Steam speed records including City of Truro and Mallard

Yes but in a separate thread please.
To contain smell of lousy loosers

 

124.25/55.54 as steady state at mp91 and the cheap power model asks for 2228 hp against air and 450 against rolling and get 1478 from gravity.
Cylinder power needed for even speed 1800 hp.

Time markings on roll indicates that Mallard started working 4 hours earlier/6 tons coal burned ?

V2 maximum test stand ,two firemen and screned coal was 1990 ihp.
Model is not a mile of but next step is to test and trim model fudge factors where acceleration was high.

 

Breaking my rule - again - but I cannot let this go unchallenged.

Again.

It is good to see Mr Andrews’ account in a more readable account.

However this is not the smoking gun he thinks it is.

Again.

And it is nice, observing as one researcher to another, the claim the dynamometer roll has a fixed rate of paper through its rollers (which was previously made) has now been modified to confirm what I have always maintained - the speed and distance pens work independently and the roll runs through proportionately to the speed it is travelling at.

So there is NO FIXED LENGTH OF PAPER PER MILE AS HAS BEEN PREVIOUSLY CLAIMED.

By Mr Andrews:

https://imechearchive.wordpress.com/2020/07/03/guest-blog-a-matter-of-time-and-space/

So let’s put to bed the issues as I see them:

1) Regular variation every mile in the paper

I went through the dynamometer roll at length and never found this “regular variation” over every mile. We need to see Mr Andrews’ working out in order to confirm Mr Andrews’ account.

I have said this before, several times.

But even if - and only if - there actually was such a variation in the roll per mile, since time and distance are being recorded simultaneously and also independently recorded by way of the dynamometer’s pens, how quickly the paper goes though the rollers is irrelevant to calculating the speeds involved.

All the paper does is act as a medium for confirming the time and distance stamps: as they are being recorded independently and also simultaneously, you can measure the time stamp against the known quarter and milepost marks by way of the two sets of stamps and you will get instantaneous speed at each point between them.

The actual length of the roll and how much paper was used to record a mile is not important.

Now, a far better question would be to ask if the chronograph for the time stamps or the ratio of ninth wheel to rail to pen for measuring distance is accurate, but we know from LNER history that the dynamometer car was repeatedly calibrated. So it seems unlikely that these two highly important parts of the equipment were in any way faulty for Mallard's 126mph run. I reject his claim that there is any issue with the equipment.

2) How much power is required to accelerate a train from 124.5mph to over 126mph

Mr Andrews has posed this question and told us it takes more than 2000 DHP.

Great. Is that calculation assuming the train is on the level?

Because we know (because we are constantly reminded about it!) that Mallard was descending Stoke Bank at that point - 1 in 240 it says on the dynamometer roll. In either words, Mallard at 124mph with the dynamometer and six streamlined coaches was accelerating downhill on a 1 in 240 gradient.

Then there’s air resistance. Has that been taken into account?

Because a fundamental issue I have with virtually all of the power calculations is that not a single researcher to date has given us an indication of how much horsepower is saved by the streamlined casing.

Only Bert Spencer has shown us any real work in that field. Here is a graph showing his calculations visually:



So a streamlined locomotive saves at 120mph close to 400 horsepower required to accelerate to that speed over a conventional steam locomotive. This is a huge saving, and it is also in Mallard's train too that the rest of the train, bar the dynamometer car (sheltered immediately behind the streamlined tender) also reduces horsepower required to overcome air resistance.

Mallard’s conjugated valve gear produces on the middle cylinder additional power at speed by way of overrun, which is more apparent at higher speeds, and has been shown in history to be up to an additional 10% on the middle cylinder. This is a much, much smaller factor but worth being in mind, given the high speeds involved (and we know it overheated due to this additional power and overrun).

So on the basis that:

• Mallard is accelerating the train downhill
• Its streamlined casing significantly reduces the horsepower required to overcome air resistance
• Ditto for the rest of the train

I can well believe that Mallard did not have to achieve 2000 DBHP to then achieve 126mph on a downward slope having already accelerated the train to over 125mph approaching the next level section.

