Fuel Gauges

[Tom 4536]

Has anyone been able to get the factory fuel gauges to work with a JPI 900. On my new JPI 900 engine monitor the fuel gauges always read full even when the tanks are less than half full. The installer said he ran the required JPI calibration procedure; I believe him because the tanks were full to the filler rim when I picked up the plane. Occasionally the gauge will read close to correct for a second but then reads full 99% of the time. Any ideas will be appreciated.

[tmo]

Are you sure you have the JPI configured for the right type of gauge (resistive)?

[Steve W]

They can work correctly. The Mooney uses 2 fuel sensors in series per wing, if they messed up the wiring and only used the inboard sensor then this could be a problem.

I'd get the installation manual and read through the calibration section, with it full go in and see what the raw readings and the table is set to are and then once you've burned off some fuel go into the settings again and see what the numbers are. As I recall the raw numbers should roughly be linear. 

[carusoam]

Depending on the age of the JPI...

It can be changed between resistance, or frequency...

Older JPIs go back to the factory to be changed... newer versions can be changed in the field...

Check that the floats are working, supply variable resistance...

PP thoughts only, not a mechanic...

Best rewards,

-a-

[shawnd]

I have been researching this topic recently for my own airplane. Do you have the stock fuel senders? Were they ever overhauled? JPI 900/930s work better(?) with digital fuel senders like those from Cies. At the least, you should be able to empty out the tanks and re-test what values the senders are reporting by adding known quantities of fuel incrementally. Hopefully that will help isolate the issue.

[Tom 4536]

I have a new JPI 900 and the original fuel tank sending units. The installer set up for the 4 sending units and I know the calibration was done because the installer said he did the calibration and the tanks were completely full when I picked up the plane. Since my original post I have found out that the sending units should usually be rebuilt or replaced when installing an engine monitor. I will be installing new Cies digital fuel sending units.

[shawnd]

Cool, I was told the same by MSC near me. Good luck!

[Eduleo]

I did not install CIES senders when I did my jpi900 install back in April. I've found that the senders are accurate within a gallon at the fuel pump and the fuel totalizer on the jpi (after adjusting the K factor a couple of times). I've flown 150 hours that way.

[ArtVandelay]

I used the factory senders, I had to overhaul both inboard senders. The trick is as the analog senders can be erratic with its readings, which doesn’t affect old analog gauges but digital sees the transient readings. If you pull the senders and test with a ohmmeter, look for consistent readings as you slowly move the armature, any jumps means they need to be serviced.

 

It’s about $170 to overhaul, or Cies are about $450, if you go with capacitance version the JPI needs to go back to the factory and you need to replace all senders. In my case that was 4x$450, I chose to stay with the factory senders.

 

Inboard senders spend all the time submerged, outboard units don’t unless you refill after every flight, so inboard units usually fail more often.

[mhrivnak]

I just went through this with an owner-assist install in my 1990 M20J. I used the factory senders, which are resistive.

We leveled the airplane, drained the tanks, and then carefully re-filled. At least twice, one of the arms was stuck enough that the value didn't rise hardly at all after adding 8 gallons, but just giving the wing a nice shake got it to free up and read correctly. It's possible that if something like that happened and he didn't catch it, your 3/4 and full readings might have actually been taken with the arms stuck closer to the 1/2 reading mark.

I agree with the suggestion of getting the install manual (it's on JPI's website), reading about how to get into the fuel level settings, and look at the table of values that are entered. (take a picture while you're there in case you ever need to restore those values) See if they look funny at all. In my case going off memory, I had readings on each side that were roughly ~15. ~450, ~650, ~850, ~1100. Not a linear progression, but significant increases for each reading. You might have very different numbers, and that might be expected, but I'd guess you should see some kind of sensible progression. Speculating as a total amateur, it would make sense if the factory gauges just indicate the "raw" reading, so it may be reasonable to expect a progression of values that wouldn't be wildly inaccurate if indicated with a needle.

One other long-shot thought; the factory wiring (on my J at least) only has a single wire run from the inboard sensor up to the gauge. But JPI insists that you run two conductors to each sender, one of them a ground, and attach the ground as close as you can to the sender. It's tempting to just re-use the factory wires rather than run new (it's a pain just getting the old wiring out), but I have no idea what symptom you might see if he took that shortcut.

If it comes to it and you want to lay eyes on the wiring, especially to see if the outboard sender is still connected, the inboard sender on the copilot side is fairly easy to get a look at just by removing the metal guard at the base of the door opening.