The breeching for a 32-pounder carronade was 8" in circumference, clinched to an eyebolt beside the gunport, and taken through the loop above the cascabel. [Not spliced, since they often needed to be replaced in a hurry.] If nothing else, this should be shown.
According to Douglas A Treatise on Naval Gunnery (1855), page 415, 'second-breechings' or 'preventer breechings' were always rigged on carronades when in action. This was because if a breeching ruptured or pulled a ringbolt, the piece was liable to be thrown on its sides and cause damage to the carriage. He says that the ends of the preventer were taken through holes in the ship's side and toggled, but does not specify exactly how the inboard part was rigged. Caruana in History of English Sea Ordnance: The Age of the System Volume 2 page 380, mentions preventer breechings, but doesn't elaborate. In any case, the modeler would be better ignoring this item.
The side tackles ran from eyebolts at the rear end of the slide to eyebolts in the ship's side. The second figure in Lavery's book show this as a luff-tackle, rigged to advantage ...i.e. with a velocity ratio of four. The other figure is less clear, since the artist seems to be showing both outboard blocks overlapping, but my best guess is that he is representing gun-tackles rigged to advantage (VR or three). With 32-pounders, luff-tackles seem more likely, but only the most enthusiastic modeler should be content with gun-tackles (two single blocks).
No train tackle was needed with a carronade, because it never
needed to be hauled inboard. [The train tackle was so named because
it was secured to the train (trail) or rear of the gun, not because it had
anything to do with training the weapon.] However, Figure G3/1 on
page 109 of David White's monograph on Diana, indicates an
eyebolt for a 'training tackle' on the rear of the carriage, just above
the truck. I don't what his authority for this was, but if it is authentic,
it suggests that tackles were used as well as handspikes to train the
gun. [I would consider White a most reliable source]. However once
gain, even if one could find support for training tackles, they would
be best ignored by the modeler.
{John H. Harland}