Tri-X Fury causing ground loop?

[SpecChum]

I fear I’ve got a ground loop issue which seems to be caused by GPU activity.

I was originally using the line out of my Xonar DX into a Fiio E11k and didn’t really notice any issues but wanted an upgraded desktop amp, so decided to get a Little Dot I+ tube amp. However since doing this I get a slight electrical sound when moving the mouse and an awful screeching sound when using the GPU; playing games is a complete no go. Even the GPU test on GPU-Z causes it. Increased CPU load doesn’t seem to affect or indeed cause the interference; it’s only the GPU.

Both the amp and the PC are going into same extension strip and I’ve changed the power leads on them both to no avail.

At first I thought it was just EMI interference from my DX as it’s not shielded and sits right above the graphics card. I borrowed a SoundBlaster X-FI HD USB sound card and the interference is still there; obviously this connects to the PC via USB so at that point they share a common earth.

To rule out the Little Dot I+, which is a hybrid amp, I borrowed an all-tube Little Dot MkII and the interference is still there. Moving the Little Dot further away from the PC doesn’t help. Interference is happening whenever the RCA cables are connected – both amps are dead silent otherwise. To confirm this, there’s no interference at all when I plug my Little Dot directly into my Galaxy Tab S using the exact same RCA cables, so it’s definitely the PC.

Incidentally, I noticed last night that whilst I get no interference at all using the E11k when it’s plugged in to charge from an actual socket charger, I do get interference once the USB cable is plugged in to the PC; this, to me, pretty much confirms a ground loop issue on the PC.

Relevant specs are:
Corsair 230T case
MSI Z97 Gaming 5 motherboard
Sapphire Tri-X Fury at stock speeds
EVGA SuperNova G2 850W PSU

Anyone have any suggestions please?

 

[Ma_Deuce]

I would try to isolate the amp and the PC with an isolation transformer if you have one available to you. Or one of the power conditioners that are designed to take the cross talk out of home theater systems.

Sometimes it will work, sometimes not but it's worth a try.

 

[Phynaz]

That's not the symptoms of a ground loop. That's the syptoms of a bad switching power supply throwing off emi under load.

 

[SpecChum]

That's not the symptoms of a ground loop. That's the syptoms of a bad switching power supply throwing off emi under load.

Cheers, but I don't think it's this for a few reasons.

My psu is a fairly well respected evga g2 plus emi interference wouldn't create a ground loop, it'd introduce noise to the actual audio signal. My audio itself is clean as a bell, even through the on board sound card when I plug headphones in directly, bypassing the external amp.

Also, I can run prime 95 and not get any interference, it's only when the graphics card is under any kind of load, doesn't even need to be a high load.

It's definitely a grounding issue.

 

[railven]

EDIT: Whoops, misunderstood what you meant by grounding loop.

 

[SpecChum]

It's fine stability wise, it's literally just the audio it affects.

 

[birthdaymonkey]

What you're describing isn't a ground loop. A ground loop is a constant low-frequency hum that you get when a system is grounded in two different places - e.g. through your house's cable TV wiring and through your electrical wiring. You lift one of the grounds and the hum goes away. What you have is interference that could be coming from the graphics card or the power supply. Only way to find out is to try different GPU/PSU.

 

[SpecChum]

What you're describing isn't a ground loop. A ground loop is a constant low-frequency hum that you get when a system is grounded in two different places - e.g. through your house's cable TV wiring and through your electrical wiring. You lift one of the grounds and the hum goes away. What you have is interference that could be coming from the graphics card or the power supply. Only way to find out is to try different GPU/PSU.

I thought the very same thing at first but the audio is crystal clear until I introduce another earth point. If it was just interference I just can't see why it doesn't contaminate that too?

I've not ruled out actual interference but the evidence does seem to suggest it might not be.

Also I'm sorry but your understanding of ground loops is flawed I'm afraid. If the potential difference is constant then yes you would get a constant hum but the differences and thus current inside a pc are far from constant and thus you can hear varying,more high pitched interference, especially when you move the mouse or scroll a web page, and especially when running 3d applications.

