Special CANBus cable or just M27500 2-conductor??

[tdragger1966]

I came across this in a couple of listings for a specialized Carlisle CANBus cable:

"GARMIN documentation recommends the use of this CAN24TST120 (CIT) cable to interface the CAN Bus on G3X..."

I can't find this anywhere in the Garmin install docs and even the Garmin how-to videos seem to show standard M27500 shielded cable. At ~5x the cost is this cable really necessary? What did y'all use for CANBus wiring?

 

[ColoradoSolar]

I just used standard M27500 cable

 

[JoopSJ]

Steinair sells MIL-Spec Tefzel wire.
https://www.steinair.com/product/22-ga-2-conductor-shielded/
Very good quality.

You will also need the 3 conductor for Rs232 connection.

 

[fl-mike]

I’m using EN2714-013B cable. Lighter weight and spiral shield. While not quite as good as braided, it is SO much easier to work with and smaller.
This is what Airbus uses on their aircraft.
I got mine from Aircostcontrol.

EN 2714-013B is pair
EN 2714-013C is triad

 

[mfleming]

So to answer your question directly.

Maybe

If your Can Bus run is less than 66ft then the 22-2S will work just fine but if your total run exceeds 66ft, then yes, you’ll need the expensive stuff.

I’m no expert, just repeating what Garmin told me when I asked the same question to them.

PS- 66’ is easy to exceed.

 

[tdragger1966]

So to answer your question directly.

Maybe

If your Can Bus run is less than 66ft then the 22-2S will work just fine but if your total run exceeds 66ft, then yes, you’ll need the expensive stuff.

I’m no expert, just repeating what Garmin told me when I asked the same question to them.

PS- 66’ is easy to exceed.

Gotcha. Thanks! My plan is to run from AP roll servo in right wing to avionics in the panel to GMU11 in left wing and finally terminate at AP pitch servo in empennage. So, yeah, might get up to 66. Could probably have saved some run by mounting the GMU in the tail also but worried about the 2 AP servos and who knows what else that might interfere. Need to get the tape measure out...

 

[mfleming]

Gotcha. Thanks! My plan is to run from AP roll servo in right wing to avionics in the panel to GMU11 in left wing and finally terminate at AP pitch servo in empennage. So, yeah, might get up to 66. Could probably have saved some run by mounting the GMU in the tail also but worried about the 2 AP servos and who knows what else that might interfere. Need to get the tape measure out...

Yes, a lot of people don't realize how much length is added when all the ups, downs and turns are added up. I was told there's no way my -7 Can Bus should add up to 61' but it does and there's little or no room to trim it down

 

[RVbySDI]

Gotcha. Thanks! My plan is to run from AP roll servo in right wing to avionics in the panel to GMU11 in left wing and finally terminate at AP pitch servo in empennage. So, yeah, might get up to 66. Could probably have saved some run by mounting the GMU in the tail also but worried about the 2 AP servos and who knows what else that might interfere. Need to get the tape measure out...

there are a lot of G3X systems with the GMU11 terminating in the tail. They don’t seem to be talking about issues with that installation location.

 

[Walt]

If you put the GMU11 either out in the tip by the roll servo or on the back deck you can just extend the CAN bus a few feet from the servos to reach it.

 

[1001001]

I came across this in a couple of listings for a specialized Carlisle CANBus cable:

"GARMIN documentation recommends the use of this CAN24TST120 (CIT) cable to interface the CAN Bus on G3X..."

I can't find this anywhere in the Garmin install docs and even the Garmin how-to videos seem to show standard M27500 shielded cable. At ~5x the cost is this cable really necessary? What did y'all use for CANBus wiring?

I saw this on Steinair as well...I don't even know how you'd use that wire to hook up the CANbus because the G3X documentation and pinouts only want two conductors and a shield. That Carlisle cable appears to be four conductors (white, blue, pink, pink) and a shield. Seems like a waste to me.


