RMweb Gold martin_wynne Posted August 29, 2015 RMweb Gold Share Posted August 29, 2015 Hi Andy, In conducting your experiments I feel the main effect of eliminating the coning is going to be the reduction in differential effects on sharp model curves. Even if the coning is insufficient to provide a full differential action, it must help a bit. Rather than simply seeing whether they derail, you will need to measure train resistance on say a loco + 8 carriages running round say a 30" curve, with and without any coning on all the wheels. Also I believe coning must help with electrical pickup. Reducing the contact area increases the contact pressure, which is more important than area in the environment of a dirty rolling contact. Of course, this does rather depend on whether the rails are vertical or inclined. regards, Martin. Link to post Share on other sites More sharing options...
AndyID Posted August 30, 2015 Author Share Posted August 30, 2015 Hi Andy, In conducting your experiments I feel the main effect of eliminating the coning is going to be the reduction in differential effects on sharp model curves. Even if the coning is insufficient to provide a full differential action, it must help a bit. Rather than simply seeing whether they derail, you will need to measure train resistance on say a loco + 8 carriages running round say a 30" curve, with and without any coning on all the wheels. Also I believe coning must help with electrical pickup. Reducing the contact area increases the contact pressure, which is more important than area in the environment of a dirty rolling contact. Of course, this does rather depend on whether the rails are vertical or inclined. regards, Martin. Oh no! I'll have to build a scale dynamometer coach now Hi Martin It might be simpler to build a variable curved incline and determine the rolling gradient for various tapers on a single vehicle. The tricky bit will be keeping everything else constant otherwise the results could be misleading. Andy Link to post Share on other sites More sharing options...
RMweb Gold martin_wynne Posted August 30, 2015 RMweb Gold Share Posted August 30, 2015 It might be simpler to build a variable curved incline and determine the rolling gradient for various tapers on a single vehicle. The tricky bit will be keeping everything else constant otherwise the results could be misleading. Hi Andy, The results of that would be interesting. But a curved gradient introduces a track twist, which then adds all the variables of vehicle suspension into the mix. Also, rolling freely under its own weight does not allow for the effects of speed, or the buffering and coupling effects of a heavy train. Nor for the difference between propelling such a train (with slack couplings and the flanges hard against the outer rail) and pulling it (with tight couplings and the possibility of string-lining across the curve at low speeds). Have fun. regards, Martin. Link to post Share on other sites More sharing options...
AndyID Posted August 30, 2015 Author Share Posted August 30, 2015 .....on the other hand we might try something like this - This one has three degree tapers each side and it happily rolls down a very slight incline without falling off the track. It swings from side to side on straight track as it goes along, proving that the self-steering is working as it should. However, the amplitude of the lateral motion is quite large. I also got the impression that it increases at lower speeds, but don't hold me to that. It is sufficiently large that, if there were flanges, they would encounter the rails before the self-steering corrected the direction. It's probably more meaningful to make a very light-weight dummy truck chassis that supports two heavy "wheelsets". The chassis could be tested with various different wheel profiles. The idea with this arrangement is to minimize axle bearing friction as much as possible by putting the mass in the "wheels" rather than in the chassis. EDIT: Our posts crossed in the Azores. Link to post Share on other sites More sharing options...
RMweb Premium 47137 Posted August 30, 2015 RMweb Premium Share Posted August 30, 2015 In conducting your experiments I feel the main effect of eliminating the coning is going to be the reduction in differential effects on sharp model curves. Even if the coning is insufficient to provide a full differential action, it must help a bit. To everyone: I have set up an Excel spreadsheet to help find out about the wheel slip on curves with different angles of coning and different wheel and track dimensions: differential.xls I chose a flat curve with a constant radius passing through ninety degrees to give a reasonably long distance of travel. The inputs are the cells shaded green - I have put some notional values for a layout in a corner of a room. I have assumed the train is travelling fast enough to push the flange of its outer wheel against the outer rail. Playing around with the values I found a nominal 11 mm diameter wheel gets virtually no slip with these dimensions. Alternatively, a 12 mm wheel needs a 3.55 degree cone angle. But overall, the slip figures stay very small on a 800 mm curve. Pondering the use of metric measurements, I think the display of three decimal places is reasonable for the intermediate calculations here, it helps you see what is going on. - Richard. Link to post Share on other sites More sharing options...
Grovenor Posted August 30, 2015 Share Posted August 30, 2015 I think using a useable tread width of 2mm is way to high for 00 models, the useable tread width should be the same as the amount of sideplay between wheelset and track which I would expect to be under 1mm. Regards Link to post Share on other sites More sharing options...
