G3X Fuel Level Troubleshooting

I’ve been troubleshooting an odd issue with a G3X system equipped with a GEA24 EIS utilizing the stock Van’s resistance float fuel level sensors.

I’m hoping someone can share the voltages that register in the calibration page for full and empty for a similarly equipped plane.

By way of background, when I initially calibrated by RV-14A, my system derived the following data points:
Empty - 0.3 volts
Full - 2.21 volts

I calibrated every two gallons and, as expected, all of the data points followed the same trend: as the tank level increased, the measured voltage increased.

At some point between pre flight for the first flight and start up for the second flight, the system seems to have inverted the readings. Now, as the tank level increases, the voltage decreases. Curious. Very curious.

I’ve been in contact with Garmin and have done extensive troubleshooting including using a dummy circuits, a VOM, etc. Garmin generously provided a replacement unit under warranty, but that did not resolve the problem.

Please let me know the following for your system:
Calibrated voltage at 0 gallons
Calibrated voltage at 25 gallons
When fuel level increases, does the voltage measured on the calibration screen consistently increase or decrease?

Thanks in advance for sharing your information.

Lance

Resistance decreases as the arm moves up towards full, so voltage observed should increase as you add fuel and mine does this.

Suspect that you may have accidentally selected a parameter in the G3x that reverses the reading. Not sure if there is one, but suspect there might be, as Garmin would want to support as many configurations as possible. Also possible that you had some data corruption on the calibration table.

Everyone's voltage readings will be different, as the builder bends the float arm and therefore everyone see somewhat different readings at empty and full.

Larry

Solved. I think.

Thanks for the quick reply. That info is very helpful.

I think I’ve figured out why the readings are inverted. For the benefit of those that may experience this issue in the future, here’s my hypothesis.

In order to have sufficient gas on hand to complete the calibration process (25 gals), I made multiple trips to the self serve pump with Jerry cans to fill one wing tank to the tabs. I then used that tank as my supply and drained off 2 gallons at a time which I then would add to the opposite tank. So starting with 25 gals in the right tank, and 0 in the left, I would add a calibration point to the left tank table for 0 gals. I’d then draw off two gals from the right tank, add it to the left, and after the voltage stabilized, add another calibration point to the left tank curve. Rinse and repeat until the right tank is empty and the left tank is full. Then repeat the process in reverse, drawing from the left tank to fill the right.

The trap I fell into was that the inputs were apparently swapped in the GEA configuration table. While I was adding fuel to the left tank, I was actually reading the voltage for the right tank. I was seeing a voltage change but it was from the change in level from the opposite tank.

So I calibrated both tanks, ended up with reasonable looking calibration graphs and then, in preparation for first flight, split the gas between the tanks and felt assured that since both tanks read half full, I must have gotten it right.

Then, after the first flight, I topped off both tanks, and now they magically read empty.

The fix will be to drain a tank, swap designations for each input in the configuration page, and then recalibrate each tank.

Had I simply started with both tanks empty, I would have noticed the discrepancy immediately. The whole process to calibrate both tanks can take over an hour, and I didn’t want to make the fuel truck guy stand there for the first 30 minutes. Instead, to avoid those trap, go buy or borrow 25 gallons of Jerry can storage so you can start with both tanks empty.

It’s been said that a smart man learns from his own mistakes, but a wise man learns from the mistakes of others. Hope my mistake helps you to avoid making one of your own.

If you think your calibration curves are correct, but just assigned to the wrong inputs, you may not need to drain your tanks or do any recalibration. First try this:

After you fix up your wiring, Take a look at "Fuel Quantity Calibration Data Backup" in section 35.4.32.12.1 of your G3X installation manual (rev AN). Export your fuel calibration data to an SD card, then find the resulting files on the card. I forget what they'll be named but they should be easy to find. Swap the file names, then re-import the calibration data from the card. The result should be your original calibration data, just swapped between left/right tanks

Update

Turns out that pins 6 (Fuel Qty 1) and 9 (Fuel Qty 2) of J244 were, in fact, swapped by the shop that built the panel harness. I couldn't reassign the pins on the configuration page (pin 6 is hardcoded as Left Tank and pin 9 as Right) so I extracted the pins from the d-sub and switched their positions.

I then recalibrated each tank. YMMV, but this is how my system is working:

The fuel sensor resistance decreases as the fuel level increases (240 ohms with empty tanks decreasing to 34 ohms with full tanks). In "Resistance sensor" mode, the GEA24 reports the voltage drop across the sensor. As resistance decreases, voltage drop decreases.

Hence, as fuel level increases, resistance decreases, and the corresponding voltage decreases. You end up with a nice smooth downward sloping fuel quantity curve (going from left (low voltage) to right (high voltage). With the tanks empty, the GEA24 reads 2.196v; with the tanks full (@ approximately 22 gals, the fuel sensor reaches its highest point of motion), the GEA reads 0.300v.

One important call out: the illustration on Page 35-168 of the G3X Installation Manual shows a rising calibration curve with voltage increasing as fuel level increases. This is the exact opposite of my final curve and, I suspect, might be appropriate if the GEA24 is in a different sensor mode (voltage instead of resistance). My initial curve was very similar to the illustrated curve, which substantiated my incorrect confirmation of the first calibration results.

Given the criticality of proper fuel measurement to safety of flight, I'm hopeful that Garmin will add some additional guidance in their next iteration of the Install Manual. Until then, I'd suggest you calibrate the first tank with both tanks empty to avoid the trap I fell in.

Looks like all would have been correct from the start if your right and left pins were correct from the start. My pins were also initially backwards because I didn’t know what side should be input 1. Give a smart guy a 50/50 and it’s wrong 90 percent of the time.