Thoughts on the PRR Q-2

Allen,

Well, that is the way I look at it too. Perhaps jg has a better understanding of the forces involved here. But it looks to me like the J is essentially a T-1, and so should have been allowed with the modified counterweights, which would have been an existing design.

The longer (or additional) side rods on the single engined alternative to the Q-2 would be rotating weight, and thus able to be counterbalanced. An additional cylinder set would require two more reciprocating rods that could not be completely counterbalanced.

Not only am I not a mechanical engineer, I'm not even a locomotive engineer anymore. Last week I retired. Now I'm just another guy with an opinion.

Les

Les, first let me congratulate you on your retirement. I still have 30 years, 2 months, 21 days till I can hang up my reverser.

Second, let me apologize for reading to much into what you said originally. I got it in my head somehow that you were originally speculating that the Q2 should have been built during the war instead of the J1. Considering the increased wheel diameter and increased boiler pressure of the J1, I doubt the WPB would have cared about updated counterweight specs... after all, those specs would have been established and confirmed before the war.

Once we entered the war, and the War Production Board went into effect, new designs were banned, period, end of story, no exceptions... except for advances in direct military technology. The builders weren't even allowed to build diesel engines for domestic use (I'm sure there were exceptions, but I don't know of any). I know that UP wanted to buy diesels to augment their fleet, but were forced to buy the last batch of big boys instead. I'm not certain if this was because the higher precision technology of the diesel might have used other restricted materials or because they wanted full diesel capacity available for military service? I can't believe that it was to save steel. The only plausible reason I could see would be to a) keep US domestic oil consumption at a minimum b) a steam engine of the day had 6 to 1 horsepower advantage? (picturing the Niagara going up against a 4-unit set of E's).

Back to dynamic augment and reciprocating masses: Using the x-8-x versus x-4-4-x arrangement as an example:

In a four-eight-four, the rods you have are: the main rod, connects the piston tail rod via crosshead to the main crankpin. you then have three connecting rods. Rod 1 connects wheel one to the end of rod2. Rod two connects wheel2 to wheel3. Rod3 connects the end of rod2 to wheel4. (Using older, non timken high-speed, lightweight, articulated rod designs). Because all four wheels are connected, you have to consider the effect of the huge amount of power, aka piston thrust, available to all four wheels. The main rod connects that piston to the main drive, it therefore carries the MOST amount of force, and must have the most amount of structural integrity. The means of increasing structural integrity is to increase the cross-section thickness of the rod. This exponentially increases weight. Main driver to (either driver 2 or 3 depending on engine setup) carries a bit less force since some is consumed by the work done by the main driver, both tractive effort, and in overcoming friction and inertia. Since there is a bit less force, the rod involved can be a bit smaller than the main rod. Now, force has been consumed by the main driver and a second driver. We have a marked decrease in power available. This means less force is transfered to drivers 1 and 4, so these connecting rods can be made significantly smaller. All these decreases in size result in decreasing weight in the rotating masses, reducing how much counterweight is required.

Now, breaking the design into a duplex, you have the main rod to driver 2, and the connecting rod to driver one. Now, because the cylinder is only powering two drive axles, the cylinder can be made smaller in diameter, reducing the resulting surface area that the steam has to effect power, reducing the actual piston thrust considerably. Since the piston thrust is cut, the main rod's girth can also be cut in half because it does not need to carry the same amount of force to the main driver crankpin. Consequently, the connecting rod does not need to carry as large a load of forces and its own girth can be reduce a corresponding amount.

Because the weights and forces involved (per engine set) are so much smaller, the engineers can make better calculations to get exacting counterweights needed.

Now we get to the crux of the theory. The reciprocating mass... the piston, the piston rod, the crosshead, the first 2/3 to 3/4 of the main rod. All this mass equates to inertia at speed. There are only two ways to handle this energy... use steam to cushion the piston, and use the counterweights to try to cushion the effect of the reciprocating masses. Steam is pretty much out. As the piston compresses the steam, the pressure goes up, as does the heat content. But not equally. This means that at some point the pressure in the cylinder will exceed the steam temperature's ability to keep it steam. Now you have incompressible water in the cylinder... and now you have no cylinder head. Guess where we MUST counteract the reciprocating masses?

