View Full Version : how do i ground my Front USB better?
farlo
03-01-2011, 12:45 PM
sorry if this is in the wrong area, but i've been noticing lately that my front usb ports arent grounded very well. I have a line6 usb preamp for my guitar, and i get a bit of extra buzz. i touch the usb ports on the front, and the buzz goes away, so im pretti sure they're not grounded very well.
i have an old coolermaster ammo 533 (http://www.coolermaster.com/product.php?product_id=5) case, and was looking for ideas on better grounding.
Drum Thumper
03-01-2011, 04:29 PM
I ran into a similar issue with a coolermaster case a while ago as well. I fixed it by replacing the case.
What I had happening is if I did not ground myself out prior to plugging a flash drive in, it'd instantly BSOD the computer.
farlo
03-01-2011, 04:58 PM
i never have issues with bsods or anything, its just a minor annoyance with audio equipment adding an extra slight buzz.
maybe if i solder a wire to the metal around the usb socket, and just screw the other end of the wire to the inside of the case that may help.
dr.walrus
03-01-2011, 06:14 PM
maybe if i solder a wire to the metal around the usb socket, and just screw the other end of the wire to the inside of the case that may help.
seems to make sense, try it with tape first to make sure you're not wasting yr time?
CorsePerVita
03-01-2011, 06:26 PM
I can't even use mine anymore. When something so much as TOUCHES my usb ports on the front there is a quick bzzt and the computer locks up and freezes. I have to then restart.
Konrad
03-02-2011, 03:04 AM
The simplest solution is to use other USB ports. What you're hearing might be 60Hz mains hum, if so, it'll occur on all the USB ports and be unresolvable without a battery powered UPS, line conditioner (isolation transformer), upgraded (PF efficient, APF capable) PSU, or 60Hz line filter (notch filter (http://freecircuitdiagram.com/2008/08/08/hum-filter-an-ultimate-weapon-to-remove-ac-line-noise/) circuits are cheap/easy builds). Obviously you want to simplify by plugging straight into the port, every additional USB hub and gizmo on the line can introduce new problems. Most USB ports have two connectors (sometimes more), don't plug a sensitive device into the same octopus as USB devices (like keyboards or fans) which generate electrical oscillation.
Original USB specs contained 5-pins, the fifth being an external/floating ground (intended to support coax USB cabling, which never happened). I would try connecting pin 5 (that is, the external shell on the USB receptacle) to the PC chassis (ideally at one of the brass mobo standoffs or the PSU shell or even a black GND wire on a PSU connector); officially you should connect to your mobo (at pin 10 in this (http://attachments.techguy.org/attachments/103260d1176642862/standard-usb-pinout.gif) diagram), though the spec is obsolete and I really wouldn't want to have ground loops flowing across my mobo. The assumption here is that the USB, mobo, PSU, and chassis all have different ground potentials; so the idea is to link them together as a common ground at a single point.
Wiring doesn't need to be thicker than 28G, though fatter ground lines never hurt; I would use 24G just because I'm neurotic that way, so many problems I don't want to deal with can simply evaporate with solid electrical grounding. Soldering isn't strictly necessary, PVC tape or zapstraps might suffice to tightly secure these extra wires.
You could also buy/build a USB port or hub with optoisolator circuits; this would be the electrically cleanest signal, though often with a bandwidth cap imposed by the opto component - probably not an issue for your audio signal. Some PCIe USB cards provide optoisolation, a bit overkill (unless you require USB3.0) but a great way to ensure independent "clean" USB service for mission-critical devices.
USB specs allow maximum USB cable lengths. These are conservative, and fault-tolerant, and everybody in Shenzhen ignores them anyways ... but there are practical limits. You just can't reliably use 20m of USB cabling when the spec demands 2m; longer cables suffer from signal degradation (LC voltage losses, EMI/RFI absorption, etc), you might actually have a cable run acting like an antenna (especially in proximity to electrically noisy amps, transformers, oscillators, and switches), it's not unheard of. Longer USB cable runs are possible (in theory) if you replace the cheap USB-spec data cables with better ones, even CAT5e stuff is vastly superior if you don't mind making them (soldering USB plugs onto the ends).
You can also pay a little premium for gold-plated USB connects. Audiophiles swear by them, although I personally think they have merit but are overhyped; electrical contact cleaners should generally suffice. Isopropyl (anhydrous) should work well enough, stuff like DeoxIT should work perfectly, although Stabilant-22 (once marketed to audiophiles in a diluted form under the brand Tweek) is definitely the best (and most costly).
I'm assuming your audio/amp equipment is impedance-matched. If not, you'll hear some kind of noise even when plugging into battery-powered laptops; this would not be a computer/USB problem. You might have issues/limits imposed by your PC software or audio hardware, impossible to say.
xr4man
03-02-2011, 09:28 AM
i have a similar problem and was thinking about grounding the usb connector shell. i'll probably do that once i have the new case ready to have my internals transplanted.
farlo
03-02-2011, 10:23 AM
Konrad, very thorough very informative post. thank you.
Konrad
03-02-2011, 10:48 AM
lol, well the short version is that USB is pretty good but it's not magic and it's not foolproof; it's still governed by the same basic electrical principles that apply to every other kind of circuitry. Modern computers are phenomenal machines, but they still build 'em cheaply and auxilliary components (like USB) are usually the first to suffer from substandard part choices when total platform costs need to be tightened; as long as it can claim minimal compliance to advertise a USB 3.0 SuperSpeed label it doesn't much matter how robust the part specs really are.
Powered by vBulletin® Version 4.2.1 Copyright © 2025 vBulletin Solutions, Inc. All rights reserved.