Announcement

Collapse
No announcement yet.

Tricky USB power question - who can tackle it?

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Tricky USB power question - who can tackle it?

    Here's the deal. I added a radioshack USB hub to my system because I was running out of ports and I wanted to hook up my USB wireless adapter. The hub can be self powered (can't support the wifi adapter in this mode - too much current) as well as take 5v from the supplied adapter. I hacked off the AC end of the cable and hooked it up to the 5v rail from the opus, so it's drawing its 5v at up to 2.4A from the Opus.

    However, I have one problem. The wifi adapter and most other devices will power on and then immediately shut off when they detect that there is not enough power available to them. I don't know how they do this, but they do, and it's a hardware thing not software - this all happens within a split second.

    When I turn on the system, the hub comes right on, as do the USB devices, then they power off just like they would if the hub was self-powered (not enough power). Now, if I unplug all the devices and plug them back in everything works fine. Best I can tell, here's what's happing. The OPUS is powering up the USB 5v a split second before it powers up the 5v rail that is supplying the extra current to the hub, and in that split second, the USB devices have time to power up, think there's not enough power (5v rail hasn't been energized) and shut off. Shortly after, the 5v rail comes on and it can fully support everything once I unplug and replug all of the devices.

    Is there any way I can make this work? I figure I just need to make the hub wait until the 5v rail is powered up before it turns on. I figured I would do this by cutting the 5v+ from the usb cable so it didn't have power until it got it from the 5v rail, but when I do this Windows doesn't even recognize the hub. For some reason it needs to be drawing power from the USB 5v line to be recognized. I verified this on other computers as well.

    Can anyone figure out a solution?

    Thanks in advance,
    Dan

  • #2
    I figured I would do this by cutting the 5v+ from the usb cable so it didn't have power until it got it from the 5v rail, but when I do this Windows doesn't even recognize the hub.
    This doesn't make sense. You said the 5V line comes up just after the 5V in the cable, how has windows loaded so fast and already detecting devices at this point?

    Comment


    • #3
      are you using WinXP or Win2k?
      2007 Honda Fit Sport 1.5L SOHC-VTEC

      Comment


      • #4
        Originally posted by MrPerfectionest
        This doesn't make sense. You said the 5V line comes up just after the 5V in the cable, how has windows loaded so fast and already detecting devices at this point?

        It hasn't. This all happens within a split second. Like I said, this is a hardware problem, not software. There is not enough power for all the USB devices once the 5v USB bus comes up, so they shut off. Nothing to do with Windows turning them off. And they won't come back on (and stay on) until the extra current is available from the 5v rail. Once it's there, they'll stay on fine. I just don't want to have to unplug and replug every USB device every time I turn the car on...

        Does that make sense?

        Comment


        • #5
          If you don't have enough power then add in a POL...

          Comment


          • #6
            Originally posted by Montanaja
            If you don't have enough power then add in a POL...
            Ugh. Did you read my post? I HAVE enough power, it's just not getting supplied right.

            Comment


            • #7
              Again... I think the easiest fix would be a 5v POL.

              Comment


              • #8
                Again, I ALREADY HAVE enough power. The USB bus would STILL COME ON before the 5v with a POL. Ugh...

                Comment


                • #9
                  You could try wiring your 5v from your opus directly to the usb connection for your usb hub (obviously disconnecting the existing usb power supply first..)??

                  It seems a bit weird that the hub doesn't work when the 5v usb supply is disconnected and the hub is running on an external supply (never tried it before).

                  Prem
                  Re-boxed Dell Optiplex GX110 PIII 733 256Mb RAM
                  Modified Sproggy PSU
                  Custom made automatic start-up/ shut-down controller
                  Dual 2.5" HDDs
                  Laptop DVD Reader
                  D-Link FM Radio
                  5" PSOne LCD
                  Installed and working. 75% complete.

