Pennsylvania RR T-1 modification?

Discussion of steam locomotives from all manufacturers and railroads

Moderators: Typewriters, slide rules

Allen Hazen
Posts: 2456
Joined: Fri Mar 12, 2004 10:14 pm
Location: Edmonton, Canada (formerly Melbourne, Australia)

Pennsylvania RR T-1 modification?

Post by Allen Hazen » Fri Aug 17, 2007 4:13 am

Somewhere, a long time ago (in a galaxy far away) I read of a proposal to modify a T-1 by installing crank axles so the forward and rear sets of driving wheels could be coupled inside the frames. It would seem that this -- if feasible -- would have done much to alleviate their wheelslip problems. A similar design idea was incorporated in the 1970s proposal for the "ACE" modern steam locomotive.

Does anyone know more about this proposal?
How far the design work went?

(And what would it have been called in Whyte notation? Somehow 4-4-4-4 no longer seems appropriate, but wouldn't have looked exactly like a conventional 4-8-4 either! Maybe a "4-8*-4"? (Grin!))

Posts: 65
Joined: Wed Jun 07, 2006 2:45 pm

Post by Juniatha » Mon Aug 20, 2007 6:10 pm

Hi, Allen

Yes, I have read about that proposal. This would have made a locomotive type of complex character even more demanding and more complex.

Now, while the Duplex concept was created in order to use more than two cylinders on wheel sets in one rigid frame, yet to avoid crank axles, it would have been a contradiction by itself to install two crank axles with two cranks each as a means to amend Duplex type difficulties!
If from the beginning the wheel arrangement would have been left 4-8-4 instead of 4-4-4-4 and a four cylinder layout had been used (unusual in America of course), then there would have been need for but one crank axle with two cranks.

The forces acting on such 'inside coupling crank axles' would - under normal conditions of adhesion - be much smaller than the piston forces and main rod mass acceleration forces imposed on the crank axle of a four cylinder type of engine (with two outside plus two inside cylinders). However it could not be relied upon such 'regular' conditions would always prevail. On the contrary, the provision of those 'inside coupling crank axles' was to deal with non-regular conditions, i.e. one set loosing adhesion while the other set doesn't.
Such conditions could impose extremely high forces on these crank axles. Just imagine the engine loosing adhesion on a 'greaser' point enough to produce a slip. Then the first drive set reaches clean, dry rails again and wants to stop slipping - but the second wants to continue. It now rests entirely on those inside cranks to stop the second drive set with all its mass inertia of wheel sets and rods! This can produce extremely high forces working on the seats of the wheels on those crank axles, wanting to turn the wheels on their axles. The axles and cranks themselves also are under high stresses from these forces.

The sort of coupling attained by such a 'detour coupling' is not a very rigid one.
Just think of the way a piston force produced by one of the rear set of cylinders has to go to reach wheels of the first drive set:
The second set of cylinders drive to the fourth driven axle. From there to third the piston force is transmitted directly via coupling rod on the same main pin. From third to second driven axle it is via pin of third axle, wheel, axle, inside crank, then via coupling rod to inside crank of second driven axle, its wheel and pin to outside coupling rod to pin of first driven axle.
There is a lot of flexion in that!
Result: this type of coupling cannot prevent micro slip which is the start of actual slipping. So, sure footedness is hardly improved by this!

The virtue of inside coupling however might be that the two sets of drives are mechanically fixed to keep a defined phase one to each other.
a) That could be used to improve overall reciprocating mass balancing: By setting the relative phase of the two drive sets so that reciprocating forces are counterbalanced between the two sets, counter weights in drive wheels could be reduced and so there would be less hammer blow effect.
However, unfortunately the very position or phase of pins of the sets relative to each other which results in best overall reciprocating mass balancing is also the one that is most unfavourable for adhesion!
So, there is little that can be gained by fixing the drive sets to such a relative phase.
b) Fixing the drive sets to a phase which is best for adhesion asks for individual balancing of reciprocating masses in each of the sets since no counter balancing between the two sets is then available. That means, balancing counter weights cannot be reduced in the two drive axles as in case (a).

To my understanding, there is thus no way around optimization of both balanced running and adhesion of each drive set individually.
That was never done on the Pennsylvania Duplexii - with the Q-2 strangely doing admirably well in very heavy freight traffic in spite of that and without the reason for that ever having been investigated : it is obvious that the Q-2 must have been much better than the T-1 as concerns adhesion , otherwise the Q-2 could never have run train weights of 10000 tons or more which is an extremely high load relative to the adhesion weight of the locomotive.

Myself I would like to get more data and information on the behaviour of the PRR Duplexii. I would like to kindly encourage anyone who has any information to let me know.


Posts: 1717
Joined: Fri Mar 12, 2004 1:04 pm
Location: Oakland CA

Post by timz » Mon Aug 20, 2007 7:03 pm

Juniatha wrote:However, unfortunately the very position or phase of pins of the sets relative to each other which results in best overall reciprocating mass balancing [i.e. 180 degrees out of phase?] is also the one that is most unfavourable for adhesion!

Posts: 278
Joined: Fri Mar 12, 2004 8:29 pm

Post by Typewriters » Tue Aug 21, 2007 9:27 am

I think it's important to remember here that the whole concept of duplex drive locomotives began with the desire to reduce dynamic augment, and that there was also a desire to reduce piston thrust in order to allow a reduction of forces and thus hopefully reduce wear. In that light, the re-coupling of the driving axles of a duplex drive locomotive would have defeated the purpose of the initial concept.