RMweb Gold martin_wynne Posted August 30, 2015 RMweb Gold Share Posted August 30, 2015 I think using a useable tread width of 2mm is way to high for 00 models, the useable tread width should be the same as the amount of sideplay between wheelset and track which I would expect to be under 1mm. For RTR wheels at 14.4mm back-to-back, the sideplay is about 0.2mm on 16.2mm gauge and about 0.5mm on 16.5mm gauge. For typical kit wheels at 14.6mm back-to-back, the sideplay is about 0.4mm on 16.2mm gauge and about 0.7mm on 16.5mm gauge. "about" because it is affected by the flange profile, and the rail section, and the weather forecast. Martin. Link to post Share on other sites More sharing options...
davefrk Posted August 30, 2015 Share Posted August 30, 2015 If I remember correctly Newcastle Science Museum have a test track which shows in practical terms how railway wheels work. The twisting/ curving track is maybe 300mm gauge and the wheelsets are hard foam. The only wheelset I could make go down the track without derailling was the coned wheelset which were WITHOUT flanges. The wheelset with 'square' flanges and parallel tread would jump about, fall off or jam in the gauge. I believe that the coning does help in the smaller gauges by reducing the friction on curves, at the right speed and curvature the coning will steer the vehicle through without the flanges touching the rails at all, okay everything needs to be 'just right' for that to happen but even if the flanges are against the outer rail coning will help reduce the friction. With a long heavy train at slower speeds though the first part of the train would be 'string lining' on a sharp curve as said previously, this can cause more friction as the wheels are hard against the inner rail and also fighting the coning, the loco might spin out but, increase the speed a good bit and help the loco to get the train moving so that more of the wagons might be against the outer rail due to the forces, the train will go round and round no problem. I did just this recently with a 117 wagon train on Wharfeside. The loco started the train on the gentle curves but got stuck when the front part the train was round the 6 foot curves into the fiddleyard, turned up the juice so the DJH Black five was flying and the train continued to circulate at what Andy R would call 100% reliability. Decent quality wheels on well built track both of which match in spec. will always work, doesn't matter what gauge it is. Just my two pennies worth. Dave Franks. Link to post Share on other sites More sharing options...
raymw Posted August 30, 2015 Share Posted August 30, 2015 Hi AndyID, Thanks for trying. One of my previous posts alluded to the same principle as a flat belt drive. I wonder if your roller would run better with a convex curve (wooden barrel shape) instead of the two straight cones. A lighter material would give more wobble, if you like. It is one of the not scalable nature situations. You can blame digital photography for all this mental stress. Now, when the wobble has been reduced sufficiently, by producing the correct profile for the roller, you can machine away the unneeded material, and then try again, (because the weight will be altered). Then Martin can design some 'point-work', so we can route it to where we want - high check rails on the outside looks about right. (nothing wrong with reinventing the wheel. Best wishes, Ray Link to post Share on other sites More sharing options...
RMweb Premium 47137 Posted August 30, 2015 RMweb Premium Share Posted August 30, 2015 I think using a useable tread width of 2mm is way to high for 00 models, the useable tread width should be the same as the amount of sideplay between wheelset and track which I would expect to be under 1mm. For RTR wheels at 14.4mm back-to-back, the sideplay is about 0.2mm on 16.2mm gauge and about 0.5mm on 16.5mm gauge. For typical kit wheels at 14.6mm back-to-back, the sideplay is about 0.4mm on 16.2mm gauge and about 0.7mm on 16.5mm gauge. "about" because it is affected by the flange profile, and the rail section, and the weather forecast. Sorry - the figure of 2mm was very wrong - but the spreadsheet does let people fiddle with numbers and see when coning does have a differential effect so I think the post has some value. My own feeling (from cleaning wheels) is that coning does keep model wheels "on course". The model trains are proportionally much lighter than the real thing and I would like to imagine the coning increases the margin before a derailment. Which ties in with the Newcastle Science Museum too. - Richard. Link to post Share on other sites More sharing options...
