Scary G3X failure mode

[PT20J]

I was just perusing the letters to the editor in the June AOPA Pilot and saw one where a reader described a G3X failure where red X's started appearing on one after another G3X instrument until nothing was available. The problem turned out to be a bad GEA 24 (engine monitor) that swamped the CAN bus shutting down communications between the G3X and the LRUs. This was on a SLSA that didn't have a standby G5 or GI 275, but presumably those instruments might have been affected also.

In my installation, I have a AV-20-S for a backup attitude indicator and it is not connected to anything common to the Garmin equipment except for pitot and static sources. I felt it important to have some non-Garmin technology onboard to keep me upright while I troubleshoot a catastrophic failure in the primary equipment. In this case, I would pull all the LRU circuit breakers and then bring them back on line one by one to isolate the bad LRU. It should be possible to get most everything working again. 

[N201MKTurbo]

I have been doing way to much work with the CAN bus lately. Using both the CANOpen protocol and proprietary CAN protocols on the same bus. If Garman uses CANOpen, it would be possible for any node to tell another node to change to the stopped operational state. This could be done if any node started swamping the bus. This would cause the node to stop sending PDOs (Process Data Objects) which are sent automatically both periodically and when a value changes. 

[PT20J]

2 minutes ago, N201MKTurbo said:

I have been doing way to much work with the CAN bus lately. Using both the CANOpen protocol and proprietary CAN protocols on the same bus. If Garman uses CANOpen, it would be possible for any node to tell another node to change to the stopped operational state. This could be done if any node started swamping the bus. This would cause the node to stop sending PDOs (Process Data Objects) which are sent automatically both periodically and when a value changes. 

Interesting. I assume that this is done in the interface chips so that there is a minimum of circuitry involved. But in the reader’s description the GEA 24 apparently did take the bus down so maybe Garmin doesn’t support this. 

[N201MKTurbo]

20 minutes ago, PT20J said:

Interesting. I assume that this is done in the interface chips so that there is a minimum of circuitry involved. But in the reader’s description the GEA 24 apparently did take the bus down so maybe Garmin doesn’t support this. 

It would take a software change in one or all the nodes to detect another node is going rogue and shutting it down.

[Rick Junkin]

21 hours ago, PT20J said:

In my installation, I have a AV-20-S for a backup attitude indicator and it is not connected to anything common to the Garmin equipment except for pitot and static sources. I felt it important to have some non-Garmin technology onboard to keep me upright while I troubleshoot a catastrophic failure in the primary equipment.

I did the same, a carry over of the “peanut gauge” from other airplanes. That reminds me, I need to practice an iPhone/peanut gauge approach next time I’m out with a safety pilot. I haven’t done that yet.

Thanks for the head’s up on the failure mode Skip. I hadn’t seen that letter yet.

Cheers,
Junkman

[Stanton R]

The certified G3X touch has a dedicated backup RS232 for the airdata GSU25 and the GEA 24. Canbus failure should only kill the heading, auto pilot and most of the navigation information coming from the GAD 29. The non certified G3X the Rs232 is optional. 

Edited May 15 by Stanton R

[Mooney in Oz]

3 hours ago, Stanton R said:

The certified G3X touch has a dedicated backup RS232 for the airdata GSU25 and the GEA 24. Canbus failure should only kill the heading, auto pilot and most of the navigation information coming from the GAD 29. The non certified G3X the Rs232 is optional. 

Am I assuming correct that the G5 should survive and not be taken out?

[PT20J]

4 hours ago, Stanton R said:

The certified G3X touch has a dedicated backup RS232 for the airdata GSU25 and the GEA 24. Canbus failure should only kill the heading, auto pilot and most of the navigation information coming from the GAD 29. The non certified G3X the Rs232 is optional. 

You’re right. I verified it from the documentation. Thanks for pointing that out.

[EricJ]

1 hour ago, Mooney in Oz said:

Am I assuming correct that the G5 should survive and not be taken out?

The AI and air functions shouldn't be disturbed (one hopes) since it has everything it needs to keep that working, including a GPS receiver for alignment, internally.    The GMU11 magnetometer is on the CANbus, so if the bus goes down the HSI may go offline since it may no longer have magnetometer alignment.

What they'd actually do may be TBD, though.

[PT20J]

2 hours ago, Mooney in Oz said:

Am I assuming correct that the G5 should survive and not be taken out?

Actually, the more I look at this the more I think Garmin thought out the failure modes for the certified system pretty thoroughly. The G5 has its own ADAHRS, so it doesn’t need input over the CAN bus from the GSU 25D, and the G5 has an RS-232 MapMX connection to the GPS. So, if the CAN bus goes down, it will lose the GMU 11, but it should then revert from HDG to GPS TRK.

[EricJ]

1 hour ago, PT20J said:

Actually, the more I look at this the more I think Garmin thought out the failure modes for the certified system pretty thoroughly. The G5 has its own ADAHRS, so it doesn’t need input over the CAN bus from the GSU 25D, and the G5 has an RS-232 MapMX connection to the GPS. So, if the CAN bus goes down, it will lose the GMU 11, but it should then revert from HDG to GPS TRK.

G5s also have an internal GPS receiver and antenna, so they don't even need the RS-232 feed as long as the unit gets sufficient signal with the internal antenna at its panel location.    I've confirmed that mine does, just sitting in the panel.   If not, it's easy to add an external antenna and put it someplace where it gets sufficient signal (e.g., on top of or just under the glareshield).

 

[Mooney in Oz]

4 hours ago, EricJ said:

The AI and air functions shouldn't be disturbed (one hopes) since it has everything it needs to keep that working

 

3 hours ago, PT20J said:

The G5 has its own ADAHRS, so it doesn’t need input over the CAN bus from the GSU 25D, and the G5 has an RS-232 MapMX connection to the GPS.

Thanks for the responses. That is just what I wanted to read.

I have the G5 as a backup AI to the G3X as per the STC. I don’t have a 2nd G5 as a HSI.