Steam speed records including City of Truro and Mallard

In his book "British Atlantic Locomotives", Cecil J Allen notes the claims of very high speeds speeds (up to 117 mph) by the L&Y "Highflyers", but expresses disbelief. The highest speeds accepted in the book were on the LNER - 93 mph from a GN Ivatt Atlantic at Little Bytham and 88 mph from a GC Robinson Atlantic at Haddenham.

Hardwicke's record-breaking 1895 run was made with a featherweight train of three 6-wheel carriages, so it was able to maintain speeds that only became normal in the electric era some 80 years later. It averaged 55 mph on the ascent from Tebay to Shap Summit, then reached 82 mph on the descent to Carlisle. It was clearly capable of moving fast, I suspect rather faster than most express engines of that era.

Moving on to later years, it is striking that Mallard and 05.002 appear to have been in a league of their own speed-wise, at least as far as European steam is concerned. The third and fourth highest steam speed runs look some way behind - 114 mph by LMS Coronation in 1937 and 113 mph by East German 18.201 in 1972.

There appear to have been a number of claims of very high speeds by steam locos in the USA, some looking more plausible than others.
Among the more likely is that a Milkwaukee streamlined F7 4-6-4 maintained 120 mph over 4 miles. But that was a stopwatch timing, so obviously comes with a margin of possible error.

 

As the trains that operate there are limited to 75mph, I doubt it would be any higher.


Sent from my iPad using Tapatalk

 

If those locos were to reach high speeds anywhere, those would have been likely places; respectively coming down Stoke bank and continuing to accelerate after the steep descent to Princes Risborough.

 

Got my copy and I have to say it’s a really interesting read thus far. I am relying on what little German I remember plus Google translate at this point (I may yet ask a friend to help too).

 

This from David Slee in Australia may be of interest

Michael Rowe

 

We need to see Mr Slee’s and Mr Andrews’ workings out, I am afraid.

My understanding was that steam was shut off and a braking application was made, which explains the decrease in speed subsequently after the speed record area on the graph.

Fundamentally there’s two issues being posed:

• Is the dynamometer car inaccurate?
• Are the accelerations/decelerations on a slope of 1 in 200 and 1 in 240 possible?

On a side note - I have discovered in the course of reading the 05 book that the original rolls for the 05’s run are not available, or may not be in existence. The evidence for the run is dependent on a graph which shows speed against steam application and a useful gradient diagram (for which I cannot see where the “level track” it was achieved on, is!) but the reproduction in the book is not of the best quality.

I intend to send a few emails out this weekend to some contacts to see if anyone over the channel may have a better copy.

There’s no doubt of the excellence of the 05 design. I am studying the unusual boiler design and noted it has some similarities with Gresley’s design approaches.

 

It’s not so much a question of whether the dynamometer car is accurate, or not. It is more about we have a statement from Mr Andrews that the distance trace was consistent and a statement from you that the gap between the quarter mile marks varied as speed increased. Both statements can’t be right, as far as I can see, but I don’t know enough about the car and how the traces were produced to reach any sort of conclusion. We also need to know that there is no doubt as to the cars accuracy. I’m sure I’m not alone in all this.
Personally, I’d like it to be a positive outcome with regard to Mallard’s 126 mph record.

 

Again, I’m going to be academically on point for this and not elaborate on that part of my findings in relation to Mr Andrews’ work, but I will say that I disagree with him.

On the subject of the speed record, I think we need to have an open mind as the data can be interpreted a number of ways, which is what Mr Andrews has done with his own analysis.

Personally, I am less concerned about the actual record than I am about seeing the full run, finally.

 

Personally I am most concerned who published the cyclic speed numbers below milepost line?

LNER 1938 or a british steam fan long after event ?

Mallard by Hermod posted Dec 5, 2023 at 10:03 AM

 

Hermod - with respect - they’re not “lotto speed numbers” - and even if you don’t agree with them, you don’t have to be rude about them.

