LMR: Ackermann Swingback Church Pew 2nd Class Carriage

[What About The Bee]

Continuing to report on 2nd class carriages of the Liverpool and Manchester Railway, I will investigate one of the earliest carriages on record for the LMR.

On the 27th of July 1829, the Liverpool Albion provides us with a description of the carriage:

"The next grade of machine is like a compartmented, oblong square of church pews; without doors, the ends neatly painted, and it is boxed in at the front and back, with panelled work, surmounted by a small railing or balustrade. The rails for the support of the back turn over, so that the passengers may face which way they please, and the machine never requires to be turned around. In other words, the Liverpool end will be the back to Manchester, and the front in returning."

As Opening Day was 15 Sept 1830, the Liverpool Albion described this carriage more than 1 year before the railway opened.  It also appears in the Ackermann long aquatint consist prints in 1834, so this carriage type had longevity.

The phrase "rails for the support of the back turnover" is quite interesting.  In current times, we call this swingback seating.  Swingback seating is used in marine applications.  Follow this link to see illustrations of the type.  https://www.google.com/search?tbm=isch&q=swingback+seat

As is typical of early 2nd class carriages, they began service without any canopy.   This illustration by Reynolds shows 2 of these carriages without canopy.

I know this is a rather poor sketch, but due to its unique portrayal is included here. 

This 1833 portrayal by Issac Shaw, known good observer, is presented by Ackermann.  It is shown side by side with the more familiar blue 2nd  canopy carriage. 

Note that the canopy is significantly lower than the blue 2nd canopy.  To account for this lowered height, there appear to be cutouts in the canopy.  Thus, passengers could enter and exit without bashing their heads. Oddly, the cutouts do not align with the doorways.

The next portrayal is the 1834 illustration presented by Ackermann.  The 1834 Ackermann print does not possess the Shaw signature, leading me to believe Ackermann had a staff artist do the work.

Here, I note the canopy cutouts now have one for each doorway, supporting the head bashing theory.  Yet the canopy is raised to the same height  as the blue 2nd canopy, thus eliminating the need for cutouts. 

From this evidence, I offer that we may detect a rare error by Shaw.   Nearly all depictions of this canopy show the cutouts.  The cutouts are only required if the canopy is lower.  Yet 3 cutouts does not make any sense.  I believe the correct answer is that the canopy was lower, yet the Shaw depiction is missing the correct quantity.

There are a few more observations I can make.  
1)The 1833 depiction shows 6 doorways, the 1834 depiction 5.  Reynolds shows 4 doorways. Shaw must have seen these carriages, they were on the railway from the earliest days.  Perhaps the carriages came in many types.    
2)The swingback seats appear to be fairly tall relative to the floor.  Passenger thighs are mostly portrayed at angles, thighs not parallel to the floor when the passenger is seated.   Tighter legroom is not just a feature of modern aviation.
3) Seating appears to be 4 abreast. There are 5 swingback seats and two fixed seats, one at either end of the carriage.  If all seats are occupied, that is (5 swingback + 2 fixed) × 4 abreast = 28 passengers per carriage.  Compare to the standard blue 2nd at 24 passengers, and we may understand the utility and longevity.

A comparison of this carriage, using the 1834 portrayal shows that even when the carriage is shown with only 5 doorways, the carriage is significantly longer than the blue second.

The wheelbase is also  longer than the 2nd blue carriage, in the exact same consist

I would like to model this carriage, but am challenged by the length.  I cannot simply use the standard Hornby chassis.  The Hornby chassis is too short and I am unwilling to compromise.  It should therefore prove exciting to model, as it includes a novel chassis.

Further, in my usual desire to over spec any model, I would very much like the swingback seats to operate.  3D printing offers the ability to have articulated parts printed, making operational swing back seats a distinct possibility.

Looks like fun!

Bee

[What About The Bee]

In the previous post, I made the assertion that we may have witnessed a rare error by Issac Shaw, known good observer.

Intellectual honesty forces me to retract that assertion. I simply was wrong. I have been putting the Swingback Church Pew into CAD.  In doing so, I must constantly reference the aquatint drawings, so as to accurately depict the details.  And those details revealed pertinent information.

My assertion was based upon those circular cutouts in the canopy, and how they did not align with the doorway openings.  

