed was unimportant and relatively low FA already a concern. It literally took an act of Congress to get stokers put on them, which perhaps shows up mistaken design priority better than any other thing. It would be highly interesting to see what a K5 modified as 5399 was would have been able to do ... just that it would neither be used to advantage or serve as an example for service conversion or new practice any more than a K4 conversion would.
Now, I still don't quite understand why PRR didn't do any poppet/lightweight rod conversions on the M1 chassis: that would promise to be groundbreaking.
I agree that a modernized M-1 fleet would have been a better investment than just about anything in steam PRR could have done, T-1s. Q-2s, or even modernized K4s.
M636CAfter the boiler improvements, it was tested against a standard K4 5341. At 40 mph, it was 16.2% more powerful in indicated horsepower. At 60 mph,17.1% At 80 mph, 22.9%
Franklin rotary poppet [valve] gear drive, possibly located on #1 axle.
Poppet_drive by Edmund, on Flickr
PRR 5399. PRR Photo E12275, Hagley Trust. According to their file.
This view better illustrates the levers that communicated the cam box with the combination lever as discussed above:
Franklin by Edmund, on Flickr
Regards, Ed
gmpullman Franklin rotary poppet [valve] gear drive, probably located on #1 axle. Poppet_drive by Edmund, on Flickr PRR 5399. PRR Photo E12275, Hagley Trust. This view better illustrates the levers that communicated the cam box with the combination lever as discussed above: Franklin by Edmund, on Flickr Regards, Ed
Franklin rotary poppet [valve] gear drive, probably located on #1 axle.
PRR 5399. PRR Photo E12275, Hagley Trust.
I think the rotary drive on the locomotive axle would be for a lubricating oil pump. Vernon Smith, who was involved in the design of the gear on 5399, says "the cams to operate (the poppet valves) were placed in a double deck cam box that required an oil pump to lubricate the upper cam and rollers".
He also said "since the drive for the gears came from the locomotive crossheads only, it was neccesary to cross the drive over inside the box to get the motion in the correct phase to drive the respective links."
M636CHe also said "since the drive for the gears came from the locomotive crossheads only, it was neccesary to cross the drive over inside the box to get the motion in the correct phase to drive the respective links."
See also the Berry accelerator gear (on Dr. Richard Leonard's site) for a different application of the principle.
M636CI think the rotary drive on the locomotive axle would be for a lubricating oil pump.
I must be getting my Franklin Type A & B confused. I know that on some installations the "rotary" drive is quite obvious, linked to the crank pin the way it is. Seeing the photo of the right-angle axle gear reducer mounted on the axle that the PRR 5399 simply had the rotational drive shaft between the drivers, essentially hidden from view.
This patent drawing shows the Type B if I'm not mistaken.
Franklin Rotary by Edmund, on Flickr
Solid documentation is certainly hard to come by! Looking at the arrangement for the Type A, I can not see where the right-angle drive as shown in the Hagley photo series would have been applied.
Franklin-Type-A by Edmund, on Flickr
There certainly must have been a great deal of head-scratching and perhaps a grumble or two from the Lima engineers tasked with mounting that cam box ahead of the steam chest. From what I've read, Franklin (Lima) covered the cost of the installation in hopes of selling the PRR on the idea.
Overmod M636C He also said "since the drive for the gears came from the locomotive crossheads only, it was neccesary to cross the drive over inside the box to get the motion in the correct phase to drive the respective links." I believe that is a necessity for any crosshead drive, going back to the earliest examples which are I believe in British practice. See also the Berry accelerator gear (on Dr. Richard Leonard's site) for a different application of the principle.
M636C He also said "since the drive for the gears came from the locomotive crossheads only, it was neccesary to cross the drive over inside the box to get the motion in the correct phase to drive the respective links."
I believe that is a necessity for any crosshead drive, going back to the earliest examples which are I believe in British practice.