A number of roads had, prior to the designing of the 4-4-4-4 T-1, studied track pounding and wear through operation of high-speed steam locomotives and a whole new controversy about the formulae used to calculate balancing forces and weights had developed. One famous example at this time was that of the Baldwin-built I-5 class 4-6-4 locomotives constructed for the New York, New Haven & Hartford which developed serious slippage and track pounding difficulties even though they were designed with the latest formulae and concepts of balancing.

Quite aside from that, there very likely would have been no room to couple the axles of the T-1 internally due to the mass of the (one piece, cast) frame; certainly, no one would have wished to weaken it in the required center area since the rear cylinders were mounted here. The freight Q-1 (4-6-4-4) proved that rear mounting of the cylinders for the rear driver set was a bad idea, so this would not be an option for a redesigned T-1 either.

The T-1 locomotives were considered slippery by crews; weight tranfer between driver sets was considered a major problem. Some locomotives were modified with reduced cylinder diameter, in order to improve factor of adhesion but the modification was not spread to more than a couple locomotives. The Q-2 (4-4-6-4) freight locomotives were equipped with an electro-pneumatically operated wheel slip arresting system which reportedly (even though it was new and high-maintenance) actually did work some of the time.

I am not certain what is meant by saying that the reason for the acceptable performance on the road of the Q-2 has never been investigated. These locomotives were fairly conventional in overall design, even if large and even if fitted with two sets of drivers; they were essentially as big and powerful as the biggest articulated 4-6-6-4's of the day with one less driving axle and no articulation. Unfortunately for their fans they were built late, suffered boiler problems, were more expensive to operate than the 2-10-4 J-1's and were relegated to the weeds with storage at Crestline beginning in 1952 for some Q-2's. (All were off the roster before the end of 1955.) In the final analysis the Q-2 was more powerful and slightly more efficient than the C&O-derived J-1 but its high operating costs caused it to be retired while the J-1 (and diesels) rolled on.

-Will Davis

Posts: 1717
Joined: Fri Mar 12, 2004 1:04 pm
Location: Oakland CA

Post by timz » Tue Aug 21, 2007 6:46 pm

Typewriters wrote:the re-coupling of the driving axles of a duplex drive locomotive would have defeated the purpose of the initial concept.
Why? The reduced piston thrust is still just as reduced, and the two engines could be coupled 180 degrees out of phase, which ... sounds too good to be true, but wouldn't the reciprocating masses balance themselves (over the whole locomotive, anyway, not the individual engines)? So, no need for overbalance, so no dynamic augment?


Post by rdganthracite » Wed Aug 22, 2007 7:59 am

Why? The reduced piston thrust is still just as reduced, and the two engines could be coupled 180 degrees out of phase, which ... sounds too good to be true, but wouldn't the reciprocating masses balance themselves (over the whole locomotive, anyway, not the individual engines)? So, no need for overbalance, so no dynamic augment?[/quote]

Since the pistons are double acting the opposite sides of the same piston on any steam locomotive are 180 degrees out of phase with each other. On a 2 cylinder locomotive the drivers are "Quartered", meaning that opposite sides are one quarter revolution out of synch with each other. That is 90 degrees. So with just 2 cylinders a "Standard" steam locomotive gets four power strokes with each revolution of the drivers. Each power stroke is 90 degrees from the one proceeding it.

In order to have the smoothest power on a Duplex drive you would want the front and rear set of drivers to be 45 degrees out of synch with each other. But that raises another possible problem. On the side rods on a two cylinder locomotive the thrust on the rod bearings and crank pins changes direction twice every revolution. On a coupled duplex the thrust direction on the inside cranks and rods will change direction EIGHT times every revolution as the power is applied first by a front cylinder, then a rear cylinder, then a front cylinder ......... I hope those cranks are strong enough to take that punishment!!!

Posts: 1717
Joined: Fri Mar 12, 2004 1:04 pm
Location: Oakland CA

Post by timz » Wed Aug 22, 2007 5:08 pm

rdganthracite wrote:In order to have the smoothest power on a Duplex drive you would want the front and rear set of drivers to be 45 degrees out of synch with each other.
And if you wanted the smoothest possible power on a four-cylinder 4-6-0 or 4-6-2 you'd want the same-- right? You remember SR (the English one) tried it with the Lord Nelsons circa 1925; did anyone else try it? Or were all the rest 180 degrees out of phase on each side of the engine?

Posts: 25
Joined: Tue Mar 16, 2004 6:49 am
Location: Northern Virginia

Post by feltonhill » Sat Sep 01, 2007 8:28 am

So far the only source I (and several others) have found regarding the T1's cylinder reduction is in Vernon Smith’s book, One Man’s Locomotives, at pg 161. There he lists the following nine numbers:

6110, 6111
5521, 5522, 5524, 5531, 5532, 5536, 5540

Assuming PRR was considering connecting the two engine sets somehow, its money would have been better spent training the crews how to operate the T1 rather than making them more mechanically complex. Based on many interviews with the men who actually operated them (See articles in PRRT&HS magazine, The Keystone), they were given no direction regarding proper operation when the T1's arrived. To a man, the crews also stated that if you operated them correctly, a T1 was not hard to handle. It would not slip "uncontrollably" as some authors have stated, either starting or at speed. But it certainly was different than the K4.

