G5 altitude calibration

[Andy]

The G5 EFIS that I fitted to my FX-3 initially read 40 ft higher altitude than the GSU 25 altitude displayed on the GDU 465. After about 2 years the difference had increased to 60 ft and all the error seems to be in the G5.

I mentioned this to a friend who works on multiple aircraft and he commented that he had seen other G5 with poor altitude calibration.

If you have a G5 installed, what difference do you see between GDU and G5 when both have the same baro setting?

The 60 ft error bugs me but there is no way to enter a correction offset. Baro settings cannot be split when the G5 is fully integrated with G3X. The G5 must be calibrated at sea level, 10,000 ft, 20,000 ft, and (optionally) at 30,000 ft. That requires paying someone with an approved air data test set or designing and building my own test set. (That project is underway and I'll report if it works.)

[kshunz]

My G5 initially showed about 20' higher than on the G3X, and after about 18 months it's now about 30' higher.

[Andy]

My G5 initially showed about 20' higher than on the G3X, and after about 18 months it's now about 30' higher.

Thanks for the report which seems to agree with other comments on G5 altitude error that I found in a web search.

I'm making progress with my air data test set. I have the basic sensor interface working and displaying on an LCD. Next step is making a vaccum tight enclosure for the sensor so it reads instrument static rather than ambient pressure.

[ve6yeq]

In preparation for delivery of my CarbonCub, I have been flying a Super Decathlon that has both a Garmin G3x (front seat) and a uAvionics AV-30 (back seat). Even after multiple calibrations, it has not been possible to get both to match and the drift between them appears to vary slightly depending upon the day. I have not looked into the pressure transducer specifications to see how much variation is just inherent in the sensors, piping, etc.. By the sounds of this thread, I will be looking into this once I take delivery of my CarbonCub (G3x + G5)…

[Andy]

The 60 ft error bugs me but there is no way to enter a correction offset. Baro settings cannot be split when the G5 is fully integrated with G3X. The G5 must be calibrated at sea level, 10,000 ft, 20,000 ft, and (optionally) at 30,000 ft. That requires paying someone with an approved air data test set or designing and building my own test set. (That project is underway and I'll report if it works.)

My protoype air data test set matured to the point that I attempted a G5 static sensor calibration. I first did a check to see if the G5 showed similar errors relative to the test set as it did relative to the G3X:

test set pressure altitude / G5 indicated pressure altitude

1315 / 1370 (ambient)
0 / 50
5000/ 5055
10000 / 10065
15000 / 15070
20000 / 20080
25000 / 20080
30000 / 30105

I then calibrated the G5 at 0, 10k, 20k and 30k holding the test alttitude within 1.5 ft at sea level, 4 ft at 10k ft, 5.5 ft at 20 k, and 6 ft at 30k.

After calibration the G5 indicated error was as follows:

test set pressure altitude / G5 indicated pressure altitude

30000 / 30010
25000 / 25010
20000 / 20005
15000 / 15000
10000 / 10000
5000 / 5000
1351 / 1350 (ambient)

Flight test showed the G5 read 10 ft higher than G3X from 1500 to 3,500 ft. I don't know if this difference is because my uncalibrated test set is in error or because the G3X (GSU 25C) is in error.

I'm pleased to have reduced the annoying 60 ft difference to an acceptable 10 ft difference but I may perform additional testing to get the difference closer to zero.

[kshunz]

I've used surgical tubing and a jug of water to test for static leaks, but that's as close to a DIY static test set as I figured I would get. I'm sure I wouldn't be the only person interested, if you were willing to share some details of your design.

[Andy]

I'm sure I wouldn't be the only person interested, if you were willing to share some details of your design.

Not really a design yet, more a proof of concept. The overview is:

1. A pressure sensor - I used a Chinese board with a Bosch (or it may be a clone) BMP390 sensor. I have 2 Adafruit sensor boards that are more likely to use genuine Bosch parts that may be used for the next version.

2. An airtight enclosure for the sensor that makes provision for a pressure port and leak free connection of the 6 wires of the sensor interface.

3. An Arduino processor and display module

4. Code for the Arduino processor

5. A pneumatic system that allows the pressure supplied to the sensor and the test article (G5) to be varied between sea level and 20,000 ft (optionally 30,000 ft).

The only parts that are available for purchase are the sensor board (1) and the Arduino processor and display (3) everything else was created or improvised at home. The pneumatic system includes 2 Dill Pickle jars used as vacuum reservoirs; a mountain bike gas shock pump to push the altitude down to sea level; an automotive hand vacuum pump to set altitudes higher than ambient; a hose clamp to act a variable bleed for setting test altitudes.

The sensor board enclosure was made from a piece of plastic cutting board with a cavity milled out for the sensor and a closure plate with rubber gasket. Wire feedthrough was improvised with a 0.1 inch pitch header with the pins epoxy sealed.

I started with the wild idea that I could fully automate the altitude test profile using a computer controlled electric vacuum pump. I found an inexpensive pump that would pull down to 20,000 ft but decided it would take far too long to create a design that could automatically set the altitude test points. I found out that even $5,000 commercial test sets control the altitude by manual adjustment of bleed valves.

By now you are probably realizing that it's not something that could be copied easily. Perhaps the idea will inspire someone to do it better.

[Andy]

Flight test showed the G5 read 10 ft higher than G3X from 1500 to 3,500 ft. I don't know if this difference is because my uncalibrated test set is in error or because the G3X (GSU 25C) is in error.

I took the prototype, without the pneumatic system, to the hangar yesterday and did a comparison of pressure altitude readings for the 4 sensors, the G5, and the G3X (GSU 25C). Based on that comparison the two Adafruit sensors seem to be the most accurate, the G3X seems to be 10 ft high, and the G5 is now 10 ft higher than the G3X. The enclosed sensor that was used to calibrate the G5 seems to be the least accurate but it did take some abuse during development.

If I used a perfectly accurate sensor to calibrate the G5 I think the G5 would read 10 ft lower than the G3X (GSU 25C) instead of 10 ft higher. It's going to stay like it is for a while but I may take another look at it when I install the new -55 GSU.