Also - FYI - on the dynamometer roll it records DBHP, and I don't know what roll Mr Andrews is looking at, but I am looking at my digital copy of the roll and at the 125mph mark it reads 1265 DBHP, down from 1290, whilst going down the bank.

Mallard accelerated the train to 123 1/2 mph at the point the gradient goes from on the level to the 1 in 240 and at that point is producing 1370 DBHP, ahead of descending the 1 in 240 slope.

3) The 05 002 speed claim

Mr Andrews makes the same points I have made about this locomotive, but I have an additional point to make. The graphs and date we have of the locomotive's run shows 200.4kph achieved whilst accelerating uphill.



And the gradient profiles indicated throughout the run are quite sizeable!

So for consistency's sake:

1) If we are saying Mallard cannot have achieved 126mph whilst going downhill because it cannot produce enough horsepower to accelerate from 124 to 126mph
2) It seems unlikely that the 05 002 with a slightly shorter but not much lighter load (that was not streamlined in any way) can have accelerated up to 124.5mph accelerating up a greater grade than Mallard is descending.

This shows a big inconsistency and misunderstanding on Mr Andrews' part of the data he is looking at.

As far as I am aware, only this, in a German language version of 05 002's history together with a larger copy (that was made available to me) exist for 05 002's run, and as Mr Andrews confirms, the original dynamometer car roll no longer exists.

4) The 100mph GWR Saint

I cannot be bothered to go through this in as much detail but that section is a total non starter. No, I reject Mr Andrews' claim of the Victorian GWR Saint 4-6-0 as having "comfortably exceeded 100mph". Everything from DBHP, to air resistance, to accurate recording of the mileposts and timing make this, amongst other GWR runs, a non starter.

So again, as one researcher to another, Mr Andrews, if you are reading:

1) Did you take into account the gradient profile and streamlining for your claim that Mallard needs 2000DBHP to achieve the acceleration to 126mph?
2) Please show us the raw data that you recorded that shows this "regular variation"
3) Please also acknowledge that it isn't relevant, as the only crucial thing is what the pens do to the paper and that they are working simultaneously, and whether they are accurately recording quarter miles and seconds
4) How can we accept that you believe Mallard accelerating to 126mph in one instance, going downhill, but believe the 05 002, accelerating to 200.4kph, going uphill? Is this not inconsistent?

When I started looking at this in February this year I happily showed all of my working and the data I was working from (and you can find it earlier in this thread, and also copied across to the Gresley thread).

Please draw your own conclusions.

Good night.

 

Strong ‘Godfather 3’ vibes here Simon…

 

I hope not, I only want to see good research and understanding being discussed!

 

How did the second markings relate or index to the quartermile markings without having pen collisions sometimes?

 

You cannot be serious.

I sent you a copy of a section of the roll. The pens are spaced apart.

Hermod - really?

 

Sorry, couldn't help myself! Carry on

 

Your yellow grafiti is in a wrong place at km 110 and the upper circle shows steam temperature .
The happenings happened at km 65 to 59 on a grade that is 1:5000 falling or rising.
Could You please move defamination of 05002 to a new thread?

 

Whoops, wrong pic. I will replace!

 

I find it worrying that two seriously inclined individuals can analyse the same roll, one of them find a regular one-cycle-per-mile variation and the other be sure that there is no such variation. Anyway for certain the rest of us on here are in no position to decide who is right.

I do have some specific comments on Simon's posting. He says
"the roll runs through proportionately to the speed it is travelling at.

So there is NO FIXED LENGTH OF PAPER PER MILE AS HAS BEEN PREVIOUSLY CLAIMED."

Those statements seem contradictory. If the paper is driven by the ninth wheel on the track, how can the length per mile change, other than possibly with some irregularity such as the regular variation detected by Andrews?