Some people learn how to handle new equipment, some don't. The former group gets the job done; the latter group does all the noisy BP&M stuff . Guess who attracts the attention.

Posts: 65
Joined: Wed Jun 07, 2006 2:45 pm

T-1 and Q-2 Duplexii

Post by Juniatha » Mon Sep 03, 2007 3:42 pm

Hi Typewriter

On your remarks of Aug 21:

I agree with your remarks concerning the proposal of inner coupling between the drive sets. The reversed position of cylinders on the rear drive set was of course a seductive proposition as concerns coupled wheel base but it proved there was more to it that just rearing the conventional layout. One thing among others is that with cylinders behind drivers, in forward gear the piston thrust transmitted by the drive rod will cause easing of axle load, no augmenting, thus easing slippage.

More and more the question proper is, actually how slippery had the T-1 locomotives to be considered in relation to a 4-8-4 of equal adhesion weight and driver diameter?
Even with former drivers as a witness the picture becomes blurred. Memory tends to gilden things as well as to darken things some times.
Points that seem sure:
a) – any engine with but four out of a total of sixteen axles driven cannot be an awesome puller; as far as 4-8-4s were used on slow heavy freight frankly it was a misuse to me that meant working these engines at the borderline of their output range on low thermal and output efficiency;
b) – without any special training to handle the locomotives at best efficiency there was little that could be expected in the hands of the average driver of a loco class much different from a standard K-4s; unquestionably, the more talented and more interested crews could get better results of an engine not plain simple to handle than did other crews;
c) – the duplex disposition of drives in a single frame is not simply a 4 + 4 = 8 coupled disposition – there is more to this special arrangement of drive sets than first thought had anticipated; this caused the engines to be sensitive in certain ways:
since there was no coupling between the drive sets, they could turn out at any relative phase; since certain relative phases would work less well than others as regards adhesion, actual adhesion limit on a given condition of rail surface tended to be less predictable, lest a driver would know which relative phases do work best and check out actual positions on his steed each time before starting and handle throttle accordingly – which is unrealistic;
d) – the characteristics of torque development and mechanics of a Duplex can be so adjusted and improved by sensible design as to overcome those adverse flaws in starting behaviour, making the 4 plus 4 Duplex disposition really equivalent to an 8-coupled design – and even better!

Your question what I meant by ‘reason for the Q-2’s performance never investigated':

Let’s first look at the relative train loads and adhesion mass of Q-2 versus T-1:
train loads: roughly 10 to 1;
adhesion mass: principally 5 to 4 (relation of driven axles)
So, the load factor was much higher on the Q-2 engines, right?
Now, the mentioned electro-pneumatic butterfly valve could only check a slip, not prevent it.
If the Q-2 had been as slippery as the T-1, then these valves would have had a lot of checking to do yet the engine could not have handled these loads because the necessary t.e. would not have been transmissible to the rails.
But as it turned out it actually was.
Why was that so in the Q-2 and not so in the T-1?
Have an answer from the PRR test department? I would highly appreciate to get to know about anything they did test and possibly did find out – yet to my knowledge it was not investigated. It was just accepted that things turned out more lucky with the Q-2 than they did with the T-1. Had the reasons for the different starting behaviour been investigated, findings might have been applicable to the T‑1.

Your remark: The Q-2 was a fairly conventional design, as big as biggest 4-6-6-4 articulateds except for that it wasn’t an articulated and had one drive axle less.

And exactly that is the point: it was not an articulated and it had one drive axle less.
Now, how could it bee, that in the T-1, the Duplex proved incapable to live up to pulling power of 4-8-4, i.e. single unit drives with an equal number of driven axles, while in the Q-2 the Duplex did prove equal to articulateds even having one more driven axle?
See the contradiction?

It has nothing to do with the rest of the design being more or less conventional. It’s the different behaviour of the drive set disposition that is intriguing.
If the Q-2 suffered boiler problems that has no consequential relation to the Duplex disposition of the drive.

Hi timz:

On your question of Aug 20:
Why is a 180° setting of #1 versus #2 unsuitable for best adhesion?

Because it has the drive sets work in opposed relative phase. I am not talking of adhesion factor as preset by cylinder size, pressure and adhesion mass, I'm talking of the actual adhesion limit that the engine can reach without slippage on a given surface condition of rails.

On your remarks of Aug 21:
Why does inner coupling defeat the initial concept?

The reduced piston thrust would still be present, right. But the concept of the Duplex was to use four cylinders on a set of drivers in a rigid frame without having to use crank axles.
Now, in a proper four cylinder engine type you would have to use one crank axle with dual throws and arrange your frame layout accordingly to accommodate the inner drives. In a coupled Duplex you would have to use two of those axles and would also have to rearrange your frame layout to accommodate those double crank axles – yet without the benefit of the proper four cylinder type of having balanced piston forces each side in a conventional 180° crank phasing outside / inside on each side. So, what’s gained then by the Duplex concept?
Drive sets of a coupled Duplex set at 180° and relying on self-balancing first against second set of reciprocating masses would put a lot of pounding for and aft on the two drive axles and would strain the whole structure considerably with alternating mass forces. Also, the pure for and aft movement is not all there is in balancing of reciprocating masses: how about cross balancing, how about the vertical components caused by the whipping movement of the main rods and imposed on the cross head?

