My understanding is that NTG actually has a minimal effect in increasing blood/oxygen supply. This is because diseased coronary vessels are obstructed and already maximally dilated, so are affected little by NTG. The primary benefit of NTG is on the demand side of the equation: myocardial wall tension is decreased due to reductions in preload and (to a lesser extent) after load. Also, flow in the smaller vessels likely is increased via reductions in wall tension.*
I used to think that the higher BP limits you sometimes see in NTG protocols were overcautious. In reality though, they make sense: if NTG doesn't actually dilate the coronaries much, then coronary perfusion in these patients is MAP dependent. You only get so much benefit from reducing wall stress, after that, by lowering BP further you are just reducing the driving pressure from the aortic root into the coronaries. Considering NTG hasn't been shown to really work, it makes good sense to be selective about who you use it in.
As for whether you should you keep giving NTG if it doesn't seem to be helping the pain, it makes much more sense to me to target a change in BP than it does an arbitrary number of doses. Say, a drop in MAP of 20%, not to go below a MAP of 70. Something like along those lines. Of course it's very difficult to "titrate" a BP with SL tabs.
*This actually may explain why studies have (so I'm told, I haven't read much about this myself) found little benefit from NTG, though it appears to many to help a lot in some conditions: most patients with severe CAD will have imbalances in Mvo2 supply:demand that can't be adequately fixed by attempting to reduce demand through NTG; you need to increase supply and nitro isn't as good at that as we used to think. On the other hand, patients with a lesser degree of disease or those who imbalance is due more to increased SNS tone may appear to benefit from NTG. Or maybe those with better collateral vasculature do benefit from increased supply due to NTG, and this is just a fairly small subset of patients. Without having read the literature, I suspect that NTG does benefit certain subsets of patients, just not the CAD population as a whole. I don't know if any of this is true; just thinking out loud here.