Carriages / Wagons for Locomotion No.1

[What About The Bee]

I continue to be intrigued by the timing chain.  Such a novel feature.  

Stephenson and Dodds, in their patent, called the part that engages the chain a "wheel furnished with cogs".  When the wheel turns round, the projecting cogs enter the links, "completely preventing the chain from slipping"

Detail
Wood, Practical Treatise, 1825 

Examining the drawing, literally centuries later, the first thing that springs to mind is a sprocket and bicycle chain.  The description of function matches perfectly.  Indeed, the sprocket in the Wood drawing could very well be a bicycle sprocket.

Why would Stephenson and Dodds not use a bicycle chain?  Well, it hadn't been invented yet.  The primitive velocipede¹ was patented in 1817, two years after the Stephenson Dodds 1815 patent.  The bush roller / bicycle chain wasn't invented until 1880, 66 years in the future.

There is a better way to engage a chain than the method that  Stephenson and Dodds used.  That would be a "chain gypsy"² typically found on a windlass or capstan for raising and lowering an anchor chain. 

This provides excellent purchase on the chain.

The "double studded chain wheel" (chain gypsy) was patented by George Daniel Davis³ , who specifically claimed the construction of the chain wheel "so constructed that every alternate link of the [chain]  may lay flat upon the edge or circumference" of the chain wheel.   Patented in 1858, 44 years in the future for Stephenson and Dodds. 

Stephenson and Dodds, lacking a time machine, could not use a bicycle chain or a chain gypsy.  Hence the 'wheel with cogs' and chain.

In examining the chain coupling, I think it important to understand the purpose.  It was to retain the quartering between the front and rear axles.  Most steam locomotives, like Rocket and successors, are quartered across the axle.  The quartering is precisely fixed at 90°, since both wheels are hard mounted to that axle.

Chain coupled axles need not be so precise. Yet on these chain coupled locomotives, as long as one piston was still somewhere in the power portion of the stroke when the other was at top/bottom dead center; the locomotive would still start.  Therefore a few degrees of quartering error would not matter.  Indeed, Staniers 3 cylinder engines for the LMS show that quartering need not be at 90°.

The drawings show no idlers to take up tension and the drawing by Stephenson himself show the  chain to be slack.   A very near contemporary book: History and Progress of the Steam Engine, Galloway,1834 states what must be an innovation post patent.   Chains wear and the affect of the wear is called "chain stretch"⁴.  The innovation described was interesting. "When the chain got worn by frequent use, or was stretched, so as to become too long, one of the [axles] could be moved to tighten [the chain] again, until a link could be taken out, when the [axle] was moved back again to its former station".  They shifted an axle to set chain tension! This has a small affect on the angle of quartering, again, it is not significant.

The number of times the removal of a link could be performed is obviously limited.  Eventually, the links would wear through.  When replacing a chain, yachties are advised to replace the chain gypsy as well.  In Stephenson's case, that would be the wheel with cogs.

Bee

¹a two wheeled bicycle without pedals or chains.  Indeed, there was an early foot powered velocipede railway car noted in 1829.  Sadly, no images of this car exist (to my knowledge).  The car is fairly well documented in the press.

https://community.hornbyhobbies.com/forums/topic/35033-an-unknown-lmr-carriage-emerges-from-the-press/#comment-372391

² the actual term.  In the US, "chain wildcat"

³  text of Davis patent
https://books.google.com/books?id=8lzbPMbBLYcC&pg=RA12-PA1

Patent Drawing 

https://books.google.com/books?id=8lzbPMbBLYcC&pg=RA13-PP4

⁴ As a chain wears, the chain apparently gets longer.

 In actuality the inside contact surfaces of the links wear away.  This wear makes the center distance between two links sepatate.  When accumulated over many links the length of chain measurably increases.  Chain stretch.
 

[What About The Bee]

I thought to closely examine the chain itself.

Wood, Practical Treatise, 1825 provides us with an unusual description.  "This endless chain, which is now solely used upon these kind of engines, consisted at first of one broad and two narrow links, alternatively, fastened together at the ends with bolts; the two narrow links were always outside of  the broad link; consequently, the distance they were separated laterally would be equal to the breadth of the broad link, which was generally two inches, and the length three inches"

Quite the sentence and difficult to understand.  I puzzled over this sentence.  

