Am Dienstag, 5. Juni 2007 09:48 schrieb Vortex: > > Increasing the capacitors make things *worse*, not better! Empty > > capacitors are like short circuits. You have to limit the current flowing > > into empty capacitors by applying a resistor or (better) a choke coil to > > the supply lines. > > But i think the usb power supply isn't switched on or off. So the > Capacitors have to be loaded only if i plug in the device. And that's > not the problem. > USB Power is usually switched, besides from non-standard ports on certain desktop mainboards. The capacitors are *not* loaded as soon as you plug in the device in that case.
As soon the host controller detects the device speed resistor -- the 1.5k resistor between D+ and +5V if you built your adapter as the DS2490 data sheet suggests -- it reports that incident to the driver and the USB host controller driver turns on power for that port (or not, if no driver is loaded). In case a device has a failure, the usb driver can turn off the power for the specific port. It may do so if the device is to be powered off due to power saving issues. E.g. a laptop could turn off all USB devices but an external USB keyboard or card reader during STR state. The power is switched off if the presence resistor is there but the device does not respond to the initiation handshake, too. > The failure happens after a few successful > transactions. > Ok. so the problem isn't strictly related to power-on. It may be a related issue, though. > If the power supply would be switched the choke coils > would make bad things too. > At least if they are large enough to store > reasonable amounts of energy. (Wasn't there something like u_L=di_L/dt? > At least there would be large voltage peaks then....) > They would induce an additional voltage, but this is low with reasonable small coils as the switch chips don't turn off too fast. > > > USB circuits are powered on on detection/driver-reload, so this power-on > > problem is more inconvenient than with other devices. > > Are you sure? Of course the peripheral device may be in some sort oft > sleep mode until it's initialized by a driver. But the power supply on > the usb-bus is on all the time. At least in my experiences. > For some mainboards, that may be true -- but it is non-standard. See above. The reason may be that absurd USB "devices" like this http://www.bitmark.com/giftmugs/sampler/various/Mug-Boss/Coffee-Mug-Warmer.html won't work with switched ports, as they have no USB device controller for cost reasons. > > The facts about burning the fuse are certainly right. But this current > peaks should be blocked by the capacitors in the usb power supply lines > (33pF, 100nF, 4,7uF). If this is not sufficient i guess i would need > a choke coil instead which is larger than the whole circuit itself :-). > Capacitors can block current peaks if they are *fully loaded*. If they are empty, they cause additional current flow. That's why you need an LC or RC circuit, not a capacitor alone. The choke coil is there to *limit* the current flowing into the empty capacitor, not to provide energy to the circuit. You could use an 1..6 Ohms resistor, too, but a ~µH coil has the advantage of >>6 Ohms resistance on power-on and nearly zero resistance as soon the power flow is at steady state. If you want to try with a series resistor: A 6 Ohms resistor will limit the current at 5V short circuit to about 833mA. That may be low enough not to trigger any overcurrent circuit in the host adapter. Why 6 Ohms? Because it will too lower the supply voltage about 0.6 Volts at 100mA. 4.4 Volts is the maximum tolerable for the DS2490 to operate. If your circuit draws lower current, you may increase the resistor, though. Kind regards Jan -- 43rd Law of Computing: Anything that can go wr fortune: Segmentation violation -- Core dumped ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Owfs-developers mailing list Owfs-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/owfs-developers
