Host USB ports are supposed to use a MosFET with internal current sense for power protection. The FET signals the O/S when its current threshold is being exceeded and then the O/S suspends the port and removes power.
If you apply +5v INTO the host port, you can create current spikes across the FET which would damage the FET, trip the O/S into thinking a power fault has occured and shut down the port. Even in this state, the reverse bias diode in the FET will continue to allow the external voltage to be fed into the the motherboard.
The FET/Motherboard external voltage will keep fighting each other as the ripple in the motherboard +5v and your external +5v criss-cross each other.
This would probably be bad....
You can diode OR all of your mother board/external supplies together and then run them into a buck/boost regulator. THat way the buck/boost will maintain a local +5v supply at your board "regadless" of external voltages.
Thoughts?
Steve Spano FLE
From: Joseph Smith joe@settoplinux.org Date: 2008/11/15 Sat PM 03:53:43 EST To: Peter Stuge peter@stuge.se CC: coreboot coreboot@coreboot.org Subject: Re: [coreboot] LPCflasher Project
On Sat, 15 Nov 2008 21:26:05 +0100, Peter Stuge peter@stuge.se
wrote:
Joseph Smith wrote:
Anyone know if USB ports have power feedback protection.
All the details are in the USB spec. IIRC electrical is a separate chapter.
What I mean by that is if I use the LPCflasher as a inline flasher with a PLCC32 socket plug,
I don't understand this. Do you mean a socket or a plug?
Both, a socket on top of a plug that is also connected to the LPCflasher. If I power on the motherboard it is going to supply voltage back to the LPCflasher, right?
flash the chip
Flash which chip?
Any LPC or FWH.
and then power up the motherboard, I don't want the VCC's from the motherboard to short out the flasher and or the USB port from the host PC.
Make sure GND is connected across and you should be OK.
Yup.
The 3.3V regulator I am using has a internal diode so I am not so worried about that. I am just worried about the power source 5V USB. I could always use a Schottky diode that will only drop the voltage by .6V, but I don't know if 4.4V will suffice???
Suffice for what? Look at the documentation for the regulator you're using to see what input voltage it needs.
I wasn't really ready to release this yet, because the power part is still a work in progress but it will be the easiest way to explain this. See the diode I have above the regulator? I put that there so power is not flowing back to the USB port of the PC doing the flashing when you power on the motherboard that just got flashed. My question is if I take out that diode and you power on the motherboard that just got flashed is power going to flow back into the USB port and damage it? And if I leave the diode it is going to drop the 5V by approx. half a volt. Will I still be able to flash 5V chips at 4.5V? Does that make sense?
-- Thanks, Joseph Smith Set-Top-Linux www.settoplinux.org
-- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
On Sat, 15 Nov 2008 16:09:58 -0500, steve@fl-eng.com wrote:
Host USB ports are supposed to use a MosFET with internal current sense
for
power protection. The FET signals the O/S when its current threshold is being exceeded and then the O/S suspends the port and removes power.
If you apply +5v INTO the host port, you can create current spikes across the FET which would damage the FET, trip the O/S into thinking a power fault has occured and shut down the port. Even in this state, the reverse bias
diode
in the FET will continue to allow the external voltage to be fed into the
the
motherboard.
The FET/Motherboard external voltage will keep fighting each other as the ripple in the motherboard +5v and your external +5v criss-cross each other.
This would probably be bad....
That's what I was thinking, thanks.
You can diode OR all of your mother board/external supplies together and then run them into a buck/boost regulator. THat way the buck/boost will maintain a local +5v supply at your board "regadless" of external voltages.
Thoughts?
Ah, so I can use a Buck-Boost regulator to compensate for the diode voltage drop. That makes sense. In that case I could just cut out the simple LED circut, use the Buck-Boost regulator to boost the 5V to 8V and put a LED inline as my diode and the output should be a solid one way 5V power line right?