But, if we have perfectly balanced the drivers, that we must inker with that balance to counter act the reciprocation. Now we through the balance out of... well, balance.

BUT, if by reducing the amount of the reciprocating masses in the first place, we can reduce the amount of mass needing to be counter balanced, we reduce the amount of which the driver is out of balance. Eventually, with the right amount of mass reduction, we get to the point where the masses out of balance are within the range that is acceptable hammer blow.

Now, keep this in mind. The counterweights are rotational weight, as is the crankpin. However, the driver rods, AREN'T. Their inertia will present as tangent forces added to the drive pins. This force is increased as speed increases. The slower the wheel, the lower the kinetic energy of the rods.

You know, the more and more we talk about this, the more and more I'd LOVE to get some time on a high-level engineering computer to do some advanced calculations and simulations to see exactly what type of forces we actually are talking about in these types of systems.

All these concerns were part and parcel to the design of the ACE-3000 in the 1980's. They tried to solve the cross-counter balancing, hammer blow and even reciprocating mass issue by locking four cylinders into position and ensuring that every force created on the engine was opposed equally by another force being created. With four cylinders locked together, they actually managed to create an 8-power stroke per revolution engine.

----------------------------------------------------------------------------------------------------------------------------------------------------------------

Additional thoughts:

I think the main problem with the Q2 was that the engineering department tried to create an engine that solved one inherent difficultly of design.. the hammer blow, but then the operating department was used to verify the effects, when in fact the people getting the benefit of the new design was the track department. The only physical benefit to the operating department would have been a gentler riding characteristic, and long term perhaps less structural damage to the frame. Where-as the track department would have benefited from much lower track pounding, hence fewer broken rails and less rail replacements.

When taken in an overall-grand-scheme-of things, the extra 10k of servicing might have generated untold amounts of savings if the entire fleet was so equipped... lower long term heavy engine repairs, lower track repair costs, fewer slow orders, less track time out of service for rail replacement, less dynamic augment induced loosening of railbolts, faster average train speeds.

jg:

So, essentially the comparison is between the effects of reciprocating mass of two large piston/valve gear/rod assemblies vs. four smaller ones? That makes sense to me.

Thanks for the congratulations. I have been working away from home for most of the time since 2005, and it is really good to be back. One of my big at-home projects is to finish my model railroad. I hope to post some pictures of it here.

The 30yr/60yrs old and out is a great policy. I took a retirement with 26 years in at age 62, but I think the reduction in RRR is worth the extra retirement time given.

I'm out of the New Mexico desert now, and back in semiarid Ogallala NE.

Les

I don’t think that it is unreasonable to assume that the Q2 achieved the objectives of lower dynamic augment and lower piston thrusts as compared with a 2-10-4, and that longer term these benefits might have shown up to advantage, had there been no other problems. I suspect that its adhesion difficulties might have been the main reason that it was sidelined early. At the operating level, where a slippery locomotive made it much more difficult to do the job, one easily could imagine Q2s being the last selected from the ready track, after the better choices (from the operations viewpoint) were all taken.

That the duplexes had adhesion difficulties surely should not have come as a surprise. As I recall the factor of adhesion (FA) for the Q2 was something like 3.9, although I do not know the individual FAs for the front and rear units.

The empirical evidence is that the FA required for successful operation decreases with the increase in the number of coupled axles, up to a point, anyway, such that 4.0 is a reasonable target for an x-4-y freight locomotive with drivers of around 69 inches. Higher drivers also require a higher FA; e.g. compare the UP FEF2 with say the Alco WWII standard 4-8-4. And the late high-drivered 4-6-4s and 4-4-2s typically had quite high FAs.