The two designs that come to mind are G.J Churchward's first "Star", no 40, later 4000 and Deeley's 999 class 4-4-0. It was not widely used apart from these eleven locomotives in total. Both these locomotive types had inside cylinders and valve gear, although No 4000 also had outside cylinders with the valves operated by rocking arms.
Inside rotary drive to poppet valve gear was quite common with British applications of Caprotti gear, including export locomotives. A number of LMS "Black 5" locomotives were built in 1947-48 with rotary drive from the second (driving) coupled axle.
M636CA number of LMS "Black 5" locomotives were built in 1947-48 with rotary drive from the second (driving) coupled axle.
Overmod M636C A number of LMS "Black 5" locomotives were built in 1947-48 with rotary drive from the second (driving) coupled axle. Can you, once and for all, resolve the difference between British Caprotti and 'Reidinger'? From what I know, Reidinger had the driveshaft only on one side, driving both cams (and I don't think they had the flexible coupling in the center that the shaft in an FM 38D OP engine does) whereas British Caprotti, like Franklin type B, has a separate shaft for each side. Differences in the actual cam follower and valve operation I couldn't make out from the small drawings I had.
M636C A number of LMS "Black 5" locomotives were built in 1947-48 with rotary drive from the second (driving) coupled axle.
Can you, once and for all, resolve the difference between British Caprotti and 'Reidinger'? From what I know, Reidinger had the driveshaft only on one side, driving both cams (and I don't think they had the flexible coupling in the center that the shaft in an FM 38D OP engine does) whereas British Caprotti, like Franklin type B, has a separate shaft for each side. Differences in the actual cam follower and valve operation I couldn't make out from the small drawings I had.
The whole thing is more complicated than you suggest...
Reidinger gear was applied to a total of seven locomotives (but one set was successively applied to two locomotives, so there were only six sets of valve gear).
Your description applied to the one set that was moved, first applied to LNER D49 4-4-0 365 and later moved to sister 361 (BR 62764).
All locomotives with Reidinger gear were built with rotary cam Lentz gear with stepped cams with limited fixed cutoff sttings.The Reidinger patent covered continuously variable cams to give variable cutoff settings.
I don't know to what extent the original Lentz valves and layout were maintained, but the external appearance of the Reidinger modified locomotives was not significantly changed.
The LNER locomotives, both RC Lentz and Reidinger were driven by a shaft from a bevel gearbox on the right hand side only. The Reidinger unit had the bevel gearbox at the cylinders mounted outside the cylinder casing while the Lentz engines had it (just?) inside the casing. These were three cylinder locomotives with 120 degree cranks.
The other five locomotives were LMS Fowler Class 5 2-6-0s which were two cylinder locomotives with bevel gear drive on each side.
LNER 365 was built with RC Lentz gear in 1934 and was converted to Reidinger (with steam actuated valves) in 1939. 365 gave trouble in service and after the bevel gear drive was damaged in 1941 it was taken out of service. Edward Thompson rebuilt it with two inside cylinders with piston valves.
The valve gear was retained and fitted to 62764 in 1949, but in this case the original valves with spring actuation were retained. the locomotive was extensively tested on the Rugby static test plant, and it returned to service and retained the Reidinger gear until withdrawal in the 1960s.
The five Fowler 2-6-0s built in 1931, were rebuilt with Reidinger gear in 1953, possibly as a result of the tests on 62764, and they also kept the Reidinger gear until withdrawal.
Since it was based on RC Lentz conversions, I assume that Reidinger valve gear had the poppet valves arranged horizontally above the cylinders while British Caprotti had the valves arranged vertically above the cylinders.
The BR standard Caprotti locomotives had steam actuated valves, and I assume the last two LMS Class 5s with outside rotary drive also had steam actuated valves. I can't say about the earlier LMS Class 5s with the inside rotary drive that I mentioned earlier.
OvermodNow, I still don't quite understand why PRR didn't do any poppet/lightweight rod conversions on the M1 chassis: that would promise to be groundbreaking.
