I came across the shunt approach today while researching my problem on the web. Basically, a shunt of known resistance bypasses the ammeter, thereby reducing the current running through it. The gauge is then calibrated based on the calculated current flow that it will measure. This was used on Detroit products in years gone by. One fellow said his shunt wire broke and the ammeter was fried by the increase in current, causing a fire under his dash.
You are right about the marine guys, too. They find ways to make things work under very adverse conditions for any electrical system.