                  Comment


                  • #10
                    it is kind of a bandaid, but you could use devcon to disable and reenable the hub (once you get into windows) which is basically a software unplug the plug. Not the best solution, but it might work until you find something better.

                    |V1 Virtual Display|

                    Hard work often pays off over time, laziness always pays off right now.

                    Comment


                    • #11
                      My powered USB hub works similar to yours. The AC adapter that came with the hub was producing about 7.5 volts DC into the hub. I thought I might get by trying to power the hub up with some existing regulated 5 VDC on my shutdown contoller but the hub behaved very eratically. What was happening was putting 5VDC into the hub was not enough overhead for the on board 5v regulation needed to drive each port so the port voltages were dropping down to about 4.3 volts and thus some devices seemed to be OK, others not so much.
                      I added a 7.5 volt rail to my controller and now the hub powers up reliably with all ports putting out 5 volts.

                      My guess is your hub with only 5 volts applied has the port voltages dropping low during initial inrush of current during power up and everything goes dead. If you plg the devices in one by one you don't have a big current draw and the hub is able to keep the port voltages high enough to satisfy the device.

                      Comment


                      • #12
                        Originally posted by dmurray14
                        Basically, the hub can act in 2 ways: self powered and externally powered. I have it set up to be powered externally from the regular 5v rail of the opus. HOWEVER,

                        Apparently, the Opus must be powering up the USB 5v first, because the hub gets powered up, the devices act like they're going to start, then they all shut off because there's not enough power (happens in a split second before the 5v rail comes up to provide extra power.) Once the 5v rail is energized, I can unplug and replug everything in and it will work fine as there is now plenty of current to go around.

                        I guess I need to find a way to make sure the hub doesn't turn on until it gets power from the 5v rail, I tried doing this by cutting the 5v wire in the USB cable (thereby making it powered only by the 5v rail) but Windows won't even recognize the hub unless it's drawing from the USB 5v line.
                        Just adding your response to my question in this thread, because this one seems to have a bit more activity.

                        This is really strange.

                        I have an Opus (150W, if it matters) and have a powered hub running off of the molex connector and it's working great.

                        jayearl, if you are correct and the situation is actually similar, then this is easily fixable. But there's no mystery here... the USB hub's transformer should tell you the output and at what potential.

                        I doubt murray made that mistake, although it is easily possible to overlook the obvious on occassion. I do it all the time

                        Anyways...

                        Montanaja, how would you wire the POL? If you wire it up to an ACC line in the car, then this should fix the problem, assuming that murray's theory is correct.

                        However, doing this will lead to some strange behaviour by the USB devices because they aren't synchronized with Windows. You'll be fine on startup, but strange things will happen if you start the car back up without turning the computer off, or if you hibernate or suspend.

                        Take the Xenarc, for example. I am redoing my wiring because it is running off its own PSU and not the Opus. I'm just going to be using a relay from the 12V rail to control its power, as opposed to having the touchscreen on whenever my car is in ACC or started (my wiring diagrams are here): http://mp3car.com/vbulletin/showthread.php?t=65163

                        Regardless, let's assume that murray's theory of USB power is correct.

                        I don't understand why this is different from my situation. Maybe the motherboard is a little quick to initialize the USB? Or the Opus is slightly different?

                        Anyways, here's a way to test the theory:

                        What happens if you take a USB device that is fine being powered through an unpowered hub.

                        Now:

                        Don't power the hub, but plug in this USB hub, and boot up the computer.

                        Note the results, and turn off the computer.

                        Power the hub via the molex connector on the Opus. Keep the same USB device plugged in, and boot up the computer.

                        Note the results.

