according to official specs, not more than 500mA should go over the wires, but many devices seem to ignore that; and on my PC, that causes the mainboard to protect itself by simply switching off the powersupply.
are there any schematics about how to build a limiter for usb? eg, get an usb extention cord, cut it and solder in a few parts?
from what I know, I cannot just reduce usb to the data wires only since it needs the +5v/gnd for handshaking or such.
you could just add a series resistor to limit the current, but you'd get a voltage drop as well.
Not sure about limiting current, but you could get a powered usb hub. Then your pc will not see any of the load
>>984117
i think you're asking the wrong question
limiting the power will prevent your usb devices from working
go get a powered usb hub.
>>984169
>limiting the power will prevent your usb devices from working
I would agree if I'd try to use eg usb-only harddrives; but the ones I use have a wallplug
my problem are mainly battery powered devices. there seems to be some agreement amongst developers which boils down to "zomg, power! let's suck in as much as fast as we can, who knows when there is more!"
>>984176
Devices which don't actually speak USB but simply use a USB connector for charging should be plugged into a charger, not your PC.
Devices which actually use USB for communication should know better than to draw more than 500mA (or more than100mA without negotiation).
Even devices which only use USB for power should still know better than to draw significant power from something which doesn't identify as a power supply. However, enough people are willing to put up with this crap that it continues.
>>984195
>should
>should
>should
that's one ideal world you're dreaming of
it's not like that though, so I'm searching for a solution
>>984117
>since it needs the +5v/gnd for handshaking or such.
in theory it shouldn't need them, but its common to use the appearance of host voltage to start the device handshake process.
you could replace the host 5v with a 5v power brick and just plug/unplug the device end last/first.
you can configure a linear voltage regulator as a current source very simply however it will act as a current source, varying the voltage to try and push the configured current down the line.
a current limit like you want is not trivial.
>>984216
> it's not like that though
That's your choice. Or is someone actually holding a gun to your head and threatening to shoot you unless you plug this crap into your PC?
If a device only uses USB for power, buy a wall-wart with a USB port. It's simpler (and probably cheaper) than trying to make a current limiter, and has the added advantage that it will actually power the device (all that a current limiter will do is stop the device from crashing your PC; it won't make the device actually work).
OTOH, if you're talking about something which actually needs to be plugged into a PC, it's unambiguously defective. Send it back, reverse the charge, give the product negative reviews wherever you find it, and report it to whatever agency deals with consumer protection in your neck of the woods. Or don't; again, it's your choice.
>>984117
>on my PC, that causes the mainboard to protect itself by simply switching off the powersupply.
Just get a 5v USB charger you double nigger. Given how fast micro USB cables seem to wear out, most people probably have one or two extra chargers around their houses that they don't have a corresponding cable or device for. Just bum one off a friend and spare your mainboard and PSU 5v rail the trouble.
>many devices seem to ignore that
I have literally never seen these devices that you speak of.
The only examples I can think of that (can) take more than 500mA are hard drives and devices with batteries.
Devices with batteries often charge at the maximum rate allowed by the controller, or not at all if the current rate doesn't match (iPhones).
Hell, USB 3.0 is specced to go to 1.5A, and USB 3.1 with Power Delivery allows up to 100W (5-12V with auto negotiation)
I don't know what's up with your motherboard but usually when I plug in a device that sinks too much current, the device gets shut down and not the PC.
>>984134
This.
>>984117
>according to official specs, not more than 500mA should go over the wires
No, that's wrong. USB spec states that USB ports must be capable of supplying a MINIMUM of 500 mA, and that 500 mA is therefore the amount of current that the spec guarantees for any one device. Any USB port that supplies more than 500 mA still meets the spec, but any device that draws more than 500 mA is not technically to spec and is thus not *guaranteed* to work with all USB ports (though it still might if the port in question can supply the needed current).
>>984266
>>984269
>>984355
I'm not charging the device; actually I don't want to charge it at all, just use it.
Examples? There's a smartphone (HTC I think) which I have to use (business wise, for emergency) which also contains an encrypted file with various data, like docs, logins etc. When I connect it, blam, my PC shuts down. It works at $job, but that PC there is not even a year old, while mine here is ~5 years. Similar experience with some mp3 players.
Cameras and sticks work fine because they are not charged via usb.
And powered hubs aren't a perfect solution either, because quite a few of them seem to backfeed:
http://electronics.stackexchange.com/questions/75654/7-port-powered-usb-hub-backfeeding-to-my-pc
http://qdosmsq.dunbar-it.co.uk/blog/2012/08/the-dangers-of-cheap-usb-hubs/comment-page-1/
In an ideal world these problems would not exist. In an ideal world there wouldn't be any attacks on computers, but we still use firewalls.