On your remarks of Aug 22:
Regarding Lord Nelson 135° crank setting

This type of engine did not prove successful in achieving any higher adhesion. It did produce some odd forces on coupling rods while they passed through dead centre with the inner pistons still out of dead centre and producing torque which then could not be transmitted to the neighboring axle by that side coupling rod: this effect caused torsion of axles and when coming out of dead centre that coupling rod had to bring that torsion back in line – a very high thrust on bearings resulted which wore the rod bearings out quickly – yes, it must be said ´the disposition depended on a certain slack in order to prevent jamming!
This is why it was never repeated although the Nelsons struggled on – just as in Britain some most weird designs did struggle on as built with but very few exceptions.

Hi rdganthracite

On your remarks of Aug 22:
Quote “On the side rods of a two cylinder locomotive the trust on the rod bearings and crank pins changes direction twice every revolution.”

Sorry – this is only true at low rpm and only true to the road bearing, not the crank pin.
At starting, when influence of mass forces can be neglected, it is only around one side of the crank pin that piston force acts on (at intervals and to varying degrees over wheel turn, depending on the piston load and thus on cut off): it is always the side that drives the wheel in the same direction, i.e. moves the engine in one direction. If the pin would be loaded alternatingly, the wheel would move fore and aft and the engine could go nowhere - *g*.
At speed however, the influence of mass forces, working of the engine at short cut off and the influence of lead intake and exhaust make things much more complicated.
In principle, loads should change four times per revolution on rod bearings, however under adverse combinations of mass forces, piston forces and valve gear characteristics, more frequent vector changes are possible and that is one reason of hard running and progressive wear and destruction of rod bearings in such a locomotive. This is also one reason why it is essential in two cylinder locomotives to keep valve gear setting well aligned and working as uniform as possible.
While this is important in any kind of engine type, the two cylinder locomotive is more sensitive in this respect than three and four cylinder engines because of the heavy piston thrusts and mass forces in the but two drive lines.

However you are correct in stating that with front to rear drive sets fixed at 45° phase in a coupled Duplex, the inner coupling rods would have a lot of push and pull to take. And because of the flexibility I mentioned in my initial comment, this would not answer high adhesion demands.

Hi feltonhill

On your remarks of Sept 01:
Crew training rather than making engines more complex:

Right! I fully agree! Also agree with your remark on noise and attention. It seems a shame the engines did not receive better backing up when entering service. In my regard, this sort of ‘build it to leave it’ attitude was also responsible for the total failure of any of the advanced steam locomotives, some of which would look not half bad in retrospect. But their derivation in behaviour and driving and maintenance demands inevitably was their coffin nail without any special care taking in traffic. In technology it just isn’t so, that you could deviate from a beaten path, build something completely different and expect it to do the job right just from the start and without any looking after.

Last edited by Juniatha on Mon Sep 03, 2007 3:45 pm, edited 2 times in total.

Posts: 25
Joined: Tue Mar 16, 2004 6:49 am
Location: Northern Virginia

Post by feltonhill » Mon Sep 03, 2007 9:35 pm

Although virtually unknown, a Q2 was tested on N&W in 1948 from Williamson to Portsmouth and return. I have copies of some of the correspondence that was developed during and after the tests. The Q2 did not fare well in comparison with N&W's Class A, primarily because the A had one more driving axle and a longer stroke. The Q2 had severe adhesion problems with the front engine at low speeds even on sand when the anti-slip device didn't work (which was most of the time). I believe the tests prove that the additional driving axle was important if the locomogive was to be useful at lower speeds, or transitioning through lower speeds at a relatively slow rate before attaining design speeds say in the 40-50 mph range. Although the Q2 was designed and used by PRR in high speed service with trains less than 10,000 tons, it lacked the flexibility to handle very heavy trains (10,000 to 11,000 tons) at average speeds of about 30 mph. For perspective keep in mind that the N&W Class A eventually handled trains of 15,000 tons and more from 1955 through 1958 at speeds up to 45 mph on the same route. The Q2 struggled with 11,000 tons. Will try to get back with more tomorrow. Too late tonight!

Allen Hazen
Posts: 2456
Joined: Fri Mar 12, 2004 10:14 pm
Location: Edmonton, Canada (formerly Melbourne, Australia)

Post by Allen Hazen » Tue Sep 04, 2007 3:42 am

Many thanks for all the comments! I don't have much to add.

(Re "defeating the concept" and also "making a complex locomotive even more complex"): Internally coupled duplex-drive was part of (one version of) the ACE-2000 scheme for a diesel-competitive steam locomotive in the 1970s. As I recall, it was claimed that the internal coupling (a diagram in an article in "Trains" showed two inside rods: the axle before and the axle after the split, therefore, both having two crank throws) would be simpler (in design? to maintain?) than the inside driving rod of a three-cylinder locomotive. I'm not sure why this should be true, but if it is...

