Dan - the only place to sense the "real" temperature is on the engine side of the thermostat. Unfortunately there is no external plumbing on the 2GM that offers visibility on the engine side of the thermostat. There is a by-path connection on the outside of the thermostat that prevents the water pump from discharging to "blocked" system when the thermostat is closed. Any hot water exiting the engine by the thermostat is mixed with this cold water. Hence the cold water sensing symptoms you saw. The cooling water temperature switch on the 2GM screws into the thermostat housing on the front of the engine. This would be the only place to screw in a sensor and drive a gauge to get an accurate water temperature reading. If you look at this link, you'll see the Yanmar service manual Chapter 8 (Cooling System): http://www.j30.us/files/yanmar-cooling-system-chapter8.pdf
Page 5 has a schematic of the closed loop system (Freshwater Cooling). The picture on the 2nd to last page of this Yanmar operation manual has the raw water cooled system: http://j30.us/files/Rhapsody/yanmar-manual.pdf

You can see how the cooling water piping and the switch screwed into the thermostat housing on page 32 of this presentation: http://j30.us/files/Rhapsody_Yanmar_Rebuild_Winter_2006-2007.pdf

Just curious - what didn't work on the dual-contact sender? You could test it in a glass of water. Hook up an ohm-meter across the gauge terminals and pour boiling water over the sensing end. The resistance should increase. The spec is 10-180 ohms. If you want to get real scientific you could build a little fixture with a thermometer and plot the resistance vs temperature. The switch on the over-temp side should close at 209 degrees F +/- 5 dgrees.

Bill