 

[Phynaz]

If it's a ground loop you can get rid of it with this http://www.amazon.com/Ebtech-Hum-Voltage-Filter/dp/B0002E4YI8

 

[Wall Street]

It isn't ground loop (a 60 Hz sin wave doesn't sound like what you are describing). How is your DPC latency? I have had similar issues in the past due to a rogue USB device messing up my latency. Otherwise, you can try ferrite cores on your cables.

 

[SpecChum]

Interesting, I'll certainly look into that suggestion. How would I test the latency?

Thanks.

 

[Phynaz]

It isn't ground loop (a 60 Hz sin wave doesn't sound like what you are describing). How is your DPC latency? I have had similar issues in the past due to a rogue USB device messing up my latency. Otherwise, you can try ferrite cores on your cables.

Yes, a ground loop produces a hummimg, buzzing sound. But since the op is so convinced he has a ground loop I posted the fix. He's convinced himself what it isn't.

 

[Ma_Deuce]

If it's a ground loop you can get rid of it with this http://www.amazon.com/Ebtech-Hum-Voltage-Filter/dp/B0002E4YI8

OP, does the mouse noise go away if you have music turned up and playing out of the amp?

Try something like Phynaz is suggesting or anything that isolates your plug from your amp and the plug from your pc.

I have not had any luck with ferrite cores and the type of noise you are describing.

 

[Ma_Deuce]

Yes, a ground loop produces a hummimg, buzzing sound. But since the op is so convinced he has a ground loop I posted the fix. He's convinced himself what it isn't.

A ground loop doesn't always have to be a steady humming or buzzing sound. Depending on the equipment being used, application ect, you may not have any symptoms until someone turns on a vaccuum cleaner for example. At which time you can see/hear the interference.

OP might even try one of the cheap ground lift adapters. These have saved my day on many occasions.

http://www.filmtools.com/grliadco41.html

 

[MrTeal]

The Fiio E11k has a battery, correct? Do you see the same behavior if you're using the GPU but the USB charge cable is unplugged? If so, it's definitely not a ground loop.

 

[SpecChum]

Yes, a ground loop produces a hummimg, buzzing sound. But since the op is so convinced he has a ground loop I posted the fix. He's convinced himself what it isn't.

Sorry about that, I've read my messages back and I do sound like a bit of d1ck, wasn't my intention. I am 100% sure this is ground loop issue tho as it comes through USB ground as well as audio. You can also hear the noise when the amp is on 0.

OP, does the mouse noise go away if you have music turned up and playing out of the amp?

Try something like Phynaz is suggesting or anything that isolates your plug from your amp and the plug from your pc.

I have not had any luck with ferrite cores and the type of noise you are describing.

The noises are amplified with the sound but louder music does drown it out; the noise created by the gfx card, that's much louder.

A ground loop doesn't always have to be a steady humming or buzzing sound. Depending on the equipment being used, application ect, you may not have any symptoms until someone turns on a vaccuum cleaner for example. At which time you can see/hear the interference.

OP might even try one of the cheap ground lift adapters. These have saved my day on many occasions.

http://www.filmtools.com/grliadco41.html

Ah, I'm in the UK, all our plugs are 3 pin

I'm going to order a USB isolator I think; looks like most of the cheaper RCA ground loop isolators cut the audio at 40hz; doesn't sound much but I can only imagine it'd decrease bass impact somewhat. I've got fairly high end headphones (DT990, HP200 and UE6000) and they can all play 40Hz easily.

 

[SpecChum]

The Fiio E11k has a battery, correct? Do you see the same behavior if you're using the GPU but the USB charge cable is unplugged? If so, it's definitely not a ground loop.

With battery cable unplugged there's no interference at all, that's what I'm using now.

As soon as you plug the USB into the PC the interference appears on the E11K, so I can only assume it's coming through USB ground.

Headphones straight into any of Xonar, X-Fi HD and on board all interference free which is why I'm pretty certain it's not EMI.

Sorry about this, I can only report what I see lol

 

[MrTeal]

With battery cable unplugged there's no interference at all, that's what I'm using now.

As soon as you plug the USB into the PC the interference appears on the E11K, so I can only assume it's coming through USB ground.

Headphones straight into any of Xonar, X-Fi HD and on board all interference free which is why I'm pretty certain it's not EMI.

Sorry about this, I can only report what I see lol

Do you see the same effect if you plug in the USB cable to a cell phone charger instead of into your PC?

