As far as I could tell from my previous BoP testing with no skill modifiers, impulse engines grant a fixed amount of speed per power at all power settings, meaning that the actual speed vs power setting graph is linear for all three engine types, and not a curve as most seem to expect. The difference comes in that combat has a higher 'base speed' than the hyper, but adds less speed per power, allowing the hyper to eventually 'cross over' and start exceeding the combats at some particular power setting.
The results from my recent testing confirmed a theory of mine from the previous testing: The different impulse modifiers on the different shiptypes also changes the crossover point.
My earlier testing found that for BoPs/Escorts(Impulse Modifier: 0.20) it was at approximately 43 power setting. The testing I just did with a Cruiser(Impulse modifier: 0.15) found that the crossover point was at 57 power setting(with a speed difference of 0.03 in favor of the hyper).
The numbers I've generally seen used were from a DSSV(Impulse Modifier: 0.17), and had the crossover point somewhere near 50-ish, which is consistent with my theory.
What seems to be happening is that the balance between the base speed and the power per speed isn't scaling identically between ships with different impulse modifiers. As far as I can tell, the higher the impulse modifier, the lower the crossover point is (meaning it's even LOWER for delta flyers!)
What this means then, is that there's actually four different sets of impulse efficiency "curves", one for each impulse modifier in use.
The effect being that hyper impulse engines are more viable for escorts, BOPs, and Raptors than any other ship type, and least viable on cruisers and carriers, with science vessels jammed in the middle somewhere.