Ground loop only solved by removing XLR Pin 1, is it safe?

[saaif88]

Hey everyone,

I'm dealing with the age old problem of ground loops, or more accurately, ground loops and EMI interference.

My signal chain is as follows: PC (HDMI) --> TV (HDMI ARC) --> Wiim Ultra (RCA to XLR) --> JBL 308P MKII

Without my PC turned on, I can hear some buzz and interference, but it is slight. With the PC on and running any GPU tasks, the noise produced overwhelms the audio signal.

What I have tried so far:

Removing Mains Earth/PE: Works perfectly, but is incredibly unsafe and not a real solution
Connecting grounds together: All of the devices are on the same power bar, on the same outlet. I have also tried using jumper wires to connect the HDMI shields to mains earth which resulted in no change. Using the phono ground adapter included with the Wiim Ultra and connecting that to mains earth does reduce the noise significantly, but it is still audible. Connecting that to the HDMI shield also works, but noise is still audible. I could not find a way to tie all of the grounds together to eliminate the noise unfortunately.
RCA Ground Loop Isolator: Fixes the problem, but ruins the frequency response. Very obvious differences in sound quality using these.
Behringer HD400/PYLE-PRO PHE300 Hum Eliminator: The Behringer was out of stock so I purchased the identical looking PYLE-PRO. This does mostly work, and maintains the frequency response better than the RCA ones, but it still degrades the sound quality. Surprisingly though, it does not get rid of the interference noise 100%, which is likely due to the fact that the shield/ground on the inputs/outputs of the unit are still tied together. I am not sure though, and I would have to measure the resistance.
Removing the bridge between XLR Pin 1 and Pin 3: So far this is the solution that works 100%. There is no degradation in sound quality at all. Most RCA to XLR cables have Pin 1 and Pin 3 tied together, and there are some discussions about removing Pin 3, but in my case, removing Pin 1, which is ground, completely eliminates the noise.

Now for the original question: Is this safe or is this a good idea? I am aware that this works because of the fact that the balanced inputs of the studio monitors can use a signal even without a ground reference present, but I am not sure what the downsides are. The PC and studio monitors are still grounded through their AC plug, so there should be no risk, but I am not sure. I have seem some recommendations to put resistors and capacitors between Pin 1 and Pin 3, but I am not sure if that will actually work.

Any advice or suggestions would be greatly appreciated!

**Further information: I decided to try disconnecting Pin 1 from Pin 3 while keeping Pin 1 connected to the XLR shield and this also completely removes the noise. I would assume that this is even safer since the XLR shield is likely connected to mains earth as well.

 

[sergeauckland]

If removing the bridge between pin 1 and 3 works, that's fine. As long as safety earths are not removed, then the system remains safe.

S

 

[AnalogSteph]

Behringer HD400/PYLE-PRO PHE300 Hum Eliminator: The Behringer was out of stock so I purchased the identical looking PYLE-PRO. This does mostly work, and maintains the frequency response better than the RCA ones, but it still degrades the sound quality. Surprisingly though, it does not get rid of the interference noise 100%, which is likely due to the fact that the shield/ground on the inputs/outputs of the unit are still tied together. I am not sure though, and I would have to measure the resistance.

Please do - I am just about 100% positive that the Behringer does not connect Pin 1 in the least. That said, transformer CMRR is somewhat limited on it as well.

Removing the bridge between XLR Pin 1 and Pin 3: So far this is the solution that works 100%. There is no degradation in sound quality at all. Most RCA to XLR cables have Pin 1 and Pin 3 tied together, and there are some discussions about removing Pin 3, but in my case, removing Pin 1, which is ground, completely eliminates the noise.

Not a huge surprise, as it (a) breaks the ground loop and (b) makes the connection at least somewhat balanced. This is basically the equivalent of a "ground lift" switch. As mentioned, it's a perfectly fine thing to do in this scenario. Just perhaps mark the cable with a note that it's ground-lifted just in case you ever want to use it with some speakers that are floating.

 

[EngineerNate]

Beat me to it re: ground lift. I have pro amps with a button on the back that does this intentionally.

 

[KSTR]

Did you use a true balanced cable between Wiim and the speaker? Looks like you didn't... and that's because the majority of RCA-->XLR cables are designed wrongly, using an unbalanced cable (for cost reasons).
Correct way to do it, see this post.

 

[EngineerNate]

Did you use a true balanced cable between Wiim and the speaker? Looks like you didn't... and that's because the majority of RCA-->XLR cables are designed wrongly, using an unbalanced cable (for cost reasons).
Correct way to do it, see this post.

Yes! True unbalanced cables should almost always be coax and even quasi balanced should be STP or starquad as shown.

 

[Speedskater]

In balanced XLR output to balanced XLR input, the shield is sometimes not connected at the input end.
But there will be some loss in high frequency RFI interference rejection.

 

[rcstevensonaz]

Did you use a true balanced cable between Wiim and the speaker? Looks like you didn't... and that's because the majority of RCA-->XLR cables are designed wrongly, using an unbalanced cable (for cost reasons).
Correct way to do it, see this post.

