I have come across an interesting old article about this ship by Strafford Morss in Warship International (2006) 43 No 3 273.
While the piece is mostly concerned with the many small scale tests that were done to determine the support of belt armour, and the plates in the conning tower and turrets, there are a couple of interesting points about details. Chief among these is that there appeared to be a weak point near the bottom of the torpedo bulkhead. This did not go down as a single plate to the ship's bottom but, as in German WW1 era ships, was jointed to an extended frame in the bottom. This joint was a little above the inner triple bottom. The joint was both welded and joined by rivetted straps both on the inside and the outside of the bulkhead. Morss says that this probably stiffened the bulkhead at this point which made it less able to deflect to absorb the energy from an underwater explosion. Another undesirable feature of this bulkhead was that its slope changed at the joint so that its inclination became closer to the vertical at the bottom. [As far as I can see the reason for that was probably the need to fit this inclined and inboard bulkhead into a space limited by the combination of the beam and the width of the machinery spaces. Perhaps the slope of the belt had earlier been planned to be less than the final version selected. If the belt had been moved out a little a straight bulkhead could have been accommodated but then the armour deck would have been wider and heavier. The North Carolina class had no such kink in the torpedo bulkhead.]
Another point is that Morss gives a different reason for the 1.25 in external plating at the waterline than the one familiar to me. Up to now, I have understood this was primarily provided as splinter protection. Morss says it was provided to activate the fuzes of shells. Of course, it would have had some value in both roles.
But most of the article is concerned with structural arrangements for armour support. Many tests were conducted with 1:11 and later 1:12 models. These small scale models were used because they allowed various alternatives of each feature to be tried and because of time constraints; they were cheaper too. While these tests were limited in their predictive power for 1:1, it was considered that tests of two alternatives would likely to be accurate for which arrangement was best. One pound shot was used, the 1:11 scale corresponding to 14 in shells and the 1:12 scale to 15 in shells. Earlier ships had unkeyed armour plates backed by heavy framing, further backed by the 2nd and 3rd decks. The thickness of the 3rd deck was fixed by the need for it not to be damaged significantly by shells striking the belt abreast this deck. The new arrangements featured keyed plates that were interlocked for which lighter backing structure sufficed as well as a thinner 3rd deck. [I had surmised this deck was thinned in compensation for the heavier decks above it.] The shells that were most damaging to structure were those that were stopped, being either lodged in a plate or rebounding. [Certainly, these shells had less energy than those that go through but all of that energy is deposited in the plate struck.] The structural soundness of the joint between the barbettes and the armour deck when either was struck by shells was also a concern. Again, keyed joints in the barbette plates allowed the barbette thickness below the armour deck to be reduced from the 3 in in North Carolina to 1.5 in in Massachusetts. There is also the comment that the design of the interlocking turret armour plates required great care as there was no armour backing. Overall, savings in structural weight in armour support helped to make this a successful design. But, as mentioned above, given a multitude of joints in such a ship, and given other sometimes conflicting requirements, not every joint could be a good one,
Neil Robertson
While the piece is mostly concerned with the many small scale tests that were done to determine the support of belt armour, and the plates in the conning tower and turrets, there are a couple of interesting points about details. Chief among these is that there appeared to be a weak point near the bottom of the torpedo bulkhead. This did not go down as a single plate to the ship's bottom but, as in German WW1 era ships, was jointed to an extended frame in the bottom. This joint was a little above the inner triple bottom. The joint was both welded and joined by rivetted straps both on the inside and the outside of the bulkhead. Morss says that this probably stiffened the bulkhead at this point which made it less able to deflect to absorb the energy from an underwater explosion. Another undesirable feature of this bulkhead was that its slope changed at the joint so that its inclination became closer to the vertical at the bottom. [As far as I can see the reason for that was probably the need to fit this inclined and inboard bulkhead into a space limited by the combination of the beam and the width of the machinery spaces. Perhaps the slope of the belt had earlier been planned to be less than the final version selected. If the belt had been moved out a little a straight bulkhead could have been accommodated but then the armour deck would have been wider and heavier. The North Carolina class had no such kink in the torpedo bulkhead.]
Another point is that Morss gives a different reason for the 1.25 in external plating at the waterline than the one familiar to me. Up to now, I have understood this was primarily provided as splinter protection. Morss says it was provided to activate the fuzes of shells. Of course, it would have had some value in both roles.
But most of the article is concerned with structural arrangements for armour support. Many tests were conducted with 1:11 and later 1:12 models. These small scale models were used because they allowed various alternatives of each feature to be tried and because of time constraints; they were cheaper too. While these tests were limited in their predictive power for 1:1, it was considered that tests of two alternatives would likely to be accurate for which arrangement was best. One pound shot was used, the 1:11 scale corresponding to 14 in shells and the 1:12 scale to 15 in shells. Earlier ships had unkeyed armour plates backed by heavy framing, further backed by the 2nd and 3rd decks. The thickness of the 3rd deck was fixed by the need for it not to be damaged significantly by shells striking the belt abreast this deck. The new arrangements featured keyed plates that were interlocked for which lighter backing structure sufficed as well as a thinner 3rd deck. [I had surmised this deck was thinned in compensation for the heavier decks above it.] The shells that were most damaging to structure were those that were stopped, being either lodged in a plate or rebounding. [Certainly, these shells had less energy than those that go through but all of that energy is deposited in the plate struck.] The structural soundness of the joint between the barbettes and the armour deck when either was struck by shells was also a concern. Again, keyed joints in the barbette plates allowed the barbette thickness below the armour deck to be reduced from the 3 in in North Carolina to 1.5 in in Massachusetts. There is also the comment that the design of the interlocking turret armour plates required great care as there was no armour backing. Overall, savings in structural weight in armour support helped to make this a successful design. But, as mentioned above, given a multitude of joints in such a ship, and given other sometimes conflicting requirements, not every joint could be a good one,
Neil Robertson
statistics: Posted by neilrobertson1 — 11:34 PM - 1 day ago — Replies 3 — Views 157