What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

G3X CAN bus not working at all

iamtheari

Well Known Member
I finished wiring my RV-14's G3X system and attached the wings today. The RV-14 wiring harness includes CAN bus termination in the right wing at the aileron autopilot servo and in the tail at the elevator autopilot servo. I thought that I had been conscientious during my wiring, and as far as I can tell all the other wires are working. RS-232 from the PFD to the EIS, to the ADAHRS, and to the GTN 650 all work okay. RS-232 from my Vertical Power VP-X Pro to the MFD works.

Unfortunately, I have zero success with the CAN bus. The PFD does not see anything else on the bus. I did check resistance between the HI and LO pins on one connector and got 60 ohms, which is correct for a 120-ohm terminating resistor at each end, in parallel. I tried powering on just the PFD and a handful of other devices, one at a time (tried GTR 20, GMA 245R, G5, MFD, and autopilot servos) to no avail.

Before I (a) pull my hair out or (b) pull all the wires out, I'm wondering if anyone has some troubleshooting tips that I can try on an incremental basis to find and solve the problem. I know I'm not the first person to go down this road so I hope to hear some reassurance this weekend. Thanks in advance!

P.S. Even if I have to rewire the whole thing, I'll still be on cloud 9: we moved the fuselage to the airport and bolted the wings on today. It's never looked more like an airplane than it does tonight!
 
Maybe this will help... G3X can bus troubleshooting from Garmin
 

Attachments

  • G3X_CAN_Bus_Troubleshooting.pdf
    110.4 KB · Views: 336
Check for ground

I had a serious headache with my can bus. During wiring, a sharp blob of solder poked through and grounded out the can bus under a strain relief clamp. It was maddening to locate the issue as it was intermittent. I split the can bus about in the middle and checked for ground on each side and repeated until I found the culprit.
 
**Disclaimer** I'm not an avionics guru.

You mentioned you used the stock wiring harness - Are you using the GMU11? If so, how did you handle that since the stock harness IS STILL being made around the GMU22?
 
**Disclaimer** I'm not an avionics guru.

You mentioned you used the stock wiring harness - Are you using the GMU11? If so, how did you handle that since the stock harness IS STILL being made around the GMU22?

I used the GMU 22, mounted in the left wing where the RV-14 has a provision for either a Dynon ADAHRS or Garmin round magnetometer to be mounted.
 
Wires..

I would start by checking the can hi and lo in the PFD, and make sure they are in the correct pin #’s by counting from the edge of the row (sometimes the stamped numbers are left of center on the sub holes, and sometimes right of center, always count. If correct, ohm from can high to all other can high pins, and same for low. I’ll bet you find one mis-pinned. The can bus can be solid, just off a few slots in one connector!
 
Thanks for the replies. My MFD was easier to unplug so I started with that. The CAN low pin reads 62 ohms to ground. The CAN high pin reads 2 ohms to ground. My takeaway is that my CAN high wire is shorted to ground somewhere along the chain.

I appreciate the tip to check the pins at the PFD. I'll start with that connector as I try to trace where the wiring fault lies.

My plan is to confirm that the pins are in the right places in the connector and then try to strip back the heat shrink over the solder joint on the CAN high wire so I can de-solder it and pull the wires apart. Once I identify which side of that connector the problem is, I'll re-do the joint and move to the next device in that direction.

Here's a question that came up while I was scratching my head. I configured my GEA 24 sensors on the PFD display. They are all faulted out. The backup RS-232 connection to the GEA 24 appears to be working. My guess is that the PFD can't transmit the configuration to the GEA 24 so it doesn't know what sensors are attached, and that's why the engine gauges are just a bunch of red X's. Does anyone know if that's a fair guess, or if I should be tearing apart my GEA 24 connectors again? (The first time was because I had used unshielded wire for the sensors where Garmin recommends shielded.)
 
I would concentrate first on the shorted wire. Work with the known issue and solve it and then see if the other problem is there. I would disconnect every LRU connector that has Can bus and see if short goes away. Also check each connector for a bent pin. Might have bent a pin while you were connecting one of them.