All I want is the equivalence of a hw-firewall to protect a PC from all those crappy engineers who save $0.01 on a product by ignoring standard.
>>984500
>>984500
>crappy engineers who save $0.01 on a product by ignoring standard.
Like the ones that designed your motherboard or assembled your PC?
Devices don't "draw current". They present a resistance, and the power supply either maintains the voltage by supplying current, or it sags.
A computer is supposed to do its own voltage regulation of the USB ports, so that a dead-short can't develop hundreds of amps, start a fire, and sag the main power supply. (USB charging spec actually requires devices trying to charge to increase their load cycle until the port starts to sag, thus discovering the maximum current it can supply.) What a computer is supposed to do, when its USB voltage regulator sags too far, is stop supplying the port until the device is removed. Not crash.
The reason the device you're looking for doesn't exist is that 99.9999% of computers handle USB overcurrent correctly.
Maybe get a PCIe USB card that isn't retarded?
>come to /diy/
>buy a usb hub
>buy a charger
>buy a pci card
wat?
>>984577
Oh, I'm sorry, were you expecting "reimplement the entire USB specification (plus ten years of errata) from scratch, after building your own IC fab out of shipping containers"?
>>984585
>Oh, I'm sorry, were you expecting "reimplement the entire USB specification (plus ten years of errata) from scratch, after building your own IC fab out of shipping containers"?
No, why? There is functional USB port.
All I wanted was some help with limiting the mA that can be drained from the port. That's why I asked here, because I assumed those coming to /diy/ know more than "buy this or that".
After some research I found this. Doesn't look like an overly complex rocket science project. Depending on T1/T2, I might already have all the required parts around here somewhere.
Alternatively, I could also buy an L200 IC (Programmable Voltage Regulator with Current Limiting)
>>984500
Out of curiosity, who manufactured your computer and/or motherboard?
Just taking note so that I can avoid them for the rest of their short, short existence.
>>984559
>Devices don't "draw current"
I get what you're saying; the term isn't perfectly fitting given what's actually going on (which is more like "passing," "bleeding," "draining" etc.), but like it or not, "draw" is the accepted "correct" term.
>They present a resistance
Not all loads are ohmic, but sure. I guess.
>>984776
DH67GD
>>984781
No wonder Intel doesn't make desktop mobos anymore.
How sad.
>>984559
>99.9999% of computers handle USB overcurrent correctly.
Macs work fine :^)
>>984766
>No, why? There is functional USB port.
I don't know how many times anons have to spell this out to you, but if plugging in a device causes your USB ports to supply too much current and shut your computer down, the PROBLEM IS WITH YOUR USB PORTS or power supply somewhere, not the USB standard and not with the devices.
Nobody else has this problem, and the "ideal" world you're talking about where this problem doesn't happen, literally exists for everyone else but you. I don't know what you aren't understanding.
The devices aren't "sucking" all the power out of your computer and killing it because of "crappy engineers".
The devices are being "fed" too much power BY your computer, that's the problem, and the fault lies with the "crappy engineers" who created your motherboard or whichever piece of circuitry in your PC you cheaped out on and isn't functioning correctly.
>Not all loads are ohmic, but sure. I guess.
Kek, so you flunked out of electrical engineering 101 and now you're trying to look smart huh retard? That has nothing to do with anything.
Why ask for advice and then shoot down everyone trying to help you? Like these guys you are arguing with even though they are right:
>>984195
>>984266
>>984355
>>984559
>>984585
>>984787
My window has done this. If you short it, it knows immediately and disables the port.
>>986282
>>984787
I think that's for dangerous shorts, not stuff like your ThinkGeek USB Mug Warmer drawing 750mA instead of 500.
>>984117
>from what I know, I cannot just reduce usb to the data wires only since it needs the +5v/gnd for handshaking or such.
If for whatever reason you only need power, you can absolutely do the opposite and your computer will gladly power whatever you connect to it.
>>984117
>are there any schematics about how to build a limiter for usb? eg, get an usb extention cord, cut it and solder in a few parts?
yes you can: google for info about "LM317 adjustable current limiter",,, but you would need a separate limiter for each specific device.
http://www.ti.com/lit/ds/symlink/lm317.pdf
see section 9.3.3
>from what I know, I cannot just reduce usb to the data wires only since it needs the +5v/gnd for handshaking or such.
The first thing I would do here is get a short USB extension cord and cut into it, so that I could use a multimeter to check what each device is drawing.
I would expect it to be way more likely that it is just one device that is the culprit.