Edited to add: my estimate for CANbus in the RV-10 based on some quick work with a tape measure is (starting in the right wing at the roll servo) 8 + 6 + 10 + 6 + 12 + 12 + 10 = 64 ft. That goes from right wing to panel to left wing tip, back to the left wing root and back to the yaw and pitch servos. I think there's plenty of extra in there for service relief and runs between bits in the panel area.

 

[N804RV]

Just use twisted pair. Steinair sells it for less than $1/ft. I believe Dynon uses unshielded twisted pairs on their Skyview network, seems to work just fine.

But, If you're worried about RF, then buy the shielded twisted pair. Steinair sells the shielded 2 twisted pair (4 conductor) stuff spec'd by Garmin for $6.65/ft.

 

[mfleming]

Just use twisted pair. Steinair sells it for less than $1/ft. I believe Dynon uses unshielded twisted pairs on their Skyview network, seems to work just fine.

But, If you're worried about RF, then buy the shielded twisted pair. Steinair sells the shielded 2 twisted pair (4 conductor) stuff spec'd by Garmin for $6.65/ft.

Dynon and Garmin use different architectures for their systems and are not comparable.

Garmin requires twisted and shielded cable for their canbus systems.

Edit: $0.80/ft from Steinair. This is what you want unless your canbus run goes over 66’, then you may need the expense stuff.

 

[jliltd]

I use Gigaflight twisted pair 120 ohm cable for CAN Bus. Garmin approved and available at Spruce. CAN Bus cable should only have two conductors and a shield so I think reference to the 4-conductor Carlisle cable above is confusing the Garmin HSDB Ethernet cable used for high speed databus with the 120 ohm CAN Bus cable.

I choose to use the quality 120 ohm cable because (1) my time and effort fabricating and installing the harness is worth more than the cost savings. (2) Same for troubleshooting CAN Bus errors after installing in the airplane. (3) when flying a glass flight deck for primary instrumentation why wouldn’t I want the best stuff recommended by Garmin? One day you might find yourself in IMC when a solar flare event hits or are near White Sands restricted area when they are playing with EMI pulses. And (4) maybe a few hundred hours down the road someone might want to upgrade to the latest offering at the time (“G5X”?) or maybe add a couple of gee whiz LRUs that will add length to the CAN Bus or be more sensitive to EMI. As the years go by it seems everything (and everyone) gets more sensitive about outside influences.

In fact I will call the above and raise you by adding more length of fancy cable. Contrary to human nature and its inherently frugality, most smart shops learned early on that the more expensive a wire or cable the more important it is to not skimp on the service loop. In fact add an extra foot when it comes to the pricey stuff. Why? Because the only thing more expensive than an extra foot or two is coming up short after pulling a harness through the whole airframe. Like a haircut you can always cut it shorter but not add any back. Coming up short renders the whole first run of expensive stuff useless for that install since these signal wires can’t be spliced ($$). Then add in the labor and time expended in replacing it ($$$) and the error cost is compounded. Nobody has time for that. The extra $6 to $12 of snazzy cable up front is a cheap insurance premium.

 

[1001001]

But, If you're worried about RF, then buy the shielded twisted pair. Steinair sells the shielded 2 twisted pair (4 conductor) stuff spec'd by Garmin for $6.65/ft.

The problem is that nowhere in their published documentation for the G3X does Garmin specify that cable.

They do say to keep the bus length shorter than 20m (66ft) and to use

Wiring used for the CAN bus should be shielded twisted-pair cable, MIL-C-27500 or equivalent. 22 AWG
or larger wire is recommended for physical robustness and ease of installation.

(Section 2.3.1.3.2 of the G3X Installation manual, latest revision)

 

[Longez]

The problem is that nowhere in their published documentation for the G3X does Garmin specify that cable.

They do say to keep the bus length shorter than 20m (66ft) and to use (Section 2.3.1.3.2 of the G3X Installation manual, latest revision)

Most EAB owners don't need or want to use the more expensive 120 ohm CAN bus cabling, so it isn't mentioned in the installation manuals for non-certified aircraft.

