Have you considered using the S&W mounting plate and a goalpost shaped loop of brass wire,

instead of fixing it across the buffer heads, it makes it easier to mount the coupling, as it's then

a single unit. Also, you can then adjust the length of loop to suit your curves (I also think it looks

better than across the buffer heads)

Thank you for the suggestion! I agree that they do look better. I just wasn't sure if they worked so well on radius 2 curves. The wire-on-buffers method looks like it should allow for more sideways movement? But I think it would be a nice thing to experiment with, and as I have just bought a new soldering iron tip I feel up to the challenge. I shall report back with my findings!

We tried a few different auto couplings before standardising on Sprat & Winkle and have now used them in 4 or 5 exhibition layouts over 20 years.

they are robust and reliable with only minor adjustment needed between shows. We have adopted the base plates and brass loop mostly (though my dad does still like to glue bars on buffers).

Thanks for that. Fathers can do silly things.. I should know as I am one. D

Further experiments have been undertaken to see if I can uncouple/couple on a radius 2 curve, such as I have besides the inner platform. It seems to be sort of possible.. just. You can push the hook past the bar, but it probably would not work too well if I was relying on a magnet under the track. I might try snipping off part of the hook so that it passes the bar more easily. What can go wrong??

Yes, it does look a bit overly bent in that photo.. I shall check the angle! They have to be quite bent to miss each other, as in one of the photos from my post further up the page. But that situation occurs only on reverse curves.. and if I can avoid such curves on my layout, which I am trying to do, then perhaps I can indeed straighten them out. I think the next step is to put the track together and try moving some trains about.

Thanks for the tip! D