Serial communication over 300ft in an aircraft

Discussion in 'The Projects Forum' started by dmusoke, Jan 12, 2010.

  1. dmusoke

    Thread Starter New Member

    Jan 4, 2010
    11
    0
    Hi all:

    I have an application where i need to transfer data over 300-ft of shielded RF cable from the tail end(B side electronics) of a plane to the nose end(A-side electronics). My data rates are low, under 200 KBits/sec between the A and B side electronics. One shielded RF cable carries data and the other carries power to the B-side electronics as shown in the attached schematic. As you may have found out already, this is NOT a differential serial communication scheme. One coax carries the data and the other carries power.

    My problem is that the chassis of the aircraft, which is used as the system ground, is not at a constant voltage. The A-side ground could be close to 0 volts but the B-side could have a chassis floating voltage as high as 10Volts above the A side ground.

    I'm concerned that I wont be able to reliably, if at all, establish communication betweenthe A and B-sides. The 28Vdc ground on the B-side goes through DC-DC converters whose common output is isolated from their input ground pins, so that circuitry wil be referenced off this common output. This should be OK, right:confused:?

    How an I prevent the difference on ground potentials from killing my communication? Should i tie the shields on the A-side to ground but leave them open on the B-side for both cables?

    Please help...

    Thanks,
    David
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You're going to have to float the shields on one end.

    I'm wondering why you're not using optocouplers/optoisolators?

    If your DC-DC converter is a flyback type (output isolated from input) you should be fine in grounding it to the airframe on the B-side - that is, if the airframe is made out of metal (doesn't sound like it...)
     
  3. dmusoke

    Thread Starter New Member

    Jan 4, 2010
    11
    0
    Thanks SgtWookie....and the reason why i cannot use opto-couplers (oh how i wish I could!) is that I'm also carrying RF carrier info(1GHz) super-imposed on both the 28V cable and data cable. I know i didn't mention this before for i thought I'd unnecessarily complicate my design to the forum members. Sorry about that.

    So yes, both cables carry RF data as well. So i have RF data + Control data(< 200KBps) on ONE cable and have RF data +28V on the other cable.

    Does this change your previous suggestion to float one end of the cable?
     
  4. ifixit

    Distinguished Member

    Nov 20, 2008
    638
    108
    A pulse/RF transformer might work at the "B" end to isolate the ground of the signal path, but is dependant on the data coding method used.

    You might need two pulse/RF transformers, one for the 1G signal and one for the 200KHz signal.

    The devil is in the details
     
  5. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Well, you're going to have to match the impedance on the RF signal, yet block the DC component. If you have an impedance mismatch, your return loss/VSWR will go right in the toilet. It won't be easy working at those high frequencies.

    Poke around MiniCircuit's site: http://www.minicircuits.com/
    They have some DC blocks here pretty cheap: http://www.minicircuits.com/products/dc_blocks_main.html
    and lots of other nifty stuff for high frequency.
     
  6. dmusoke

    Thread Starter New Member

    Jan 4, 2010
    11
    0
    Thanks ifixit and SgtWookie. The low speed data and RF separation along with the RF termination have been taken care of using bias tees. I get Rf and data completely separated for further processing. The DC blocks from mini-circuits only pass RF BUT thanks for the effort to steer me in the right direction:D...

    Any other issues you can forsee that I need to take care of?

    Thanks:)!
     
  7. CDRIVE

    Senior Member

    Jul 1, 2008
    2,223
    99
    I don't understand this at all. Are you saying that if you had a DVM with 300' of test leads with one lead connected at GND A and the other lead connected at GND B you will read 10 volts?? The nose of the fuselage would have to be insulated from the tail for this to happen.
     
  8. dmusoke

    Thread Starter New Member

    Jan 4, 2010
    11
    0
    Yes CDrive...good point,

    It depends on the type of aircraft. Unfortunately for the older ones, this is the case. The newer ones less probably so. By newer, I believe these are ones built in the last 10 years or so. Unfortunately, I cannot dictate what type of planes the design would go into, so I have to consider all possible cases.
     
  9. 3ldon

    Active Member

    Jan 9, 2010
    82
    3
    This better not be a critical system....:eek:

    Its ok until lightning strikes it. It should be referenced to something, that way the bulk capacitance can absorb spikes, unless the breakdown voltage between the two is measured in tens of kilovolts, or you have plenty of MOV's protecting it.

    You have a 1 ghz link working already, is there any way you can inject the information into that signal? or send it over the 28 volt bus using off the shelf ethernet over powerline?
     
  10. dmusoke

    Thread Starter New Member

    Jan 4, 2010
    11
    0
    Thanks 3ldon:

    The 28V bus is already carrying a separate RF stream on it already. Same with the data bus. So both busses are carrying RF streams injected with their respective DC and low frequency data streams.

    The data stream is protected via a 16-bit CRC code so hope that provides enough security for the system. The B side ground will be referenced to the bulk chassis in the tail section of the plane so the A & B grounds will not be connected at all to avoid ground loops problems.

    Thanks for the MOV's idea. I'll look into that as well.
     
  11. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    Questions like this are really annoying. You have a mystery application on a 300ft long airplane with really tricky wiring that *already* carries multiple RF signal and DC voltage converters etc etc.

    Any aviation electronics on a highly regulated 300ft long passenger or commercial airplane will be designed by very highly paid, uber-qualified aviation electronics engineers working in teams for $$$ companies.

    Those uber-qualified aviation engineers do things in very simple conservative SAFE ways (unlike anything you are proposing) and then they spend big $$ afterwards on independent testing and certification etc etc before there is any chance it gets installed on that 300ft airplane.

    And they one thing they are NOT going to do is get on a hobby electronics forum and ask strangers how to design something that is really difficult and probably prone to faiing and/or causing the OTHER systems to fail.
    :(
     
Loading...