Also: "But even if - and only if - there actually was such a variation in the roll per mile, since time and distance are being recorded simultaneously and also independently recorded by way of the dynamometer’s pens, how quickly the paper goes though the rollers is irrelevant to calculating the speeds involved".

If I have understood correctly (please tell me if not) there are actually two entirely independent records of distance travelled by the train. One is distance along the paper as that is driven by the ninth wheel. If that were not of interest, the ninth wheel would not have been needed and the paper could just as well have been driven by clockwork. The other record of distance is the marks made manually when passing the quarter mile posts.

Both of those records are subject to possible errors: distance along the paper caused by any imperfection in the drive transmission, such as the possible once per mile variation; and random human error of a fraction of a second in making the quarter mile marks. Therefore neither is absolutely trustworthy, but they are entirely independent, so calculation of speed should take full account of both.

 

The Copyrigth protected pictures up til now does not show it clearly.
I think there are four equally important quater mile marks made from four ramps on side of paper drum.
Paper drum rotates once per mile to within 1part in 2000 according to Andrews.
To relate these quarter miles to earth it was job of some humans to press a button passing real mileposts and write milepost number on moving paper.
The accuracy of this operation has no relevance for record claim.
1904 sea clocks were tested to within 7 seconds on a day or 7 divided by (24 times 60 times 60) or around 1 part in 12345.
It is not clear from public material to se how far pens for seconds , automatic quarter miles and manual earth mile marks are separated in longitudinal ,running paper direction and with fast moving paper this can be of relevance.
It has not been written to what accuracy paper drum quartering works.
The qartering accuracy needed to defend one view or the other in this lovely debate is better than .1 mph to 124.2 or one in 1242.
Makeable in 1904 but a close shave.

 

Let’s clarify this, hopefully for good.

So the paper goes through the rollers, driven by the ninth wheel directly affixed to the running rail.

(We are all in agreement).

There is one pen for distance and it records quarter miles, and full miles, by way of a specific shape it makes when the wheel has completed a specific amount of rotations.

On the actual graph paper, there is, if you like, a “secondary record” of distances.

You have the quarter and full mile records done automatically by the dynamometer car, and you have mileposts marked manually by hand by way of identifying the mileposts as they passed in the train.

I have calculated instantaneous speed at each automatic, quarter mile, and I have also checked the LNER’s 5 second intervals over the 125mph using my spreadsheet and calculations.

That means my data - which is all from the automatic notation - is consistent.

It will make no difference to the automatic notation of the two pens.

So even if there is an imperfection in the drive transmission and you were getting the paper passed through quicker or slower than it should be, it doesn’t have any effect on the pen’s movement and notation.

The thing you are looking for when using the automatic notations is that, because they are being recorded independently and simultaneously, is using in your distance time equation either a quarter mile (which is possible) or if you like the full mile (and thus less data points, not more).

Which means you measure the seconds between each quarter or full milepost on the roll. Even if the paper was going more quickly, or more slowly, through the rollers, it doesn’t change the price of fish: you will still get a notation for seconds between each quarter milepost that is consistent to both pens.

The difficulty is in how you measure the exact seconds on the graph paper and for that I have used digital measuring tools when recording how many seconds I think each quarter and full mile have been achieved. Because as the train speeds up, as Mr Andrews observed, the seconds pen markings stretch out due to the speed of the paper going through the rollers.

As long as the pen is recording one seconds accurately, and the ninth wheel and the distance pen are recording accurately, the two independent recordings can be taken verbatim and you can then calculate the instantaneous speeds as I have outlined.

Now, I fully accept that the measurements made digitally will be open to interpretation. However, from my side, this has been peer reviewed, and my supervisor has seen the working out too: I am satisfied with my approach, I intend to publish it as part of the PhD.

And again, I have shown my working repeatedly in this thread and the Gresley thread. I have invited Mr Andrews to make available his dataset he has created so we can see his working out.

As for the mileposts notated - they are useful for observing the gradients. I have separately calculated the speeds using the milepost data in addition to the automatic notations and I have observed that there is no regular variation “per mile” - because the only way you get that is if there is a set distance per mile on the roll.

