Saint Class 135 ish mph

William Kennedy?

 

So.... the LNER chappie spotting mileposts and pressing a button is corroborated as accurate by the charts. This suggests that a man can be trained to accurately see and react to mileposts.

However, Mr Rous-Marten, recognised by his peers as the foremost train timing expert, you do not credit with the skill to accurately see and react to mileposts.

LNER mileposts are correct, as the human observer's sightings are corroborated by the dynamometer chart. GWR milepost positions in 1904 cannot be considered accurate.

I am certain I read CRM "used a pair of Swiss chronographs"

Your argument relies on making assumptions about man and equipment- just negative assumptions.

 

Ross how do you think the LNER dynamometer car works?

Clue: it’s not a man spotting mileposts pressing a button...

It’s an ink line dragged across a continuously moving roll driven by an additional wheel, pressed to the railhead.





Your move.

 

Read what I wrote
"So.... the LNER chappie spotting mileposts and pressing a button is corroborated as accurate by the charts"
referencing Courier, post 368

And don't be patronising

 

Fact checking is key. https://en.m.wikipedia.org/wiki/Dynamometer_car

My apologies.

 

Mallard - official
Flying Scotsman - official
CoT - unofficial
2903 - extremely unofficial!

That's what history says. Unless someone produces a smoking gun, that won't change, but the debate is awesome! Any conclusions outside of the above historical perimeters at this stage is still conjecture or possibly belongs in the realm of faith for those like me who have not delved as deeply.

 

How does a dynocar get calibrated in 1938?

 

I'm not sure it does. I don't believe there was any independent body cross checking the data and setting conditions for what constituted a record as there was for aviation and road. That's not surprising since that function was (and usually is) performed by the world association/governing body for the relevant sport and there was and is no sport of racing rail vehicles. If there were I suspect it would throw out just about all the records as being too much apples and oranges and gravity, but if it did allow gravity assist it would most likely authenticate COT at somewhere round 98 or 99, and Mallard at 125.

 

I did wonder about that - exactly who, if any, is the authenticating party as terms such as "official" or "authenticated" are widely used. Or, was it a case of "we did this, this is our proof, now you can go and report it/write about it in books". .???

 

How about:

Official ................... there was a dynamometer car
Authenticated ........ there was a dynamometer car and a bloke with a stopwatch
Disputed ................ "Well you don't believe my claim"
Dubious/doubtful .. Pretty much all of the above

Does that help?

 

A very fair point. Worth thinking about.

 

You are both correct. The LNER speeds were derived automatically - pen marking half second pulses on paper driven by wheel on rail. However there were also marks created by observers spotting mileposts. That was done to record where exactly on the line the dyno record was taken. D R Carling stated he was in a seat in the off-side bow window looking forwards and using a bell-push to mark on the record any mile posts, gradient boards, bridges etc that would serve to locate the train at any time. Percy Dobson was doing likewise on the near side.... (Carling's papers are held by the IMechE)

That means the Mallard dyno roll can be used to indicate the likely accuracy of clocking mileposts - not something I had thought of, thank-you Ross.

So this lunchtime I plotted out the time errors of the mileposts that were recorded during Mallard's run. That is possible by comparing the time difference between the milepost marks and the automatic quarter mile pulses.

The results were:


The error shown is the sum of milepost location error, human error (reaction times) and my measurement error. Ideally if you are trying to detect an error of circa 0.1 seconds you would be measuring with an accuracy of 0.01 seconds - but that is just 0.1 to 0.2 mm on the paper (depending on speed). I can't guarantee that - and so it is possible some of the scatter in the graph has been introduced by me.

I would be interested in hearing people's thoughts on this data. One observation - it appears the higher the speed (mileposts 90 to 94) the better the accuracy - probably because at lower speeds the milepost takes more time to pass the window.

 

- The dyno roll doesn't show 125 mph. I've intentionally not said what I believe the max speed was as I'm still rechecking the data - but it won't be 125.

- FS vs CoT with trailing loads in proportion to their weights would be an interesting comparison (a real world contest might also show that a 95 ton loco doesn't have a fireman twice as fit as a 55 ton loco - and in a longer run he or she might be the weakest link). One other difference - CoT's all out effort lasted less than 5 minutes, all the LNER ones were >10 minutes - that makes a big difference to the max possible power for any loco. (BTW my life would be much simpler if Truro had been a Great Central or L&Y loco - easier for all to be objective about the evidence and what is likely to have happened)

- Mallard vs 05 002 dhp is an interesting comparison. At 100 mph Mallard was producing more power at the drawbar.

 

Interesting you say that, it was not something I was aware of. More apologies due to Ross from myself on that point.

However you are indicating that the dynomometer car's roll is again more accurate than human using a bell push.

Thereby introducing a different variable in the LNER results to the GWR ones - you can work out the actual distance between "mileposts" from plotting the time errors of the mile posts...

I would be hesitant to comment at all on the likely accuracy of the mileposts from the human hand. Actually, based on what you have given to us above, I would ignore the positions of the mileposts supplied by the human hand altogether.

The Dynamometer car roll allows you to work out the top speed, and gives you distance and time, thanks to the automatic mark. That is as you have demonstrated in your graph above, far more accurate than the human reaction giving a milepost plotting.

This is because the likelihood is that, even if the dynamometer had some variable in terms of the slowing up and speeding up of the additional wheel on rail/paper through the rollers, it is consistent in its potential inaccuracy whereas the human reaction time is consistently inconsistent in its inaccuracy...!

 

Oh gosh no, they are useful. The automatic quarter mile tick is fixed in stone, but the manual ticks are reset at each one. That means that they provide a calibration of the automatic 1/4 mile ticks. Whilst the relationship between any individual manual mark and its adjacent automatic mark is going to be meaningless, if you take a rolling average of say the last half dozen manual ticks, then if the offset between the rolling average and the automatic tick changes according to a regular pattern then that indicates a calibration problem, either with the wheel or with the mileposts. If there is no such trend though, then that gives you confidence the calibration is good. If there is a definite and significant trend, though, that is tricky, because you then need to find out whether the wheel was badly calibrated, or whether the mileposts are badly set out. That could be fun!

 

Exactly. There is indeed a small difference between the automatic mark and the milepost data - perhaps the measuring wheel was slipping very slightly - but the difference is small enough to ignore when quoting speeds. When I created the plot of timing error for the mileposts I factored that in - ie put a trendline through the points and then looked at the difference between each point and the trend line expressed as a time in seconds.

However you are indicating that the dynomometer car's roll is again more accurate than human using a bell push.
Over short distances you use the automatic ticks to check the errors in the manual ones. Over long distances (say milepost 105 to 85 - 40 feet of paper) you use the milepost ticks to check the automatic ones - because over such a long distance the errors in milepost position and human reaction time are insignificant.

 

Gravity assistance or mortgaging boiler as mr Cox used to express it?

 

and a more sophisticated front end

 

Ok. So LNER chappie's marks vary (if I'm reading the chart accurately) from -0.2 to + 0.1 sec, so a variance of 0.3 sec. If CRM was using a 1/5 second chronograph, then that would add up to+/- 0.5 second error. So that might give a speed range for CoT of between 96.77mph and 108.43mph.

So,(with tongue firmly in cheek) as we are talking GWR, we have to assume the higher speed was in fact the true one. And as a Saint has twice as many cylinders as a City, and 6 driving wheels rather than 4, that proves that the Saint must have been doing about 325 mph