Hi Frank,
that's what I needed to know before I typed a book. I just went through this with my DMI knotmeter, and after a nice chat with them, I have some troubleshooting steps to follow:
1) Disconnect the BNC (input coax) from the back of the instrument. Take a paper click and bend it into a U so you can put one end into the center of connection on the instrument and then tap the other end of the paper clip on the outside of the connector. The faster you tap it, the higher the speed reads. So, if you get a reading, it's probably your sending unit or in the cabling and your head is OK.
2) if it responds to the above, hook the cable back to the instrument and locate your transducer. From the transducer end, work back up the cable about 10 feet. There's another BNC connector. You can disconnect it, and using the same paper clip (what a versatile tool), short out the center lead to the outer part of the connector by tapping. Again, you should get a speed reading, and the faster you tap, the faster you go. If this checks out, your cable integrity to this point is OK. and you are now down to the sending unit. The sending unit is held into the thru-hull with a large cotter type pin. Make sure you have a blank plug at hand, and remove the pin, gently twist and lift, and see if you can't withdraw the core of the sending unit, of course being poised to put the plug in. Water will come in. You will get wet, but after you've done it 1/2 a dozen times, you'll get used to the shower and cleaning up the mung on it.
Of note, your symptoms are very similar to mine, which required a new sending unit. I would have real readings, and then nothing, and then perhaps .2 and so on.
) Then it worked it's way down to 0, where it has been for a while. 
