What's new
Van's Air Force

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

Dynon Servo Twisted Pair Wiring Question

PhatRV

Well Known Member
I am in the process of wiring my Dynon roll servo for my RV8. The diagram specifies two twisted pair cables for each of the servos.

Does this mean I have to use the shielded twisted pair cable or just twist the cables together when I create the wiring loom?

The diagram doesn't show the twisted pair symbol connected to ground so I am assuming the cable pair doesn't have to be shielded?

What have you guys done for your airplane?

Thanks

Phat Vu

47995471142_7d11d8746d_b_d.jpg
 
Also buy my distribution hub at Stein to simplify the wiring...:)

Gill -
Is Stein selling those yet? I do not see them on their website. I just sent in my PDA360EX for a wiring harness and I'll add one to the order if they have them.
 
What happens if you don't twist the wires?

This question has real-world ramifications for this boy who pulled multi-conductor wire early in the project and just now realized Dynon wants some twists in the cabling. :(
 
I will second SteinAir for doing the harness. They did a great job on mine and simplified it.
 
Also buy my distribution hub at Stein to simplify the wiring...:)

Dynon sells a ready made servo cable for a very reasonable price. One of the goals in this RV8 is to build as much of the electrical and electronic stuff as possible. I always want to learn how these things work .
 
What happens if you don't twist the wires?

This question has real-world ramifications for this boy who pulled multi-conductor wire early in the project and just now realized Dynon wants some twists in the cabling. :(

I see two possibilities:
1. No real world ramifications at all

2. Dynon was woefully inadequate in providing proper specs for the required wiring.

If you look at wiring where twisting is critical (ex: CAT5, CAT6 etc computer cable), the number of twists per foot, the direction of twist for each pair, etc, are all spec'd.

In my limited experience, most 'twisted pair' is more for convenience and wire ID than function. Exception might be something like a balanced line pair operating with relatively high source & load impedance.

edit: A closer look at the wiring diagram shows that they also spec'd wire color. My bet is that they're telling the installer what to look for in their harness.
 
Last edited:
Gill -
Is Stein selling those yet? I do not see them on their website. I just sent in my PDA360EX for a wiring harness and I'll add one to the order if they have them.

He has them in stock but hasn't added them to his web site yet....:)
 
Dynon sells a ready made servo cable for a very reasonable price. One of the goals in this RV8 is to build as much of the electrical and electronic stuff as possible. I always want to learn how these things work .

The distribution hub is slightly different from the Dynon one and you change the ends (sex change) to keep the servo cables unique.
 
You all make this too hard.

You need a single, four conductor (#22) shielded wire for the data. A #20 for power, a #20 for ground and a #22 for disconnect. Done.

No hub, no magical harness. A standard 9 pin D connector for the four conductor shielded to the network, then power to your autopilot breaker, ground and your momentary autopilot disconnect switch.

Carl
 
Did I miss the requirement for a shield? I don't even see that, in the drawing he posted.

No - but a four conductor shielded wire fits the bill for twisted pairs, readily available and is durable.

Used this in two RVs. Will do the same in the next project.

Carl
 
I see two possibilities:
1. No real world ramifications at all

2. Dynon was woefully inadequate in providing proper specs for the required wiring.

If you look at wiring where twisting is critical (ex: CAT5, CAT6 etc computer cable), the number of twists per foot, the direction of twist for each pair, etc, are all spec'd.

In my limited experience, most 'twisted pair' is more for convenience and wire ID than function. Exception might be something like a balanced line pair operating with relatively high source & load impedance.

edit: A closer look at the wiring diagram shows that they also spec'd wire color. My bet is that they're telling the installer what to look for in their harness.

Dynon does spec the number of twists, it is in the EFIS installation manual.
 
You all make this too hard.

You need a single, four conductor (#22) shielded wire for the data. A #20 for power, a #20 for ground and a #22 for disconnect. Done.

No hub, no magical harness. A standard 9 pin D connector for the four conductor shielded to the network, then power to your autopilot breaker, ground and your momentary autopilot disconnect switch.

Carl

If you are using two servos spread out (rear fuselage and wing) the the hub helps. It also makes the pwr/grn connections easier.

But it's all personal choice...:)
 
I see two possibilities:
1. No real world ramifications at all

2. Dynon was woefully inadequate in providing proper specs for the required wiring.

If you look at wiring where twisting is critical (ex: CAT5, CAT6 etc computer cable), the number of twists per foot, the direction of twist for each pair, etc, are all spec'd.