In July 1940, Lima approached the PRR with a proposal to rebuild an M1 into a 4-8-4 equipped with Franklin Oscillating Cam Poppet Valves. Will Woodard from Lima felt that the M1 had much room for improvement. PRR wasn't interested so Lima got no further with this proposal. A comment was made in Hirsimaki's book that M1s were freight engines (nope, they were dual-purpose locomotives) and this wasn't the type of service poppet valves were suited for (sounds like an excuse).
Steam-actuated popple valves designed by Engineer of Tests L.B Jones were applied to #5346 in June 1940. They proved unsuccessful.
I would love to see a fully upgraded and streamlined PRR M1s fleet hauling heavy named trains like The General and Trail Blazer on the west end of PRR's system.
Jones 3D Modeling Club https://www.youtube.com/Jones3DModelingClub
Part of the problem could have been the Pennsy's symbiotic relationship with Baldwin. Alco tried, but gave up on doing business with the PRR after several attempts. According to the George Drury book on American steam Alco had done some steam developmental work for the PRR only to have the contract given to Baldwin. It wouldn't surprise me if something similar happened to Lima.
Flintlock76According to the George Drury book on American steam Alco had done some steam developmental work for the PRR only to have the contract given to Baldwin.
To me this was one of the more appalling injustices in locomotive history!
Jones1945In July 1940, Lima approached the PRR with a proposal to rebuild an M1 into a 4-8-4 equipped with Franklin Oscillating Cam Poppet Valves. Will Woodard from Lima felt that the M1 had much room for improvement.
In any case, I think the more capable superheater, combined with a type A system sized to deliver the mass flow of the larger valves in the M1a, would have produced dramatically better high-speed horsepower; a little better balancing (the stumbling block here being that the expense of Timken roller rods was still perceived as a 'passenger-only' requirement -- and while this is before the 610 testing, I saw no indication in the postmortem comments by Cover on that testing that would indicate he saw value in using that approach on an engine with a 72" wheel)
Note that PRR while they tested lightweight roller rods on a K4 did not put a set of 5399 at any time -- which would have been the most logical of any PRR engine at the time to receive them. Not on the Q1 either, despite 77" drivers, and perhaps for the best not on the Q2, which would probably have had exponentially more buckled lightweight components...
Would Lima have gotten the contract to modify additional M1s if the test unit had proven as phenomenal an improvement as 5399? My guess is, as above, that Baldwin would "underbid" (perhaps nudge, nudge, wink, wink, say no more) and do the fleet mods ... just in time for the wartime traffic, perhaps?
OvermodThere's a little more to that story than you indicate.
You're absolutely right Mod-man, I was just trying to keep it simple!
OvermodHad Woodard made the far more sensible suggestion of improving the M1's valves and superheater along the lines of what proved successful on 5399
I thought Woodard's proposal in July 1940 included the improvement of the M1s' superheater, larger steam passages connecting the cylinder, and the front-end throttle as well because #5399 got these treatments in early 1940. But I don't know the detail of his ideas on modernizing the M1s. A modernized M1s would have been a game-changing project, I believe duplex believers in Baldwin and PRR foresaw that. The success of #5399 wasn't good news but an eyesore to Baldwin's plan (which was using PRR as a stage to demonstrate and to prove the rigid-frame duplex concept would work well), let alone the modification of the M1s. It is a reasonable presumption that Baldwin played a role in stopping Lima's proposal from happening.
OvermodNote that PRR while they tested lightweight roller rods on a K4 did not put a set of 5399 at any time -- which would have been the most logical of any PRR engine at the time to receive them.
If #5399 equipped with lightweight roller rods, and even newer drivers, that would have made this engine a perfect upgrade example. But at the time, PRR and Baldwin had already been working hard on rigid-frame duplex like the S1, later the T1 and Q1, something so much bigger, powerful, and faster (including acceleration rate) than an upgraded K4s like the #5399. They probably didn't want to spend extra money to make #5399 even better.