4-6-6-4s, with 6-coupled driving units, required higher FAs than x-8-y freight locomotives, because of the greater “slipperiness” of 6-coupled engine units as compared with the 8-coupled type. Also to be considered was that the front unit of simple articulated and duplex locomotives was more inclined to slip than the rear unit, and this differential seemed to be more marked for x-6-6-y locomotives than it was for the x-8-8-y types. The early UP Challengers were I think, a bit on the slippery side. The D&H derivative addressed that problem partly by having increased adhesive weight for the front unit and partly by going over to a single-plane articulation joint and concomitantly a fully equalized front engine unit. The UP “big” Challengers, which looked more like a three-quarter edition of its 4-8-8-4 than a lineal development of the early design, had a higher FA, something around 4.2 or so, as compared with the 4.0 of the 4-8-8-4.

Now the Q2 was evidently intended to do something close to the same job as the UP big 4-6-6-4, but with 5 instead of 6 driving axles. Looking at the general evidence on FAs, then one might have expected 4.2 or even a bit higher for the rear, six-coupled unit, and something noticeably higher for the 4-coupled front unit. But the Q2 FA of 3.9 was significantly more optimistic, and very much so for the front unit.

The workable FA curve seems to flatten out at the 4-coupled point, with but minimal further FA reduction coming with additional coupled axles. I wonder whether lateral friction effects come into play here. If say the some of the drivers tend to bind when their flanges are pressed hard against curved rails, then the “braking” effect of the binding might be intermittent or sporadic, thus leading to TE variations sufficient to induce a slip. If so, then the Q2, with its very long rigid wheelbase, might also have been prone to this effect, amelioration of which could have required even higher FAs. Surely, the lateral cushioning devices would have offset this effect, but that would depend upon their disposition. The ultimate Alco/Blunt approach was to have lateral cushioning devices on all but the last pair of drivers in any set, with progressively decreasing lateral maxima from the foremost pair. But I don’t think that the Q2 was like this, and maybe it could not have been, as lateral for the second driver set, main drivers for the front unit, would surely have been limited to less than the Alco/Blunt progression would have indicated.

So maybe the Q2 would have worked better not only if it had sane FAs, but also if it had been an articulated. More complex, true, but the UP big 4-6-6-4 of 1942 had demonstrated that with appropriate running gear such could track safely at any speed likely to be contemplated for freight operations.

The N&W 2-6-6-4 testing might have given a clue. I think this locomotive had a quite low FA for its type, but as stated the N&W mostly kept it on the flat, and had a combination of such well-maintained track and roadbed and general conditions that it could get away with lowish FAs. Put it this way, the N&W 2-6-6-4 might not have looked so good in UP 4-6-6-4 territory, where the grades and often adverse weather conditions pointed to more conservative FAs. Maybe the PRR would have come up with a different answer had it tested a UP 4-6-6-4.

Cheers,

There are some concerns with this however. One is that the "slippery-ness" of the T1 is now being called into doubt, re the article in Trains a few months back that specifically looked at the issue.

A direct comparison of FA of the Q versus the UP-Challenger is not equal. The Challengers and other articulateds were inherently a different issue because the front assembly was not part of the main locomotive frame. Where-as the Q's had one single solid frame with extra cylinders. Because of this, I would hypothesize that the FA you mentioned for the Q was a flat calculation, not a weighted average of the two units.

The Factor of Adhesion is supposed to be a direct ratio of the tractive effort to the adhesive weight. The reason FA doesn't matter to diesels is that they have almost perfectly uniform torque (almost perfect: since the magnetic fields have beginnings and ends and overlaps, there will be some variation. Most will be nullified by the advanced computer control traction software and the gearing between the traction motor and axle).

If the FA needs to be 4(-to-1) for a steam engine, it stands to reason that this is to overcome the highs in the parabolas of the torque of a steam engine generated by four power strokes tied directly to the axle.

At slow speeds, the unequal torque can be handled by a skillful engineer adjusting the throttle and cutoff of the cylinders. At higher speeds, this becomes an all or nothing proposition: you pretty much have to shut the throttle until the wheels catch again, then try to throttle back up, without ripping it out and ripping knuckles and drawheads.

Now, if we factor in that inaccuracies of the counterbalancing of the wheels and reciprocating masses, it makes sense that the reduction of moving mass would only have a favorable impact to the engine by helping to even out the torque. This suggests that duplexes SHOULD have actually been LESS slippery than an equal singleton. That they weren't suggests that there is something else in the design causing problems. Perhaps the calculations for how power should be divided up between the cylinders is a far more complicated algorithm that imagined and therefore something better calculated by computers that by people. Whatever the reason is, it is obvious that computer simulations SHOULD be carried out by anyone interested in this topic.

