Saint Class 135 ish mph

Wasn't aware of anyone saying anything else.

 

It's presently away being rebuilt by Thompson.....

 

Ahem!

I allowed myself a smile reading that.

 

Sorry Jim, maybe I missed the few pages of the thread where Scotsman, Papyrus and Mallard's records were all called into question but somehow CoT and the Saint weren't...!

 

Not followed this thread in its entirety, have we discovered yet whose dad is harder?

 

Definitely not mine. Gets misty eyed when we catch spiders to put outside.

 

And here's the result…

Simon

 

The roll has two types of distance measurements marked on it. Each quarter mile there was an automatic pulse. Also each time a milepost (and station, tunnel, etc) was passed one of two observers pressed a bell push to make a mark on the roll. One of these observers was D R Carling. Over a distance of twenty miles you can see there was a very small difference between these marks. Perhaps caused by the measuring wheel slipping on the rail. An insignificant error.

So I can measure between these marks - and they are as near as I can determine the 24"/mile that they are supposed to be.

I can't independently check the time pulses. But if they were better than 1/5000 they don't effect the speed (if you are measuring to 0.1 mph) and if they were worse than that I think it would have been apparent at the time. However there are inaccuracies in the individual half second pulses - perhaps caused by vibration or varying friction between pen and paper.

That is why the 126mph claim - based on a 1 second measurement - is not robust.

 

I don't think anyone is disputing that:

- any claim to speed will come with an associated error, so the speed can only be claimed as X +/- y
- the error bar on an LNER claim measured with a dynamometer is likely to be smaller than the GWR CoT claim measured with stopwatch and mileposts.

The issue though is to think about why a particular claim is significant, and therefore the impact of the error on the confidence with which any claim can be held. Mallard's significance is that it is the fastest steam locomotive, but it holds that title by a very narrow margin - so the larger the error bars, the less confidence there is in the claim. Had the second highest authenticated speed by a steam loco been, say, 120mph, then there wouldn't be much controversy about whether Mallard had the record, and whether it was 125 or 126 would be pretty academic. But because the the second highest authenticated claim is 124.5mph, suddenly whether Mallard did 125 or 126 becomes significant, especially if the likely error bar on the measurement is of the order of 1 mph or greater. Likewise, the (somewhat greater) error on CoT's record is probably sufficient that you can't be completely confident about it definitively being the first loco past 100mph. Had the claim (backed up with the same source of evidence) been, say 105mph, there wouldn't be much debate because even an error of 3 or even 4mph would still put it clearly over 100mph. Equally, had the claim been, say, 95mph, people would probably have been happy about accepting a then-record speed and not worried too much about whether it was actually 92 or 98 or whatever.

In both cases, the degree of scrutiny is because of a combination of (a) a claim that marks a significant milestone but (b) the milestone is exceeded by a sufficiently small margin that you can't have complete confidence in the achievement that leads to the fame.

Tom

 

But then there's the question of did the German's really get 124.5?

 

That concerns me also - comparing a modern interpretation of Mallard's speed against the historic one of 05 002 is not apples to apples.

The original roll of the German run is not preserved as far as I know. It is known that the method was very similar. In some ways it appears less accurate (the paper speed was much less, 150mm per km I think, about 10"/mile), in other ways more accurate (it appears that the 200.4 kph claim is based on an average over 5 km) - but without seeing the original roll who knows how accurate their messwagen was at such high speeds.

I suspect that the most you can say is that the two locomotives achieved speeds too similar to be distinguished using the measuring methods of the time.

 

Agreed, and why I said in an earlier post that any attempt to reanalyse Mallard's record, especially to try to determine whether it genuinely holds a record, would need to apply similar analytical rigour to the German claim as well so as to make a fair and proper comparison.

Tom

 

I've never claimed that stopwatches and mileposts are as accurate as a dyno car. What I have said is that the evidence points to Truro's 100mph being a stronger claim than Mallard's 126. As others have pointed out, that statement does not rely on the two methods being equally accurate.

For 3440 CRM claimed the quarter mile from 171.25 to 171.00 in 8.8 seconds, 102.3 mph. Somewhere just after mp 171 the brakes came on. The definition I have used for "did City of Truro reach 100 mph" is 100 mph at mp 171, which could be equivalent to an average speed over the previous quarter of about 99 mph - or 9.1 seconds, 0.3 seconds more than Rous-Marten's timings. A likely error from a fifth second stopwatch would be 0.3 (0.2 for the rounding of the stopwatch itself and 0.1 for human reaction times). So the 100 mph is feasible - even when you apply that large possible error. On its own that quarter mile timing wouldn't be very convincing - too easy to make a gross error on one stopwatch reading. But you also have the trend of the previous timings and the additional timing over the mile (mp 172 to 171) as supporting evidence - plus the 99-100 mph claimed by W Kennedy. So a strong claim - even including the possible error of 3%.

3% off 126 mph would be an error of around 4 mph - and I am not claiming the LNER record was that far out. The error is actually <2% (that may seem high but basing a speed record on a 1 second time is very dodgy. Even if the dyno car was perfect the difference between 125 and 126 is only a few thousandths of an inch on the dyno roll.)

The LNER dyno car was accurate enough that if you calculated a speed for Mallard based on the average over a mile the result would be within about 0.1% - but that average would be less than 124 mph.