 

So, a few preliminary thoughts on the 05 002 loco:

• It’s an obviously carefully thought out design, both from a streamlining and a basic mechanical design point of view
• Baltic type, 4-6-4, so much bigger than the A4
• The frames are huge. This is a big beast
• The Germans used a wind tunnel too to help design their streamlining and one of their photographs has given me pause for thought - they set up a test rig with two models of their proposed locomotive stacked one on top of the other, wheels to wheel - it’s a really bizarre looking thing, and makes me wonder what sort of results they would have got with that. They must have been a reason for this, because there’s nothing we did in Britain that is comparable to this
• The streamlining is split into two sections, effectively. The skirting around the wheels, and the boiler casing. Between them is a running plate, not immediately obvious in photographs
• The 05 series had their skirting go to the rail level, which reduces overall drag significantly around the driving wheels and valve gear. Stanier opted to leave the driving wheels and valve gear open for maintenance purposes, Gresley mostly covered it up on the A4 but stylised it to be in line with the tender running plate
• The 05 series boiler is huge, and appears to have two domes. The length of the tubes has really taken me aback, you could remove the front section with the first dome and overlay the rest onto the A4 boiler and almost have parity. I think though that a Bulleid Light Pacific boiler would have parity as the sections with domes are shorter than on a British boiler
• The grate size is not huge, it’s not a wide firebox like on the A4. It has straight sides, so it doesn’t taper in like a GWR locomotive boiler. That makes sense, it’s a Baltic and has a bogie set underneath it
• The boiler has a slight taper from the firebox front - and I really do mean slight, the drawings seem to indicate that the two sections with domes are dimensionally identical and are connected by a riveted strip
• So my initial thinking is that this additional length of boiler allows the 05 002 to have more steam in reserve, which makes sense when you look at the speeds it achieved and how well they are being maintained
• The smokebox will look familiar to LNER enthusiasts as it has a sloping front. There’s a rally good photograph of the locomotive under construction where you’d be forgiven for thinking you were looking at an A4 under construction, the round topped boiler and smokebox setup is very familiar
• The 05 002 only has a single blastpipe, but the dimensions are significantly bigger than, say, the single chimney setup on the original A4s
• The 05 002 is also a three cylinder locomotive, but with two sets of valve gear driving outside onto the centre driving axle. This setup when you look at the drawings shows three valve gears exactly in line at the slide bar end, but the centre cylinder drives the front driving wheels.
• From the outside, a Brit would look at the 05 002 and think it was a Gresley setup, with the longer outside connecting rods, but the best way of describing this is that it’s a Peppercorn divided drive setup and to maintain the wheelbase length without over elongating the locomotive, the Germans made the outside connecting rods longer than in the inside one. Which is the reverse of what Thompson did on his (equal length connecting rods) and the same thing as he did on Great Northern, but managing to have the bogie under the cylinders. (As an aside, I really like this setup and it makes a hell of a lot of sense).
• The tender looks small compared to the loco, but I think is actually a tad bigger than the LNER tender, certainly has more carrying wheels.
• It also continues the streamlining down to railhead level, so the whole of the 05 series design is completely streamlined
• It also has a corridor, a setup really familiar to LNER enthusiasts)
• The 05 series needed smoke deflectors by the time of its speed record run, so the streamlining at the front end gains perhaps a negligible to small amount of drag with the large deflectors added around the chimney. This is speculative.

Overall, it’s a really clever piece of locomotive design, effectively maximising dimensions where necessary. I feel convinced this is a racehorse that is also a marathon runner, you can see how it was optimised for high speed runs and also to be able to maintain high speeds for long periods of time.

Some thoughts on the test train:

• The German dyno car has some outside similarities to the LNER one by way of being a clerestory vehicle
• Whereas the LNER dyno car stands out in the LNER test train (6 streamlined vehicles behind it) the German train is only three coaches behind the dyno car and they are all clerestory vehicles, not streamlined.
• But this makes a uniform train of vehicles that are all narrower than the streamlined locomotive in front, and more or less the same dimensions in width and shape throughout
• The German train is much shorter than the LNER train and is lighter - it looks like 200 tons to Mallard’s 240 tons, I need to check this again

Some thoughts on the record:

• We have a set of graphs which show a range of stats including speed and gradient
• This isn’t the primary evidence but secondary evidence derived from primary evidence
• The original dyno roll appears to not be available to examine, and we now only have this secondary evidence as a record of the events that happened on the day
• The dyno car did a record very similar to the LNER one and appears from the translation to work identically
• The locomotive had its own speedometer in the cab and two recorders - but this stops at 200kph (!)
• The claimed speed is 200.4kph which equates to 124.5 mph
• The text suggests that the actual record was not attained on the level, I need to check the precise point on the graph but it appears to be on a downhill section like Mallard
• I think British observers may have misread the paragraph in question, I have had to re-read it a number of times but the translation I have states there are “not many level sections” throughout the run, and certainly the graph suggests there either aren’t any, or the gradient is slight uphill or downhill
• This doesn’t appear to be a problem anyway as the 05 maintained some amazing pace throughout its runs including 124mph sustained, which I think is the biggest achievement

Overall, the whole project including locomotive and test train and operation looks to be of the highest quality.