Nicholas Wood was in a prime position to understand the Killingworth engine.  He was appointed an assistant colliery viewer at Killingworth, in 1811.  His superior?  Ralph Dodd.  Dodd, of course, the other patentee with Stephenson in 1814.  Indeed, Wood became Killingworth Colliery Viewer in 1815.  In 1818, Wood and Stephenson performed experiments with a dynamometer they designed, at Killingworth. ( https://community.hornbyhobbies.com/forums/topic/33593-railway-oddities/?do=findComment&comment=368416 ).  

Nicholas Wood was there and had genuine first hand knowledge. Wood was in a unique position to evaluate and describe the chain.  I do think this is the solution.

The blue links are 2" wide, 3" long.  Two yellow narrow links are on the outside of the blue link, as specified.  The red bolts go thru the blue and yellow links.  I've sketched in gray nuts, but these are not mentioned.  Take a moment to read the description from Wood as you study my solution to the chain.  The chain fits the description.

I was studying the (unfortunately low resolution) patent drawing for UK Patent 3887 Dodd and Stephenson.  I turned the drawing sideways, and wow, there it was.

UK Patent 3887
Dodd & Stephenson 

There is a long thin object drawn there.  Examining it in detail, I can see the side plates (yellow) and the blue links.  The ratios of width and length are not as described in Wood but there nonetheless, along with the red bolts.  That is typical of many patent drawings.  Deliberate proportional errors to throw off your competition.

So the real question now is the shape of the blue links.  Are they made of rod, like standard chain or are they cylinders, as I have drawn.

The rear elevation view of the "Killingworth Locomotive" (per Wood) shows the wheel with cogs and the chain.  The asymmetric object to the right on the axle is the eccentric for driving the slide valve.  Wood's Slip Eccentric, if you please. Returning to the wheel with cogs, we can observe the yellow side plates.  The blue links between the side plates show straight lines side to side, not curved.   So from this I conclude that the blue links are indeed cylinders

It is quite tricky in this end view to make out what is the wheel with cogs, and what are the blue links.  The yellow narrow links are clearly present however.

As the blue links are two inches wide, then the wheel with cogs must be just under two inches wide.  That is very robust sprocket indeed.  This chain can not twist and fall off the sprocket.  Once again, the resemblance to a bicycle chain is startling. 

The chain is such a unique railway feature, that it forms part of the patent claims and drawing.  

Bee

[threelink]

@What About The Bee  Hi Bee. Given that the blue links are, in effect, tubular and the red bolts appear not connected to the blue links in any way I wonder to what extent wear would have been a significant problem notwithstanding adjustment of the axle (what a job!) to maintain tension. I would imagine that the red bolts and the blue links must have moved in relation to each other quite substantially as the chain worked, particularly when passing round the sprockets, creating friction not necessarily overcome by the lubricants of the day. Your investigations of this early railway technology continue to fascinate.

[What About The Bee]

Hi ThreeLink 

16 hours ago, threelink said:

... when passing round the sprockets ...

Thank you for a thought provoking statement.  

There is a difficulty in the chain geometry that I am attempting to resolve.  When the chain is straight, there is no difficulty. My CAD matches Wood's verbal description and drawing, perfectly.  Yet when the chain goes round the wheel with cogs, there is an unresolved issue in matching the drawing.  

I will need a few days to reason my way through.  Please do bear with me.  I will make every effort to resolve the geometry issue and with it, answer the friction question.  

Bee

 

Edited Tuesday at 02:17 by What About The Bee
A word was missing

[threelink]

Hello again Bee. This is me "thinking aloud" so to speak but it occurs to me that as the sprocket rotates the load on the chain is concentrated momentarily only on the lower or upper teeth of the sprocket , depending on the direction of travel. That being the case the remainder of the chain round the sprocket may be under far less load than I had first envisaged and therefor less prone to wear. Indeed, depending on the degree of adjustment in the axle position the chain links may go slack as they rotate about the back of the sprocket. I am sorry that this is couched in less than technical terms but I am a bit of a duffer when it comes to certain aspects of engineering and tend to work on instinct rather than knowledge.

I look forward to reading the results of your cogitations on the chain geometry.

[threelink]

And slackness would allow the blue links to rotate randomly so preventing localised wear by the red bolt constantly bearing against the same section of the blue link.?

[What About The Bee]

The Chain Dilemma

How, exactly, is the chain constructed?

Choice 1: The bolt is free to float inside the blue link, as I have sketched.  The bolts are 180° apart when the links are straight because of the tension on the chain.

Choice 2: The bolts are captive in bores located on the inside wall of the blue link.  The bolts are  180° apart when the links are straight because of the two bores.  In fact, the bolts are always 180° apart.

So what happens when the chain goes round the wheel with cogs?  I superimposed light pink circles with a center dot for each blue link to help see the solution.