The added benefit of using a properly wired RCA to XLR/TRS cable — e.g., 2-wire STP instead of coax cable, as shown in @KSTR's wiring — is that the balanced input device (i.e., JBL speaker) will be able to substantially eliminate any noise introduced into the cable between the output device (unbalanced WiiM) and the input device (balanced speaker).

 

[saaif88]

Hey everyone, thanks for all of the replies! I decided to do some more in depth testing and report back my findings. There is quite a stark difference between some combinations, though I am not entirely sure why.

My setup is certainly a "worst case scenario" in the case of noise, but then again, this is just me using my PC for its intended purpose.

I will say this, the Wiim Ultra alone does not produce any audible noise at all when not connected to the TV or PC. It is completely silent which is a good sign that it does not contribute to any noise in the system by itself.

Now onto the testing; I am running a GPU stress test on my PC to ensure a constant load and constant noise. I allowed the system to warm up for 10 minutes and then started testing. I am using my iPhone 13 Pro with the case removed running Decibel X. I do not have a proper SPL meter or a calibrated microphone, so these are just relative tests so hopefully they can highlight the difference between each method. I have these ranked from best to worst performing in terms of noise produced with the microphone 1ft from the JBL 308s.

PYLE-PRO PHE300: My original testing was incorrect because I was mistakingly using TS cables which for some reason still produce noise. With proper TRS cables there is absolutely no noise, and as a matter of fact, there is even less noise than with the monitors unplugged from the rest of the system! These JBLs are known for a bit of hiss and background noise, but with the PHE300 plugged in, the noise goes down even lower than without any cables plugged in, which is surprising. That being said, even though this does get rid of all of the noise, I am not sure how much it affects the frequency response. I will try to use REW through my audio interface and compare it to the loopback to see if I can get some useful information.

Unplugged: This is just the noise of my room, my PC fans running almost at 100%, and the background noise from the speaker. This should be the lowest value, but as mentioned, the PHE300 actually lowers this even further.

RCA to XLR Cable with Pin 1 disconnected from Pin 3: This is the solution proposed in my original post and it certainly works great. There is still some very slight noise produced in the speaker, but from a reasonable listening distance it is completely inaudible. It is also low enough that my phone microphone cannot pick it up.

These are the cables I am using: https://www.amazon.ca/dp/B084DFD31M

Generic RCA cable to RCA to XLR Adapter: I grabbed a random RCA cable I had laying around and hooked it up to a $5 RCA to XLR adapter I got on Amazon. Surprisingly, this had lower noise than both the Star Quad Cable and the more expensive Kirlin RCA to 1/4" cables. I am not sure why, but there is without a doubt a very audible difference between this and the other options. At this point though the noise is loud enough to be noticeable and annoying.

Monoprice Star Quad XLR cable to XLR to RCA Adapter: As pointed out by some others, a star quad cable is even better for noise immunity and provides a quasi balanced cable. In my case, this is only slightly better than the worst offender, and truthfully, I tried a standard XLR cable and got the exact same results, so at least for my testing, this does pretty much nothing to reduce the noise compared to the options above. I am just using a 6 foot cable though, and maybe for longer runs it would do something.

Kirlin RCA to 1/4" Cable: Finally, the worst offender, which just so happens to be the starting point and the whole reason why I was trying to solve this problem in the first place. While this looks like a very well made cable, and not very expensive, it was by far the noisiest cable in the whole experiment. I have no idea why this cable is so noisy, maybe it is just poor construction, or it is a TS cable instead of a TRS cable, I am not sure, but the noise with this cable is at the point where it overpowers a significant portion of the audio from my PC.

**BONUS** RCA to 1/4" through Douk L3: I was not aware that the noise could get any louder, but I have also been A/B testing some speakers using this Douk L3 RCA switcher, and under normal conditions it is not very noisy, but running this test it introduced such a significant amount more noise even compared to the worst case offender. This device does not come with a USB power supply so I had to find one and use it, but even running it off of a battery bank, the noise is the same. Not sure why, but interesting regardless.

I hope that someone finds this information useful. I would like to test a DRV134 board and see how that compares, but I will do that in the future. Thanks for all of the help guys! I will also report back the frequency response of the PHE300 when I measure it.

 

[AnalogSteph]

PYLE-PRO PHE300: My original testing was incorrect because I was mistakingly using TS cables which for some reason still produce noise. With proper TRS cables there is absolutely no noise, and as a matter of fact, there is even less noise than with the monitors unplugged from the rest of the system! These JBLs are known for a bit of hiss and background noise, but with the PHE300 plugged in, the noise goes down even lower than without any cables plugged in, which is surprising.

These kinds of isolators tend to be using fairly low-impedance transformers (think around 600 ohms) as otherwise permissible cable lengths on the output would be truly pathetic or even negative. (It's still a good idea to keep lengths between the isolator and speakers as short as you can, no more than 1.5 m if you can help it.) Given that the JBLs have a bit of a reputation for power supply noise getting coupled into the input, seeing a low source impedance may help keeping that at bay.