In my limited experience, most 'twisted pair' is more for convenience and wire ID than function. Exception might be something like a balanced line pair operating with relatively high source & load impedance.

edit: A closer look at the wiring diagram shows that they also spec'd wire color. My bet is that they're telling the installer what to look for in their harness.

I grew up in the telephone business. They learned well over 50 years ago, that without twisting the pairs that are in close proximity that one audio conversation would bleed into the other due to the magnetic radiation from the electrical signal. The same effect occurs in data transmission and twisted pairs are typically spec'ed to eliminate this bleed over and the errors that come with them. The faster the oscillation of the binary data signal, the greater the chance of error. Slower speeds, like RS-232 have very few problems with this bleed over. 100 Megabit LAN cables have lots. CAT 5 cable has as much or more to do with the capacitance of the insulators as it does with the twisting, to deal with the same issue.

Cables that aren't in a tight bundle, like a multi-pair cable, have significantly fewer issues, as do short cable runs. The Dynon DSAB uses an RS-485 transport. You can do some research on the issues to be expected without using the 485's required twisted pair. RS-485 resembles a CAN bus at the physical layer.

Others should know that most aviation wire is not twisted pair. All of the wires in the bundle are twisted together and not in pairs. a typical mil spec four conductor cable should never be used for two pairs when twisted pairs are specified.

Larry
 
Last edited:
This gives me pause. If I end up purchasing harnesses from Rob Hickman (which do appear to incorporate some twisted pairs), I'll use my existing wire runs to fish/pull the new wire though the airframe (and then trash the old wire runs as useless to me). Tempting to put connectors on the wires I've already pulled, enter phase one and see if there are issues, but who wants to find out the hard way that his servos don't work reliably some of the time? Not me.

Not sure what to do.
 
Last edited:
What happens if you don't twist the wires?

This question has real-world ramifications for this boy who pulled multi-conductor wire early in the project and just now realized Dynon wants some twists in the cabling. :(
The twisted cables are all for data lines with high level signals. The twisting prevents noise from being generated by the wiring. The wires, if not twisted will put out electromagnetic noise that could affect other systems, in particular audio quality. Twisting is highly recommended. Shielding would be required to prevent to prevent the databus from being affected by the indirect effects of lightning inductive coupling) or HIRF (high energy radiated fields) from things like microwave antennas and certain high energy radar emissions.
Twisting the high and low wire causes the fields to be neutralized. If you are wired up and no issues, good for you. It just means that your wire routing separated the noise from potential victim systems, either by design or luck.
Theoretically, without twisting, you could cross couple the EMI into other data lines, particularly if they ran parallel for some distance.

The noise could also couple into antenna cables causing problems.
Again though, if your wiring is completed, and no issues don?t rip it out to twist the wire.
 
My wiring is far from complete, but the runs to the trim and A/P servos are already pulled in the wings and fuse. Not going to change the trim servo cables because they are not likely emitters or victims of RFI, but the autopilot servos are a completely different beast, signal wise. I'm a bit reluctant to test this and have to re-pull wire at the start of Phase One, but also reluctant to believe it's going to make a real-world difference in system behavior.
 
When it comes to avionics why do people presume to know better than the manufacturer?

If it says use twisted pair then just use it and save brain-power and brownie-points for when there are real problems.

The number of issues I have helped with where the installer has said “I read that but I didn't think it was important“...
 
I grew up in the telephone business. They learned well over 50 years ago, that without twisting the pairs that are in close proximity that one audio conversation would bleed into the other due to the magnetic radiation from the electrical signal. The same effect occurs in data transmission and twisted pairs are typically spec'ed to eliminate this bleed over and the errors that come with them. The faster the oscillation of the binary data signal, the greater the chance of error. Slower speeds, like RS-232 have very few problems with this bleed over. 100 Megabit LAN cables have lots. CAT 5 cable has as much or more to do with the capacitance of the insulators as it does with the twisting, to deal with the same issue.

Cables that aren't in a tight bundle, like a multi-pair cable, have significantly fewer issues, as do short cable runs. The Dynon DSAB uses an RS-485 transport. You can do some research on the issues to be expected without using the 485's required twisted pair. RS-485 resembles a CAN bus at the physical layer.

Others should know that most aviation wire is not twisted pair. All of the wires in the bundle are twisted together and not in pairs. a typical mil spec four conductor cable should never be used for two pairs when twisted pairs are specified.

Larry

Just in case too much information is required, telephone cables also have what is called 'roll' so that bundles of 25 twisted pairs do not run in parallel with each other for long distances. This reduces noise between adjacent pairs. Twist and Roll, sounds like a 60's dance. Not a shield in sight.

V
 
Back
Top