Which, as we have established, there isn't! The dynamometer car was not originally designed to be going at such speeds, in my view it was at the absolute limit of its ability in the 120mph speeds.

The ninth wheel driving the paper in my view only showed any significant change in length of the paper at high speeds.

That should be enough to explain my position, I think?

 

Thank you for this very clear exposition. As a layman, the phrase that leaps out at me is:

In the context of your subsequent comment about what the dynamometer car was designed to do, I have then to ask about the reliability of the calibration of the equipment for this extreme usage.

We can't revisit precisely how the calibration worked some 86 years ago, but we can interrogate this roll (and possibly others?) to assess the consistency across speed ranges. Do these support the case for consistency?

 

We could, if we had other rolls where we could see similar speed ranges. But we don’t, so we can’t.

There’s a law of diminishing returns I feel in going to the nth degree of accuracy in terms of “how accurate was the dynamometer car at recording distance and time”.

In 1938 it was accepted by everyone - literally everyone - including other railways’ CMEs in Britain that the efficacy of the Dynamometer car’s recording was not in doubt.

We know it was regularly calibrated - everyone to a man within the LNER both contemporary, prior to the run, and after who were involved have never once doubted the accuracy of the instrumentation, only their interpretations of the data it gave out. The LNER dynamometer car was calibrated against the Midland Railway equivalent, only for the latter to be found to be lacking in calibration! I have related this on this thread before.

To me, after speaking with other researchers who have seen this, and discussing this at length, we’re satisfied that the dynamometer car was in good order on 3 July 1938, and therefore the question of whether it accurately recorded the data can be dispensed with, by way of the caveat that we have no way of proving otherwise.

This is part of the “assumptions” regarding speed records and will be fully reported in that section of my PhD when I go to publish in a few years time.

 

Thank you - the clarification of what is known and not known is appreciated.

 

It is not a roll problem but machinery
There are more possible ways to distort evaluation .

The red Gresley 1938 curve has one per rev variations of speed plus minus .6%.
The paper drum has a radius of 97.02mm.
If centerline was/is 0.6mm eccentric we need look no further.
It can be tested with a dial indicator.
But can of course have been repaired since 1938
The snail-cogged wheel drive can be suspect but have the habit of growing better with use.
A little but not much more difficult to test.

 

I never thought a thread could get more tedious than the WSR one.

Still, I just have to learn to avoid it.

 

Yes please, but we're not quite there yet.

OK, I had misremembered about how the quarter mile marks are made.

However, for those to be made when the wheel has completed a specific amount of rotations that must be by means of the same drive train (possibly a different part of it) that moves the paper. They are thus potentially subject to any random or systematic error in that drive train, such as the putative once-per-mile variation.

One check that should be easy to make is whether those quarter-mile marks are spaced evenly on the paper. If they are not, that proves a flaw somewhere in the drive train. If they are spaced evenly, that could mean that the drive train was perfect or it could mean that the same flaw affected both.
Edit: quotation sorted out.

 

MellishR, with respect I am going to refer back to my original statement. It explains why.

And again, no, you won’t find each quarter mile measuring exactly same.

That was my original starting point when I was looking at the roll, and I very quickly realised that the length between each quarter mile stamp, as with the length of the seconds on the graph paper, varies with the speed and changes as the speed increases.

The general rule might have been that the dynamometer normally gives a set amount of paper per mile in a run, given the setup (but in Mallard’s case, in order to get better results the usual one foot of paper allocated was changed to two foot). But that would be with the dynamometer car travelling at much lower speeds than was found in Mallard’s (and Flying Scotsman/Papyrus’) exceptional high speed runs, which for my money are why we see the variations in paper length at the higher speeds, but even that is immaterial given the notations of speed and distance recorded simultaneously are what actually matter to make the calculations required.

Don’t worry, I am returning to the ether now!

I suggest anyone seriously interested in discussing this further with me gets in contact with me directly through the contact details in my signature. Happy to do so.