Balancing full size locomotive wheels

Whilst reading David Clarke's book on the Riddles Class 7 Standard Pacific ("Britannia" Class) he shows a close up of the balance weight on a main driving wheel. He points out that these were actually not solid, but merely pockets that were filled with molten lead to achieve the correct balance.

My question is how did they achieve this back in the day? Before accurate dynamic balancing machines using strain gauges and such, was it just a static balance (maybe balance it on a point and add lead till it sits level as some cheap home automotive balancers do?).

Thanks for any information, as you've probably guessed I don't have any full size loco wheels to balance but would love to know how it was done effectively! I guess it didn't have to be perfect as the speed of rotation was relatively low even for high speed locomotives with smaller wheels.

Jon.

A look at the recent ME 4586 may help, I think there was a thread about the article at the time.

Edited By JasonB on 20/07/2018 13:12:35

Before computers, wheels only had a static balance using a pair of horizontal blades standing on edge. The sharp edge of the blade has little rolling friction. Not perfect but good enough for slow rotaing shafts and wheels.

I remember that car wheels where balanced by placing on a jig with a pivot point in the centre and a spirit level bubble used to check balance.

Paul.

The real problem with these wheels was not a matter of getting the individual axle assembly in perfect balance. There was a further need for deliberate imbalance of the axle to counteract the effect of the to and fro (but not up and down) of the pistons, con rods, and valve gear.

Ignore this need, and the whole train will judder too and fro at all speeds, with serious effects on the crew as well as the passengers and the rolling stock. A driver who cannot safely hold a cup of tea, well, that won't do at all. So, while the use of knife edges helped in achieving the ideal, the question was always 'how much extra mass to add, and where'. The issue is complex because the two* pistons (etc) do not work in opposition (at 180 degrees) but at 90*, and this adds another problem, a rocking or swaying effect. Add too much extra mass, or in the wrong place, and this can add a new shaking - up and down - so the driving wheels can lift off the track at every revolution at high speed, which as well as being worse for the rolling stock and the track and the personnel, defeated the object of getting the speed really high, as a wheel in the air does not drive. You can see (I hope) that there was always a lot of experimental work involved, to work out what was needed, as well as the 'production' problem of getting the balance right on a run of identical locomotives.

* other numbers and arrangements of cylinders are available ...

Regards, Tim

Edited By Tim Stevens on 20/07/2018 18:20:04

Not unlike the photo on the front of ME I mentioned earlier

The GWR certainly had something better than knife edges, but I'm not sure how it worked.

**LINK**

It was definately pre computer, even the modern ones built by the model gas turbine lads don't always employ computers. I think you can do it with valve amplifiers and a cathode ray tube, they were a commercial product by the 1920's

Nothing wrong with static balancing, if done right. Racing motorcycle flywheels were all done that way for over a century and still are on many high-performance Harleys etc. (No Virginia, that's not an oxymoron.) and classic racers such as Manx Nortons etc. Usual procedure for balancing motorcycle flywheels is setting the mainshafts on a pair of level knife-edges and attaching a weight to the crankpin. The exact weight is calculated by a dark art involving formulae taking into account weight of the reciprocating parts and the weight of the revolving parts (eg  piston vs big-end bearing) and those parts in between that do a bit of both (con rod).

Now train wheels are a lot bigger than motorbike flywheels but they don't rev as high as an 8,000rpm bike engine so same principles could apply.

Edited By Hopper on 21/07/2018 00:27:02

We should all remember that the railway companies, especially the Big Four, were far in advance of anything known outside their industry. However they were very coy about sharing their knowledge. It is on the record that the military were amazed at the competence of the railway workshops when searching for engineering contractors when preparing (?) for WWII..

This attitude was prevalent across europe, forum members will recall that Churchward had to buy locos he wanted to test from France and Germany. See the history of his interest in compounding.

Incidentally there are published photographs of the use of Leitz equipment to align the drive mechanism in Colletts time when the normal workshop practice elsewhere relied on the use of mark one eyeball and string.

There seems to be a lot of mis-understanding about balancing round here.

