Steam speed records including City of Truro and Mallard

The reason I have looked at the quarter mile marks is precisely because they are made automatically as the dynamometer car travels each quarter mile.

I’ll send you a PM later on.

 

Edited wording to clarify question.

 

My embolding.

I'm not an engineer or a mathematician, but by my reckoning that statement means that an accuracy of 99.9995% was achieved by two men pressing a button (averaging their clicks?).

Really?

99.9995%?

That's as good a reason as any in my book for someone to go right back to Mallard's source material & look at it afresh.

Pete S.

 

Observation on both sides of the train. Does that mean that they both had to press the button for a mark to be made? Or does that mean they pressed a button and called out what they were observing? What were they lining up with on the outside of the train to judge when they passed an object outside? Was this a fixed point or a “sight” of some sort?

Also how was the roll calibrated to the start point? And how was the actual distance from the start point to the end of the roll calculated to provide a comparison?

I struggle to believe there was a correlation to anything like 99.9995% between the marks on the paper and either the actual distance travelled or the externally observed landmarks recorded on the paper.

 

I don't think that is what Andrews is saying. In particular note that he says there are "automatic ones made every quarter mile and ones made when the train passed lineside features such as mileposts. " In other words, two different sets of marks: one made automatically, and one made manually.

As I interpret it, on the one hand there are automatic marks made every 1/4 mile. Those derive from the dynamometer mechanism and require no other manual intervention. Their accuracy is purely a function of the dynamometer mechanism. Andrews suggests that the relationship between those marks and the length along the paper is 99.9995%, from which he concludes that the paper is driven at a very consistent rate relative to the distance travelled.

The other marks that are made by button clicks by observers cannot have any great accuracy for the reasons stated (i.e. human reaction time, what precisely are you lining up on etc). However, they are useful for providing a set of known features that aid later analysis of the roll by providing a correlation with known land marks. (so for example, easier to say "the peak speed occurred shortly after passing under Much Meddling Bridge" than "the peak speed occurred between the 27th and 28th quarter marks on the chart".

That's how I interpret it at least.

(And as an addendum - generally mileposts are only on one side of the line. So I interpret it that one observer looked out one side and had the job of clicking at every mile post; and another observer, entirely unrelated, looked out the other side and noted other specific landmarks).

Tom

 

You and @K14 have got this back to front. Andrews is claiming that paper drive has a very consistent feed rate of 24 inches per mile travelled.
Edit. Cross posted with Tom who has explained it more fully.

 

Someone needs to put their money where their mouth is to settle this once and for all.

Get the Dynamometer out on the ECML, tack it on the back of an Azuma and see what its measurements say…



Simon

 

We continue to ask questions about the paper drive and the quarter mile marks because we seemingly don’t know how the car operated. IF the chart and pen were both directly driven from the fifth wheel I would pretty much expect them to be 99.995% in agreement. However, Simon has said on more than one occasion that the quarter mile marks get further apart the faster the train goes.
I’m still confused

 

Sadly, on today’s railway, I think you’ll find that making them couple together will be the showstopper. And if as I suspect it’s vacuum braked, you couldn’t even slip it into an Mk4 set.


Sent from my iPhone using Tapatalk

 

Yes, which I am saying, and have been saying for some time, that I am not finding on my copy of the roll.

I need to be clearer in what I say.

I have found that at the high extremes of speed, the quarter mile marks are getting closer to each other.

We are not talking by significant amounts - but it is noticeable and where this happens is significant.

I am looking at each quarter mile mark individually - Andrews approach is different to mine.

It’s why I have been saying “limitations of the equipment” against what has been said by others, which is “defect”.

 

A stationary test can be made by disconnecting the clutch (red handle?) between the inside worm that drives the 24 tooth wheell on paper feed drum and the thing down close to fifth wheel and then use some modern digital angle resolvers and indicators to worm shaft and feed drum.
24 turns by hand of wormshaft can give information of systematic errors

 

"I need to be clearer in what I say.".
If you are commenting about mathematics then you must use mathematical terms such as "significance" consistently with their conventional mathematical meaning.