If you want to see the call out for this wire used in the generally larger certified aircraft G5 and G3X Touch installation manuals, you can download the G5 Installation Manual for Certified Aircraft (page 26/27 for Rev. 28 manual).

Steve

 

[1001001]

Most EAB owners don't need or want to use the more expensive 120 ohm CAN bus cabling, so it isn't mentioned in the installation manuals for non-certified aircraft.

If you want to see the call out for this wire used in the generally larger certified aircraft G5 and G3X Touch installation manuals, you can download the G5 Installation Manual for Certified Aircraft (page 26/27 for Rev. 28 manual).

Steve

Thank you for that!

Here's what I found in there...

There are two types of wire approved for CAN bus installations as part of this STC, 120Ω CAN cable or
shielded twisted-pair cable, MIL-C-27500 (see Section 3.2.1).
It is recommended to use the 120Ω CAN cable for the following scenarios:
• If CAN bus lengths are expected to reach the maximum allowed. (see section 3.4.5.2)
• If there are future plans to install Garmin STC’s SA01866WI, Installation of Garmin GFC 500
Autopilot with Electronic Stability and Protection or SA01899WI, Installation of Garmin G3X
Touch Electronic Flight Instrument System
If none of the above mentioned scenarios are a factor in the installation, use of shielded twisted-pair
cable, MIL-C-27500 is acceptable.

I haven't been able to find the Part 23 STC installation manuals online, but I did have a friend who runs an avionics shop provide me with the G3X STC information and it says:

The electrical architecture of the CAN bus takes the form of a linear “backbone” consisting of a
single twisted wire pair with an LRU connected (terminated) at each end. The installer should
attempt to make this “backbone” as short as practical. The maximum overall length of the CAN
bus from end to end must not exceed 100 feet.

3.7.2.2.2 Aircraft with Existing CAN Bus Wiring
If the aircraft has already existing Garmin LRUs connected to a CAN bus, some of this existing
CAN bus wiring may be retained. Use the following guidance to determine if the wiring can be
retained:
a. Measure the current length of the CAN bus wiring
b. Determine the part number of the CAN bus wiring
Revision 7c. If the wire part number is M27500, only 60 FT of the existing wiring may be retained with
the installation of the G3X. Shorten length of installed wiring as necessary to support the
installation of the G3X system.
d. If the wire is one of the two listed CAN wire part numbers in Section 3.1.2, then all the
wiring may be retained

3.1.2 Materials Required but Not Supplied

...

120Ω CAN bus wire:
o GigaFlight Connectivity P/N GF120-24CANB-1
GigaFlight Connectivity Inc.
6180 Industrial Ct.
Greendale, WI 53129
United States
Toll Free: +1 (844) 303-1093
Phone: +1 (414) 488-6320
Email: [email protected]
o
Carlisle IT P/N CAN24TST120(CIT), (must be ordered from the following Carlisle
IT facility):
Carlisle Interconnect Technologies
5300 W. Franklin Drive
Franklin, WI 53132
United States
Toll Free: +1 (800) 327-9473
Phone: +1 (414) 421-5300
Fax: +1 (414) 421-5301

I would take all that to mean that more than 20 meters but less than 100ft requires the special CAN Bus wiring called out.

 

[g3xpert]

CAN Bus Wire

The Carlisle wire is called out on page 1-5 of the G3X Touch Installation Manual. If you were to end up with a bus length beyond 20 meters, and did not use the Carlisle wire specifically, some bus communication errors may result.

The Carlisle part number called out in the G3X Touch manual, Carlisle IT P/N CAN24TST120(CIT), contains 2 conductors.

Thanks,

Justin

 

[1001001]

The Carlisle wire is called out on page 1-5 of the G3X Touch Installation Manual. If you were to end up with a bus length beyond 20 meters, and did not use the Carlisle wire specifically, some bus communication errors may result.

