Point rodding: compensators

[philsandy]

I've read (on another forum) that compensators are normally every 10 yards in a run of rodding. Is this correct?  ie.  a compensator every 120 mm in 4mm scale, seems rather a lot.

I searched for photos of point rodding but they do not seem to show compensators as frequent as every 10 yards.

Also that  if there are 2 cranks in  a run, ie. where the run diverts 90 degrees to cross under a track, then a straight length run,  and then diverts 90 degrees again, a compensator would not be required as the cranks would do the job of a compensator?

[Miss Prism]

What you seem to have read on another forum is gibberish - a point rodding run should require a maximum of one compensator.

It is true that a compensator might not be required if the push-pull equation can be achieved by suitable crank orientations.

 

[DCB]

Dont forget there are both soundly engineered GWR compensators and rubbish ones used by other folks.......

[TheSignalEngineer]

20 hours ago, Miss Prism said:

What you seem to have read on another forum is gibberish - a point rodding run should require a maximum of one compensator.

It is true that a compensator might not be required if the push-pull equation can be achieved by suitable crank orientations.

 

Unless you have more than one drop off for multiple ends, then there will be one between the box and the near end, followed by a second between the near end and far end.

[TheSignalEngineer]

This is a sketch I posted in a thread a while ago, can't work out how to link yet but the title was 'Midland Railway - Crank connection to stretcher bars'

[Miss Prism]

17 hours ago, TheSignalEngineer said:

Unless you have more than one drop off for multiple ends, then there will be one between the box and the near end, followed by a second between the near end and far end.

 

Of course. I was trying to keep it simple, in the spirit of the OP's enquiry.

 

[philsandy]

On ‎2‎/‎2‎/‎2019 at 5:19 PM, TheSignalEngineer said:

Unless you have more than one drop off for multiple ends, then there will be one between the box and the near end, followed by a second between the near end and far end. 

 

Thanks for your reply. So would the rodding run below need a compensator?

Also the goods yard has 2 points operated by a ground frame, both have 10 yard straight runs, would compensators still be needed even for these?

 

[TheSignalEngineer]

Compensation simplistically has to cover the whole length of the rod from the signal box lead-off to the connection on the point stretcher. Depending on whether the cranks change the action of the rod and other factors such as a long run in the shade, length of cross rods and lie of the points there will be one compensator roughly in the middle of the 49 yds stretch.

For a ground frame the rod is often so short as not to need a compensator, especially if the cranks are set to change the rod action from pull to push. A longer run to double ended points may need compensation

[steve4rosegrove]

On 02/02/2019 at 17:14, DCB said:

Dont forget there are both soundly engineered GWR compensators and rubbish ones used by other folks.......

More gibberish.

[ikcdab]

Alluded to above but not explicitly mentioned is that rodding runs are set up so that half the rodding pulls and the other half pushes. The compensators make that change of direction. Thus this allows for expansion and contraction. 

[steve4rosegrove]

That's taken as read, but still no answer as to maximum length without a compensator.  I cannot find my reference material.

 

[micknich2003]

Steve, max length uncompensated is 30ft.

[micknich2003]

Steve, more pages.

[2ManySpams]

So, a couple of notes to try and clarify what's been said...

 

Metal expands and contracts with temperature. Rodding runs are metal and some are long enough that the expansion and contraction could start having an interesting effect on how turnouts work. That's particularly the case if the rodding all moves in the same direction (all push or all pull) when actuated.

 

To overcome this problem the idea is that the design of the rodding has near equal lengths of push and pull in a run. This can sometimes be achieved just through the use of cranks to change push into pull at 90deg changes in direction. 

 

Where a balance can't be achieved just with cranks and the difference between lengths of push and pull is >30' or 10 yards, a compensator is used. This is used in line with the rodding run. 

 

The extract below is from The Signalman's diagram above. 

Signal box is bottom left. Lever 7 is pulled and the rodding moves towards the box (pull). It reaches a crank which then turns anticlockwise and pulls the next length of rodding. There's then a compensator in the rodding run which changes pull into push. The next crank is pushed, turning it anticlockwise, which pushes the next rodding run which throws the turnout blades. 

 

In this case the location of the compensator is that at which the total length of pull is equal to the total length of push. 

 

So you don't need a compensator every 30', just where the difference between push and pull totals is >30'. 

[Michael Hodgson]

On 03/02/2019 at 23:44, philsandy said:

 

Thanks for your reply. So would the rodding run below need a compensator?

Also the goods yard has 2 points operated by a ground frame, both have 10 yard straight runs, would compensators still be needed even for these?

 

If it isn't already clear from the answers above, yes a compensator is needed, and its position is important. 

It will be approx half way along the 49 yard length. 

The position is half way so that the expansion of one half (push) is offset by the expansion in the other half (pulling). 

 

For simplicity I am ignoring the rodding crossing the track which tends to cancel each other out because

• they are relatively short, and
• with a compensator half way, one of them is pushing when the other is pulling 

 

In practice you would first determine where cranks were needed, then add up the length of the pushes and the lengths of the pulls to calculate the ideal half way point to fix the position of the compensator.

 

It doesn't work perfectly, because part of the rodding run might be in shade (under the platform coping stones) whilst another part of the run is in full sun, so expansion is not necessarily uniform,

Edited August 24, 2021 by Michael Hodgson

[Grovenor]

Quote

It doesn't work perfectly, because part of the rodding run might be in shade (under the platform coping stones) whilst another part of the run is in full sun, so expansion is not necessarily uniform,

Which is one of the reasons to use facing point locks.