The 1834 Ackermann aquatint of this carriage has a glaring error.  We know this to have swingback seating.  The benches should be rows of passengers, facing either Liverpool or facing Manchester, as the Liverpool Albion reports.  Yet the 1834 depiction  shows passengers seated back to back, which conflicts with the report.

Shaw properly depicts the passengers all facing one way, like church pews.  

The 1834 canopy height and the 1834 passenger seating have caused me to discount the 1834 depiction.  The 1834 may be an accurate depiction, it may not.  Shaw depicts what the Liverpool Albion states.  If that is correct, then the canopy must also be correct.

Shaw is the authority for the details of the carriage and canopy.  As I studied the canopy, I realized that every row did have access to a circular cutout.  

While not ideal, this is one possible solution of where to put your head as you get yourself off of or onto the carriage.

This is the current state of the CAD model.  The carriage shell, canopy posts and canopy are all depicted; as well as the benches.  No chassis or swingbacks yet.

This is compared to the carriage to the right in the 1833 depiction 

I think this a reasonable match.  It will obviously be further tweaked into submission as I go.  

I will now shift into the chassis development.  I will use pinpoint axles, like all of Hornby's Era 1 rolling stock.  I note the axle ends are a conic, they are shaped like an ice cream cone.  The depression that axle end fits into is a conic, also shaped like an ice cream cone.  I did a one time, one axle, one location measurement, and got side to side play of 1.1 mm.  That is, the axle can be moved side to side relative to the carriage, and it moves a total of 1.1 mm.  Now obviously, I can measure many of these Era 1 carriages, many times and take more care in the measurents, but this may be a data point a correspondent knows from experience

What is the nominal side to side play for pin point axles?  I would be most appreciative you letting me know!

Bee
 

[81F]

Hello Bee,

Might I suggest a some pure speculation about the cutouts - could they be there for some other purpose connected with the uprights or could they have been second hand originally intended for a shorter three compartment coach which might explain why they are not the same length as the coach body.

Also a swing back seat means that the distance between the "doors" would not need to be as great as that required for back to back seating, so each compartment could be around 2/3rd the size of those in the blue coaches. Consequently the entire swing back coach would only need to be around 1/3rd longer than the blue coach.

Note also that nobody is sat facing at the end of the coach in the 1833 image which also suggest to be a shorter compartment.

Edited April 18 by 81F

[What About The Bee]

1 hour ago, 81F said:

swing back seat means that the distance between the "doors" would not need to be as great

I've got the width at the top of the elliptical openings at just over 24 scale inches (0.622 meters) and the solid between the openings, where the seats reside, at just over 15 scale inches (0.395 meters).  The depth of the seat is shallower than the depth of modern seats.  If it is back to back seating, each passenger gets 7½ inches.  Cramped isn't the word.  

 

1 hour ago, 81F said:

Consequently the entire swing back coach would only need to be around 1/3rd longer than the blue coach.

You are correct.  It is about ⅓ longer.  This image, by Shaw, takes the two carriages from one consist.  With pixel addresses, I get the chassis at 1.27 times longer which is very close indeed to ⅓.  

You will also note, in the upper left hand corner, how low the canopy is on the swingback church pew carriage.  Exactly as Shaw drew it.

1 hour ago, 81F said:

 

Note also that nobody is sat facing at the end of the coach in the 1833 image

That is a difficulty I have yet to resolve.  The "seat" is somehow attached to the endplate.  When all the swingbacks are arranged so that passengers face to the left of the image (as shown), the last opening on the right is wide open.  A shallow seat could be in that rightmost compartment and still provide leg room.  Yet if there were passengers on the endplate on the left, all those passengers in the left "compartment" will have cramped leg room.  

Now flip the swingbacks over.  All the passengers in swingbacks face right.  The opening on the left is free, and the cramped area would be on the right.  

Perhaps it is a shallow folding bench on each end?  Folded into position for use but only when suitable?  I have not addressed this in the model.  Sharp eye there @81F  I do think the design is symmetric.  It almost begs to be by the description in the Liverpool Albion.