The Carlisle part number called out in the G3X Touch manual, Carlisle IT P/N CAN24TST120(CIT), contains 2 conductors.

Thanks,

Justin

Justin,

Thanks for that! However, I think a lot of this confusion comes from the later sections of the G3X installation manual (for experimental, not certified, use), specifically Section 2.3.1.3.2, which states (as I quoted above):

2.3.1.3.2
CAN Bus Wiring
Wiring used for the CAN bus should be shielded twisted-pair cable, MIL-C-27500 or equivalent. 22 AWG
or larger wire is recommended for physical robustness and ease of installation.

It appears I missed the earlier reference to the special CAN wire and was mostly referring to the later section.

 

[1001001]

Just for everyone's information, the Gigaflight cable called out in the manual is easy to get from the manufacturer at an exceptionally reasonable price, significantly less than from other suppliers.

Go here and click the "Contact/RFQ" button and submit a question about the cable. I got a response from them in less than 5 minutes! The guy I spoke with, Jeff, was very personable and helpful.

 

[rongawer]

I agree with Dave, they are very responsive and shipped quickly. My total was about $2.80 per foot. Definitely more than the 80¢/ft Stein charges for the 2-conductor shielded.

Like others, I calculated over 60 feet before I installed and was concerned about signal quality, so I spent the money for the high quality stuff that Garmin recommends.

I ended up with just over 61 feet of installed CAN bus in my RV-10 with the GMU11 in the tail, roll, pitch and yaw servos, and just about every other gizmo Garmin sells for the G3X, with redundancy. So you should meet the 66 ft rule (20 meters) with no problem.

 

[PilotjohnS]

Just me

I used 26 awg twisted shieled pair standard tefzel wire. I like the smaller wire as my feed thru holes are tight. However, using 26 awg is not for the faint of heart; it takes special techniques in order to get a good crimp at the pins. YMMV

 

[1001001]

I agree with Dave, they are very responsive and shipped quickly. My total was about $2.80 per foot. Definitely more than the 80¢/ft Stein charges for the 2-conductor shielded.

Like others, I calculated over 60 feet before I installed and was concerned about signal quality, so I spent the money for the high quality stuff that Garmin recommends.

I ended up with just over 61 feet of installed CAN bus in my RV-10 with the GMU11 in the tail, roll, pitch and yaw servos, and just about every other gizmo Garmin sells for the G3X, with redundancy. So you should meet the 66 ft rule (20 meters) with no problem.

How does the GMU11 work out in the tail? I was planning a wingtip installation but if it works in the tail, so much the better. Can you share your exact location?

 

[Walt]

So I thought I would give the gigaflight CAN bus wire a shot on my current panel install. Takes a little different technique to work with but overall I like it.
The main thing I like is no more soldering the CAN wires outside the back shell as the two 24 ga wires can be crimped into the size 20 d-sub pin.

 

[PilotjohnS]

Also

So I thought I would give the gigaflight CAN bus wire a shot on my current panel install. Takes a little different technique to work with but overall I like it.
The main thing I like is no more soldering the CAN wires outside the back shell as the two 24 ga wires can be crimped into the size 20 d-sub pin.

Walt
If you are going to do this, then you can also bring the shield ground termination into the backshell; there is no reason to leave it outside the connector. YMMV

 

[Walt]

Walt
If you are going to do this, then you can also bring the shield ground termination into the backshell; there is no reason to leave it outside the connector. YMMV

Don’t think there is really anything to be gained by doing that, and it makes the wiring inside the back shell to crowded IMO. The pigtail still has to brought outside and then attached to the back shell so the length of the pigtail is probably about the same, which means no real gain in noise rejection.

 

[PilotjohnS]

well, yes

Don’t think there is really anything to be gained by doing that, and it makes the wiring inside the back shell to crowded IMO. The pigtail still has to brought outside and then attached to the back shell so the length of the pigtail is probably about the same, which means no real gain in noise rejection.

