I'm not so sure this problem is very easy to fix. I think (and please correct me if I'm wrong here) that the current damage model is something like this: Every piece of ordinance, and every bullet has some kind of damage value. Lets assume that a .50 cal bullet has 1 in damage value. Then every target, or target surfance has some kind of damage resistance (cant come up with a better word). When it comes to aircraft, every part of the plane has its own damage resistance. When enough damage points is inflicted upon a target surface, that surface breaks (and we get to see a wing fall off etc).
I suspect there is some kind of randomizer in there somewhere too, so that that .50 cal bullet sometimes inflicts 2 damage points, and sometimes 1/2 damage points.
Now, if I'm correct in my guessings here, that would mean that everything can be destroyed by everything. If you shoot long enough on a Hangar with the 7.92 rear MG on a 110C4, eventually the hangar will be destroyed. It will take alot of time, because the Hangar can absorb the equivalent of 3k lbs worth of bombs, but eventually you will be able to shoot the hangar to pieces. Every single 7.92mm bullet will make .25 points worth of damage to the target or something like that. Then it only becomes a question of time and ammo before the hangar is destroyed.
Over to our PzIV, like every other target in the game, each target surface on the pz has a damage resistance value. And since the panzer is armored, I suspect that this is reflected either by giving it a high damage resistance value, or by reducing incoming damage values by some factor.
The same should apply here. If you shoot enough 7.92mm bullets on the frontal armor of the Pz, eventually it will be destroyed.
I havent had time to test this theory yet (because I just came up with it, and Im at work). But if I'm correct, or at least close to the truth, then the whole .50 cals taking out armor-issue would need a complete rewrite of the damage model to fix.
As I pointed out earlier, it is not realistic to ask for a fully realistic damage model. The number of calculations that would be required for a realistic damage model in tank to tank combat is astounding. And besides, it should not really be neccessary. What we want is a realistic simulation, at least that is what I want. And 90% of the calculations that would be required in a realistic damage model for tank combat can be reduced to a couple of randomizers within certain variables to achieve as good a result as any.