The problem basically is that no matter how much you make a K4 better, it's still only a six-coupled locomotive, and there are hard limits on how large a train you can pull. Remember that there is very little that an articulated locomotive with twelve 80" drivers and 27"x28" cylinders can't pull, even at 205 pathetic psi, and that is the zero-added-cost locomotive alternative available to PRR (thanks in part to its massive over-ordering of them very late in the game) any day. ALL the duplexes presumed less overkill (in fact the T1 bent over backward to cost much less than a 4-8-4, which may make us snigger a bit in hindsight but is certainly visible in design analysis) and if you REALLY want overkill, imagine a couple of 5399-modded K4s needing to be doubleheaded. (Perhaps, if all went well, on a Cincinnati service...?
As soon as you say "4-8-4" with reference to a Lima M1 proposal, you're already talking much more rebuild than any straight eight-coupled is worth to the PRR. And since both Snyder preheaters and the benefits of better waterleg circulation were very well known to Woodard by 1940, it's almost inexcusable to think that all that money for an 8-coupled rebuild, when a 10-coupled duplex had far more available horsepower for far less augment, could possibly be justified.
Keep in mind this was also the era in which the direct-drive steam turbine a la 6200 was being evolved, and a 4-8-4 version of this was carried all the way into Westinghouse promotional literature in 1948. That (with at least the two-speed reversible planetary Westinghouse patented for the arrangement after the war) would have been the thing to build if you wanted eight-coupled that made good use of a two-axle trailing truck. (Not to mention what would happen if you put a Bowes drive on it, instead of playing with Too Much Horsepower For Your Water with 9000hp single-unit V1 variants...)
While I certainly wouldn't put it past Baldwin to try a bit of skulduggery, remember that PRR was NOT happy with them by 1940, having detected the hand in the cookie jar (in retrospect, I think part of the eventual failure of the BCR coal-burning turbine scam was related to institutional memory of the severe burning PRR took on the S1). Since on first principles alone the Lima proposal wasn't likely to fly, I don't think we need to follow the conspiracy-theory very far even if it were objectively true...
Overmod PRR was NOT happy with them by 1940
Probably more like a spat within the family than a knock-down drag-out barroom brawl. Kind of like when your younger brother borrows your best sport coat for a hot date, or worse, like borrowing your car without asking permission for the same date.
Yeah, you're plenty PO'd, but you get over it.
I don't think I answered Overmod's question about Reidinger valve gear in my last attempt, so I kept looking. Here is the patent, issued in 1940:
https://www.freepatentsonline.com/2210328.html
I found it quite confusing at first, until I realised that it was a camshaft that allowed the cams to rotate independently of eachother. thus allowing conventional cams and followers to be used in conjunction with "infinitely" variable cutoff.
It does appear to be an extremely complex solution to the problem, but it must have worked, at least by 1949 when the D49 4-4-0 was tested on the testr plant at Rugby. One of the five 2-6-0s was tested there in 1954, so someone must have thought that it was a practical solution.
The design looks like it was intended as a a "drop in" replacement for the fixed step cams in conventional rotary cam Lentz valve gear. The mechanism for rotating the cams used a transverse rack arranged exactly like the rack for shifting cams on Lentz RC valve gear. In fact Reidinger gear was sometimes called "type RR" suggesting it was marketed by Davey Paxman, the licensees of Lentz gear.
https://www.paxmanhistory.org.uk/lentzloco.htm
Scrolling down gives a good view (if inverted) of the Lentz RC gear as applied to the LNER D49.
While on the subject, here is a photo of the second Reidinger D49
The LMS 2-6-0s had a rotary drive on each side...
I haven't found anything on the steam actuated valves on LNER 365, but it appears that the other locmotives had spring actuated valves.
Login, or register today to interact in our online community, comment on articles, receive our newsletter, manage your account online and more!
Get the Classic Trains twice-monthly newsletter