However (!)

• The 05 002 isn’t being run with what could be described as something akin to a service train, but Mallard was pulling something close to an in service train
• The length of its high speed runs shows a marked difference in acceleration compared to Mallard, particularly in runs uphill
• The lack of primary evidence (at this point) means we have to take a lot of this on trust, whereas with Mallard we have the primary evidence and can analyse that to the nth degree
• The text indicates special arrangements for braking the train on its run - the Germans don’t seem as confident as Gresley that the train will brake well!
• The streamlined casing really needs examining, I think there are elements of its design that are not as optimised as they could be, but the general approach is sound

Overall - we really need in this country to know more about this particular class and its speed record, I have come away feeling very impressed by the German design. I think Mr Andrews is right about one thing in particular: the 05 and the A4 classes probably do have honours even as examples of 1930s engineering towards high speed rail, and ultimately even if Mallard turns out to be faster (or not, as the case may be), more needs to be said about the German loco in British media.

 

Mallard was travelling faster than 124mph at 90.5, from what I am seeing in my copy of the dyno roll. But what matters is the acceleration - is she still accelerating throughout those sections? The dyno roll to me suggests yes, she was continuing to accelerate throughout those sections.

I do not believe that the acceleration required is out of the realms of fantasy, in particular when you look at the 05 002 run which seems to indicate similar levels of acceleration on its run.

All of this will be in my write up and the numbers published together with the formulas used, the methodology and the base numbers from the primary evidence I have investigated.

One point of note - all of the positive appreciation for virtually every speed record locomotive so far aside from Mallard seems to be based on secondary sources and not primary evidence - whereas Mallard appears now to be unique in her primary evidence surviving to the modern day.

This to me is something significant. We are able to investigate Mallard’s run to an nth degree more than every other locomotive, some of which have nothing more than secondary write up, after the event, with some fairly questionable time keeping records.

The lack of consistency and rigour being applied across the board is interesting to note.

 

IMO it is correct that the dyno rolls for 4468 are analysed but there does seem to be a strong wish amongst some to disprove her record claim.
Parochialism at its best as I'm sure that some of her detractors would be singing a different tune if the loco had been designed by their favourite engineer and run on their favourite line. Then there are those who knock British loco design and achievements while singing the praises of all things foreign. Good luck with your research paper.

 

How accurate are those gradients that Hermod keeps quoting? After this length of time we will never know. They may be the official gradients but experience tells us that, when track is re-laid, it rarely conforms to the sub base on which it is laid. Except under bridges and perhaps at platforms, the general trend is to lift track over time. As an example, the NYMR is famed for its 1 in 49 gradient but the reality is that it is not the case. It is merely an average over several miles. Thanks to Ploughman of this parish, we know that the gradient at Darnholm between 21M 45ch and 21m 60ch varies between in in 42 and 1 in 43, which is significantly steeper than the official 1 in 49. It's not a huge length of track but is getting on for a quarter of a mile, and is two train lengths so not insignificant in its effect on the train and in calculating horsepower and tractive effort. The following half mile varies between 1 in 47 & 1 in 49 and the preceding 15 chains is 1 in 55 so you need to be looking at about a mile to get that average 1 in 49.
After years of relaying, it is highly likely that the gradient down Stoke bank (and elsewhere) is a variable figure and the nominal can only be used as an average over a significant distance. Given that, in terms of the effect on the train, changes of gradient are not instant and require to have the whole train on the gradient to be a valid value in a calculation, then I'm wary of any such calculations based on nominal gradients.

 

It’s interesting because the German book makes no mention of Gresley or Mallard, but goes into explicit detail into the group of English locomotive engineers who visited in order to take a trip on one of the last test trains (which did not achieve 124.5mph) including one William Stanier and one Harold Holcroft, both greeted on board the train together with Herr Himmler and other notable representatives of the Nazi Party.