"We are not talking by significant amounts - but it is noticeable and where this happens is significant.".
And then you directly contradict yourself in the same sentence!

 

Thanks. All that confirms my understanding of the system. I haven’t read Andrews so the bit I didn’t know was that his claim related only to the paper feed and automatic marks, and though that makes more sense, I am sceptical that the machine in question would achieve that level of accuracy. Best available at the time I don’t doubt, but accurate to that degree over all speed ranges? Unlikely in a mechanical device no matter how well made.

 

I’m not doubting what you are saying. What I am saying is that we need to understand why they are getting closer together. This would tend to indicate that the paper is moving more slowly. (Or that the quarter miles are getting shorter!)
As I’ve said, unless someone knows better, we don’t know exactly how the car works. ‘We’ being those taking part in this discussion.

 

Another thought on the subject of paper distance and quarter miles. I do not believe you have a true copy of the trace , merely a set of photographs of it. How were these taken? Ideally they want to be by a camera a fixed distance from the trace but, if (say) it was simply a hand held one then any slight variation in distance would lead to a slight variation in length. Are there printed lines on the paper, as we would get with todays plotter papers, which would provide a reference?

Edited to change an auto suggest which I missed.

 

What is needed is a defence quality scan between mp 92.5 and 89.5 analysed by a submarine accustic technician expert.
Plus assurence that no LNER employed touched the machinery or paper during the period.
The last is difficullt
The first can be crowdfunded .
Im in with 10£
An expert can easily find best fitting third order polynomium and first and fourth order sinewaves.The sines are systematic errors as no 400tons steamtrain can accelerate and retard like that.

 

In regards to the photographic copies I have:

• They were taken by the skilled in house team at the NRM's search engine facility
• They are scaled and I am satisfied that this creates clarity in measurement
• The photographs were taken from a set fixed distance as per the NRM's in house team's usual work

If I was not satisfied with the quality of the photographs, I wouldn't be attempting this.

In regards the original roll:

• It is a plain roll, likely cream or white when originally made
• It has 8 line traces that give the various measurements in red ink

I hope that clarifies.

Incidentally, I have taken the decision to place on ignore a couple of members.

I'm happy to be subject to scrutiny when I provide results, but I am not prepared to have to answer for my decision making on the subject of any research I am doing. Nor am I going to argue back and forth on "clarity" when it is quite clear one member has an axe to grind, regardless of what I say.

 

Whilst for many reasons this is not practical it would seem to be much easier to rotate the 5th wheel using a specially designed equivalent of the well understood rolling roads used in many applications.

There will clearly be differences in the friction between dynamometer wheel and rolling road compared to the wheel rolling on track (I've not noticed anyone questioning whether the dynamometer drive might not be true speed at 120+ if there is any slip between it and rail caused by imperfections such as track joints or whether its own mass acting like a flywheel smooths out such things) but it could be used to demonstrate the operation. Like rotating the wheels of 35029 it may also be useful to the museum as a way of providing public demonstration of how the record was claimed.

I am assuming that we know that the equipment was not changed after the Mallard run in any way which would mean its current state can't be compared to events many, many, years ago.

 

You intruded here in a very old thread that was reactivated more or less accindentally by Hirn and me.
We are now fully informed that you are working on a Ph.D and have a high value ,very secret copy of Mallards roll .
Thank You and good luck.

 

@Steve 's last two posts ( #634 and #635) triggered a memory. My first degree was civil engineering and some time in the 1960s I had a lecture on photogrammetry. My notes are long gone but I was prompted to search the internet for "errors in photogrammetry" and found many technical articles of interest. Reading one or two, persuades me that, despite @S.A.C. Martin 's response above, the photographs he has of the original roll cannot be relied upon to provide measurements free of distortions that will allow valid analysis at the required resolution to challenge Andrews.

https://www.google.com/search?q=errors+in+photogrammetry