|
|
Post by Jim Scott on Mar 3, 2015 23:43:25 GMT
Julian, What is the approximate load carried by each spring and what are the dimensions of the steel leaves?? Also, do you know the designed deflection under working load? Thanks Mike Hi Mike A LBSC Terrier in 5" gauge is a very small model and Martin Evans' 'Boxhill' as designed has an operating weight of just about 60 lbs. I'm anticipating being able to get 'Earlswood' up to about 75 lbs which, distributed equally over six driving wheels, gives a loading of a little over 12lbs per wheel. The deflection required at this load is that required to bring the axle centre height to that of the designed working height of the valvegear, in this case a distance of 1/8". The axleboxes have a further 1/8" clearance before 'full bump' although I doubt if any instantaneous loadings would result in an increase of even 1/16". Screw adjusters are fitted to allow for initial adjustment during testing. The prototype has 13 leaves, which when scaled is 12 leaves each 1/32" thick x 0.350" wide plus one short 0.040" base leaf. The master (top) leaf has its holes spaced at 3.188" and is backed by a full length support leaf, both leaves being hardened and tempered spring steel. The following 10 leaves were made from heat formed Tufnol with the bottom spreader 0.040" spring steel. The whole stack is fixed in the buckle by a 1/16" dia screwed pin passing through all the leaves. I arrived at the above spring makeup after deflection measurements with different combinations of steel and Tufnol. Some folk have used used spring steel in its hardened and tempered state, but this is hard to work with ordinary HSS tooling. Also other types of 'springy' steel such as parcel strapping or hacksaw blades have met with some success. Further, it seems that in these quite small springs the Tufnol makes very little difference to the overall rate (stiffness) and is only required to hold its shape. There is the option of swapping Tufnol for spring steel should I need to stiffen it up further. Let no-one underestimate the amount of repetitive work required to produce six sets of leaf springs...! Cheers Jim S |
|
|
|
Post by miketaylor on Mar 4, 2015 13:32:22 GMT
Thanks for the data Jim,
I asked because I have been roughing out some springs for the Garratt I am designing. 0-8-0 + 0-8-0.
I have assumed 20 kilos per axle - say 12 lbs per wheel.
With the basic spring 4 inches long and using 3/8 wide springs of 20 gauge - .036 inches - I calculate that the deflection under 12 lbs load should be 0.122 inches for a spring with 8 leaves, all steel.
This seems very heavy compared with the steel tufnol composites to which I keep seeing reference.
The effect of the tufnol is small compared to the steel in terms of material stiffness, except, I suspect that it may have some effect in a case like yours by broadening the central support length for those two steel elements and thereby, effectively, shortening the spring. This could be significant as the deflection is proportional to the cube of the length.
Mike
PS should we perhaps move this discussion to the Leaf Spring thread, rather than cluttering up Julian's Boxhill.
|
|
|
|
Post by miketaylor on Mar 4, 2015 15:02:59 GMT
That post made me look at my calculations again.
Nothing wrong with them EXCEPT the load at 20 kilos per axle is 22 lbs per wheel - not 12. Which makes the calculated deflection nearly 1/4 inch!!
It also shoves the maxm. steel stress up to around 68,000 psi which is too high.
Thicker leaves needed. According to the calculations. So why all this tufnol??
Mike
Ignore thia lot. My original 12 lbs is correct - see my next post.
|
|
|
|
Post by Deleted on Mar 4, 2015 15:21:42 GMT
That post made me look at my calculations again. Nothing wrong with them EXCEPT the load at 20 kilos per axle is 22 lbs per wheel - not 12. Which makes the calculated deflection nearly 1/4 inch!! It also shoves the maxm. steel stress up to around 68,000 psi which is too high. Thicker leaves needed. According to the calculations. So why all this tufnol?? Mike Hi Mike I haven't properly calculated what each axle should support on my loco yet ( need to subtract tender weight from the full designed weight first) but thought I'd reply in regard to ' why all the tufnol' My loco also has 1/4" deflection and total weight with tender is 230lb....I'm guessing that the tender weighs 30-40 lbs but not checked yet so lets say the loco weighs 190lb over 6 axles..that is 31lb per axle. I tested what weight i needed to compress the trailing axle springs down to their correct riding position...this took iirc 106lb that was all sitting over the trailing axle....ok take some off for the bogie axles at the other end ( no mains yet) I reckon that it still took far more than the 31lb to get the springs down level which is where they should be on an A1. iirc there are 11 leafs and just under half are tufnol...I suspect more will need to be replaced with tufnol when the loco is complete. My point is if all steel they would be far to strong for the weight involved, I doubt that I could compress mine if all steel, it's hard enough as they are. Regards Pete |
|
|
|
Post by miketaylor on Mar 4, 2015 16:03:18 GMT
Hi Pete,
Spring dimensions are a highly critical part of any comparison.