 

[SpecChum]

Do you see the same effect if you plug in the USB cable to a cell phone charger instead of into your PC?

That produces zero interference, it literally only happens if I plug it into the PC.

EDIT: Actually, that was just from memory, lemme check shortly to confirm.

 

[SpecChum]

Confirmed, there's no interference with the E11k plugged into a mains powered USB charger.

 

[birthdaymonkey]

If it's a ground loop, where is the other ground that your PC is connected to? If everything is connected to one electrical outlet without a cable TV or antenna (or a connection to something else), it can't be a ground loop.

http://www.epanorama.net/documents/groundloop/

According to this site, the way to diagnose a ground loop is to disconnect everything from your system and reconnect things one at a time till you find the source of the ground loop. Have you tried unplugging everything except the PC, monitor, and headphone amp? (Including network I guess, although I haven't heard of ground loop from network cables before.)

EDIT: Was just doing some additional reading - apparently you can get mouse movement noises if your motherboard isn't properly grounded to the case. Have you tried removing your board from the case?

Also, this thread seems relevant:
http://arstechnica.com/civis/viewtopic.php?f=8&t=1163183

 

[tweakboy]

wtf is that

 

[truckerCLOCK]

Yes, a ground loop produces a hummimg, buzzing sound. But since the op is so convinced he has a ground loop I posted the fix. He's convinced himself what it isn't.

Gotta love how people ask for advice....then don't take it.

 

[MrTeal]

If it's a ground loop, where is the other ground that your PC is connected to? If everything is connected to one electrical outlet without a cable TV or antenna (or a connection to something else), it can't be a ground loop.

http://www.epanorama.net/documents/groundloop/

According to this site, the way to diagnose a ground loop is to disconnect everything from your system and reconnect things one at a time till you find the source of the ground loop. Have you tried unplugging everything except the PC, monitor, and headphone amp? (Including network I guess, although I haven't heard of ground loop from network cables before.)

EDIT: Was just doing some additional reading - apparently you can get mouse movement noises if your motherboard isn't properly grounded to the case. Have you tried removing your board from the case?

You could still have a ground loop, just not a typical earth ground one. The primary idea with a ground loop is that you have two points that should be at a common potential, connected with two wires. Normally you see it with the shield of your signal cable and earth ground, but it's just as possible to see with with the cable ground and USB DC ground. The idea is that when you draw a bunch of power over the USB cable to power the audio, you get a voltage drop over the cable, half being a drop on the 5V line and the other half being a drop on the ground. Since both the signal cable and USB cable are connected at both ends, you get current flowing in the shield of the audio cable as well.

The obvious difference is that with a typical audio system that ground current is 60Hz, which gets injected as hum on the audio signal. On the USB cable, the ground current should be DC and (depending on amp topology) you shouldn't see a lot of audible noise. I'm not really sure why it would be GPU usage dependent though, or why you'd see it at 0 volume unless the amp has a large quiescent current. The common point is that they're both PCIe though. OP, does the amp have issues if it's plugged into the MB's onboard audio?

 

[SpecChum]

You could still have a ground loop, just not a typical earth ground one. The primary idea with a ground loop is that you have two points that should be at a common potential, connected with two wires. Normally you see it with the shield of your signal cable and earth ground, but it's just as possible to see with with the cable ground and USB DC ground. The idea is that when you draw a bunch of power over the USB cable to power the audio, you get a voltage drop over the cable, half being a drop on the 5V line and the other half being a drop on the ground. Since both the signal cable and USB cable are connected at both ends, you get current flowing in the shield of the audio cable as well.

The obvious difference is that with a typical audio system that ground current is 60Hz, which gets injected as hum on the audio signal. On the USB cable, the ground current should be DC and (depending on amp topology) you shouldn't see a lot of audible noise. I'm not really sure why it would be GPU usage dependent though, or why you'd see it at 0 volume unless the amp has a large quiescent current. The common point is that they're both PCIe though. OP, does the amp have issues if it's plugged into the MB's onboard audio?

Yeah, it happens through xonar analogue out, on board analogue out and soundblaster usb.

The interior of my case is painted, I'm wondering if that's isolating the ground to the motherboard? I've not had opportunity to delve deeper as it'd pretty much be a system rebuild.