AndyID Posted August 31, 2015 Author Share Posted August 31, 2015 Hi Richard, Thanks for the spreadsheet. Assuming a maximum 0.4 mm usable tread width, for a 3 degree cone angle the minimum "no slip" turn radius seems to be a bit greater than 14 feet. I suspect that's not far off the scaled number from the prototype. Assuming that is correct, there has to be a considerable amount of differential slip even with generous model railway curves. In model terms a 14 foot radius is often a pretty good approximation for a straight line Although it might make an entertaining party trick, I realized my double-cone experiment doesn't really tell us much other than demonstrating the undesirable hunting effect. An effective model must have constrained axles which I think means it ends up looking like some sort of four wheel truck with axle bearings. The tricky bit is minimizing bearing friction to prevent it swamping wheel slip friction. I'm thinking of something modeled along the lines of Fred Flintstone's car. It also occurred me that it might be fairly easy to quantify how much friction "no cone" wheels produce by allowing free rotation between the wheels on an axle and comparing the result with the case where both wheels are locked on the axle. Cheers! Andy Link to post Share on other sites More sharing options...
Andy Reichert Posted August 31, 2015 Share Posted August 31, 2015 Sorry - the figure of 2mm was very wrong - but the spreadsheet does let people fiddle with numbers and see when coning does have a differential effect so I think the post has some value. My own feeling (from cleaning wheels) is that coning does keep model wheels "on course". The model trains are proportionally much lighter than the real thing and I would like to imagine the coning increases the margin before a derailment. Which ties in with the Newcastle Science Museum too. - Richard. With respect, I'm sorry but I'm getting very different results working from first principles. Has anyone else checked the spreadsheet algorithms? Andy Link to post Share on other sites More sharing options...
RMweb Premium 47137 Posted August 31, 2015 RMweb Premium Share Posted August 31, 2015 With respect, I'm sorry but I'm getting very different results working from first principles. Has anyone else checked the spreadsheet algorithms? Andy Obviously I'm not "someone else", but there was a typo in the formulas for similar triangles. Here is a second spreadsheet with corrected formulas and also a diagram to help show what I am doing. differential.xls If you set the value of coning angle to 45 degrees you can see how the arithmetic is working. The spreadsheet makes two important assumptions and I'll declare these now in case they disagree with convention: 1 - the track is flat, with no super elevation, and the outer wheel rides up onto its larger diameter during cornering 2 - the wheel diameter is measured to the centre of its tread If the track has super elevation and the wheels slide down so the inner wheel is on its larger diameter, the results will be completely different. - Richard. Link to post Share on other sites More sharing options...
Andy Reichert Posted August 31, 2015 Share Posted August 31, 2015 Obviously I'm not "someone else", but there was a typo in the formulas for similar triangles. Here is a second spreadsheet with corrected formulas and also a diagram to help show what I am doing. differential.xls If you set the value of coning angle to 45 degrees you can see how the arithmetic is working. The spreadsheet makes two important assumptions and I'll declare these now in case they disagree with convention: 1 - the track is flat, with no super elevation, and the outer wheel rides up onto its larger diameter during cornering 2 - the wheel diameter is measured to the centre of its tread If the track has super elevation and the wheels slide down so the inner wheel is on its larger diameter, the results will be completely different. - Richard. Yes, now it agrees with mine, but I didn't double check Martin's figure for "slop" and assumed that it was the total L-R possible movement, not slop on each side. So if I double that, as you have, then we both show that the minimum radius for coning to actually work (0 slip) for standard RTR 00 is around 9500 mm, approx 30 ft. For 00-SF of course, the both ways "slop" with the same RTR wheel must be 0.3 mm less. In which case the minimum radius for coning to work rockets up to 37000 mm, or approx 120 ft. Interesting. . . . Andy Link to post Share on other sites More sharing options...
AndyID Posted August 31, 2015 Author Share Posted August 31, 2015 Yes, now it agrees with mine, but I didn't double check Martin's figure for "slop" and assumed that it was the total L-R possible movement, not slop on each side. So if I double that, as you have, then we both show that the minimum radius for coning to actually work (0 slip) for standard RTR 00 is around 9500 mm, approx 30 ft. For 00-SF of course, the both ways "slop" with the same RTR wheel must be 0.3 mm less. In which case the minimum radius for coning to work rockets up to 37000 mm, or approx 120 ft. Interesting. . . . Andy Yikes! I'll take a look later. Funny how the brain works. Last night I was thinking about a test rig to make things a lot more quantitative rather than qualitative, but I was not very happy with it. When I woke up this morning I came up with a far simpler (and probably more consistent) test rig, but you'll have to wait till later to find out 'cos I have to run into town now Link to post Share on other sites More sharing options...