Deflection goes up proportional to the cube of the spring length and it goes down in proportion to the cube of the leaf thickness, so quite small changes in spring dimensions can have a large effect on deflection. Eg, add 20% to the length and take 20% off the thickness and theoretical deflection will go up by a factor of 3.4.
I had to look up the basic calculations and lo and behold - I recognised them from 50 odd years ago at uni.
By the way - my previous post was wrong and my original 12 lbs was right. The way I was doing the calculation I needed to consider the load on one end of the spring only - not the wheel load.
In practice, at scale sizes especially, I have no doubt that trial and error is the best approach. Calculation will give a start but then one needs to make a trial spring and check the deflection characteristics and then modify as necessary. One thing to look out for may be consistency of material thickness - as noted above, a small change in thickness can have significant effects.
Mike
|
|
|
|
Post by miketaylor on Mar 4, 2015 16:22:58 GMT
Jim, The calculated deflection for your 2 full length leaves, ignoring any effect of the tufnol comes out to 0.15 inches. So, quite good agreement with your measured 1/8 inch. The loadspreading effect of the tufnol and the steel bottom leaf would definite;ly cause some reduction over the calculated deflection. Six springs eh? When I think that my monster will require 16 the heart sinks. For anyone wishing to pursue the issue further, the basic calculations are given here: www.engineersedge.com/material_science/leaf-spring-design.htm. Mike |
|
|
|
Post by Jim Scott on Mar 4, 2015 22:28:19 GMT
.......... In practice, at scale sizes especially, I have no doubt that trial and error is the best approach. Calculation will give a start but then one needs to make a trial spring and check the deflection characteristics and then modify as necessary. One thing to look out for may be consistency of material thickness - as noted above, a small change in thickness can have significant effects. ........... Hi Mike I agree entirely with this. The calculations get you to the ballpark figure but you need to test various spring combinations until a satisfactory outcome is achieved. The prototype spring should be made from the steel batch to be used and subject to a heat treatment procedure which is repeatable and consistent. Great care must be taken during the heat treatment as its all too easy to to get this wrong, especially with thinner leaf sections. Don't ask how I know... I'm conscious of the fact that we're drifting away from the subject of Julian's thread by discussing general leaf spring construction but if you want to start a thread re your Garrett you might get some good suggestions for mass production methods. I found making six sets of 13 leaves very time consuming and a bit tedious at times, just the thought of making 16 sets...  Cheers Jim S |
|
jma1009
Elder Statesman
Posts: 5,922
|
Post by jma1009 on Mar 4, 2015 23:07:07 GMT
i am very grateful to Jim and Mike for their contributions here as i havent made my springs yet!
as Jim knows ive a couple of locos with spring leaves made out of old hacksaw blades which i have found work very well and arent quite so stiff as hardened and tempered spring steel. grinding down the blades is a bit of a pain and done on a sunny day with a stiff breeze blowing down wind and the bench grinder set up in the garden on a portable stand/bench, and when i havent got anything better to do!
cheers,
julian
|
|
|
|
Post by Jim Scott on Mar 7, 2015 13:25:43 GMT
......... The calculated deflection for your 2 full length leaves, ignoring any effect of the tufnol comes out to 0.15 inches. So, quite good agreement with your measured 1/8 inch. The loadspreading effect of the tufnol and the steel bottom leaf would definite;ly cause some reduction over the calculated deflection. ......... Hi Mike Just a few more observations before putting the subject of Stepney’ springing back in Julian’s hands.... I did a deflection test on each of my spring sets at the working load of 12 lbs resulting in the following values: a) 0.150” b) 0.143” c) 0.0141” d) 0.140” e) 0.140” f) 0.128” My ‘calibrated weight’ was made up of various bits of heavy metal to give 12 lbs, which is probably the top end of the individual wheel loading I’m going to achieve. The most accurately made spring is the one with least deflection (f), having both steel leaves perfectly matched and in full contact. The others have the top leaf at a slightly smaller radius (occurred during heat treatment) and so the top leaf has to ‘take up a bit of slack’ before the combined effect of both leaves are felt. In practice, these figures don’t mean a lot, most folk will set their springing by trial and error anyway and designs usually have provision for a small amount of adjustment. I for one have not a lot of experience of just how much movement actually occurs above and below the mean axle height whilst the loco is operating but I have it in mind its probably not more than +/- 1/16” on a reasonably good track? I also did an overload test on spring b) to simulate a momentary load transfer of about 25% to the front or rear axle. With 3 lbs added to give a load of 15 lbs, deflection increased from 0.0143” by 0.034” to 0.177”. Also, to cover the case of the loco weight falling on two axles only, as might occur say when unloading from a car, I loaded the spring to 18 lbs which added 0.067” above the mean ie 0.210”. There is about 1/8" space above the axleboxes when normally loaded so not much danger of hitting the top of the horn slot. In all these tests the dial test indicator returned to zero after removal of the load, showing that no further ‘set’ was being added to the spring. Guess I had better get on with the adjustors now instead of just thinking about them... Cheers Jim S ps you have a pm. Photo showing test of spring b), 12lb load and 0.143" deflection  |
|
|
|
Post by miketaylor on Mar 7, 2015 15:32:42 GMT
Just to say thanks for posting that Jim.