(Re performance of Q2): We have "the Q-2 strangely doing admirably well in very heavy freight traffic" (Juniatha), "equipped with an electro-pneumatically operated wheel slip arresting system which reportedly (even though it was new and high-maintenance) actually did work some of the time" (Will Davis), but also "had severe adhesion problems with the front engine at low speeds even on sand when the anti-slip device didn't work (which was most of the time)" (feltonhill). I fear we have only anecdotal evidence to go on: no statistics! So (sob!) we'll never really know how successful the Q-2 was as a design. The PRR retired them well before the J-1, but that's not conclusive: they were a minority type (26 as opposed to 125 J-1), and the added complexity came at a bad time economically: railroads post-WW II had a harder time recruiting maintenance people than before... so the Q-2 might well have been retired first even if it was a noticeably betterr performer than the J-1.

(Re "Q2 was tested on N&W in 1948 from Williamson to Portsmouth and return") Déja vu all over again! C&O tested a T-1 and thought it wasn't as good as their heavy Hudsons. ... As a couple of people have pointed out, though, getting the best performance out of a PRR duplex took a skilled driver, so maybe we shouldn't EXPECT a Q-2 to do very well when driven by a N&W engineer who was used to power with different characteristics. ... The Q-2 certainly LOOKS like a locomotive designed for fast freight on reasonably low-grade routes: not what one would expect to be good on coal trains. Still... the N&W A also had 70 inch drivers (and the PRR, in 1956, was fairly happy with the work rented ATSF 5001-class, with 74" drivers, did on ore trains!).
I'll be interested in anything more you can tell us, Fentonhill.

(re: Adhesion weight) As I recall, the Q-2 had VERY high axle loadings (over 75,000 pounds on some axles), so it almost certainly had more than a 5:4 adhesion weight advantage over a T-1. Not sure what the N&W A was like in this regard: a bit lighter per axle than the Q-2 would be my GUESS, though with 6 driving axles still greater TOTAL adhesion weight than the PRR engine.

(re: the TOPIC): Just to keep it clear, the only project for internal coupling of the axles I have seen reference to was for the T-1. (Which doesn't mean I don't find the discussion of the Q-2 fascinating!)

Posts: 278
Joined: Fri Mar 12, 2004 8:29 pm

Post by Typewriters » Tue Sep 04, 2007 4:35 pm

Well, there sure is a lot to keep up with here.

First, the statement I made about the boiler problems the Q-2 had of course has nothing to do in and of itself with being duplex-drive locomotives but does relate to the fact that the PRR very carefully monitored operating costs of all its locomotives.. and when it came time for storage, the high-cost per mile Q-2 and VERY high cost per mile T-1 locomotives were the prime candidates. K-4 and J-1 locomotives (4-6-2 and 2-10-4) rolled on. That right there causes many people retrospectively to label the duplexes as "failures." I don't think it's that simple.

The Q-2 had 393,000 lbs on drivers, or about 78,600 lbs per axle average. So, Allen, you're about exactly right. Seeing that the Q-2 had 100,800 lbs tractive effort plus another 15,000 lbs from the Franklin trailer booster, we can see that it could roll anything it could start pretty fast (ie near the Pennsylvania Railroad's 50 MPH freight speed limit) given the extremely high HP available (as we know, 7987 indicated horsepower at 57.4 MPH with about 40% cutoff.) 63.5% of engine weight on drivers.

The T-1 had an adhesive weight of about 273,000 lbs by my figures which is about 68,250 lbs per axle average. Testing figures I have are 6550 indicated horsepower at 85 MPH at 25% cutoff; also, on the road with dynamometer I have a mark of about 4600 drawbar horsepower at 55 MPH and another of just over 5000 horsepower. Tractive effort of the T-1 I have at 64,700 lbs. 54% of engine weight on drivers.

Note that the Q-2 is still increasing in HP as it approaches the system speed limit of 50 MPH for freight. Note also the T-1's horsepower peak at 85 MPH, where you'd like it to be for use west of the mountains. One of the mentioned N&W tests using the T-1 did note that the T-1's horsepower remained pretty constant up to 90 MPH, again exactly what you'd like for an engine designed to operate where PRR wanted them.

It's pretty obvious that the T-1 COULD perform when in good shape, on good track with a good (experienced) crew and good coal and water. It's also pretty obvious that the T-1 locomotives tended to be shop queens, what with the Franklin poppet valve gear and complaints of inaccessibility for maintenance. Couple that with the complaints of slipperiness (compared to a K-4, a T-1 is slippery!) and you can see why the engines got a bad rap in some quarters.

One mention was made of front engine slip on the Q-2, which E.T. Harley in his book notes was a problem due not only to that engine operating with unfavorable weight transfer and with worse track condition but also because (on test, it was proven that) it developed not 40% of the total power output but closer to 44%. Obviously, the locomotive was designed for equal power output per axle (front engine, two axles, cylinders 19 3/4 by 28 inches and rear engine, three axles, cylinders 23 3/4 by 29 inches) but due to steam flow / piping arrangements they didn't get it quite exactly perfect. (Front engine had 12" piston valves, and the rear had 14".)

However, his descriptions of slippage with the T-1 make it sound far worse - he describes not only violent high-speed slip of drivers but also an incident in which he witnessed a T-1 stall when it slipped - - moving onto a turntable!