RMweb Premium 47137 Posted August 31, 2015 RMweb Premium Share Posted August 31, 2015 Yes, now it agrees with mine, but I didn't double check Martin's figure for "slop" and assumed that it was the total L-R possible movement, not slop on each side. So if I double that, as you have, then we both show that the minimum radius for coning to actually work (0 slip) for standard RTR 00 is around 9500 mm, approx 30 ft. For 00-SF of course, the both ways "slop" with the same RTR wheel must be 0.3 mm less. In which case the minimum radius for coning to work rockets up to 37000 mm, or approx 120 ft. I have used TAN(RADIANS(angle)) * (thickness / 2) where "thickness" is equivalent to Martin's figure for slop - so I reckon I've halved the slop figure to get the movement for each direction, to either side of the centre line of the tread. Coning begins to provide a differential effect as soon as the track begins to curve, and while we can identify a sweet spot this is surely not the whole story. Somehow, we need to find out how much wheel slippage is acceptable on curves, and I haven't got a clue how to do this analytically. We might do experiments with curves on gradients and to keep this simple (easy to define) it might be a flat board with a circle of track (flat all round) tilted at an angle to find slippage on one side then slippage on both sides i.e. wheel spin (*). But I've got doubts whether this will find much - when I was experimenting before I built my "Shelf Island" layout, I found wheel spin was much more dependant on gradient than on curve radius. In essence, I got down to a gradient where everything stopped and slid downhill, but easing the gradient a little got everything working again. I ended up happy with 15-inch curves at 1:20 as my limit to build, knowing 1:18 would work and 1:17 would not, but this is for two-wagon trains. And it says nothing much about coning. And, my finished railway has some super elevation, which looks pleasant (beauty in the eye of the beholder here) but defies definition. - Richard. Edit: (*) what I'm trying to say here, is when wheel slip happens on one side, traction is dramatically reduced. Link to post Share on other sites More sharing options...
Andy Reichert Posted August 31, 2015 Share Posted August 31, 2015 I have used TAN(RADIANS(angle)) * (thickness / 2) where "thickness" is equivalent to Martin's figure for slop - so I reckon I've halved the slop figure to get the movement for each direction, to either side of the centre line of the tread. Coning begins to provide a differential effect as soon as the track begins to curve, and while we can identify a sweet spot this is surely not the whole story. Somehow, we need to find out how much wheel slippage is acceptable on curves, and I haven't got a clue how to do this analytically. We might do experiments with curves on gradients and to keep this simple (easy to define) it might be a flat board with a circle of track (flat all round) tilted at an angle to find slippage on one side then slippage on both sides i.e. wheel spin (*). But I've got doubts whether this will find much - when I was experimenting before I built my "Shelf Island" layout, I found wheel spin was much more dependant on gradient than on curve radius. In essence, I got down to a gradient where everything stopped and slid downhill, but easing the gradient a little got everything working again. I ended up happy with 15-inch curves at 1:20 as my limit to build, knowing 1:18 would work and 1:17 would not, but this is for two-wagon trains. And it says nothing much about coning. And, my finished railway has some super elevation, which looks pleasant (beauty in the eye of the beholder here) but defies definition. - Richard. Edit: (*) what I'm trying to say here, is when wheel slip happens on one side, traction is dramatically reduced. I think we are now in sync on the slop values. However, without any significant coning effect the model wheel will hit any model railway sized curve coming off a straight section by staying close to the same original straight line. And after that, will be solely guided by the flange against the outside rail. So I have to disagree about any proportional effect below the minimum working radius. Slip is just basically turned on or off. FWIW, my analysis of this many years past was aimed at estimating curve friction and whether model coning (or anything else) would prevent it. Interestingly, the extra distance travelled by the outer wheel (same diameter ) is simply the track gauge times the angle the (track) turns (in radians) . The track radius cancels out of the equation, so has no effect on the amount of slip. For confirmation, it's the same old adage about passing a string tight around the equator. Add 2 x"PI" inches to the string and it will float 1 inch above the equator all around the globe! In the case of 16.5 mm, ( and regardless of coning) the outer wheel has to travel (or slip) an extra 16.5 mm x PI/2 further than the inner wheel for a track turn thru 90 degrees. ( say 26 mm). Now PI cancels out of the equation as well if you convert that 26 mm as a wheel circumference to an equivalent one turn of a wheel diameter. (that gives an 8.5 mm wheel in this case) . So a 12 mm wheel will have to rotate 8.5/12 of a turn for each corner of a circular layout, regardless of the track radius. . I'm not up to estimating or easily measuring slipping (moving) friction. Because it's not generally linear to speed , etc. The extra load on the loco has to be found some other way. Andy Link to post Share on other sites More sharing options...
trustytrev Posted August 31, 2015 Share Posted August 31, 2015 Hello, Question."Do we need coned wheels?" Answer."Only if you are a conehead" https://en.wikipedia.org/wiki/Coneheads_%28film%29 trustytrev. Link to post Share on other sites More sharing options...