-J.D.

Gentlemen, a most interesting discussion - over a two year period.

Given the failure of the Q1's design, helping to lead to the design of the Q2, I've been pondering a design change for years. Perhaps you can tell me if this would be practical (for a steam locomotive, at least).

I've always wondered if a slightly different duplex design would work, especially to reduce the balance and hammer blow problems, if the locomotive was designed as so:

For sake of argument, the design flaw of the cylinders and firebox combination of the Q1 has been solved. It works. The rear engine in this design backpedals as did the Q1.

My thoughts are to couple the drivers in this way: Right side, front engine, three drive wheels coupled. Right side, rear engine, two drive wheels coupled, crank pins 90 degrees apart, at say, 12 and 6 o'clock from front engine to rear engine.

Left side, front engine two drive wheels coupled, rear engine three drive wheels coupled, crank pins 90 degrees apart, at 3 and 9 o'clock.

In essence, a 4-5-5-4.

As far as the balancing of the drivers, the combined coupled sets would be at 12, 3, 6, and 9 o'clock to one another. The center axle would be connected to the front engine on one side, and the rear engine on the other, eliminating a slippage problem, and possibly helping to reduce the balance issues.

Given the technology of steam locomotives as we know it today, would this have been plausible?

Okay, I'm no expert in steam tech, just a fan interested in steam locomotives and fascinated by the duplex concept. Considering your "4-5-5-4" concept -- how would you deal with the problems of "backward" facing cylinders that plagued both the Q1 and B&O's Emerson, which supposedly were affected by abrasive materials kicked up into the crosshead and piston area? (And I have long wondered how that affected the SP's cab-forward articulateds that ran "backwards.") Also, in the case of something causing a portion of the drivers to "want" to slip, the slip is impossible due to the shared third wheel set and axle. But that wheel set then has to bear the stress of preventing slippage, and in fact the stress is going to come via an unbalancing of rod thrust between the two wheels connected to that axle, which means the rod bearings are going to be the first parts to absorb the stress and pass it along via the wheel to the axle and then the opposite wheel, etc. I wonder if a connection of this sort would just tear the third wheel set apart over time. Not that I really know what I'm talking about -- just saying.

GSC and RLsteam--
Thanks for reviving this discussion!
I had (before your post, GSC) wondered about the "4-5-5-4" configuration, but (not being an engineer, or even a physicist) didn't feel that I was in a position to evaluate its possibilities. Thinking about it some more, I went back to read up on some steam conceptions... and found that the two-and-a-half/two-and-a-half driving axle configuration had actually been proposed by a professional!

Bill Withuhn had an article, "Did We Scrap Steam Too Soon?" ("Trains," June 1974, pp. 36-48), in which he speculated about possible (and possible at the time: no technology that wasn't available in the late 1940s) steam designs which might have been competitive against early diesels. One was a ten-drivered engine, with cylinders places as on the Q-1 and G.H. Emerson, with side-rods connecting drivers 1-2 and 3-5 on one side but 1-3 and 2-5 on the other! Not QUITE a 4-5-5-4 "Q1mod": one major and several minor differences:
---The asymmetrical side rods aren't the only connection between the forward and rearward driver subsets: there are also inside connecting rods connecting cranked (second and fourth) and doubly cranked (third) driving axles. (Inside rods are, of course, a maintenance pain, but Withuhn at least was of the opinion that inside connecting rods wouldn't be as much of a pain as the inside driving rod of a 3-cylinder steamer).
---There are no less than FOUR trailing trucks: this is a 4-10-8. (From the accompanying picture, it looks like a four-wheel trailing truck followed a four-wheel "trailing-trailing" truck. One of the objections to the placement of the rear cylinders on the G.H. Emerson and Q-1 designs was that they were close to the firebox, where radiant heat could be bad for their lubrication: the multi-axle trailing truck arrangement allows the aft cylinders to be put in a slightly cooler location.)
---It has 60-inch drivers (instead of the Q-1's 77-inch... what WERE the Pennsylvania's designers thinking of when they used drivers that high on a 10=-drivered freight locomotive?)

