The latest Arms and Armour Journal from the Royal Armouries has an article regarding some testing of warbows against plate.
"A report of the findings of the Defence Academy warbow trials Part 1 Summer 2005." By Paul Bourke and David Whetham. pp.53-82.
Edit: The paper can now be downloaded
It also has some criticisms from Kelly DeVries along with rebuttals from the authors.
I like some of it and dislike the rest. It seems to me (and to DeVries) that these guys had preconceived ideas and set about designing a test to validate them. DeVries reckons that the bow they used was too heavy but, after reading Strickland and Hardy's book, I can't agree with him. IMO the bow they tested was typical of the time (around 150 lbs). I think that the arrows they tested were reasonable reconstructions too although I would have liked to have seen more aspen shafts and less ash. By the time of Henry V, aspen was by far the most common wood used for arrows, not ash. I also have a problem with the extremely short range they used (10 m). Hopefully the next part of these tests will involve more realistic battlefield distances.
The arrowheads are flawed. Although they use accurate reconstructions of the typologies, the hardness is far greater than that found in any bodkins to date. The authors state "it is unknown how hard average period arrowheads were or whether they would have been routinely surface/case hardened..." IMO they say this to try and avoid the fact that none of the bodkins so far analysed were anywhere near the hardness of the arrowheads used in this test. The hardest arrowheads so far analysed are broadheads (e.g. type 16), not bodkins. The one bodkin analysis they cite (their own unpublished work) indicate a hardness between 105-158 Hv which was well below the hardness of the arrowheads they used. Their first arrowhead (type 7) had a hardness of 190-200 Hv with the tip being 300 Hv. Their second arrowhead (Type 10) had a hardness of 230-250 Hv. The third (lozenge) 480-500 Hv. No surprises which one performed the best against the plate target.
Regarding the target, the authors correctly state that the Victorian wrought iron used in the past for these sorts of tests is inferior to what was available for contemporary plate armour. They reckoned that charcoal-rolled iron would be a closer match to what medieval armourers had available and I agree with them (but only for munitions-grade armour). They decide to shoot a flat sheet of this material, rather than a worked breastplate because they reasonably argue that it would produce more consistent results. This makes sense when testing the angle of impact but ignores the fact that armour was fluted and ribbed in key areas to reduce the likelihood of penetration. The thicknesses chosen by the authors seem reasonable. They test three plates: 1.15mm (206Hv), 2mm (180Hv) and 3mm (172Hv). They acknowledge that these hardness readings are in the lower half of extant examples. Williams noted that the best plate in this test is only equal to munitions grade 15th century plate. They ignore the fact that the thinnest sections of plate armour are often overlapped by another plate so that an arrow would have two thicknesses of this material to punch through. They also completely ignore the fact that plate armour was not worn against the skin. I would suggest that an arming doublet worn under armour would greatly reduce injury to the wearer.
The results. All three arrows easily penetrated the 1.15mm plate. The 2mm plate is penetrated by arrowhead 2 (9mm) and 3 (16mm). The 3mm plate defeats all the arrows. Angle of impact is 90 degrees. They note that adding wax to the arrowhead made no difference in penetration.
Conclusion. Dispite this test being heavily biased against the armour, the plate seems to have performed well. Even without an arming garment behind the plate, none of the arrows would have killed a soldier wearing 2mm of plate (the deepest penetration was only 16mm). If the padding is added, I think that there would be no injury at all. If the plate was hardened as with 15th C Milanese examples, there would be no penetration. If the 1.15mm plates were overlapped as would have been the case in a suit of plate then these too are likely to have prevented injury. The only injury that might have occurred would be the rare arrow that managed to hit a thin piece of plate that was not covered by another plate and did not have mail underneath it. These arrow strikes might hit an arm or a leg but certainly would not kill the wearer as is implied by some longbow enthusiasts. The authors then try to weasel out of these results by claiming that non-fatal arrowstrikes were likely to prove fatal anyway because of the unsanitary conditions, dispite plenty of evidence to suggest that soldiers regularly survived arrow wounds - even in the face. They also concoct a ridiculous theory that even though the armour resisted the arrow, the blunt trauma from the impact might have killed the victim anyway.
This test could have been very good. By itself it is only moderately useful. It is the best I have seen so far though. Note that the results are only valid for munitions plate, not the finer quality Milanese plate. My only major complaints are the extreme hardness of the arrowheads used and the fact that nothing was used to simulate an arming doublet. Hopefully Part 2 will build upon this data.