 

I think that's a fair statement as it goes, but it is a bit apples and oranges, which is probably what's causing all the heat.

If we take the actual highest claims for Mallard - 126, and COT 102.3, then there is far far better evidence for Mallard doing 126, than City of Truro doing 102.3.
If we take round number claims - what was the first locomotive to respectively 125mph and 100mph, then if I understand Courier correctly then the evidence is again better for Mallard than for COT.
But, the point is that those are still apples and oranges. 100mph is ~97% of Truro's measured max speed. There is enough data to suggest its more likely than not that speed was attained. But 97% of Mallards max recorded speed is 122.something, and I submit there is not the slightest doubt she went faster than that.

What Courier is suggesting is that the evidence that Mallard attained precisely the traditional measured speed is less strong than the evidence that Truro attained 97% of the measured speed. All I can say on that is that its not unfeasible. And as Tom/James^2 says, that's because of the nature of the specific marks. The other problem is that we are always tending to compare apples and oranges because of the hills. I don't think anyone would doubt that on level track with identical loads FS would be a good deal faster than COT. Its just that until sanity asserted itself COT was being flung in a madcap manner down a 1:80. FS would have easily exceeded 100mph with that much gravity assistance if a suitable1:80 existed.
Mind you, make the same comparison between 4468 and 05-002...

 

It would have been calibrated to be fit for purpose. Generally when there are references to calibration of dynamometer cars it means checking the drawbar pull is accurate. That is most important as it affects performance (dbhp) and efficiency (eg lbs coal/dbhp.hr) results - the things most of interest to the CME. If the speed was normally accurate to within +/-1% that would be more than good enough for general use.

(BTW in the 1920s the LMS car was in error and the published efficiency figures for the Royal Scots when new were very optimistic)

Some issues with the accuracy of Mallard's speed are:

- the 126mph is based on a speed over 1 second. The dyno car was not capable of that and as others have pointed out HNG did not accept that speed.
- accuracy is reduced at high speeds, I'm sure the NER never imagined it would ever be used at 2 miles per minute.
- however well calibrated the dyno car was, the scale that was used to read off the speeds was not calibrated. The error is <1% but it is in addition to other errors I have described previously.

 

But that was not the full extent of the calibration procedure that the LNER followed, my notes show. In any event, +/- 1% of a carefully calibrated tool is still in my eyes a damn sight more scientific than the "data" you are suggesting is more accurate which still at its heart is a man (with questions on his reaction time) observing mile posts (which as pointed out earlier may in fact not be in the right places) using a stopwatch (for which we do not know make, model, how accurate or reliable the timepiece was).

We are indeed comparing apples and oranges. One method can be scientifically compared and contrasted, tested and analysed. You are using the "data" of Rous Marten's run and giving us +/- figures plucked out of thin air effectively and there does not seem to be much concern from your side about the human element of the "data" you are giving us.

Because having seen reaction times physically measured there is an extraordinary variation in the human element. Giving a reaction time of 0.1 seconds for example is generous. We have no idea of how good Rous Marten actually was - a range of 0.1-0.3 seems more likely (but only if we base that on modern day human beings and their reaction times).

In man with stopwatch observing mile posts versus a fully functional and specially designed tool for measuring a variety of factors including speed, I know where my money lies and it is not with the GWR.

Yes it was - and the LNER and LMS helped each other with that as it happens.

When the dynamometer car came into LNER use several parts of its instrumentation were upgraded. As speeds increased, it was overhauled several times as Gresley sought higher speeds. One of these, unsurprisingly, was the wheel mechanism that actually drives the rolls from the rails.

This came about because the LNER dynamometer car in the 1920s was found to be more inaccurate when compared with the Midland Railway dynamometer, which had its own problems in other areas.

Let's say that the dynamometer car is +/- 2% inaccurate. I would happily accept that. Scotsman would not have done 100mph, but Papyrus would have done and easily so.

Mallard might not have done 126mph, but it would have been in or around the German record of 124mph. If you applied the same consistency to the German record, it would have been below 124mph (and what device, pray tell, did they use to measure the speed?)

With Rous Marten I am skeptical with the specific different factors on offer that you can claim the GWR "data" can be as accurate as the LNER's dynamometer car.

In fact I would go further and say, even if you factor in the issues with the LNER dynamometer car, it is still degrees of amplitude more accurate than a man, with a stopwatch (of unknown quantity), watching out for mileposts, that might not even be in the right place.

But whatever makes you sleep better at night for the GWR records.

 

I've had a think about this statement - and I would challenge that on the basis that you can measure and test the accuracy of Mallard's dynamometer car readings, but you cannot possibly do that with Rous Marten's.

At the end of the day there are several big variables in the GWR "data" and you are taking those figures on trust alone.

And for consistency's sake, if we are accepting two different ways of recording data, and we are analysing them, then a level of consistency towards how you are approaching your critique of the data must also follow, surely?

Everything you have said for the GWR side of the argument relies on making assumptions of the man with the stopwatch. That he had the best reaction times, that all of the mileposts were in the right place when he clicked the stopwatch, that the stopwatch was reliable and accurate to within negligible time loss per second, that all of the mileposts were visible, even.

I would therefore disagree with your statement entirely. The evidence for the GWR record does not in any way make the case stronger for CoT than Mallard's does for its own. Your own argument is to me making me doubt CoT's record further.

The 135mph Saint is clearly nonsense. Just for completion's sake.