Dynamic balance is not just a better version of static balancing. A perfect cylinder with equal weights attached at 12 o'clock one end and 6 o'clock at t'uther is in perfect static balance. It is not in dynamic balance because each end is out of balance and the two ends will vibrate 180 degrees out of phase.

That is why Hopper can balance a motorbike flywheel statically (it is short and flat) so the two ends are close enough together for any dynamic imbalance not to matter much. Similar for most grinding wheels.

Static balance was also good enough back in the day when we all drove round on narrow tyres. Now that we have 9" wide tyres on the humblest shopping trolley (and top speeds of 80+!) dyamic wheel balance is essential.

Computers don't enter into it, although they make the job of dynamic balancing simple enough for a tyre fitter.

If you look at the GWR balancing rig in the link by Duncan you will see each end of the axle is supported on 4 springs in a "cross" pattern. What you cannot see is a dial gauge type assembly at each end which would measure the deflection when the wheelset is spun up. You then recorded the deflection at each end. In the absence of computers, you then added a trial weight at one end at a known position, span it up and recorded deflections again. If you were clever you then drew a vector diagram and deduced the effect of your known weight allowing you to calculate how much mass you needed to add or subtract and where. You then repeated the whole saga at the other end. I was taught how to draw the vector diagram at uni and have used the technique combined with a vibration meter in field trials - but that was all quite a while ago now, so my description might not be "spot on".

Martin

Annoyingly my copy of 'Steam Locomotion', Poultney, Caxton Press London 1951 goes into locomotive balancing in depth without detailing how it was done in practice. It's implied, but not positively stated, that most balancing was done by calculation in the drawing office. In the shop, small adjustments were made to the engine in light of practical experience.

As often in engineering it turns out that, at least in 1951, balancing a steam locomotive was very much a compromise. Two sources of 'Disturbing Forces' are identified by the book: those due to revolving parts, and those due to reciprocating parts, both considered separately per cylinder.

Revolving parts consist of: Part of wheel boss forming the crank in all wheels; crank pins with collars, in all wheels; portion of coupling rod adjacent to each crank pin; portion of connecting rod next to main crank pin; return crank and crank pin; part of eccentric rod at the return crank pin end. (Compared with this assemblage, balancing a motor car wheel is easy.)

Reciprocating parts are: Piston; Piston Rod and nut etc; Crosshead with slippers and crosshead pin; Part of connecting rod next to crosshead; Union link, complete; part of combining lever.

Balancing the revolving parts is relatively simple, though there are complications. For example, the piston centre axis is offset from the centre of the driven wheel creating a transverse imbalance.

Piston unbalance appears as a tendency to twist the engine on the track ( 'nosing' ) while the endwise movements cause variations in drawbar pull, tending to rock passengers backwards and forwards ( 'shuttling' ). 'In practice, the only simple means of balancing reciprocating masses in the case of ordinary two-cylinder locomotives is to add some further revolving weight to the balance block in the rims of the driving wheels'

The need for compromise comes about because balancing to reduce unwanted engine movements damages the track! Allowing imbalance transfers forces into the train to the discomfort of passengers. Conversely, totally correcting imbalance transfers the forces into the track instead. For a given locomotive, track and service,  designers compromise between 'overbalance' and 'underbalance', ie comfort and safety versus track stress and 'hammer blow'. Examples show that different railways varied considerably in the percentage of reciprocating balance they applied, example given are:

• At 'one time', when engines were light - 66%
• LMR 2-cylinder engines - 50%
• SR 4-6-0 2-cylinder engines - 30%
• Large 4-8-2 engines on South African 3' 6" gauge - 20%
• USA - 25% to 30%

One of the main reasons steam was dumped so quickly was because keeping track in good order is gob-smackingly expensive. Hammer action cracks culverts and bridges as well as damaging track. Does anyone know how Tornado is balanced such that it's allowed to run on the mainline? Possibly computer methods make it possible to get a more satisfactory compromise than possible with pencil and slide-rule.

I suspect that model steam locomotives are lightweight compared with the strength of the track they run on. Also the track is very lightly stressed in comparison with a busy main line carrying mixed traffic. Furthermore lost revenue due to track maintenance is small. To me that suggests model steam locos should be overbalanced to increase stability rather than underbalanced to avoid giving the infrastructure a good kicking.