One thing that pops to mind here is a simple relation from later locomotives that I learned from one of my foreign friends. That is this: Many foreign builders will credit a locomotive with higher starting tractive effort capability the more axles are coupled together. For instance - at a time when US diesels were electric drive (single axle) the best you could hope for was 25 per cent. However, diesel-hydraulic locomotives were credited with 30 per cent since you had two or three axles coupled together (or in some single-engine units four axles.) Now, can we relate this to steam locomotives? Well, on a conventional 4-8-4 if the weight unloads from one driving axle, it might well load onto the others and no slip occurs. On a T-1 the chances of it unloading from driving axles to non-driving axles is higher than it is for a conventional 4-8-4, don't you think? Highly speculative here, but something that interests me.

I do note that the similarly powerful 4-8-4 NYC Niagaras had very much the same well-over-6000 indicated horsepower, roughly the same size drivers (79 inch) and about 274,000 lbs adhesive weight but were never regarded as slippery.

In the final analysis it may simply be that the T-1 proved that about 3000 indicated HP to two driving axles without adequate provision for prevention of weight transfer off a set of drivers was something to avoid. However, that's minor and in fact not relevant if the locomotives spend most of their time in or near the shops and cost much more to run per mile. THAT is much more the legacy of the T-1 in my mind than its actual tendency to slip, whether high or not.

-Will Davis

Posts: 25
Joined: Tue Mar 16, 2004 6:49 am
Location: Northern Virginia

Post by feltonhill » Wed Sep 05, 2007 8:35 am

After reviewing many years of research, I found no evidence that PRR considered linking the two engine sets of the T1 via inside connecting rods. It reduced the cylinder bore, thereby reducing TE, increasing the FA and making the locos so modified easier to handle. I believe that this idea may have originated in articles by authors outside PRR or perhaps on discussion boards like this, but I believe this is the extent of it. If a proposal was made either by or to PRR, it's not been found, not yet anyway.

There's been a lot of interest on this thread regarding the T1 so I thought it may be a good idea to post a bibliography of articles and books currently available, that reflect recent and extensive research.

The following highly biased, personally selected list will be overkill, but there are many different pieces of the puzzle to consider depending on how much you want to learn about the T1. There are many more sources than those listed here, but I figure this is way too much as is! And a word of caution. After the 11/59 Trains magazine article, T1 history devolved into anecdotes that were exaggerated, out-of-context, wrong, or had significant extenuating circumstances that were overlooked. They were interesting locomotives in interesting times.


The two books I recommend below give a good perspective in one volume. However, the T1 story is only a minor part of each book, so you have to buy the whole thing to get the pages you want. Atkins presents the best story, comparing the NYC Niagara with the PRR T1, and only makes one misstep, one that everyone made for 40+ years. It is as unbiased as you’ll find. Hirsimaki presents the best perspective of the T1 within the entire PRR saga. He shows the incredible complexity of events that surrounded these locomotives. However, his goals with the book go far beyond just the T1, so the story is a relatively small part of the entire volume.

Atkins, Phillip. Dropping the Fire, Irwell Press (1999), ISBN 1-871608-89-9, pp14-21

Hirsimaki, Eric. Black Gold, Black Diamonds, Volume 1, Mileposts Publishing, 1997

Check any of the book websites (e.g. Amazon...) for these two.


The best sources for recent research dedicated entirely to the PRR T1 are the series of articles published in PRRT&HS’ magazine, The Keystone, and two articles on the PRR T1 tests on other railroads published by C&OHS and N&WHS. The two authors have amassed a huge file of original source documents. Burnell has interviewed the crews that actually operated the T1's over the road and written en extensive series of article for The Keystone. The other two articles explain two relatively unknown tests that were conducted on C&O in 1946 and N&W in 1948. Large amounts of information survived on each of these tests.

Burnell, Neil. “An Appreciation of the T1 - The Enginemen’s Perspective,” The Keystone (Autumn 2001, pp 19-59)

Burnell, Neil. “The ‘Slippery’ T1,” The Keystone (Winter 2001, pp57-62)

Burnell, Neil. Response to 2 letters, The Keystone (Winter 2002, pp11-13)

Burnell, Neil. “A Reassessment of T1 6110 and 6111", The Keystone, Vol 37, No. 1, pp18-39

Burnell, Neil. "The Case for the T1a #5547." The Keystone, Vol.39, No. 3, pp40-52

Stephenson, David R. “Chesapeake & Ohio Tests the PRR T1". C&O History, May 2005
Part 2 of this article is awaiting publication at C&OHS now.

Stephenson, David R. “T vs. J”. The Arrow, November/December 2006

The Keystone is published by the PRRT&HS. Back issues are available, but they don’t have a very active program. Check their website.

C&O History is published by the C&OHS. Back issues are readily available, see their website. The T1 article text is available on ... 42634/pg_1

if you want a free sample of the level of detail. If the link doesn't work, put

C&O tests the PRR T1

in Google. That should get you there. However, there are no graphics, which makes some of the points hard to follow.

The Arrow is published by N&WHS. Unfortunately, back issues are available only as 1-year sets. This may be changed shortly.


Crosby, John R. “Last Chance,” Trains (August 1993), pp 54-56

Lamb, J. Parker. Perfecting the American Steam Locomotive, Indiana University Press, 2003, pp152-160

Lamb, J. Parker. “Supernovas of Steam,” Steam Glory, Classic Trains Special Edition No. 2, Fall 2003

Dr. Lamb does an excellent job of putting the T1 development in perspective with the rest of U.S. steam locomotive history. His conclusions are very well thought out and reasonable in both his book and the Classic Trains article. The Crosby piece in Trains is the best first-person story ever written about the T1.