Jay
 
CAN bus Troubleshooting

Thanks for the replies. My MFD was easier to unplug so I started with that. The CAN low pin reads 62 ohms to ground. The CAN high pin reads 2 ohms to ground. My takeaway is that my CAN high wire is shorted to ground somewhere along the chain.

I appreciate the tip to check the pins at the PFD. I'll start with that connector as I try to trace where the wiring fault lies.

My plan is to confirm that the pins are in the right places in the connector and then try to strip back the heat shrink over the solder joint on the CAN high wire so I can de-solder it and pull the wires apart. Once I identify which side of that connector the problem is, I'll re-do the joint and move to the next device in that direction.

Here's a question that came up while I was scratching my head. I configured my GEA 24 sensors on the PFD display. They are all faulted out. The backup RS-232 connection to the GEA 24 appears to be working. My guess is that the PFD can't transmit the configuration to the GEA 24 so it doesn't know what sensors are attached, and that's why the engine gauges are just a bunch of red X's. Does anyone know if that's a fair guess, or if I should be tearing apart my GEA 24 connectors again? (The first time was because I had used unshielded wire for the sensors where Garmin recommends shielded.)

It sounds like you are on the right track. Typically (but not always) if there is a single pin incorrectly assigned somewhere, you will still see some of the CAN bus LRU's come and go offline. For the entire bus to be crashed as you describe, the CAN bus wiring is usually shorted to ground somewhere.

You will want to get the CAN bus up and running before further troubleshooting your GEA 24. Feel free to give us a call at 1-866-854-8433 if you continue having trouble getting the CAN bus up and running.

Thanks,

Justin
 
Thanks, Justin. Given that the CAN high wire is definitely shorted to ground, I do suspect that's the only problem I have. So unless you guys have some magical device that can isolate a segment of wire without cutting it, I don't think I'll need to call you up just yet. If you do have such a device, please let me know how much it costs to rent and where to send my check. :)
 
That magical device is called a TDR. Time domain reflectometer. The new digital ones can find an open or a short within an inch. Don’t know where you can find one to borrow.
 
That magical device is called a TDR. Time domain reflectometer. The new digital ones can find an open or a short within an inch. Don’t know where you can find one to borrow.

Good thought but the TDR would show a fault at every node.
 
I had a similar problem with one node shorted to ground on my CAN bus. Even after identifying it I couldn't even see the short under a magnifying glass, so a visual inspection might not find the problem.

The easiest (not easy, but easier than other methods that I am aware of) approach is to cut out a node at about the middle of your CAN Bus. Then use your meter to determine which side of that node is still shorted. (If you find that neither side is now showing a short, you got very lucky and found the shorted node on the first try!) Next cut out a node in about the middle of the side where the short is still indicated, and again use your meter to determine which side of that run is still shorted. Continue this process until you have isolated the shorted node. In selecting which nodes to cut out during this process, pick nodes that are close to the middle of the run, but are also the easiest to get to since you will need to replace them, as that can help reduce the number of difficult nodes to replace that you will cut out.

It isn't a fun task, as you will need to replace each node cut out. But this will typically allow you to isolate your bad node with the least number of cuts, as it narrows down the possible number of shorted nodes by half with each cut.

Good luck
 
Good thought but the TDR would show a fault at every node.


Couldn't you use it between each node to rule out a short in the harness? Plug it in at the magnetometer in the wing tip and it should show good until the roll autopilot servo. Plug it in at the autopilot servo and it should show good until the next node down the line. It would rule out a harness ground fault, which means you could then start disconnecting things one at a time until the problem went away.
 
I’ve made some progress in tracing the shorted CAN wire. Unfortunately, my wire twisting skills were too good and my desoldering skills too weak, so I ended up cutting the leads more often than desoldering them. That leaves me with a re-assembly question:

How sensitive is the G3X CAN bus to a difference in the length of the high and low wires between nodes? If I do not cut the low side and just strip back what I have on the high side, I estimate that each node will have a 1/4” shorter run between nodes and a 1/4” longer node length on the high wire than on the low wire. Am I going to regret that?
 
