Coupling rod knuckle joints

[Andy W]

I'm trying to get my head around how critical it is to have the holes in the knuckles at exactly the same spacing on both rods. As the knuckles don't directly relate to the wheelbase I assume they don't have to match exactly. What's the theory?

[grovenor-2685]

Unless you are talking of really large discrepancies any effect will IMHO be to small to notice. On the flat track there is no effect, when the rod is bent at the knuckle the effective length of that rod is the third side of the triangle which could be calculated for each case to see what if any effective change in length occurs. I would not bother but maybe Russ is looking for some amusement.
Regards
Keith

[steamraiser]

As long as the slight difference in the knuckle holes does not in any way alter the relative positions of the crank pin holes, then I would say you will be alright.
One of the recognised approaches is to sweat the layers for the left hand and right hand side rods between the same two axles together, then drill your holes. If done correctly all the holes will be in the same place.

Gordon A
Bristol

[Tim V]

Quite honestly I've never considered it.

I've used knuckle jointed rods for nearly 30 years. The knuckle is where it is etched (if lucky). The only dimension that is important is that the wheelbase exactly matches the coupling rod dimension.

Don't forget, we are not building locos that can climb 1mm discrepancy's. Compensation/springing is there to improve pickup, not to cope with 1mm track jumps. The wheels move up and down by very small increments, so any displacement that involves the knuckle is effectively minuscule and not worth bothering about.

[Horsetan]

The knuckles may not have to have the same spacing, but I take the view that it is where you rivet / pivot it which matters, and it helps if these are the same.

[Russ Elliott]

The OP is rather bizarre, but a little bit of Pythagoras will allay any fears.

Imagine a and b are the nominal wheelbase dimensions, and x is the distance from the centre axle to the knuckle. If the centre axle is deflected y vertically from its nominal horizontal axis, the difference between the sum of the hypotenuses and the nominal overall rod length (a + b) is:

[(a + x)2 + y2]0.5 + [(b - x)2 + y2]0.5 - (a + b)

rod-knuckle.png

If x is zero, i.e. the rods are jointed on the pin, the hypotenuse difference is far less than the normal running clearances (blocks to guides, axles in blocks, and rods on bushes) we require anyway, i.e. it really doesn't matter, even if taking an unrealistically large value for y. Here's some example values showing how the hypotenuse difference increases slightly as x is increased, but again, even with absurd values of x, the hypotenuse difference is far less than our necessary running clearances, so it still doesn't matter. Similarly, if the x values are different between one side of the chassis and the other, it still doesn't matter.

rod-knuckle2.png

[Andy W]

Thanks chaps. You've put my mind at rest. I'll try and get the joints in the same place, but it was the theory that my brain couldn't cope with.