I assembled some additional stuff on the Q2. Will post that later today.
Last edited by feltonhill on Wed Sep 05, 2007 4:16 pm, edited 1 time in total.

Posts: 25
Joined: Tue Mar 16, 2004 6:49 am
Location: Northern Virginia

Post by feltonhill » Wed Sep 05, 2007 3:31 pm

I recognize the screen names of most of the participants in this thread, and they’re serious bunch! As a result, I want to present some of the recent information found about PRR’s Q2, and also make a preliminary attempt at putting it into perspective. This material is a bit disjointed because I’m heading to Roanoke for three days starting tomorrow, and the heat was on to get this completed. When you’re facing a total of eight hours on truck-infested I-81, you never know.....

Q2 Design

The Q2 not only uses two different bores (required by the different number of axles in each engine set), but also two different strokes, 28 inch for the front and 29 inch for the rear. The different bore sizes are understandable, but different strokes? I wonder why.

The Q2 was hampered at low speeds by control problems with its dissimilar engine sets, particularly the 2-axle front engine. Keep in mind that when one contact point of a 3-axle engine set loses traction, there are five remaining contact points to absorb the cylinder forces. On a 2-axle engine, there are only three remaining contact points to absorb the cylinder force. This engine will be more prone to slip. With 2-6-6-4's and 4-6-6-4's the engine sets have the same number of coupled axles and react to rail conditions in a similar manner.

As noted earlier, the Q2 carried 393,000 lbs on five driving axles. It had an average axle load of 78,600 lbs, with a maximum of 79,780 lbs on #2 driving axle. The A carried 432,000 lbs on six driving axles, and had an average axle loading of 72,000 lbs. The As greater adhesive weight made all the difference at lower speeds because the As tractive effort/drawbar pull was considerably greater than the Q2s.

N&W Tests of Q2 6180

Information has been donated to the N&W Historical Society archives which gives considerable insight into the road tests of Q2 6180 on N&W. The information consists of daily reports from Guy Harding, PRR Asst Road Foreman of Engines, to J. A. Warren, PRR Road Foreman of Engines, 8/2/48 through 8/19/48, and several PRR internal memos involving Carl Steins, W. P. McDonald, Hal Cover, and others. Unfortunately the N&W test report is still missing.

The condition of the Q2 on arrival was poor:

Water pump not working, N&W installed new pump.
Heater valve leaking on water pump
Engine not lubricating
Injector not working properly, N&W replaced it
Slip control not working. PRR personnel eventually repaired it
Sand pipes stopped up and not lined up with rail. N&W made adjustments
Back end of left No. 2 main rod pounding
Reverse gear creeping
Only two sand pipes under lead engine. N&W added two more sand pipes.

PRR’s H. T. Cover was not amused by the lack of initiative when field personnel did not report problems in a timely manner. The Master Mechanic’s office apparently did not carry out the instructions from the higher-ups. PRR management wanted 100% and the MM gave 6180 a lot less!

Slip Control

Most of the time the slip control was not working. Maximum output was being achieved with the device malfunctioning or not functioning at all. It was repaired by PRR personnel for the last run 8/19/48. But why was an advanced locomotive so dependent on an unreliable device?

PRR personnel raised many objections to the tests. Mechanical Engineer Carl Steins and W. P. McDonald offered the most extensive critique. Steins’ objections covered:

1. Probable bias on both N&W and PRR for their own locomotives.
2. Slip control was not working.
3. Coal and water were estimated, not weighed, and the estimated quantities may be biased.
4. Q2 boiler efficiencies not the same as on the Altoona test plant.
5. Test service favorable to class A and not to the Q2.
6. Steins believes that a comparable Altoona test is the only way to make an accurate comparison between the A and Q2

W. P. McDonald presented the most detailed commentary about the tests. There are at least six major topics plus three tabular sets of performance figures:

1. Tractive effort comparison
2. Momentum grade at Kenova
3. Actual tonnage comparison
4. Effect of the number of stops
5. Resistance of 12,500 ton trains being greater than the 13,500 ton trains
6. Slipping effect on coal consumption

However, many of his points are difficult or impossible to evaluate because we don’t have access to the detailed data he had. The search for the original test report is ongoing.

N&W’s Robert Pilcher’s Comments (taped interview 1975 by Lewis I. Jeffries, author of the book N&W Giant of Steam)

Mr. Pilcher’s words, paraphrased:

Q2 had excellent boiler. Boiler efficiency may have been slightly higher than the A’s. Q2 used about 50% more steam for same output. Big difference was the short stroke against the longer stroke of the A. PRR thought you couldn’t get steam in and out of the cylinder at high velocity necessary at high piston speed. Piston speed increases as length of stroke increases.

