What is a Serial Wire?

Banjo33

Line Up and Wait
Joined
Sep 21, 2011
Messages
917
Location
Southern Indiana
Display Name

Display name:
Banjo33
What is a “serial wire?” When you have a single wire that serves as a “serial wire,” can its length be extended? Is there anything special about this wire vs standard wire?
 
Early thread drift: Is it wire with a track record of doing the same thing the same way. As in, a serial killer?

I'm gonna sit and laugh. I hope someone answers seriously. (serial - serious Not the same)

But a Serial connector is typically not a single wire in computer speak.
 
In aviation, a serial port wire can be a single wire if it is a transmit going to a receive. If it is a two-way handshake, then it needs 2 wires. A ground is often used, so a shielded pair. In practicality, it is common for a one-way serial to be a single wire, and can be extended as needed. If there is noise in a nearby system, then it may need to be shielded. Also, a serial transmit can go to many receives, but there can only be one transmit on that wire. For example, one mouth can talk to many ears, but two mouths at the same time and nobody can understand.
 
In aviation, a serial port wire can be a single wire if it is a transmit going to a receive. If it is a two-way handshake, then it needs 2 wires. A ground is often used, so a shielded pair. In practicality, it is common for a one-way serial to be a single wire, and can be extended as needed. If there is noise in a nearby system, then it may need to be shielded. Also, a serial transmit can go to many receives, but there can only be one transmit on that wire. For example, one mouth can talk to many ears, but two mouths at the same time and nobody can understand.

Great explanation, thank you! It is a single wire transmit. Basically transmitting engine data from one box to the display. However, it’s only about 4 feet long and I need it to reach about 6. It looks like a pretty standard wire, however, a lot of random data is going through it (CHT, EGT, Oil Temp/press, RPM, fuel flow/press, etc), so I wasn’t sure if there was something “special” about it.
 
Just a normal wire. I have seen wires like this get erroneous data when run close to a com transmit cable when transmitting. If something weird shows up, consider shielding it.
 
What is the mechanism for ensuring upload and download data doesn’t collide? Is it running a protocol, frequency shifting, timing, or something else?
 
Data is unidirectional; so in typical serial wire (RS232 standard), you have one wire for transmit, the other for receive, so no real chance of "collision," ala coax ethernet.
 
Just a normal wire. I have seen wires like this get erroneous data when run close to a com transmit cable when transmitting. If something weird shows up, consider shielding it.
Jesse, I’ve run a lot of comm or ‘serial’ lines in my ‘10 panel. Thanks for the excellent explanation of how they are used.

The practice I ended up following was to always run a shielded cable and to attach the shield to ground on one end. Never detected any data problems. Any thoughts about that practice?

(For those with no experience with say 20AWG shielded cable - it’s as easy to use as regular wire if you ignore the shield and not much thicker. Teasing the shield out to ground it requires just a little technique)


Sent from my iPad using Tapatalk Pro
 
When I worked at Cessna we had a project with a strain gage on a spar cap sending data to a flight recorder on an A37-B. When the Military FM radio mic was keyed the data signal pegged to the limit. It was shielded cable that was grounded on one end of the shield but the shield was just the right length that RF would build on it from the FM signal. We had to change the grounding on the shield wire so that 1/3 of the length of the wire was grounded and the other 2/3rds of the shield wire had its separate ground connection as I recall. RF problems can be interesting to chase sometimes.
 
It’s neve
Jesse, I’ve run a lot of comm or ‘serial’ lines in my ‘10 panel. Thanks for the excellent explanation of how they are used.

The practice I ended up following was to always run a shielded cable and to attach the shield to ground on one end. Never detected any data problems. Any thoughts about that practice?

(For those with no experience with say 20AWG shielded cable - it’s as easy to use as regular wire if you ignore the shield and not much thicker. Teasing the shield out to ground it requires just a little technique)


Sent from my iPad using Tapatalk Pro
It’s never a bad idea to use shielded wire. It’s not always needed, but it can be used all the time. Usually only ground one end of the shield.
 
It’s never a bad idea to use shielded wire. It’s not always needed, but it can be used all the time. Usually only ground one end of the shield.
RS232 serial lines have such a large voltage swing and aviation speeds are typically so slow that I'd be surprised if anyone saw any genuine noise on a point-to-point connection behind the panel.
 
RS232 serial lines have such a large voltage swing and aviation speeds are typically so slow that I'd be surprised if anyone saw any genuine noise on a point-to-point connection behind the panel.
I remember the official voltage swings are large, as you mention. But I've only have seen 0-5 volt swings, and now only 0-3 V on some devices. How would these lower voltages affect your answer?
 
I remember the official voltage swings are large, as you mention. But I've only have seen 0-5 volt swings, and now only 0-3 V on some devices. How would these lower voltages affect your answer?
It still needs to go negative to signal a zero bit, so it can't be 0-3V. While I haven't put a oscilloscope on it, I also think avionics stay in the -5/+5 range or higher. After all, a brand new GTN750Xi is still expected to talk to all sorts of ancient stuff. Plus, there's still the fact that typical usages are so slow and behind the panel isn't a particularly long run.
 
It still needs to go negative to signal a zero bit, so it can't be 0-3V. While I haven't put a oscilloscope on it, I also think avionics stay in the -5/+5 range or higher. After all, a brand new GTN750Xi is still expected to talk to all sorts of ancient stuff. Plus, there's still the fact that typical usages are so slow and behind the panel isn't a particularly long run.
You are correct when discussing the strict definition of RS-232. I'm describing what may be better described as TTL serial communication, but many people will often use the same hardware (wires and plugs) as true RS-232. If someone is just communicating to their own devices (as is my experience with serial), they will just use the UART built into their processors, so this is just 0 to 3 or 0-5 volts. This may be the case here, transmitting engine data from one box to another.
 
You are correct when discussing the strict definition of RS-232. I'm describing what may be better described as TTL serial communication, but many people will often use the same hardware (wires and plugs) as true RS-232. If someone is just communicating to their own devices (as is my experience with serial), they will just use the UART built into their processors, so this is just 0 to 3 or 0-5 volts. This may be the case here, transmitting engine data from one box to another.
The difficult part is that there isn't a simple answer. TTL serial is definitely more prone to interference than RS-232 (you are correct that TTL does not go negative), but they didn't go building things for airplanes without thinking about it. A well designed TTL serial implementation would be pretty damn noise resistant without shielded wire. A poorly designed TTL serial implementation would be worthless without shielded wire. Devils in the details and we aren't looking at circuit diagrams.

I would sure the hell think that the manufacturer of the devices would specify what wire would be required, and if they say you don't need shielded, then I wouldn't worry about it. If they do, well then, I'd worry about it.
 
In my case, there is already almost 3 feet of wire on the receiving end and the transmitting end, neither of which is shielded. There’s no mention in the installation manual about whether shielded is necessary or not if extending it. But, a few feet of shielded wire isn’t much more expensive and it is a little more robust than unshielded, so I’ll be going that way since it’ll be running from an engine to the cockpit through the Wing root (twin).
 
Back
Top