Julian Roberts wrote:Going back a bit Will, your post "The Final Factor" on Aug 27 835pm showed that because we can put an unprototypical amount of end float on centre axles no gauge widening is needed at all in most circumstances.
On this analysis I am unclear as to what the P4 reasoning is for any triangular gauge at all?! - rather than using anything other than a plain gauge that simply gives 18.83 regardless of curvature, at least down to what is normally regarded as a minimum, around 3' 6"?
My opinion, for what it’s worth, is that nobody thought too hard about it at all. Even the 3'6" minimum is really just a bit of "is P4 really practicable" propaganda.
When Scalefour was new, the ethos of the time was that following prototype practice was the way to go. I’m pretty sure I remember that there were originally at set of gauges which implemented the prototype stepped approach. I see no sign anybody checked the maths and, as it all worked when done well, why would you?
When I was producing 00 track in the 70's roller gauges were the norm, despite the fact that it is quite hard to roll one down a bit of accurately completed track, let alone a bad bit. Once somebody shows you a three point gauge you realise it is definitely an easier tool to use, as it is much better at holding the rail in the right place while you solder it, you just needed to remember to use it the right way round on corners. As a bonus it automatically produced gauge widening which we needed, says Martin’s quote from the rule book.
Use of the society 3 point gauge produced track which most of us found worked well, assuming you had the patience to debug it and the stock you tried to run on it. We didn't get a gauge that was good for sliding round the track to check for tight spots*
until the mint gauge turned up relatively recently. It was not what it was designed for but it did do the job, and you did have a point when you suggested that in this role it would be smart to make it match the 3 point gauge.(*Adding gauge widening may be optional but tight spots are definitely a no no.)
My point is that this is an evolved situation not a planned one, developed on the basis of what the prototype did, and regardless of its true applicability to P4 models with their, in this regard, significant differences from the prototype. (I now need a soap and water mouth wash.) Then new methods were adopted when they proved their worth practically.
It is also true that the full story still isn't here. The impact on the problem of the relationship between the wheel and rail profile is still questionable, as is the implications of the steering effect of the coned wheel. I am also aware that, while I can see the rationale behind Malcolm Cross's 3 degrees limit on the angle of incidence and I'm happy it's about right, I would have thought the exact figure would vary depending on the size of the wheel as well as the flange and rail profile at Martin suggested. I haven't the foggiest idea as to how you might approach calculating the answer. I wonder if there is anything in the prototype literature on this (Keith/Martin?).
The other thing that I don't really get
is how to use the graph that Keith posted, to work
out where you begin to have a radius that is too tight. Is it possible to show those of us less
mathematically inclined how to use the Malcolm Cross graph - perhaps taking Noel's LNER CCT
van with its long wheelbase as an example
I spent a happy hour this morning working out how best to explain this, and given the graph is a bit indistinct, if it would be worth trying to reproduce it from my spreadsheet. The versine/curve radius/vehicle wheelbase falls strait out, the interesting bit is adding the angle of incidence lines and I would be more confident if I had a better understanding of the angle of incidence/wheel size issue.
Anyway Keith got there first. I can go over the thing slower and in greater detail if you like, but would it now be worth it?