No announcement yet.

psu LED blinks WHITHOUT ground connected??

  • Filter
  • Time
  • Show
Clear All
new posts

  • psu LED blinks WHITHOUT ground connected??

    hello. i have the strangest issue with my M4-ATX. (althyough it might not be so strange..) when i connect the possitive wire to the M4, before i connect the ground, the LED starts blinking. if i disconnect all the cables going from the PSU to the motherboard, it stops.

    in my build i have 2 screens. BOTH are powered from the 12, and 5v rails of the M4-atx.. if i have all the power lines connected BUT the vga cables from the graphics card disconected, nothing happens.
    if i plug thei power of the screens, nothing happens as well. it's when i connect their VGA that the led starts blinking (Although i cant seem to find any ground. only a ground plane on the screens panel maybe? but the motherboard as well is grounded to metal plate)

    also, i have a usb sound card. if its connected nothing happens. if i connect its outpouts, to the amplifier, the led turns on and stays on (dimmed). but this does not happen if the outputs have the ground loop isolator connected (normal)

    dafaq is up with that? is it normal? my build has LOTS of noise. noise on the screen, noise on the amp... anyone got any idea?

  • #2
    I can tell you where the blinking comes from, your vga cable has a ground to , and all grounds are connected inside any device.

    I can power my screen with only a positive wire and the vga cable , provided that my pc is grounded, wont recommend it though :P

    same thing goes for the ground from your amp to your usb sound card
    View my worklog here


    • #3
      yes but this is the point. the PC is NOT grounded. (except from the metal part it screws on)


      • #4
        interesting .. maybe some of the electrical guys on here know what the issue is
        View my worklog here


        • #5
          well... i found out that my M4 , if connected to just positive (12v), with no ground, will have a +12v at ALL its rails, without having it connected to anything ... (fact that goes away if you conenct it the ground). that would be ok, but, isn't the fact that the LCD's provide ground (although they are suppose to be grounded only in the PSU) troubling?


          • #6
            And hence why GND circuits should never be fused, switched, broken or too high a resistance. (Except where a single GND path exists.)


            • #7
              I had an issue similar to this within my own car pc. I was using a Mini-Box DC-DC PSU at the time and found this same effect occurred because my VGA screen was grounded straight to the vehicle and not the power supply. It's actually pretty amazing how the same ground will transfer through literally everything, which requires you to be extra careful that none of your equipment is coming in contact with the a vehicle ground anywhere. I have since switched to a CarNetix P2140 PSU, but powering the screen off the PSU then corrected the problem. If your car PC has a metal case you need to make sure even that isn't coming into contact with the vehicle chassis somehow. The whole case acts as a ground, since the motherboard is actually grounded to it.

              With that being said if your case does happen to be contacting the chassis somehow, it wont necessarily be a bad thing for your computer, AS LONG AS THE MAIN PSU GROUND IS CONNECTED as well. Absolutely DO NOT try to power on your computer without the main PSU ground connected even if that light is flashing. Electricity will always take the path of least resistance, so even through without the PSU ground connected it still seems to be grounded somewhere, it doesn't mean it's a good ground. It's actually probably a very weak one that couldn't handle much current. That weak ground will not matter if your main PSU ground is connected, since the PSU ground will be the optimal path of least resistance. The other ground may still exist, however not much current-if any at all, will try to flow through it.

              Providing the main PSU ground is connected, the worst that could result of the "unknown" ground is that it will open the door for unfiltered alternator noise as you were describing. In my car pc I get the alternator noise despite the system is totally isolated. Because of that I set a 25 second delay (though it could be more like 17ish) on my PSU for the amplifier remote turn-on line (on of the many reasons I love the P2140) so I wont hear any noise since when the Windows sound drivers kick in the noise is 98% eliminated. The delay allows enough time for my car pc to get out of hibernation and onto the main GUI before kickin in the amplifiers.

              Best of luck tracing that ground!


              • #8
                omg. i found out what was causing the LCD to provide ground. after allot of reasearch i found out , that in the spot where i put the LCD, it had scratc the dashboard, and the metal chasis of the LCD was touching the metal skeleton of the dashboard. that is what was providing the ground!! . i will fix it now, and maybe this is what was causing the noise in my car from the begining...


                • #9
                  While they tell you to hook the ground up to your car battery last this is because of the potential of sparking... Sparking around battery can cause it to go boom.

                  And you are doing this with the ignition off... You should NEVER attach your battery if power is applied to anything. However your devices won't find a ground without the ground wire attached. (Although I wouldn't suggest touching the body of your car while attaching the ground wire.. YOU might become the ground wire...) With the ignition off you may still have some items powered but they are very low amperage and shouldn't affect anything.

                  Otherwise as far as electrical components go you should NEVER hook them up with power applied and you should always hook up the ground first. If you hook them up with power applied and any part of the device touches the body of the car you can get a partial ground. This is bad because that 12 volts could be seen as a bunch of voltage or as a negative voltage. You are lucky you didn't burn up your power supply. You should have had the fuse pulled on it, hook the ground up first, then the positive wire and then plug in the fuse.


                  • #10
                    Alas the reason for connecting the ground last is not for sparking (that occurs no matter which is connected last). It is so that if your spanner or metal tool touches ground while working on the hot side (ie, the non-ground or +ve terminal etc in modern vehicles) you don't cause a massive short.

                    Hence the simple rule - whenever working on the hot side, the ground shall be disconnected ie, -ve/GND off first and on last. (I have seen up to 12 "complex" rules to convey the same simple rule.)

                    And a human being the ground is not an issue. We are too high an impedance. If not, you would not be able put yourself across a 12V (or 24V or 48V) battery safely.

                    As for loads, yes - their power grounds should be solid before applying power else load damage can occur, but that was covered by my earlier reply.

                    Settra - it'd be nice if that LCD ground was the cause of your system noise. Signal (loop) noise is usually far easier to solve than power caused noise.


                    • #11
                      Hmm OS, have you ever accidentally grounded yourself against the car? I certainly have and that 12 volts is enough to knock you on your butt. You can weld with a 12 volt battery so don't think you can't harm yourself...

                      And while it is true it is easier to hit ground while working on the positive side this is not why it is a bad idea. Since the electrons run out of the battery at the negative terminal you are likely to spark much more from the negative post than you are the positive post. But of course you are going to tell me my training and experience mean nothing or that somehow electons just don't flow the same where you are...


                      • #12
                        You are confusing a low voltage battery with the inductive kick given by loads. But that also happens on the +ve side (thanks to duality or symmetry or whatever it's called).
                        However I suspect you are also talking about holding a connector whilst disconnecting it - eg, the +ve or -ve battery terminal. You won't get a kick by connecting yourself in series with the battery - unless perhaps you have another power source present (another battery powering the circuits, or a welder, etc). But hold across connectors when (relay) coils de-energise etc and yes, you'll get a nice kick.

                        And welders work on current. What your saying is like saying that because a battery can melt a screwdriver, it can electrocute you or give you kick.

                        As to non-symmetry because "electrons from the -ve terminal", that is plain wrong (not to mention the crap about current flow being merely "electron movement").