Dave

Edited By SillyOldDuffer on 21/07/2018 10:35:20

To partly answer SOD, a typical 5"g 0-6-0 might weigh say 150 lbs, so it's axle load is 50lbs. A fully loaded riding truck (5 off 12st men+truck) could be 1008 lbs on 4 axles, so 252 lbs axle load. Who cares what the loco axles are doing. The loco wheels are probaly bigger as well, certainly on heavier locos, which helps to reduce the wheel/rail contact stress. However, I at least find shuttling annoying, the connecting chain can be seen to got tight/slack at a fairly high frequency on some locos

Yes the two wheels on a common axle four feet apart would definitely benefit from dynamic balancing to settle the axial imbalances not so critical on Manx Norton flywheels! Quite the rig photographed on the cover of the mag there. I'm sure it was a wonder to behold in action, with no safety guards, plonked on the floor in the middle of the factory by the looks. The lack of belt guard would have been the least of their worries if things got out of hand and those wheels took off across the floor. The days when men were men and WHS were too scared to leave their cigarette-smoke filled office.

The most difficult bit of balancing any piston engine is coping with reciprocating masses, as any addition to the rotating masses to cancel it out produces a similar effect at 90 degrees. That means you can only do away with shuttling by producing nosing, or vice versa.

It is - really - possible to eliminate these forces using rotating masses, but it requires two shafts rotating in opposite directions. There is hardly any room on a locomotive (large or small) to add another shaft with bob-weights, but I wonder if it has ever been done (large or small) ?

Tim

While we are talking about balancing I think a McFarlane type wheel balancer must be very easy to emulate in a home workshop, probably not the tool for loco wheels though.

Mike

Fairly commonish practice on motorcycle engines, going for some years now. IIRC the modern-day Triumph twins have small contra-rotating balance shafts to iron out vibration. They also use 270 degree crankshafts on their parallel twins to lessen vibration. Harley-Fergusson with their gawdawful 45-degree V-twin balance problem use a balance shaft on models with a rigidly mounted engine. They are small in comparison to the engine, more the size of a camshaft than a crankshaft, so could possibly have been applied to locos if they had the knowledge back then.

Yes, the recip vs revolve weight is the dark art of balancing, the percentage factor being a dark secret held by Harley performance guys who variously swear by 50 or 55 or 60 per cent. I don't think any one of them eliminates vibration, just changes what rev range it occurs at!

Great old film footage below of making and balancing loco wheels. Not for the faint-hearted! Balancing at about 1:20. One would not want one's grey dustcoat to flap about just that extra little bit next to teh whirling wheels.

Edited By Hopper on 21/07/2018 13:21:31

Hopper you mentioned “high performance Harley’s” what are they used in “lawn mower racing”? 🤔

Dave W

Hopper's British Pathe clip shows the GWR King wheelsets at Swindon being balanced.

One would think that with all of Churchward's experience of the French Compounds he might have got the balancing of his big locos right, but no one has yet mentioned hammer blow, and the restraints imposed by the Civil Engineer.

So, even at Swindon balancing was a compromise, not helped by the divided drive on the 4 cylinder locos, and each pair of outer and inside cylinders working 180 degrees out of phase.

Harold Holcroft's 'Locomotive Adventure' Vol 2 describes how Ashford had only parallel balances, and wheelsets being send to Swindon to be balanced on the machine seen in Hopper's clip. The net result was that the Swindon machine gave no better results compared to Ashford. (Incidentally this shows that what Swindon did was not kept secret, and was shared with at least the SECR and SR).

Anyone who has set up a miniature locomotive chassis and run it on air at speed will be aware of the considerable fore and aft vibration, which has always indicated to me that balancing 3.5"g and 5"g miniature locomotives ought to require much greater attention. And as we don't have to worry about hammer blow, even more reason to do so.

Cheers,

Julian

As a follow up to my article in ME 4586 I've written a second chapter, which goes into cross balancing and 3&4 cylinder locos. I bet you can't wait!