For choice 1, the bolts aren't captive.  If we draw an orange line from blue link center to blue link center, then the yellow links should be on those orange lines.  

For choice 2, the bolts are captive.  Therefore, the pivot point is at each bolt.  The blue and yellow link lines, corresponding to blue and yellow links, are drawn to the pivot points.

Choice 2 appears to be wrong.

Examine blue link 5, on the straight.  The pivot points are 180° apart and the blue line passes through the center of the link.  Yet examine link 3.  The red line attaches to one pivot point and proceeds directly through the center of the pink circle.  If the bolts were captive, the yellow link would intersect the end of the red line, nearest link 4.  It does NOT.  Thus, the bolts are no longer 180° apart for link 3.  Similarly for link 2.

It appears, therefore, that choice 1 is correct. But! The yellow links are not exactly on the orange lines.  Go back to see.

With the chain geometry resolved, what about friction, and specifically, wear?

Assume the chain and wheel with cogs are perfectly sized.  The center distance from blue link two to blue link three fits exactly in the space in the cogged wheel for those links.  Precisely as drawn by Wood's illustrator.  The blue links nestle into the cogged wheel and the yellow link, being a precise fit, is drawn to the orange line.  Again precisely as drawn by Wood's illustrator.

The bolt, therefore, slides along the inside surface of the blue link, under the tension force.  I do not think the bolt rolls along the inside surface but I can be persuaded otherwise.  There will definitely be retarding friction and resultant wear, due to the sliding.

Sliding members are notoriously hard to lubricate.  The bolt will act as a wiper, pushing any grease or oil to either end of travel.  Thus, within a short time, the bolt will slide along the blue link, unlubricated.  The inside of the blue link will wear, as will the side of the bolt held against it.

As those surfaces wear, the chain stretches.  The center distance between blue links lengthens.  The yellow link will not be perfectly on the orange line, as it can and will droop away.  Similarly, the blue link will droop.  The wear on the inside of the chain will be reduced as members are not held under load.   This exactly matches your instinct, ThreeLink.

The outside of the chain will then wear, as will the cogged wheel, as the arc length in the cogged wheel and the chain center distance no longer match.  

Maintenance would be constant.  Replacing links is straightforward.  Wood mentions re-tensioning the chain which does not appear overwhelming.

Fitting a new cogged wheel, because it too wears, will be the most labor intensive.  Unless that cogged wheel is two halves that clamp around the axle, there aren't many other easy choices.   Replace the entire axle, wheels and all? Slide the cogged wheel off one side?    

Bee
 

Edited Tuesday at 16:06 by What About The Bee
Apologies for the somewhat blurry images. I am working with what I have. These are small details, highly magnified. If I had better images, I would use them.

[LTSR_NSE]

@What About The Bee since neither of your choices perfectly fit the image (admittedly image could be to blame) is there a 3rd choice..?

Bolts are captive 180° from each other, however the (outside) narrow links aren’t perfectly straight, but actually describe a radius/arc, allowing the (inside) broad links to perfectly fit the cogged wheel.

[What About The Bee]

It cannot be that way @LTSR_NSE.  Captive bolts cannot move relative to each other on the circumference of the blue link, the point of the illustration with numbers.  How can the pivot point move, if the bolt is captive.  It cannot.  QED, the bolts are not captive.

Other simple explanations are

The illustrator was depicting what he saw, and that included wear.

The illustrator wasn't concerned about mechanical accuracy when it comes to a minor point.  The chains are fascinating to me (us?) but may not have held the same interest for the illustrator.  The sketch is reasonably accurate but imperfect.

Bee's pink circles aren't inserted to the nearest angstrom and this leads to angular errors.  

Any combination of the above.

When @threelinkasked about friction and wear, the exact detail of construction  became a critical input.  If the chain was constructed with captive bolts, the wear would be substantially different from free floating bolts.  That needed to be resolved first.  

I am more than happy to try other explanations.  That alternate must fit all the facts we know or can derive.  Sketch some blue and yellow links.  Ignore my interpretation, but do examine the Wood's illustration, the patent drawing and the verbal description.   Explain how the bolts move on the circumference from chain straight to chain on cogged wheel.