1 hour ago, 81F said:

pure speculation about the cutouts

My comments about the purpose of the cutouts are also entirely speculative.  The entire thing is confused.  Later depictions, not by Shaw, show one cutout per opening, but the canopy is too high.  Shaw shows this weird arrangement and the canopy is incredibly low.  We only have a handful of primary depictions of the Swingback Church Pew type.  Canopies were installed starting in 1832, but this carriage type was fielded starting in 1829.   A canopy on this carriage type was therefore an afterthought by the LMR.   Ackermann published this Shaw drawing in 1833, but the drawing could easily have been created in 1832.  Perhaps the cutouts are a function of early experimentation.  

I'm not sure we will ever have a definitive answer to the purpose of the cutouts, just these handful of strange depictions. 

Bee

 

 

Edited April 18 by What About The Bee
Clean up grammar

[What About The Bee]

Making the chassis has two parts. 

Part 1 is the artistry of horn guides, horn blocks, springs, buffers and the like.  I do believe that part 1 is just judgment and artistry.  The only critical nature is that it is representative of the carriage.  I must only satisfy myself in this representation, so this really isn't a worry.

Part 2 is the pinpoint axle and pinpoint socket bearings.  This is a geometric issue.  If I get the geometry wrong, the carriage may not function at all.  It may function but track crazily.  It may ride tilted or askew. 

The first thing to consider are the lengths.  In image 1, observe the length of the axle and the bearing distance between the bases of the pinpoint sockets.

It should be clear to you that the axle length should be long enough to extend into the pinpoint sockets on both sides, but not be longer than the bearing distance.  If too long, it binds.  If too short, it may fall out or permit the axle to take an angle to the block, creating that tilted carriage.

In image 2, I have placed the point of the axle into the point of the socket in the pinpoint bearing on the right.  Without a groove, it can be observed that to force the axle into the other socket, the block must flex quite a bit. 

There is also the issue of the pinpoint of the axle interfering with the bearing socket, because the angle of the socket is too shallow.  I could never achieve this position without bending the axle or shattering the bearing socket.

In Image 3, I have added a groove.  The first thing of note is that the axle is much closer to insertion.

There is angular clearance for the pinpoint axle.  Keeping the angle of the pinpoint socket close to the angle of the pinpoint axle minimizes skew and tilt.

In image 4, the critical interference is shown.  

The axle is just a tiny bit longer than the insertion distance.  This only requires a tiny bit of block flex to enter.  There must be a tiny bit of interference, otherwise if the axle is too short, it may fall out.

In image 5, the axle is in place.  Observe that the axle is just shorter than the bearing distance.  The angles are fairly close. This is the goal position.

With these requirements in mind, I am determined to use a closely matched pair of wheel sets.  The axle length is within 0.010 mm of each other.  They are on hand.  I expect these to be the experimental set, to see if my design criteria is acceptable.  I intend to print the yellow block and test.

With a set of working criteria, I can easily redesign to Hornby wheelsets for future carriages.  

With part 2 off to Shapeways, I can go back to part one, leaving the inside of the block blank, until the experiment is complete.

I am not saying "this is how to do it".  I am saying "this is what I did".  Further, until I have a pinpoint wheel set functioning in a pinpoint block, there will be no numerical values to what you are seeing here.  I will update these when I do have a working model.

Bee
 

[81F]

Making allowances for the axles presents some interesting problems. My approach evolved out of the type of component made and its choice of materials.

Some of my initial models were coach bogies made to replace the BR Mk1 type used on the Lima Siphon G and Hornby short clerestories. As the former used a flexible plastic I chose Shapeways "versatile plastic" to enable the axle boxes to be sprung out therefore the grove into the bearing was not necessary.

Indeed my initial models were designed to use Romford "Top hat" bearings so a grove would have been not help. To this end I used a stub axle as below. The 4mm diameter was to accommodate the bearing flange while to two projections either end were 2mm diameter and stick out 5.5mm beyond the flange. This shape was then deleted from the finished bogie frames whoes inside faces were 23mm apart, thereby giving a 0.5mm recess for the bearing flange.

The nature of Versatile plastic is such that when some wheels ran a bit tight, I could slightly flex the side-frames outwards around two dozen times to slightly stretch the print until the wheels ran smoothly, being careful not to break anything. 

However, when I experienced difficulties obtaining bearings, I adopted the approach that you have taken by replacing the axle model with a straight sided 'rod' with pinpoints on the ends. Overall length 26mm with the points being cones 2.5mm x 2.5mm diameter (see below). Although, as I still used the flexible material I did not need the grove.