Note that Withuhn was invloved in the design of the proposed "ACE-3000" new-generation, diesel-competitive, steam locomotive of the late 1970s. (Remember that the price of oil increased very fast in the mid/late 1970s: many people at the time thought that alternatives to the diesel for railroad motive power made good economic sense!) That design incorporated some of the thinking behind the speculative "hight-have-been" designs in the 1974 article, including the use of cranked axles and internal rods to connect the forward and aft driver subsets, but the ACE-3000 design had only four driving axles, so there was no need for the asymmetry of the 4-5-5-4.

(Thinking about this got me thinking about the Pennsylvania Railroad's most famous steam catastrophe: I'm about to start a "Lower drivers for the T-1?" string on the PRR forum.)

rlsteam, for the sake of my discussion, the mechanical issues of the firebox / rear engine had been solved. For now. I was just concentrating on the method of coupling the drivers.

One thought is that we have essentially a ten-coupled locomotive, with four cylinders instead of two. Without scientific calculation, let's assume that the two engine sets are each producing half the power of a two cylinder arrangement to move ten drivers. With power coming from both directions into a non-duplex design, the problem of wheel slippage would be no more severe than convention two cylinder power. Is it necessary to have a "fuse" in that half the drivers can slip?

I'm sure the center axle of a "4-5-5-4" would take some different kinds of stress, but would it work? Possibly a twisting might occur, the axle getting pulled and pushed at the same time. But with half the power coming from the two sets of cylinders, might the stresses involved be lessened somehow?

Allen, thanks for the reference to Withuhn's article. I remember reading that a long time ago. Maybe that's where I got the idea for a 4-5-5-4.

Gary

Correction to my original post above - each side would show the crank pins at 180 degrees apart - one side would show pins at 12 and 6 o'clock, other side at 3 and 9 o'clock. I had posted 90 degrees apart. The pins would be 90 degrees apart on each end of each axle, as was the standard.

When it comes to the Pennsy, when all else fails, go to the Keystone! Here's where the mystery of the J1 balance issues is explained. Notice that the track damage problems didn't appear immediately. Seems the steel used in the driver suspension and equalizer parts (pins) was not properly heat treated. C&O would have discovered this shortly after the T1 was built, and of course, corrected as a matter of development. Too bad the Penn did not get the latest spec on this when they acquired the plans. Pennsy engineers had to dope this one out, and I'm kind of surprised that they didn't "ask" the C&O about this. But then not all the J1 locomotives developed this destructive trait. Anyway, the material and heat treat were changed, and the revised machines worked perfectly ! It seems strange that this would have even unfolded in the manner in which it did, but there were other concerns about the J1 that were being resolved at the same time. One of these was the reduced crew vision ahead given the large size of the J1 boiler. FWIW, The driver centers used on both T1 and J1 were eccentially the same...Penn went 70" to get extra mileage before replacing the tires. Eventually all this stuff got sorted out, and the J1 went on to become the Pennsy's best. As has been stated, The Q2 was simply too much chooch for the circumstances.

Re-reading, a decade later...
Withuhn's proposed 4-10-8 and the ACE 4-8-8 both look a bit strange to eyes used to conventional American steam locomotives: too much trailing truck! (Instinctive response: surely not enough adhesive weight for the power!)
But... I think adding additional carrying axles, even though it reduces the adhesive proportion of the locomotive's total weight, is something at least some steam locomotive designers in the late 1940s would have been o.k. with. Lima's proposal for a "double Belpaire" boiler was a 4-8-6. The anticipated gains in efficiency from having a very large firebox and grate were apparently enough to compensate for having to haul around extra weight on non-motored axles.
A 4-8-6 (or Withuhn's 4-8-8) could do the same train-hauling as a more conventional 4-8-4 with only a little extra cylinder horsepower (to haul the increased locomotive weight), but maybe using LESS coal and water.

Pity the ACE project didn't get to the stage of road-testing a prototype!