I'm back on the case with this little project
I wasn't too happy about the accuracy of the pressure sensor measurement approach. For my tanks it ends up being +/-6%. So I decided to use the Maretron ultrasonic senders that direct connect to N2K. I ordered up everything, including focus tubes, cables
, Tees, etc.
Today I went to start installing and hit one road block after another. The bottom line is that this approach is not going to work.
The first issue is that although my tanks have standard 5 bolt sensor fittings, the tank hole is much smaller than the standard size. The standard is around a 2" hole, and mine are more like 1". This means the focus tube won't fit and even just the sounder without the focus tube won't fit either. The only solution would be to remove the tank covers from all 4 tanks including all the associated hoses and vents, drill/cut the hole larger, then reassemble. It's certainly possible, but MUCH more work and complexity than I was planning on.
But the next issue was the final nail in the coffin. My fuel tanks are trapazoidal with the outboard
side sloping inward with a Vee bottom. The sounders don't like a sloped bottom, which is part of why I got focus tubes, but I thought the location of the existing sender would reach well into the Vee to give an acceptable reading. After all, the existing sender manages it somehow, right?
Not so. The fitting is very much over the sloped part of the tank, and looking straight down the sounder would be reading empty with only about 1/2 a tank. Not good. The existing sender actually deals with this is a clever way. It's a rod with a sense wire barber-polled around it, and some smarts in the probe to take some sort of capacitive measurement according to the manufacturer. What Grand Banks
did is bend the tube so that it reaches to the side and all the way to the bottom Vee in the tank. With proper calibration, it works fine.
Since there is no way to get the sounder to measure at an angle like that, it kills them as viable replacements
The brings me back to the pressure sensors since they couldn't care less what the shape of the tank is, and what twists and turns it takes along the way. It just measures fluid head
, and with proper calibration (which Maretron supports), can handle any shape tank. I also could use the existing sender wiring
to bring the signals all back to a common point at the helm
for the N2K interface. I still don't like the +/-6% accuracy issue, and might investigate a 0-1.5 psi sender which would bring the accuracy up to match other methods.
I've also stumbled on another possible approach and have an inquiry off to Maretron to see if it will work. Their TLM100 is an N2K adapter designed to connect to a standard 240-33 ohm resistive sender. Sounds good, except the Centroid Computank senders are not resistive. They take a 12V input and output a voltage signal of 0V to 4V according to tank level. But it turns out the TLM100 has something called Analog Gauge Mode where it can be connected in parallel to an existing gauge. My presumption is that rather than putting out a constant current
and measuring the resulting sender voltage (which I think is how these gauges work), in the Analog Gauge mode the TLM100 just monitors the sender voltage produced by the original gauge and puts out no sense current
of it's own. Assuming the expected voltage is between 0V and 4V, then I should be able to use this mode of sensing in conjunction with the existing Centroid senders. I've got my fingers crossed.
By the way, does anyone know for sure how these gauges work? I know there are a couple of "standards" for the senders which are just variable resistance depending on fluid level. But how so the gauges work with the sender? As I said earlier, my best guess is that the gauge sources a constant current and measures the voltage across the sender. I'm further guessing that it sources 20ma which gives a voltage range of .66V to 4.8V across a 33-240 ohm sender. I've got my fingers crossed.
Anyone know for sure?