As I moved onto complete wagon kits where versatile plastic appeared too grainy, I used a composite material approach making the body out of fine detail plastic while I still used Versitile plastic for the under frames (or at least the parts that carry the axle.

Unfortunately this approach has not worked for my 009 GVT slate wagons and I soon found that the versatile plastic too brittle despite introducing grooves. I have therefore being experimenting in making these in Bronze via Shapeways. As a result you may wish to consider getting that firm to makey our W irons and axle boxes out of either bronze or brass and then fix them into a fine detailed plastic body.

I hope you find the info useful particularly the dimensions as a starting point although they may need some adjustment as some of my prints have prooved a little too tight.

Edited April 20 by 81F

[What About The Bee]

Thanks 81F, I will study this response.

I have been looking at the elasticity numbers in the materials section, in a combined materials model.  Axle bearings one material, carriage body fine detail plastic.

Bee

[LTSR_NSE]

Bee - could it work to make the chassis in 2 halves? (left & right)
or make 2 side frames that attach to the chassis separately?

With either of those options the material wouldn’t need to flex at all, since attachment to body/chassis would hold the axles in place, & if attachments were screws (rather than glued) then removal would also be possible, if desired.

[What About The Bee]

Hi LT&SR_NSE 

Yes, indeed.  There are quite a few multi-part solutions available.  I did consider the two halves solution.  It will certainly function and so I have not rejected it.

My thought is to walk down the well trodden path that Hornby uses, to wit:  socket and groove.  It is simplicity itself, with the complexity in the materials science and mechanical engineering. 

How far can I deform a shape, without permanently deformation (yield).  That is, will it spring back, or have I changed its shape.  In some materials, yield is much lower than rupture (ultimate yield). In others, it is not and the material breaks.

In the case of fine detail plastic, the material is relatively stiff, so there is little deformation before yield.  Yield is close to ultimate yield, meaning it shatters.  It is not an ideal candidate for the socket and groove.  It might be made to function with fine detail plastic, but control of parameters is required.  

Its an interesting problem.  One that I hope to resolve!

Bee

[What About The Bee]

This is a multi-part update.  The first part is how to engineer a wheel clip to hold a pinpoint axle.  If that isn't your thing, feel free to skip onwards to Part 2.

Part 1: Elongation at Break

In a previous post, I discussed the the modulus of elasticity and other mechanical engineering criteria to determine if the yellow wheel clip block would break under the insertion force.  

Shapeways also offers a different data point, which represents a simpler way¹ to calculate if the wheel clip will break.  The data point specified is Elongation At Break.  For fine detail plastic, the EAB is a mere 4%.  That is, if the material is elongated by 4% or more by bending, it ruptures.

When you examine bending, one of the first things you will note is that the outside of the bend is elongated and the inside of the bend is compressed.   The neutral axis is in the geometric center of those two.  Material is neither elongated nor compressed at the neutral axis.  It remains the same length.

Fundamentally then, it is straightforward to obtain the arc length of the neutral axis (because it remains unchanged) and the arc length of the elongated surface (outside of the bend).  Divide the elongated arc length by the neutral axis arc length to yield a ratio.  Subtract 1 and multiply by 100 to get a percentage to compare to the EAB.

This inevitably brings us to the hinge.  Hornby Era 1 rolling stock has a very clear hinge.  Look at your model from the buffer end, along the side.  Notice that the spot immediately below the chassis is quite thin, with the horn block and horn block guides substantially thicker just below.  The hinge is right there, in plain sight, but if only you know where to look.

I've installed a hinge into my model.  You can clearly see it just above the wheel set.  It is the narrow section, with curves (fillets) above and below to eliminate stress concentrations.

You will recall the insertion interference shown in this diagram.  

This interference forces the horn guides outwards.  With the interference  greatly exaggerated, we will get the horn guides to deform to the shape of the red lines.  Take a moment to examine this diagram. 

As the axle is pressed in, the horn guide is forced outwards.  Just as the red lines show.

I have shown, in this diagram, both the arc along the neutral axis, as well as the arc along the path of maximum elongation, at the outer surface of the material.

Given the amount of interference during insertion, I can obtain the radius and angle subtended for both those curves, and thus compute the respective arc lengths.  With two arc lengths forming a ratio, the percentage of elongation is obtained.

Voila! Compare the percentage of elongation to the percentage Elongation At Break, to see if a hinge ruptures, or not.