I did it to reduce the amount of unshielded wire exposed to noise. But what i didn't realize is that it is impossible to remove the contacts from the shell. Like most folks, I think I have learnt that leaving a pigtail makes rework/ re-pinning much easier. Live and learn they say.

 

[jliltd]

I used 26 awg twisted shieled pair standard tefzel wire. I like the smaller wire as my feed thru holes are tight. However, using 26 awg is not for the faint of heart; it takes special techniques in order to get a good crimp at the pins. YMMV

In the case of 26 AWG I would strip twice the normal length and twist the strands tight together and then bend and fold the bare wire in half, effectively doubling the thickness of the conductor that inserts into the pin. Then set the crimper for 22 or 24, whichever seems closest.

 

[rongawer]

How does the GMU11 work out in the tail? I was planning a wingtip installation but if it works in the tail, so much the better. Can you share your exact location?

It passes the magnetometer interference test with no issues noted.

My GMU 11 is mounted on a shelf hanging from my tailcone. More precisely, it's mounted on the forward side of the F-1008 bulkhead on a composite shelf that is attached to the upper skin (ShowPlanes composite tailcone). The aft end of the GMU has 3/8" clearance below the upper tailcone skin and 1" clearance from the bulkhead, centered on the shelf with the connector aligned forward.

Don’t think there is really anything to be gained by doing that, and it makes the wiring inside the back shell to crowded IMO. The pigtail still has to brought outside and then attached to the back shell so the length of the pigtail is probably about the same, which means no real gain in noise rejection.

I agree. The real issue is to ensure you have no more than 1.5" of total unshielded wire measured from the back edge of the shielded back shell. If you're not using shielded connectors from Garmin, then you should at least be using a shielded back shell.

 

[Traash]

Added a GHA15 to an existing system that used MIL-C-27500 for CAN bus. System became overloaded and experienced CAN issues. Now rewiring with GF120-24CANB.

LESSON LEARNED: Use the "expensive"wire from the beginning.

This system has 17 LRU's on the CAN bus in a RV-10. MILSpec wire doesn't handle the data packages as well as the more expensive stuff. The "120" in the name means that it is manufactured to have 120ohms impedance which is what the Bus wants. That's why the resistors at the two ends are 120ohms. A balanced system runs more efficiently. MILspec is generic and impedance isn't always 120.

Of note. My problem only shows up on the "CONFIGURATION/DIAGNOSTIC" pages (hold "NRST" and "MENU" while powering up) and only on the LRU at the termination (GMU11). Depower any of the last three LRU's on the bus (GHA15, GSU28, GMU11) and the problem disappears. It's common for the termination to show a problem furthur up the line.

Interesting explanation: https://youtu.be/YBrU_eZM110?si=RoVIrGKVxsZ0dLVj

 

[msires]

Added a GHA15 to an existing system that used MIL-C-27500 for CAN bus. System became overloaded and experienced CAN issues. Now rewiring with GF120-24CANB.

LESSON LEARNED: Use the "expensive"wire from the beginning.

This system has 17 LRU's on the CAN bus in a RV-10. MILSpec wire doesn't handle the data packages as well as the more expensive stuff. The "120" in the name means that it is manufactured to have 120ohms impedance which is what the Bus wants. That's why the resistors at the two ends are 120ohms. A balanced system runs more efficiently. MILspec is generic and impedance isn't always 120.

Of note. My problem only shows up on the "CONFIGURATION/DIAGNOSTIC" pages (hold "NRST" and "MENU" while powering up) and only on the LRU at the termination (GMU11). Depower any of the last three LRU's on the bus (GHA15, GSU28, GMU11) and the problem disappears. It's common for the termination to show a problem furthur up the line.

Interesting explanation: https://youtu.be/YBrU_eZM110?si=RoVIrGKVxsZ0dLVj

I was going to say this, but you beat me too it. The 'right' wire may be expensive, but not as expensive as debugging a problem caused by the 'wrong' wire, and then ripping out and replacing with 'right' wire. When I installed some Garmin Canbus equipment in my Piper, I used the 'right' wire at $6.50 a foot without hesitation. I only needed 25 feet since I wasn't installing the Garmin autopilot, only Garmin dealers can do that on certified planes. I've spent enough time chasing problems on computer networks that turned out to be bad wires to skimp on it.