 

[rcstevensonaz]

Monoprice Star Quad XLR cable to XLR to RCA Adapter: As pointed out by some others, a star quad cable is even better for noise immunity and provides a quasi balanced cable. In my case, this is only slightly better than the worst offender, and truthfully, I tried a standard XLR cable and got the exact same results, so at least for my testing, this does pretty much nothing to reduce the noise compared to the options above. I am just using a 6 foot cable though, and maybe for longer runs it would do something.

Note: I don't think that a star-quad cable is better than a 2-wire STP cable and I find it much harder to work with when I am making my own cables. And for your level of testing precision and cable length, there should be absolutely zero difference between the two.

However, this cable arrangement does not match what was recommended by @KSTR in this post since you are using an RCA adapter. Moreover, since your wording implies an audio chain of (Source XLR) -> XLR -> RCA Adapter -> (RCA Speakers), it sounds like you are using the XLR to RCA adapter at the speaker side of the chain which would essentially eliminate all benefit of the XLR cable.

The challenge I have in reviewing your summary is that I still do not know the exact type of wire used in every scenario -- whether Coax or STP, except for the Starquad, but then details of the adapter are completely missing.

 

[EngineerNate]

Also pretty normal for a totally disconnected amp to.be a little noisier.

 

[jonfitch]

I think iFi SilentPower - GND Defender might be able to help, I don't think it affects the frequency response.

The other way to break a ground loop are optical HDMI audio cables. You need one that has no ground/power copper line that runs through the cable. It's the one's with external power cables on the output end like these (https://ruipro.store/products/ruipro-8k-gen4ak-detachable-full-fiber-armored-hdmi-cable) that are fully isolated.

 

[bigguyca]

Use a GFCI

Might as well start an "interesting" discussion. Before someone comments on this suggestion, do enough searches and AI queries to discover the parts of the NEC that apply here and get your AI tool's input on this subject. I don't want to post the results of pages of Ai responses.

Do the following at your own risk!!!

o Decide that the electric service safety ground to the duplex outlet you are using is not working properly.

o Have an electrician replace the duplex receptacle where your equipment is plugged-in with a GFCI.

o Don't connect the electrical ground to the new GFCI.

o The GFCI now provides protection against electrical shorts by comparing the currents flowing in the hot and neutral likely as well as the safety ground and circuit breaker combination.

o Optionally cut the electrical ground that goes to the duplex receptacles that follow the GFCI electrically. The hot and neutral for those receptacles is chained from the GFCI.

o Apply the labels that likely come with the GFCI to each of the duplex receptacles that now don't have the electrical ground.

- Thoughts? -


Note: MOV based surge protection won't work as well since it can't dump the surge to ground, only to the neutral. Inline surge protectors such as SurgeX and ZeroSurge will function normally.

 

[neRok]

The other way to break a ground loop

There are other-other ways...

 

[obsessive9119]

Can you just use optical between the TV and WiiM?

 

[sam_adams]

- Thoughts? -

Dig through the NEC and find the section that states that if an EGC is present at the location that the GFCI outlet is installed, the EGC must be connected to the GFCI outlet.

 

[Speedskater]

Note the the EGC (Safety Ground / Protective Earth) system has one main task and a GFCI receptacle or breaker has another.
The EGC task is to trip the circuit breaker if there is a Ground Fault (short circuit).
The GFCI task is to trip if it detects a small amount of leakage current (maybe thru a human).

 

[EngineerNate]

Dig through the NEC and find the section that states that if an EGC is present at the location that the GFCI outlet is installed, the EGC must be connected to the GFCI outlet.

Yes, while what was suggested would be “safe”, and is often done to bring 2 wire homes “up to code” and allow grounded plugs to be used without having to violate code by installing standard grounded receptacles without the ground hooked up (what the previous owners did in my house! In a room with perfectly accessible wiring! Gaaah!) or using cheater plugz, I had doubts that doing so was allowed if the ground is in fact present.

On the cheater plug note, anyone doing audio should have a few handy just to diagnose ground loops. I'd try to eliminate them by other means "permanently", especially in a home setup that is semi-permanent vs a portable live sound situation where you don't have time to chase gremlins before the show and lifting a ground for a few hours is not a big deal.

 

[sergeauckland]

On the cheater plug note, anyone doing audio should have a few handy just to diagnose ground loops. I'd try to eliminate them by other means "permanently", especially in a home setup that is semi-permanent vs a portable live sound situation where you don't have time to chase gremlins before the show and lifting a ground for a few hours is not a big deal.

Absolutely NOT! Cheater plugs should NEVER EVER be used. Anyone doing audio engineering will have some 1:1 audio isolating transformers available to break the common connection whilst maintaining audio connectivity AND not removing any safety earths.

I absolutely abhor removing safety earths for any reason.

S.