[Tim D]

I have an USB audiophile attached to a Dell Laptop (3-prong power supply). I get some ground noise with the power supply attached which is removed when disconnected using battery or with a cheater plug. Since the external unit shouldn't be drawing usb power anyways is it possible to modify the USB cable and disconnect or break the ground connection at that point rather than at the power supply and still have a functioning usb cable?

 

[chaostic]

IMHO, you can try and be pretty sure you won't break anything besides the cable.

 

[Tim D]

Welp, I did it myself, put electrical tape over Pin 4 of B adapter and it works with no noise. It seems to be much safer to lift the ground from this side as opposed to the laptop power supply as the external device is 2 prong anyways and it should have a ground connection from the IC's anyways which I imagine is quite a bit safer than having the ground supplied by only a hot-pluggable usb cable if I lifted the ground from the laptop power supply.

I'll either stick with electrical tape or see if I can remove the pin contact entirely somehow.

From what I've seen firewire interfaces can suffer from the same ground issues but I'd imagine its more difficult to meddle with since its a smaller connecting interface (but you could hack the actual cable I suppose). It'd be interesting to know if a similar solution works for firewire.

 

[Tim D]

With the ground pin removed (a good tug of needle nose pliers), the laptop power supply is required for the usb device to be seen. I guess since there is no ground reference for the usb connection anymore with the pin removed and working off battery. I'll see what happens with my desktop but it should work since its grounded.

 

[Tim D]

Works on my desktop as well. The ground is supplied by the analog unbalanced interconnects for the usb audiophile side. I am unable to perform a loop back test with a 3 pin cable since actually if the outbound cables going to the pre/amp are removed the device won't show/work since there is no ground apparently.

However I am guessing the performance is theoretically now at what the loop back would have given previously with a full 4 pin usb cable since in that scenario there is only one reference ground as well. I was getting shoddy results when introducing a grounded pre-amp/amp into the chain and this mod gives me a good 10db improvement across the boards in noise and distortion, etc.

Anyways its pretty trippy to see the usb device pop into and out of recognition by adding/removing the analog IC's! But given the way the USB audiophile is designed with hotplugability, it seems much easier/safer to lift the ground at the usb cable than the power adapter, or stereo/headphone equipment (i.e. considering you can power on the audiophile without usb cable or interconnects this should be just as safe).

Just my experience with an USB audiophile, YMMV.

Now that I think about it, my Rio Karma dock with power supply, usb cable, network cable, and RCA cables connected may also benefit from this mod, but the battery life is good enough, and the usb cable uses the mini B connection which is less common to not bother. Using RMAA the Karma + dock actually outperforms my Denon CD player when running off battery and achieves near 16 bit /44 theoretical performance lol.

Followup - It seems like I can remove both power pins (1 and 4, i.e. the longer ones on the male B end), and still have it operate. However on my old Dell Laptop there are some intermittency issues where the device drops after awhile. This does not happen on my desktop which operates perfectly. It could be something with the usb controller? The USB support on my laptop was always a bit iffy in the first place (i'd always have pops and crackles for previous usb audio class devices). The power supply/grounding noise was even much higher on my Dell Latitude c600? compared to my desktop as well. Anyhow I'm long due for a new laptop anyways, meanwhile I'm stuck floating the laptop mains or using battery or having a grungy signal. Oh well, at least its cleaner from the desktop (ignoring the extra fan noise of course

).

Gah final followup - It seems the intermittency was just the oxidized/worn usb connector on my laptop. Not sure why it was 'revealed' with the no power/2 pin usb cable compared to the 4 pin. Anyhow an application of pro gold and repeated unplugging/plugging did the trick. For all the promises of easy external digital interfaces, it seems that the oldest tricks in the books with grounding and pro-gold/cleaning contacts still apply

.

 

[pburke]

very interesting findings. I am fighting a similar issue with a different USB interface right now. I do have the M-Audio Audiophile USB and that really never was an issue with the laptop on battery, while running it on the charger pretty much was a mess. I can actually hear that charger in my audio system not even connecting it to the USB DAC at all, just by the amount of crudd the cheap Dell switching supply is feeding back into my AC lines.

But back to the new DAC, where things are much worse than with the M-Audio. I already have found a cable I'll be rigging up as a "cheater cable tonight to see if I can replicate your results with that unit (it is battery powered, but the designer said he could not drop the power pin as it connects to the "HOST" pin on the PCM2707 chip. According to the datasheet, that's what is needed to have the device show up on the USB bus in self-powered operation (makes no sense to me - what the heck do you want to self-power it for if you still need to connect it to the USB power to do so? or am I reading this the opposite way from what it is used for. Is self-powered equivalent to powered by USB bus?)

http://focus.ti.com/lit/ds/symlink/pcm2707.pdf

page 6, pin 3

The M-audio gear uses very different chips, so there may not really be any correlation between the two interfaces.

I guess I will find out what happens when I remove that pin on the cable tonight... Maybe if it's just the USB chip that wants to see power on that pin, I can build a small 5V source on the battery supply and feed that to the pin to fake the computer connection. Well, this is definitely something I need to work out before I consider building a dedicated music server.

Peter