 

[JDA_BTR]

On my 14 I spliced each canbus connection at each backshell. All of my “stubs” were thus very short like 3 inches. Seems that I could have longer stubs and a shorter overall wiring run. Each device had a canbus going in and a canbus going out and that added many feet I’m sure since they all have some amount of service loop.

Tempting to think about having a run behind the panel with a bunch of stubs coming off one for each device and letting the stub go out to the device instead of the whole bus.

 

[mburch]

On my 14 I spliced each canbus connection at each backshell. All of my “stubs” were thus very short like 3 inches. Seems that I could have longer stubs and a shorter overall wiring run. Each device had a canbus going in and a canbus going out and that added many feet I’m sure since they all have some amount of service loop.

Tempting to think about having a run behind the panel with a bunch of stubs coming off one for each device and letting the stub go out to the device instead of the whole bus.

Definitely don't do this until you understand the limitations stipulated in section 2.3.1.3.1 of the installation manual:

Daisy-chained LRUs (LRUs not at the extreme ends of the CAN bus) connect to the CAN backbone through short “stub” or “node” connections. The length of each node connection should be kept as short as possible, and should not exceed 0.3 meters (1 foot). The best way to connect devices between the ends of the CAN bus while maintaining short stub node lengths is to splice the connections as close to the device as practical. Unshielded wire sections should be kept as short as practical.

Multiple devices must not connect to the CAN bus backbone at the same point. Rather than splicing two or more stub node connections together, the CAN bus should instead be daisy chained from one device to the next.

The layout of the CAN bus must be a single linear backbone with exactly two distinct end points. Other layouts such as “star” or “Y” arrangements must be avoided.

 

[gjh22]

I used the Gigaflight cable. It is worth mentioning that it was several times more time-consuming to prepare and to terminate than the standard MIL-C-27500 2-conductor shielded twisted pair. This is partly due to some features in the cable for impedance control and partly due to idiosyncrasies of their construction (e.g. the ribbon used on the shield being less flexible and wetting solder sleeves more poorly than the braided wire on plain 2-c STP).

I would probably do it again? One-time cost and all that.

The only cable more irritating to work with was a triple-braided aerospace-grade Cat 6 I used for Ethernet on the panel.

 

[Traash]

Added a GHA15 to an existing system that used MIL-C-27500 for CAN bus. System became overloaded and experienced CAN issues. Now rewiring with GF120-24CANB.

LESSON LEARNED: Use the "expensive"wire from the beginning.

This system has 17 LRU's on the CAN bus in a RV-10. MILSpec wire doesn't handle the data packages as well as the more expensive stuff. The "120" in the name means that it is manufactured to have 120ohms impedance which is what the Bus wants. That's why the resistors at the two ends are 120ohms. A balanced system runs more efficiently. MILspec is generic and impedance isn't always 120.

Of note. My problem only shows up on the "CONFIGURATION/DIAGNOSTIC" pages (hold "NRST" and "MENU" while powering up) and only on the LRU at the termination (GMU11). Depower any of the last three LRU's on the bus (GHA15, GSU28, GMU11) and the problem disappears. It's common for the termination to show a problem furthur up the line.

Interesting explanation: https://youtu.be/YBrU_eZM110?si=RoVIrGKVxsZ0dLVj

Follow up info.
I ordered CAN wire directly from Gigaflight. Very nice company to work with. Be sure and download their data sheet that explains the wire and additionally the directions on stripping methods. I found the wire easy to work with.

I ran ~30' of wire from my wingtip (GMU11) to the roll servo, to the GHA15 under the R seat and then to the panel where it joined the existing CAN. No need to replace entire CAN. This reduced my MILSPECwire length to less than 60'.

