05-24-2008, 05:24 PM #1
A call for help from any electrical guru's!
I have been fighting a bad gremlin for a few weeks now. It has now taken over my life and is consuming me at a high rate! Please help me!!!! It is a complicated issue that has me and David stumped.
I have 3 phidget 0/16/16 cards in my overhead utilizing their own power supply for the input switching. I am using all real components in my sim.
Here is my issue,
I have had serious phantom/ghost inputs coming from the phidgets when i turn on the 28Vdc power supply to all my lamps and indicators. The led's on the phidgets will flash on/off in a split second like the switch was moved but they are not being moved.
Fast forward after many late nights, i have narrowed it down to the lamps in the TQ (park brake and stab trim) If i take the bulb out, no problems!! Put them back in, interference like crazy! I tried redundant gnds, swapping power and gnd on the bulb sockets, resistors on the phidget inputs, and seperate PS. No luck! Any other ideas?????????
05-24-2008, 05:55 PM #2
Some sort of inductive coupling or capacitive coupling on wiring that relates to the Phidgets inputs that is run in close proximity and parallel to the 28V wiring during the initial power up surge as the filter capacitors in the power supply charge up?
Look at your cable runs and try to get the cables going to the throttle quad for the lights and the cables relating to the Phidgets away from each other no matter. Is there somewhere in the wiring bundles that the two are in close proximity to each other? The best situation for this is to have then at 90 degrees to each other. And to run shielded cables.
05-24-2008, 08:37 PM #3
Thanks for the input John. Funny thing is, nothing in the TQ is even remotely near or wired to phidgets other than the common GND for the entire cockpit, which by the way is my shell.
I also just found that the 28Vdc annunciators in the MIP are causing the issue as well.
I will go back over my wiring but i still don't see how the MIP and TQ are doing this. They are not hooked to the phidgets, they are controlled by OC boards and SIOC on a seperate computer. )
05-24-2008, 08:58 PM #4
I had similar problems which was caused by the Phidget wiring being near fluorescent bulbs I used for back-lighting, matter of fact I could not get the wiring far enough from the bulbs. I tried shielded cables, didn't help. I went back to my incandescent bulbs for back lighting.
Rob, it may not be the same problem.... but just to help confirm what John said, 0/16/16 are a bit sensitive to noise.
05-24-2008, 09:12 PM #5
- Join Date
- Apr 2008
- Frankford, Ontario
These circuits are very very fast, and will respond to RF induction at the drop of a hat if you don't take care to filter all you circuits. I know you don't want to hear this, but you just have to go to every board and filter the DC supplies for RF at every chip. Seriously. It's common practice for TTL for a very good reason. And that includes your power wiring - 110v lines can easily carry RF into your circuits.
If your boards are off-the-shelf, you would think that that has already been done, right? Good luck with that - look at the mess Saitek are in with their yoke. At the very least you need to filter the outputs of all your PSUs at the point of output, and also filter the inputs to your various boards at the point of input to these boards. Yes, really.
This is something which no amount of simulating in platforms like MPLAB can possibly take account of, as I have discovered - it's entirely up to you!
05-24-2008, 10:34 PM #6
Well i found the problem(with the help of D. Allen) and all of your suggestions as well.
I have 1 28Vdc power supply taking care of all the original 28v stuff. The shell is the common GND just like in real aircraft.
I have the 64output card from opencockpits, it PWM the gnd so whenever it was turning on an indicator, depending on which one, current related i'm sure, it was introducing noise into my GND plane!! No pun intended with the GND Plane.
I noticed on the directions that it showed both + and - form the source going to the card even though it only 'sinks' the output to turn it on. SO added the 28V + wire to the input on the board and so far it is working great. BUTTTTTTT, if i turn all the outputs on like during a bulb test, it will wreck havoc with phidgets again. Under normal conditions it is now ok.
SO what i am going to try next is a sperate PS for the 64 output card to futher isolate it from phidgets. We shall see what happens now.
Thanks to all and especially David for the phone calls.
05-25-2008, 04:12 AM #7
Did you know that LED's emit loads of RFI? Without actually looking and getting to it with a meter, I suspect that your adding the power to the other side of the board helped to balance the impedance. But when you power up all the lights all the RFI they produce overwhelms your system.
Try building some very small Faraday cages around your components from now on to prevent cross interference from RFI. This is why some manufactuers build case grounds into their projects.God's in command, I'm just the Pilot.
05-25-2008, 12:40 PM #8
05-25-2008, 07:40 PM #9
- Join Date
- Jan 2007
Hi, i experienced the same thing using the OC USB Outputs card in my OverHead when i fired it up my speakers started to make buzzing!
After some trial ended up on removing it, put an FDS SYS card and it worked good!
So for me the culprit at the moment is the USB CARD from OC, have someone experienced the same?
05-26-2008, 07:37 AM #10
By Viper90 in forum lekseecon Q&A and DiscussionsReplies: 28Last Post: 04-04-2011, 04:37 PM
By Joe Cygan in forum Pilots Lounge - Let your Hair downReplies: 8Last Post: 10-29-2009, 11:01 AM
By hurl2space in forum Welcome to MyCockpit New here? Introduce Yourself!Replies: 3Last Post: 10-11-2008, 09:04 AM
By Jackpilot in forum General Builder Questions All Aircraft TypesReplies: 4Last Post: 07-19-2008, 12:34 PM