Holcroft made a reference to Mallard’s speed run, according to the book, in the German press without mentioning Gresley or Mallard or any other details than:

“…instances of ultra high speed have been recorded in England on test runs in which speed has been authenticated through the presence of a dynamometer car on the train with Pacific types, both LNE and LMSR. These were, however, peaks attained by running down gradients averaging about 1 in 200, and under somewhat hectic conditions. It was no parallel to sustained high speed on level track (190kph)”.

190kph = 119/120mph. This appears to be the source of comparison for the 05 002’s speed record compared to Mallard.

Thank you, hopefully it adds more than it takes away.

I think Steve the natural conclusion on the gradients you have to take is something of an assumption, which is one such section that will be part of my write up.

There will undoubtedly be assumptions and clarifications needed throughout in order to make the best interpretation (or interpretations - more likely in my view) of the event.

I don’t believe we can absolutely prove where the gradients were in relation to the train’s achievements, other than by looking at the dynamometer record and comparing it to known gradient diagrams for the time (point of order: I have one for this stretch of line around this time and it makes an interesting comparison).

So it’s perhaps a bit fruitless to try and work it out to the nth degree. I think you accept it’s not possible to be absolutely exact and you explain in the write up that a level of assumption has needed to be taken, because otherwise you are chasing something likely unable to be proven in that specific factor.

 

Not just that, but the makeup of the train, relative weight, gradients, wind resistance, etc etc - I have always had a bit of a question mark over the claim for CoT and a massive question mark hanging over the (frankly ridiculous) 130mph Saint claim.

That’s before we talk about the personalities involved and Rous Marten was treated with a level of distrust by the GWR, IMO - I think much of the reasoning for the GWR not originally claiming anything is absolutely because he’s involved. Something of a smoking gun.
He had outlandish views on a range of other railway topics and his claims on speeds are undoubtedly questionable, and have been questioned by better experts than I in a range of publications.

For me, he’s an unreliable commentator and that alone, ahead of the mathematics, puts CoT’s record in the bin.

It is, and it all needs to be considered in the context of the time, and the relative worth of each timekeeping log. However it is obvious that 100mph was probably achieved by some of the Pacifics of the SR by the sheer body of evidence we have.

You have the edge on me on this Al, I don’t know enough about it but it is something I am looking into for my Bulleid book.

 

(deleted)

 

The counter argument of course is that it is precisely because there is plenty of primary evidence that Mallard's claim is submitted to so much more forensic analysis.

If you take the claim of City of Truro, then it rests on about 7 or 8 quarter mile times recorded by stopwatch. Essentially there are only really two paths you can take with analysis.

The first is that which has been laid out before by @Jimc - which is to take the readings at face value, and then have a discussion about the likely error bars on hand timing. That leads you to the conclusion that there is a fairly wide spread (several mph) between lowest and highest likely values, but > 100mph lies within that possible range. It could be as low as ~ 97mph but equally could be as high as ~ 105mph.

The other option is to take your view which is that it definitely didn't achieve that speed (and, from some of your other statements, I suspect you are sceptical that it even got to 90mph). But in which case, you still have to explain the data that do exist, i.e. the stopwatch times. Essentially you are then left with two possible conclusions:

1. Rous-Marten misread his watch, and times that he read as 8.x seconds were actually 9.x. It would certainly be possible to make that mistake with one data point, but it would stick out like a sore thumb relative to those around it. It is much harder to assume he consistently got every time wrong by 1 second, because it would lead to a cumulatively noticeable error over a duration of a minute or so.
2. Alternatively, he lied, i.e. he just made the data up - but if so, why? It wouldn't seem to be for sensationalism (even though he had form in that area), since - AIUI - the times weren't published for many years afterwards.

I think at the very least, if you are going to discount CoT's claim, you have to give a convincing explanation as to why the data upon which that claim is made are wrong. It's not sufficient to say "there isn't sufficient power available for the speed": that might lead you to be critical of the data, but you still have to explain why that data are wrong.

With regard Mallard - you should consider it a benefit that the primary data is there to analyse! But the fact that it does exist inevitably means that any claims will generate more scrutiny. Think of it as a feature, not a bug.

Tom