Always nice to see a reasonable correlation between theory and practice.
Now we really had better let Boxhill have his post back!!
Mike
|
|
|
|
Post by Deleted on May 12, 2015 9:50:25 GMT
Hello all, what a great resource for this fine Loco -
First post for me - building 2 off 5 inch Boxhills and quite early in the process - approx 1 year! however I thought I was doing well with the crank axle until I saw the comments that the eccentric throw was too large (0.375") as designed by ME, so what I am looking for is the correct throw for the standard expansion link, the reason I ask is that the straps are done and webs machined etc but it's not pressed together yet so as much as I hate the idea - I would prefer to re make the eccentrics (all 8!) as I have the water cut expansion links (4 off) at approx £80 just waiting for when I get to that point
Alternatively are we saying that by using redesigned expansion links and using the using 0.375" throw it's an improvement over a smaller throw and standard expansion link... hope that makes sense...
Great to hear back from anyone who can advise.
Thanks
Dilip.
|
|
|
|
Post by Deleted on May 12, 2015 10:35:48 GMT
I built a complete Boxhill chassis to the original Martin Evans drawings .
It actually functioned quite well but had a problem common with Martin Evans designs in that at certain positions of the cranks it was impossible to move the reversing lever - the expansion links were over at such a steep angle that they would not slide over the die blocks .
There is certainly plenty of scope for improving the valve gear - undoing some of Martin Evans' improvements would be a start !
The full size engine drawings show a very neat well engineered arrangement of cylinders and valve gear and Martin Evans could just have adapted this .
In answer to OP question I would certainly advise modifying the valve gear and if that means new eccentrics then that would be worth doing .
The amount of improvement possible in valve gear design may be limited by the Boxhill arrangement of valves and ports and the required valve travel . Ideally these details should be modified as well to get a really good result .
Bear in mind though that plenty of people have built Boxhill's straight off the Martin Evans' drawings and got them to run . How far you want to go with improvements to the basic design is really just your personal choice .
|
|
|
|
Post by Deleted on May 12, 2015 13:42:14 GMT
Thanks Michaelw, yes this seems a common comment about prototype vs scale - I saw a really nice terrier at the Harrogate show on Sunday on the TSMEE stand and was really impressed with the details/workmanship. very inspiring stuff.
Thanks,
Dilip.
|
|
|
|
Post by Jim Scott on May 12, 2015 20:36:50 GMT
Hi Dilip
I have sent you an email directing you to this thread - but I see you are already here..!
This thread contains a wealth of information from a number of contributors and I'm sure you will find answers to most of your questions. I would suggest you start at the beginning and enjoy the up's and down's of Julian's excellent build.
'Boxhill' as drawn by Martin Evans is a 50 year old design and really in need some updating, especially in the boiler and valvegear areas. Also, with not a lot of effort it can be made quite close to scale. If you haven't fitted your eccentrics yet it is worth sorting out the throw which, as Michael notes, produces too much link swing leading to difficulties with the reverser at certain wheel positions. Don Ashton kindly sorted out Julian's gear, mine has the as-designed dimensions (alas already built) but optimised via Prof Hall's simulator.
Also, I would advise that you choose a prototype and stick with it. There are a lot of variations between locos, the greatest being the rebuilding to A1X, including many different colour schemes (not just the traditional Stroudley yellow ochre). Martin Evans 'Boxhill' as drawn is a 'Bitza' but many will point out that 'Boxhill' at the NRM York is not completely original either...
Good luck with your two Terriers...
Cheers
Jim S
|
|
jma1009
Elder Statesman
Posts: 5,922
|
Post by jma1009 on May 12, 2015 21:27:20 GMT
hi Dilip,
i am not sure i can add much to Jim's excellent resume above, and you have obviously examined Jim's own excellent 5"g Terrier at Harrogate last weekend.
Don Ashton very kindly re-designed the gear for me using the original eccentric throw which both Jim and i have of 0.375". Don also put the original martin evans' designed gear through it's paces on his computer simulator and the results were far from good.