I think you will find the simple explanation on offer to represent a logical choice 

Bee

 

[threelink]

Hello again Bee. Taking a break from restoring some signage purloined over 50 years ago from the the abandoned and derelict Great Central Railway I took the opportunity to re-read this thread. I realised that I had failed fully to appreciate the function of the chain. I fell into the trap of thinking that its function was to transmit power so that it would be under constant load. If I have correctly understood the position its function was purely to maintain timing or quartering, the loco being an 0-2-2-0 rather than an 0-4-0. If I have finally got it right I am guessing that for so long as the wheels did not slip the chain would not be under any great tension, merely idling about the sprocket wheels - hence the slack shown in the various drawings. In those circumstances wear would be less. Clearly, if a wheelset did slip load would be imposed suddenly if not violently until timing was a re-established but presumably that state of affairs would persist only for a relatively short period. Perhaps wear was less of an issue than I imagined. I did say that I am a bit of a duffer, having had no engineering training... 

Thank you for yet more fascinating insight into early railway history. 

[What About The Bee]

Hi @threelink

The chain was replaced with coupling rods, when the second "Traveling Engine" made its way from Robert Stephenson & Co engine works.  Likely by very late 1825, or by 1826.  The S&DR complained that the second engine did not perform as well, in writing.

This is what leads me to believe the chain was present on Opening Day.  Active hauled an impressive consist, without complaint.  Hundreds of people and tens of loaded chaldrons, not to mention the Experiment passenger carriage.  That tremendous consist.  It beggars belief to suggest that was done with an experimental engine.  It had to be a Traveling Engine with understood mechanics.  The Hetton Colliery engines, the predecessors, all had chains.

I consider the Killingworth engine to be an 0-4-0, with the chains taking the place of the coupling rods.  The chains were to take care of quartering.

Bee

[threelink]

 HI Bee. Just to clarify my understanding of rods/chains and Whyte notations, I always considered coupling rods to be a means of transmitting power from a powered wheelset to an unpowered set. Hence a loco with 2 wheelsets, only one set powered but coupled by rods to transmit power to the unpowered set, would be an 0-4-0. A loco with 2 independently powered wheelsets (as I believe was the Killingworth loco and its cousins)  with a chain intended purely to maintain quartering and not to transmit power, would be an 0-2-2-0. 

I am sure you are right about Active being chain fitted on Opening Day. 

[What About The Bee]

I accept that I am likely wrong about the Whyte notation for these early Stephenson Traveling Engines.  The problem is that they do not neatly fit into that notation.  The two pistons really aren't independent, nor are they completely dependent.

This makes a muddle of fitting it into the notation.

Example 1:  the forward axle slips.  The chain, being connected to the rear axle which is not slipping, halts the forward wheel slip and all the power of both pistons is applied to the rear axle.

Example 2: The Traveling Engine is stopped. The rear axle is at top dead center.  Thus, the rear piston cannot apply power to the rear axle.  The forward axle, being ~90° out of phase, is at full power stroke.  Apply steam.  Does the chain merely keep timing, or is it applying torque to the rear axle?

Example 3: The chain breaks during operation.  Can the Traveling Engine make it back to the engine house?  I say yes.

Bee

 

[LTSR_NSE]

Example 1 & 2 would definitely correspond to an 0-4-0.

However (imo) Example 3 is more complicated… a more modern 0-4-0 (quartered wheels not axles) with broken coupling rod(s) would become a 2-2-0 & struggle more than the (chain-less) 0-2-2-0 except if wheel slip caused both of the latter’s pistons to be at 180° (or even worse at 0°) to each other!

Edited yesterday at 17:46 by LTSR_NSE

[What About The Bee]

A further confusion is evident in the patent drawing. 

Look at the side elevation.  The Traveling Engine is hauling a chaldron.  There is a strange little pony cart between the TE and the chaldron.  There is a chain!  It runs from the TE to the pony cart.

In the patent, Dodd and Stephenson provide the answer to what this is.  They explain that the TE is expected to haul ~60 tons, but if more is desired, the pony cart  is actually the tender and it can be made to increase the tractive effort.  

Applying the friction of the bearing wheels in this way, the engines propel 60 tons or upwards upon an iron railway.  If a greater burden is to be moved, the friction of other two wheels is added, and they are made to carry the water that supplies the engine.  A groove is made in each, and a groove in the last two bearing wheels of the carriage, into which an endless chain falls.  By the propelling force of the engine this is moved along, together with the bearing wheels of the water carriage attached to it.

The only other drawings of this pony cart are clearly derivative of the patent drawing.  

Caution is always advised with non-period drawings, but I find this one to illustrate the powered tender adequate.

Bee

 

 

Edited 23 hours ago by What About The Bee
I hope this addendum provides insight into the chain. Dodd & Stephenson intended it to increase the tractive effort. The only way this can be so is if it transmitted power and not just timing.