Balancing of inside cylinder locomotives

I've only recently noticed that inside cylinder six coupled locomotives have the wheel balance weights in the 'wrong' place. I don't know how I missed it all these years.
With an outside cylinder locomotive, the connecting rod big end is always concentric with the coupling rod bearing, so the balance weight is opposite.
I understand that with inside cylinders, there is no such restriction and the designer can put the big ends anywhere on the crankshaft/axle in relation to the coupling rods, subject to each pair of cranks and coupling rods being quartered to each other.
So why are they designed the way they are?
If say, when all viewed from the left, the left hand coupling rod is at 12 o'clock and the right hand coupling rod is at 3 o'clock, why not put the left hand connecting rod at 6 o'clock and the right hand connecting rod at 9 o'clock?
It seems to me that this would improve static balance, and with weights added to the wheels and crankshaft/axle opposite each bearing, the dynamic balance would improve and hammer blow would be reduced. What have I missed?
I realise this could be a complex subject. I'd be happy with a link to an article or book rather than expecting someone to type out a long explanation here.

 

My understanding is that the arrangement is to reduce the stress on the frames.

 

I apologise for coming to this a bit late, I only just spotted this. The convention on connecting rod positions for inside cylinder 2-cylinder engines is indeed a bit variable. I'm not going to get into cross balancing or anything like that, although this link is an interesting read, if focused on a rather unusual example in the 9F.

For a long time, the convention was that the connecting rods would be opposite their adjacent coupling rod, for exactly the reasons you state: it reduces the mass required for balancing. We'll call this out-of-phase (180 degrees) and the alternative in-phase. If you search for photos of 30120, you can clearly see the balance weights on the same side of the wheel as the leading crank pin, indicating that the crank axle is out-of-phase.

The exception for this was that man William Stroudley, who firmly opined that the crank pins should be in-phase on their respective sides. Look for photos of Gladstone, and although the balance weights are slightly hidden by having spokes painted on them, the ones on the driving axle are very large and opposite the crank pins. His reasoning was that this arrangement reduced loads on the bearings, and I think it can be justified. Having the coupling rods balancing the crank axles inevitable causes a side-to-side rocking moment because they aren't in plane. Having two large masses balanced by a weight in a plane somewhere between them reduces this effect.

Now, back to Drummond 4-4-0s. At some point after the T9 class(I think the L12 or S11 class), he introduced the balanced crank axle, where the crank webs include their own balance weights. This has the advantage that it reduces the balancing loads that need to be transmitted through the axle, at the cost of increased weight. Look for pictures of a D15 4-4-0 and you won't see the balance weights adjacent to the crankpin like on a T9.

A very clear example of why you might do this can be found rather later in the Bulleid Pacifics; as built they had unbalanced crank axles to save weight, with the outside motion serving to provide balancing as you suggest. A few years later, 35020's crank axle famously failed passing through Crewkerne station due to fatigue cracking. One of the least visible changes introduced on rebuilding was a balanced crank axle (which is incidentally the reason the rebuilds have much bigger balance weights outside; I'm pretty sure that no hammer blow was introduced, despite popular myth).

Now, 8572 is a probably the most prominent inside cylinder 6-coupled engine I can think of, and it clearly has quite large balance weights on the driving axle, opposite the coupling rod crank pins. I'm not sure exactly what the arrangement is; I think they aren't big enough to balance the crank axle alone, compared to those on the other coupled axle. Given that B12s post-date Drummond's work with balanced crank axles, I suspect they have at least partially balanced in-phase crank axles with maybe a bit of additional balancing outside, but I can't find a picture of a B12 crank axle in isolation. I have found pictures of the Y14/J15 crank axle and on that the crank axle is out-of-phase with the coupling rods.

EDIT: I've remembered some further examples/counter-examples. On the GWR the 5600 class have visibly massive balance weights on the driving axle in a position indicating an in-phase crank axle. I believe this scheme was adopted due to trouble with the bearings (see Stroudley). However, if you look at a 5700 (or indeed the various Welsh pre-grouping 0-6-2T classes) the balance weight positions clearly show an out-of-phase crank axle.

The Stroudley E1 0-6-0T class were built in the Stroudley fashion with in-phase crank axles. However, years later some of them were converted by Maunsell to 0-6-2T engines for use in Devon and Cornwall. Complaints about the ride led to Bulleid re-balancing some of them with out-of-phase crank axles, which I suspect allowed for more reciprocating balance within the confines of their relatively small wheels, since the out-of-phase scheme normally requires smaller weights. So 32608 (among others) had balance weights adjacent to the crank pins, while 32697 (again, among others) kept the Stroudley layout with them opposite.

 

On the E1s, and E1Rs, I think the rebalancing was actually under Maunsell (Bradley gives dates starting in 1936).

More generally on that scheme, I've always understood (but more by empirical feel than analytical calculation) that you don't get anything for nothing in balancing, and the Stroudley scheme tends to minimise hammer plow (up and down) but at the expense of a fore-and-aft motion. Certainly the E1 tanks, as you mention, had a bad reputation for shaking passenger carriages about, something that didn't get changed in the 1880s when they were only used for passenger trains for a brief period while more passenger tanks were being built, but was picked up and changed in the 1930s after conversion to E1R. My brief footplate experience on a 56xx was that it was most uncomfortable, and from a passenger point of view you could frequently feel the surging a couple of carriages back. Again, not something of concern in their original use hauling coal wagons up and down the Welsh valleys, when presumably other advantages came to the fore.

It would be interesting to know what a Gladstone was like at speed!

Tom

 

I think the thing with the Stroudley scheme is mostly that it demands big balance weights. On a large-wheeled Gladstone, that's fairly easy to accommodate. Small wheels (like on a high tractive effort goods tank) are much less accommodating of effective balancing, because everything's against you. The maximum radius is limited, so more mass is needed to balance a given set of motion, and diminishing returns kick in very quickly; all the additional material gets closer and closer to the centre of the wheel (particularly on the 5600). That's why the 9F design had the combined reciprocating balance scheme which ignored the yaw component; it was the only way of getting enough reciprocating balance within the constraints of the mass which could be accommodated.

But yes, I would still love to experience a Gladstone at speed. Purely in the spirit of research, you understand...

 

I am told by someone who will know, that motion components in Stroudley locomotives were standardised between classes hence the massive connecting rods on the surviving E1. The exception is the A1/A1x class which had their unique, much lighter, pattern of motion. They give no problems in respect of surging.

Rides behind G.W.R. 0-6-2Ts confirm the remarks above about their proneness to surging. This was dreadful at times.

 

Given the disparate schools of design, it'd be still more interesting to know how a Stroudley Gladstone compared with an Adams Jubilee.