I’ve made some progress in tracing the shorted CAN wire. Unfortunately, my wire twisting skills were too good and my desoldering skills too weak, so I ended up cutting the leads more often than desoldering them. That leaves me with a re-assembly question:

How sensitive is the G3X CAN bus to a difference in the length of the high and low wires between nodes? If I do not cut the low side and just strip back what I have on the high side, I estimate that each node will have a 1/4” shorter run between nodes and a 1/4” longer node length on the high wire than on the low wire. Am I going to regret that?

In my opinion a 1/4" won't make any difference. The CAN bus is not that high speed.
 
I am 38 years old. On my next birthday, I will turn 40.

At least that's what I thought on September 26, when I wired the CAN bus to the DB50 connector. The roll servo is on pins 34 (high), 35 (low), and 36 (shield). The yaw and pitch servos are on pins 40 (high), 41 (low), and 42 (shield). I left 2 open pins between those groups instead of 3, so I ended up with my CAN low connected to the high on the pitch and yaw servos and my shield connected to the CAN low on them. Meanwhile, my CAN high at one end was connected to pin 39, which is the shield for the RS-232 autopilot data (unused in my installation). Thus the high wire was grounded through that shield.

I wish I had checked that before I disconnected 8 CAN high pigtails that I now have to re-connect in the airplane without being able to stand next to it since the wings are on. But the mystery is solved and it's only a matter of putting things back together so I can test and see if anything else is not working.

It's also a convenient time to wire in my taxi lights, OAT, and pitot heat wires that are now attached courtesy of the wings being on. So it's not all bad.

Happy building, everyone!
 
Glad you found the problem. For others that might read this later a more efficient way to troubleshoot would be to pull the connector off a unit, then a second unit. Place uncrimped pins, or pins with test leads as is convenient, in the correct location of the connector on each unit and continuity test from device to device. Continue on to each in turn. One will likely be pinned wrong, and this one gets fixed.

This method saves you the hassle of pulling pins or messing with the wires until after the fault location is identified. You could have used a pin removal tool to pull the pin and reset to the right hole. Key is to use the extra pins you have to make nice test points instead of jamming a multimeter into the connector on either side.

It is not common to have a ground on a bus when the wiring is all new. There is a presumption that the pins get checked for that problem prior to inserting in the shells. I check for that at any rate.... I'm also super careful to make sure every power pin is NOT grounded and every ground pin IS.

Same principle might apply to testing the can bus as it is assembled. At the end of the bus take a pin with a 100 resistor connected to ground on the hi line. Take a pin with a 50 resistor to ground on the low line. Every time you make a new splice the hi pin will be 100 to ground and the low 50 to ground. After all your connectors are correctly set, pull the test leads and connect all to their instruments.

You could do the same thing with some of the bus wires connected to instruments but I think that the results are more variable, and perhaps you would want lower ohm test leads like 10 and 20 ohm. But the principle could work. I prefer testing the harness without connections to instruments.
 
Last edited:
I survived the re-connection process and my CAN bus works now. Hooray for that.

But there’s a mystery. My high-to-low resistance is 60 ohms as expected with both ends terminated. However, the resistance to ground is not an open circuit. Rather, my multimeter reads 12.5Kohm on the 20K range, 61.3Kohm on the 200K range, and 472Kohm on the 2M range. It reads open circuit on both the 200 ohm and 2K ranges.

Is that somehow a function of the units all being plugged into the bus and powered off? Or did I forget some kind of quantum mechanical tool attached to the bus?
 
I survived the re-connection process and my CAN bus works now. Hooray for that.

But there’s a mystery. My high-to-low resistance is 60 ohms as expected with both ends terminated. However, the resistance to ground is not an open circuit. Rather, my multimeter reads 12.5Kohm on the 20K range, 61.3Kohm on the 200K range, and 472Kohm on the 2M range. It reads open circuit on both the 200 ohm and 2K ranges.

Is that somehow a function of the units all being plugged into the bus and powered off? Or did I forget some kind of quantum mechanical tool attached to the bus?

This is normal. Each device on the CAN bus has a few high impedance paths to ground.
 
Back
Top