Jim has achieved very good valve events by adding a 60 thou suspension offset on the suspension brackets on the expansion links where they connect with the bottom of the lifting links. martin evans showed no such offset.
there is also an error in the length of the lifting arms and their length on the martin evans' design.
my Don Ashton valvegear is a bit different but looks the same! the expansion links are longer, the length of the eccentric rods corrected for equal leads, the lifting links and arms are 'correct', and the suspension offset is a bit more than Jim's. the valve lap is a bit more.
i have also incorporated a crossbar through the ends of the lifting arms, and no forked end to the weighshaft arm, as per fullsize. both Jim and myself have made scale Stroudley type reversers fixed to the left hand cab box again as per fullsize.
my valve rod spindles are closer together in the steam chest and mine has the correct Stroudley channel section slidebars.
aside from the valve gear but allied to it would be to incorporate Jim's bolt on valve rod gland assembly which i can strongly recommend.
whether you reduce the eccentric throw will rather depend on how far you are advanced with everything else. some adjustment may be required to the ports and spacings but this is not something ive investigated as in my case it wasnt necessary, thanks to Don Ashton's help.
if i can be of any further help please let me know.
cheers,
julian
|
|
|
|
Post by Deleted on May 12, 2015 22:32:40 GMT
I've often wondered why the Terriers were fitted with condensing gear . With those tiny tanks it could never have been used on continuous runs of more than a few minutes and the engines were not used much on the underground system except on through services and at interchange stations .
Be nice to see a rake of Stroudly coaches behind a Terrier - either a close coupled rake of very short suburban coaches or a single 'Balloon' coach .
For anyone not familiar the 'Balloon' coaches were big and dwarfed the tiny Terriers completely . Apparently a very lively combination though !
I've got lot's of Stroudley drawings from the old OPC . My most favorite LBSCR engine of all is the smaller wheel 2-2-2 tender engine . I've certainly got the drawings for that - be nice to build one .
One drawing in my collection shows an engine even smaller than a Terrier . It's part basic sketch and part detailed . Probably a tentative drawing for an engine that was never built . You can see the Terrier design evolving in it though - just doesn't quite get there .
|
|
jma1009
Elder Statesman
Posts: 5,922
|
Post by jma1009 on May 12, 2015 22:57:21 GMT
hi michael,
one of these days i will build a miniature version of Stroudley single wheeler Grosvenor or Abergavenny.
all Stroudley locos had 'condensing gear'. there are very detailed records of the terriers and coal consumption from 1872 onwards - see the RCTS book and 'Engineer' in 1873. as a result the side tanks had a double layer, and injectors couldnt be used hence the use of crosshead pumps.
by what ever standard you use the terriers packed a big punch into a small loco and were extremely efficient plus powerful for their size and were probably even better when converted to A1X with Marsh boilers, as well as being aesthetically perfect like all Stroudley locos in most people's eyes. pre-1920 before the GWR publicity machine cranked up there was a considerable 'Brighton' following.
the IOWSR Havenstreet terriers are particularly well looked after and finely tuned and highly regarded. unfortunately the Bluebell examples are both now knackered and were never given the attention in the 'shops' that they deserved.
cheers,
julian
|
|
jma1009
Elder Statesman
Posts: 5,922
|
Post by jma1009 on Aug 23, 2015 22:31:25 GMT
after many many months ive started work again on my 5"g version of Terrier Stepney. a new central heating system and redecorating 5 rooms in the house and kitchen work have got in the way! ive been after some accurate angle to make the running board valances, but have failed to find any! the valances contribute quite a bit to supporting the weight of the side tanks as there are precious few brackets to support the running boards. anyway ive cut out with a hacksaw the valances from 1/16" steel sheet. here are the results so far. im not quite finished with them, and quite a lot more filing and profiling required to get them the correct depth, and shape on the ends. plus attaching them to something to fix to the running boards. ideally i want to have them removable for painting as painted 'claret' with intricate lining.   in fullsize the depth of the valances (apart from the ends) is half the depth of the buffer beams. this isnt helped by martin evans' buffer beams for Boxhill being too deep. mine are a bit more to scale, and Jim Scott's 5"g Terrier Earlswood has them absolutely correct. my buffer beams are 1/8" thick, but i milled the ends to give the scale buffer beam thickness protruding on the ends of the valances. once the valances are completed i can make a start on the side tanks. ive already cut up all the brass plate required, so hopefully some more progress soon. cheers, julian |
|
|
|
Post by runner42 on Aug 23, 2015 23:21:02 GMT
Hi Julian,
very pleased that your back to completing Boxhill, you must have accumulated a lot of brownie points with your nearest and dearest for the house renovations you have undertaken. I suppose it's a good thing that you can take time out to involve yourself with other than ME work, because the time off will renew your enthusiasm to complete a very fine model.
Brian |
|
|
|
Post by Deleted on Aug 24, 2015 7:27:45 GMT
Good to see Stepney back on track Julian.....pun fully intended....  looking very nice sir... Pete |
|
looking very nice sir...