PROBLEM SOLVED!!

All is now well. Next install gets all Gigaflight wire for the entire CAN bus.

 

[JDBoston]

For those that used the Gigaflight cable, how much did you actually order for a typical G3X install? I am building a 14 but let's just assume a typical install which seems for most to be around 60' total. With waste and various other needed service loops, etc how much cable did you order?

 

[Traash]

For those that used the Gigaflight cable, how much did you actually order for a typical G3X install? I am building a 14 but let's just assume a typical install which seems for most to be around 60' total. With waste and various other needed service loops, etc how much cable did you order?

Every aircraft will be different and always turns out more than expected.
You could pre-wire the plane with string and measure but even that method has it's limits. Things always change in the end. You have to expect overage. The cost is an extremely small percent considering the entire build cost. CAN bus operation is the heart of the G3X system.
Gigaflight was easy to work with, able to ship small quantities and much less expensive than other sources. For a -14 or smaller I'd start with 80'. For a -10 bump that up to 100'.

 

[gjh22]

I used the Gigaflight cable. It is worth mentioning that it was several times more time-consuming to prepare and to terminate than the standard MIL-C-27500 2-conductor shielded twisted pair. This is partly due to some features in the cable for impedance control and partly due to idiosyncrasies of their construction (e.g. the ribbon used on the shield being less flexible and wetting solder sleeves more poorly than the braided wire on plain 2-c STP).

Someone DMed asking about my solder sleeve comment here, so I figured I'd add the supplementary detail to this thread.

I occasionally re-did the solder-sleeved shield connection if it looked especially bad. It still seems to wet it somewhat and every electrical test I'm able to perform seems to show it being fine, but there is a clear visual difference. For reference, because I assume different manufacturers use different fluxes and solder chemistry, the solder sleeves I was using were the Sumitomo L-C type sold by AS. I tried the Raychem ones as well but their performance was worse.

I considered adding some additional flux or pre-wetting the sleeve with a soldering iron and some other solder. This provided the right result visually but I wasn't sure about the long-term compatibility of the various residues and so did not.

 

[Pilot8]

So I thought I would give the gigaflight CAN bus wire a shot on my current panel install. Takes a little different technique to work with but overall I like it.
The main thing I like is no more soldering the CAN wires outside the back shell as the two 24 ga wires can be crimped into the

So I thought I would give the gigaflight CAN bus wire a shot on my current panel install. Takes a little different technique to work with but overall I like it.
The main thing I like is no more soldering the CAN wires outside the back shell as the two 24 ga wires can be crimped into the size 20 d-sub pin.

is 20 gauge dsub pin a typo?

 

[georgemohr]

We used 22GA shielded Tefzel 2-conductor from Stein. Zero issues. RV-14A with a GMU-22 in the wingtip, so not an extremely long canbus length.

P.S. my friend who's an EE noted "at the baud rate Garmin is running this canbus, you could use string and cups and it would work "

 

[Walt]

Yes, the female pins are size 20, still have to some soldering/splicing when connecting to the high-density pins (audio panel and new coms).

 

[JDA_BTR]

My 14A has 22750-2 can bus wiring with zero error. But the price of shielded wire has changed and I’m putting a Gmu11 in my -8 wingtip so a little longer run than in the 14A. The price of the Gigaflight from the vendor is reasonable when purchased direct and they will ship smaller quantities so I decided to go that way in the second build. I’m also looking forward to not having to splice out the wire before pinning as two wires can be put in one pin and crimped to make the splice. I will still use a solder sleeve to connect the two grounded cables to a ground lead at each junction.