On a stationary engine you could adjust the cylinder clearance and hold to close, maybe 1/4" clearance and not knock a head out. Need about 1" on a railroad engine, otherwise maintenance of rod bushings and things that wear will cause cylinder heads to be knocked out too often. You had a [unintelligible, probably referring to the distance between the piston and cylinder head] clearance in the cylinder plus all valve ports and passages, and that causes a very high percentage of the expansion stroke in the short stroke engine. But it was a lesser percentage if you had a long stroke engine. It would make a big difference in steam consumption

The efficiency of a steam cycle varies considerable with exhaust clearance. The proper thing was no clearance and the smaller clearance ratio to the piston displacement, the greater the expansion ratio of the steam. Clearance volume is part of the entire cycle up to the ports of the valve. Passages have to be large. When you had the longer stroke engines, they behaved very much better in cylinder efficiency and that’s the reason the A’s and the rest were as high as they were overall.
End of Mr. Pilcher’s comments

Misc Observations on the Q2 and A

IMO the duplex concept was only beneficial at higher operating speeds. Asking the Q2 for maximum output at an average speed of 30 mph wasn’t something it was designed to do, and what we know about the N&W tests certainly indicate it wasn’t good at it. There were constant problems with the front engine slipping, even on sand and after the sanders were repaired. In contrast, the T1 did not exhibit this trait when tested on C&O in 1946 (5511 and 5539) or N&W in 1948 (5511). The Q2's coal consumption was high, perhaps caused by the short stroke design operating at lower-than-optimum speed.

To get its power to the rail at lower average speeds, the Q2 needed the additional driving axle the A had. This enabled the A to work effectively in the 30 mph average speed range when necessary as well as the 55-60 range on time freights. Unfortunately, we have no idea what the Q2 would have done with high speed assignments east of Roanoke. IMO, the Q2 would have given the A something to think about in this type of service which, of course, is exactly what it was designed for. Unfortunately it never happened.

On the other hand, the N&W tests are almost certainly an example of the Q2 being tested in an inappropriate service. On PRR it was almost always used on merchandise service if photo records are representative. Any negative observations derived from the N&W tests must be tempered by this observation.

The Q2 was likely very effective in merchandise service, and that’s where PRR used it. However, the N&W Class A was a more flexible design and could work both low speed coal trains as well as the high speed time freights. Also consider that PRR’s J1s worked well at both low and moderate speeds, and could turn a decent 45-50 mph when required without tearing up themselves or the track. Both the N&W Class A and PRR J1 could do a variety of things well and economically. Although PRR used the Q2 to the best advantage it could, there wasn’t enough high-speed service to justify its extra expense. The Q2 was more specialized and this inflexibility contributed to its early retirement. For the remaining steam years, the J1's were good enough.

Hope this helps

Posts: 278
Joined: Fri Mar 12, 2004 8:29 pm

Post by Typewriters » Wed Sep 05, 2007 9:24 pm

Very VERY good stuff, Feltonhill and thanks a bunch for all that! A few points I'd like to add to what you've said:

I'm with you on that piece in Trains if it's the one I think it is - isn't that the one where the crew got called out on the carpet after the run, getting asked "which one of you clowns has a pilot's license" or something like that (due to EXCESSIVE speed)? I remember an excellent Gil Reid illustration for that one too.

I'm also with you on another more important point as far as this string's topic goes - I have NEVER read or heard of any project to couple the axles of the engines of the PRR T-1 at any time, anywhere other than this message forum. Period. And although I haven't been around forever, I've been at this for many years. My brother who is even more into steam than I am also recalls NO such thing ever. I believe it's pure fantasy.

I've wondered about the differing stroke measurement front-to-back on the Q2 myself. Note: The Q1 4-6-4-4 also had differing stroke front versus rear, rear being two inches less (of course less diameter too.)

Many times people are tempted to compare the N&W A-class with the PRR Q2 since these were tested against each other -- but you are dead on when you say that this is not necessarily worthwhile. In fact, it's somewhat meaningless -- I mean, after all, we're all quite well aware of the largely-custom-designed nature of the steam era and of course you wouldn't really expect every engine to go everywhere and do everything. Heck, they still don't do that NOW. (Amtrak has no Dash 9-44CW units and BNSF has no monococque bodied P42's.) A locomotive designed for a particular operating environment and envelope should be evaluated for its ability to perform WITHIN THAT ENVELOPE. If you wanted it to run somewhere else -- well, you'd have designed it for THAT somewhere else and not where it is. But hey, that's just my two cents.

Notes from Railway Mechanical Engineer on Q2 boiler: "Combined boiler and furnace efficiency of 49.8% (at max rate) and 64.4% at lesser rate." Max rate was 137,479 lbs of water per hour. Total evaporative surface 6725 sq ft plus 2930 sq ft for the superheater; grate area just under 122 sq ft. Working pressure 300 psi. Coal use between 2.75 lb to 3.6 lb per drawbar HP hour depending upon cutoff at approx 50 MPH. I don't know if that's enough to compare directly with the A's boiler as regards the question of efficiency or not but it's about all I have here on that particular data.

In the front matter of Don Ball's book on the PRR there is related the story of the testing of a T-1 on the C&O which includes violent slipping and stalling and restarting and slipping and backing and just a bad time during some of the tests. This was engine 5539 during September 1946 on The Sportsman and The George Washington. This book was written with the help of E. T. Harley and apparently official documentation; it's one of the few mostly-picture "railfan books" that includes good hard FACT as part of its printed matter.

Thanks again, Feltonhill for all that great info.

-Will Davis

Post Reply

Return to “Steam Locomotives”