                        I would think that if the theory is correct, the device should work in both situations, because there always *should* have been enough power.
                        '03 Intensa Blue Pearl Lexus IS300

                        Specs: Xenarc 700TS, Opus 150W PSU, CarCPU Case, 2.0 GHz Celeron D, DFI PS-35-BL, SB Audigy2 NX, Centrafuse 1.4

                        Progress: Car PC up and running!
                        To Do: Navigation, WiFi, AM/FM Radio, DVD

                        Comment


                        • #13
                          Originally posted by zPilott
                          it is kind of a bandaid, but you could use devcon to disable and reenable the hub (once you get into windows) which is basically a software unplug the plug. Not the best solution, but it might work until you find something better.
                          I wasn't aware of this, but I think that I may use it for my situation. I have an external HDD that won't be recognized by Windows unless it is plugged in after the OS loads. I believe that the problem is because it does not get enough power via USB upon initialization. I don't want to directly power it, so this seems like it would be a good solution.

                          Would one just write a BAT file with a couple of commands for devcon and run it on startup? What if sometimes the device is actually not connected?
                          '03 Intensa Blue Pearl Lexus IS300

                          Specs: Xenarc 700TS, Opus 150W PSU, CarCPU Case, 2.0 GHz Celeron D, DFI PS-35-BL, SB Audigy2 NX, Centrafuse 1.4

                          Progress: Car PC up and running!
                          To Do: Navigation, WiFi, AM/FM Radio, DVD

                          Comment


                          • #14
                            Originally posted by azzurro
                            Just adding your response to my question in this thread, because this one seems to have a bit more activity.

                            This is really strange.

                            I have an Opus (150W, if it matters) and have a powered hub running off of the molex connector and it's working great.

                            jayearl, if you are correct and the situation is actually similar, then this is easily fixable. But there's no mystery here... the USB hub's transformer should tell you the output and at what potential.

                            I doubt murray made that mistake, although it is easily possible to overlook the obvious on occassion. I do it all the time

                            Anyways...

                            Montanaja, how would you wire the POL? If you wire it up to an ACC line in the car, then this should fix the problem, assuming that murray's theory is correct.

                            However, doing this will lead to some strange behaviour by the USB devices because they aren't synchronized with Windows. You'll be fine on startup, but strange things will happen if you start the car back up without turning the computer off, or if you hibernate or suspend.

                            Take the Xenarc, for example. I am redoing my wiring because it is running off its own PSU and not the Opus. I'm just going to be using a relay from the 12V rail to control its power, as opposed to having the touchscreen on whenever my car is in ACC or started (my wiring diagrams are here): http://mp3car.com/vbulletin/showthread.php?t=65163

                            Regardless, let's assume that murray's theory of USB power is correct.

                            I don't understand why this is different from my situation. Maybe the motherboard is a little quick to initialize the USB? Or the Opus is slightly different?

                            Anyways, here's a way to test the theory:

                            What happens if you take a USB device that is fine being powered through an unpowered hub.

                            Now:

                            Don't power the hub, but plug in this USB hub, and boot up the computer.

                            Note the results, and turn off the computer.

                            Power the hub via the molex connector on the Opus. Keep the same USB device plugged in, and boot up the computer.

                            Note the results.

                            I would think that if the theory is correct, the device should work in both situations, because there always *should* have been enough power.

                            Your test works exactly as you predicted, boots up fine. HOWEVER, here's the thing...I cannot power the hub from the 5v rail only, as Windows WILL NOT detect the hub unless it is drawing power from the USB 5v. I have NO IDEA how it can tell whether power is being drawn or not, but it can, and that is preventing it from working off only the 5v rail. Of course, as long as the USB 5v is plugged in, I still have the same problem - the usb 5v is powered up before the 5v rail.

                            The only thing I can think of at this point is to maybe try and wire in a resistor and see if I can fool the motherboard/windows/whatever into thinking there is something there - but still, I would have no idea what size resistor to use, and for that matter whether that will "fool" it.

                            Dan

                            Comment


                            • #15
                              Originally posted by zPilott
                              it is kind of a bandaid, but you could use devcon to disable and reenable the hub (once you get into windows) which is basically a software unplug the plug. Not the best solution, but it might work until you find something better.
                              I am using devcon, but it will do nothing for this situation. Again, the problem isn't in Windows or any software - the devices themselves are shutting themselves down right away, long before windows even boots.

                              Comment

                              Working...
                              X