 

[fl-mike]

I really like EN2714-013xxxxx cable. I was exposed to this on an Airbus program. It is lighter weight than 27500 and uses a spiral shield rather than braided. Very easy to strip back, but you can lose some shield coverage in tight bends if the spiral opens up. You can even twist the shield and cover with heatshrink to make your shield termination pigtail rather than using soldersleeves. I was a little leery of doing that, concerned it might break, so I used soldersleeves.
Specifically, I used EN2714-013B002F 24AWG two conductor shielded for my CANBus wiring and doubled up in the contact pins at the LRU "Y" connections to avoid stubs. I sourced my small spool from AirCostControl, who were very accommodating to my relatively small order.
Link to datasheet with the part number breakdown: EN2714-013

 

[bbaggerman]

Lots to unpack here and lots of interesting, ummm... theories about what is important to CAN Bus and data buses in general. Here are some things to ponder.

This is probably tl;dr.

The first thing to keep in mind is that the CAN Bus bus is a linear bus which is a single linear wire run with a termination at each end. This is most important! You should be able to put your finger on one terminating resistor (or the box that has it inside) and run your finger along the wire until it reaches the other terminating resistor at the other end of the wire. You should not encounter any other terminating resistors along the way and you should not encounter any Y or T cable offshoots to other boxes. Someone mentioned a 3" stub and I don't think that would be a problem at all. But a 10' stub spliced into the middle of the bus should be avoided. For the low speed CAN Bus this is much less important, though.

What about the cable itself? Cables have lots of parameters. People talk about 120 ohm cable. Is that important? If you put an ohm meter across your wire pair it won't read 120 ohms. It will read infinite. What gives? Cable is infinite impedance at DC and only looks like 120 ohms when it becomes a transmission line. When does it become a transmission line? There is no exact criteria but a good way to think about it is to think in terms of wavelengths. Most CAN Bus systems run at 1 Mb/s. Some run slower at 125 kb/s. Other run faster at 5 Mb/s. In the RF world "interesting" things usually start happening at about a wavelength, give or take. Ignoring cable velocity factor and such, a 1 MHz signal has a wavelength of 300 meters. The CAN Bus is a digital bus with square wave signals so there are also frequency harmonics (at odd multiple of 1 MHz) that need to be kept in mind. Even so, the wavelength of the CAN Bus signal is much longer that the length of the cables so I would expect transmission line effects like cable impedance to be pretty minimal. In the RF world signal reflections happen at impedance discontinuities. If your VHF com with a 50 ohm output impedance tries to transmit into a cable that isn't 50 ohms, or hits a place where the coax is mashed (messing up the cable impedance in that area), or into an antenna that is the wrong one then a portion of the signal will be reflected back. That reduces the amount of RF power that gets out. In the digital bus world a reflected signal goes back across the bus, messing up the signal that is already there.

What else might be more important for a good cable? Series capacitance could be important. Capacitance across the signal could slow rise times, generate phase shift, attenuate the signal or it's harmonics, or do other bad things. Keep an eye on any cable series capacitance requirements. In general less would be better. Shielding is very important and not all cable shielding is created equal. Some manufacturers quote the effectiveness of their shield. In this case, of course, more is better.

What about grounding? My worry is two things. One is you want to make sure there is no current running through your shielding. If your shield offers a better ground path than the usual ground then you are doing something wrong. Second, remember those old UHF TV antennas? You young guys can google it. Rather than being a whip antenna the UHF antenna is a loop of wire. RF is composed of two orthogonal field components, an electric field and a magnetic field. A whip antenna above a ground plane is an electric dipole antenna. A wire loop is a magnetic dipole antenna. For a closed loop any RF signal going through the loop induces a current in the loop. So if shielding or other wiring forms a closed loop (your aluminum aircraft structure can be part of the loop) then you run the risk up picking up stray RF or EMI by virtue of being a loop antenna, just like on the back of that old TV we grew up with. BTW, induced signal is proportional to the area of the loop. Keep your wires close to the aluminum structure and you can reduce this type of induced signal.

Back to the CAN Bus, I would say the most import thing is to make sure you have a single linear wire run and a single termination at each end. Next I would look at ensuring the cable shield is good and properly designed.

Lastly, I think the Dynon SkyView databus runs at 1 Mb/s also, at least that is what I measured. They don't seem to worry about any of this stuff and it seems to work just fine!