The pulse width is the time the pulse current is on. Increase the width and the coil current will increase on many if not most coils. This will or can result in some depth increase but does so at a price. More coil current takes longer to decay so the minimum delay that one might be able to use can increase. Thus, a longer pulse width can result in increasing the depth on large objects but make it more difficult to detect the smallest gold.
Pulse width is not generally something that is adjustable on many PI's. On the GS 5, at least on some versions, one could switch between 100usec and 200 usec pulse width. The longer width may have been 250 usec. I don't remember for sure. I never did see any improvement in depth with the longer pulse width but I was only testing some of the more common nuggets I have found in the past with none of them much over 1/2 oz.
Now, the 200usec width should do a little better on much larger gold and silver but would not usually will not allow the minimum delay to work on some coils that would work on the 100 usec setting.
So, part of the decision of what would be a single setting if there were no pulse width control was determined by what was to be the primary reason for the detector. For gold, generally, 100usec worked fine. Actually, 50 usec worked great for really small gold because it would allow for a little shorter delay to be used. However, one could see some depth loss on larger targets, so the 100 usec setting is a good compromise and is common on detectors such as the TDI.