AndyID Posted August 31, 2015 Author Share Posted August 31, 2015 Hello, Question."Do we need coned wheels?" Answer."Only if you are a conehead" https://en.wikipedia.org/wiki/Coneheads_%28film%29 trustytrev. See post #116 Link to post Share on other sites More sharing options...
AndyID Posted September 1, 2015 Author Share Posted September 1, 2015 Rather than run the wheels down a curved incline, all we need to do is run wheels with different diameters on a common axle down a straight incline. The wheels should have no flanges which requires some sort of carriage to center the test wheels on the inclined track. Except that it's a waste of time. It's clear from the above that prototype cone angles don't help our models negotiate the sort of curvatures we have to live with. Some amount of slip friction is inevitable and coned wheels do not help much if they help at all. I'm not convinced slip friction is a real problem, but if we believe it is, we would be much better to allow relative rotation between the wheels on an axle. That is not very hard to do on model rolling stock although it would be much more difficult, but probably not impossible, to apply to the driven wheels on locomotives. Link to post Share on other sites More sharing options...
The Bigbee Line Posted September 1, 2015 Share Posted September 1, 2015 I read these posts regarding different track standards and wheel profiles with interest. Much of the maths goes over my head. My method is pure trial and error. If it works, stay with it. For my own comment I would point out that rail and wheels have a profile 'as new' and this will wear in various ways until either of them become unfit for purpose. For example: Rail can wear until the head is quite flat. Wheels can wear until the profile is like a pulley, with a false flange starting to form at the front of the wheel. A particular problem when the bogie steers too well. There is no requirement to check the detail of the profile when examining vehicles; the basic checks are to check the wear in the tread, by a measurement of the flange height. Then a check on the flange width. Other things to look out for are the result of metal migrating across the surface of the wheel (caused by the constant rolling of the wheel). Metakl rolling towards the front gives 'rollover' and metal rolling up the flange produces 'Flange Toe Radius Build Up'. TRBU in railway speak, that when exceeding certain limits will interact with worn point blades and cause flange climb. Interestingly, in europe they measure flange angle as that too can contribute to flange climb. Have I bored you all to sleep yet........ The point I'm making is, this is a hobby..... fun.... relaxation..... I get enough grief with real trains. I'm all for these discussions, but keep it relaxed please, not too much of the 'You don't want to do it like that, you want to do it like this...' My answer is "F*** O**". That is a very commonly used term amongst railwaymen, useful in most situations.. Link to post Share on other sites More sharing options...
AndyID Posted September 1, 2015 Author Share Posted September 1, 2015 Hi Ernie, Thanks for the information regarding the real stuff. Cone angles do a very important job by preventing rapid wheel wear on real railways. You've probably noticed that rather than wearing down, model wheels tend to "wear up" when they collect a nice solid layer of crud on the treads. If it's allowed to go too far, that can cause running problems. It can be quite difficult to remove too. I started this thread because there was a lot of debate in another thread about the effect the cone angle has on a wheel when it's running through a point frog. It struck me that there might be no need for the debate if there is really no need to cone model wheels in the first place. After a lot of messing around I've come to the firm conclusion that, while It might make them look more like the real thing, I really isn't doing much of anything. It would not surprise me if a lot of people think I'm completely bonkers and violently disagree with me. That's happened before Cheers! Andy Link to post Share on other sites More sharing options...
Ravenser Posted September 1, 2015 Share Posted September 1, 2015 ........ You've probably noticed that rather than wearing down, model wheels tend to "wear up" when they collect a nice solid layer of crud on the treads. If it's allowed to go too far, that can cause running problems. It can be quite difficult to remove too. ............. Cleaning wheels is normal good practice. Rolling stock wheels should not be allowed to get into the state you describe , and if they have, they should be cleaned up as a matter of urgency. (On the British exhibition circuit there are many layout owners/groups who make a practice of cleaning all wheels including rolling stock before every show) Quite apart from anything else the crud gets rolled back onto the rails where it interferes with current collection on the locos I know iyt's more difficult to tackle on plastic wheels - but that's another major reason why plastic wheels are a Bad Thing Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.