I still have not presented my design numbers.  Yet given my current parameters, and the analysis presented above, I obtain less than 1% Elongation At Break, and fine detail plastic defined as 4% EAB.  In other words, the wheel set should insert without breakage.

The predominant criteria are:
1) the insertion interference.  This should be minimized.  The smaller the better, accounting for axle length tolerance.
2) the hinge thickness.  As the distance between the neutral axis and the elongation surface increases, the elongation ratio increases; without changing any other parameter.  The thickness of the hinge should be minimized.  Thick hinges are a no-no.  The Hornby hinge thickness, as a point of reference, measures 1.3 mm.  This makes the distance from neutral axis to elongation at 0.65 mm.

I will update this with numbers when a complete solution, including successful empirical test, is found.

Part 2: Current State of the Model

Swingback chairs now have a functional swingback.  To paraphrase the Liverpool Albion, the rail will turn over on my model, in yellow.

I carefully examined the images for any evidence of the swingbacks.  

If the rail came up much above the top of the carriage's shell, we should be able to see the diagonal members supporting the rail.  We do not.  Thus, the swingback is fairly low relative to the shell. Further with the top of the shell at waistline, the swingback was only lower back support, it did not reach shoulder blades. 

The chair is pure speculation on my part.  Based on the tops of passenger thighs, I have a seat height.  How the swingback was made, the chair supports, & etc is all just guesswork.   

Many images of this carriage present the shell with frame and panel construction.  Getting the concentric elliptical frames was an interesting challenge.

Part 3: About passenger capacity.

As this was a very early carriage,  it was subject to the early weight limit.  4 tons.  Even the locomotives were subject to this limit².  Therefore, this carriage must  have complied with the 4 ton limit.

If we permit back to back seating, as some illustrations show, then the passenger capacity is 48 passengers.  5 benches of back to back, 4 abreast or 40 passengers plus the two end benches at 4 each, for a grand total of 48.

The carriage itself was included in this weight limit.  Assign an arbitrary weight to the carriage, say 1500 pounds.  The remainder, therefore, is 3¼ tons.  

3¼ tons / 48 passengers is 135 pounds each.  A value too small for the average weight of a member of the public. 

If I restrict the passenger capacity to 28 for the swingback church pew carriage, as originally estimated, the average passenger can weigh 232 pounds.³  Far, far more reasonable.

Part 4: But what about those images with back to back passengers?

The 4 ton weight limit was only for the very earliest of days.  The rail in use was of the fishbelly type, 35 lbs to the yard.  This was clearly inadequate.  As early as 1832, broken rails were reported throughout the LMR.  By 1833, rail was 50 lbs per yard.  Whilst fishbelly rail was judged to be stronger (and it was, with yet another nod to beam theory), parallel rail was far more convenient for curves, points & etc, not requiring chairs at specific locations.  By 1837, rail as heavy as 70 lbs to the yard was in use in certain parts of the line.  The Lime Street Tunnel, the main terminus in Liverpool, was laid in 60 lb rail.

With the heavier rail came an increase in weight limits.  Perhaps the 48 passenger was acceptable.  Perhaps the swingback chairs became fixed, with a paltry 7½ shelf to rest passengers on, back to back.  The images certainly suggest so.

Which leads me to the speculative conclusion that there were two separate and unique carriages.  The first carriage type being the one with the low canopy, definitely with swingback seating.  The second appears to have dispensed with the swingback seating in favor of back to back seating, and a far more logical distribution of circular openings in the canopy, with one per door.

Bee

¹does not require extensive knowledge of beam theory. It does involve fairly complex geometry, especially as the beam is to flex under various insertion interferences, so as to measure angles and radius.

²The concern was wear of wheels and rail.  Further, rails deflect under load (more beam theory) leading to locomotive inefficiency.  We even have Professor Barlow's "Deflectometer" for measuring the instantaneous and maximum deflection of rail.

The thumb screw is brought into light contact with the underside of the rail.  With zero established, any downward deflection of the rail is magnified on the scale, the maximum recorded by the sliding member.

³The ubiquitous blue 2nd carriages held 24 passengers.  This permits 270 pounds a passenger, or far more likely, an increase in the weight of the carriage itself.

Edited Saturday at 00:18 by What About The Bee
Add word fillers for clarity, spacing for footnotes