An Active Crossover Design 545 see at once that two 1600 resistors in parallel is going to be very close, and the combined resistance of 800 is in fact only 0.23% low. We also get the maximum possible improvement in accuracy, so we look no further. 19.13 The MID Path: Time Delay Compensation The time delay required in the MID path is five times longer at 400 usec, but on the other hand it covers the 400 Hz crossover point so the delay does not need to be constant up to such a high frequency as in the HF path. Let's see if that makes thing any easier. Since a first-order allpass filter designed for a delay of 80 usec has the delay down by 10% at 2.4 kHz, we would expect the same filter designed for a 400 usec delay to be down by 10% at 480 Hz, and simulation confirms that this is so; the relationship is just simple proportion. 480 Hz is much too close to the 400 Hz crossover point so once more we are going to have to look at a more sophisticated solution. We will aim for 10% down at two octaves above crossover, in other words at 1.6 kHz, because this gave pretty convincing results in the HF path. Looking at the options, we can say that: