Ground loop problem?

[HobGoblyn]

Easier would be to buy a power adaptor that already has the bonding built in. But that can/will cause ground loops as soon as you introduce another protectively earthed devices into the system.

Thanks, if I understand correctly, using balanced audio cables usually stops any ground loop hums?

[Nominal Animal]

Unbalanced audio is a voltage signal referenced to ground.  A typical 3.5mm TRS jack has left audio signal on the tip, right audio signal on the ring, and the reference ("ground") on the sleeve.

The power supply on your active speakers has that class Y interference-suppression capacitor between the ground and the AC.  It only leaks about 1mA and isn't dangerous, but you do feel a tingle if that passes through your body (to true earth ground).

Connecting to the speakers via say a cable with 3.5mm TRS male jack on the Mac end, and to two RCA male jacks on the powered speaker end, will connect the ground on the active speakers to the ground on your laptop, which allows the tiny leakage current (allowed by that class Y capacitor) to pass through the active speakers' ground to your Mac's audio ground to Mac chassis to your body and to true Earth ground, causing the tingle.  It is leakage, not a "real" ground loop, if that distinction matters any.

Typical audio mixers have continuity on the unbalanced audio connector "grounds", so won't change anything for HobGoblyn here.

Balanced audio uses a pair of connections, a differential pair, for each audio channel.  The differential pair does not need a ground connection per se.

Your Eris 3.5 active speakers have both unbalanced and balanced inputs.  The balanced inputs are the larger, 1/4" TRS ("stereo") connectors, with tip and ring being the differential pair, and the sleeve the optional ground.  Note, however, that this ground is almost certainly connected to its power supply ground, and thus will convey the "tingle" if connected to the ground on your Mac.

Typical passive DI boxes (passive as in not requiring any power) are basically just an audio transformer to galvanically isolate balanced audio lines.  Essentially, across each transformer, one audio channel is coupled over using a varying magnetic field.  The properties of the transformer define what kind of frequency response you get at a given input and output impedances, so there are a huge variety of such audio transformers, with the "undetectable effect on the audio signal within extended human hearing range" costing several hundred USD apiece.  The ones I'd use are about USD 6 apiece.

Conversion from unbalanced to balanced is trivial with such an audio transformer: on the unbalanced side, the "ground" and the signal is connected to one coil; on the balanced side, the coil is connected to the differential pair.

Many balanced audio cables have three conductors, often with the third being a shield around the differential pair.  DI boxes that can break that connection between inputs and outputs usually call this breaking "Earth lift".

In HobGoblyn's case, the need is to break the ground connection from the active speakers to the Mac, to stop the tingle, while allowing the audio signal to pass with high fidelity.

I personally would do this DIY, using components that cost less than USD $30 off Mouser (and enclosure from a local Hammond importer, Uraltone, here in Finland).  The enclosure would be placed close to the active speaker, perhaps on/in the speaker stand, or on the floor nearby.  The connection from the enclosure to the active Eris speaker would use a pair of shielded audio cables (often called "microphone cable"), most likely Adam Hall 5-star microphone cable or something like Belden 8451010100 (from Mouser) – a pair of conductors, and a separate shield – with two 1/4" TRS ("stereo") jacks at end, both connected to the active Eris 3.5 speaker.  Their ground (sleeve) would be connected to the aluminium box, so that might give a tingle if touched.  From the enclosure to your Mac or any other sound source with a 3.5mm TRS audio connector, I'd do a custom 3-conductor shielded cable, maybe Tasker C115 or Belden 22-24AWG 3-conductor shielded cable (like this, link to my local store), where the shield is connected to the enclosure, but not to anything on the 3.5mm TRS male connector.  This way, the audio signals should be pretty well shielded from typical electromagnetic noise that could otherwise couple to the audio signal, and cause annoying hum or buzzing, while breaking any direct electrical connection that could conduct the tiny leakage current that causes that tingling.

The commercial option is to get two mono or one stereo DI box with "ground lift", and suitable cables.  It/they will do the same thing, allowing the signal to pass without direct electrical connection across that box.  While the problem here is not a ground loop per se, it is similar: both problems are fixed by breaking the direct electrical connection between the audio source (Mac) with a 3.5mm TRS audio output, and the active speakers.

Because I'm poor and not interested in high-end audio stuff, I cannot help with links to suitable DI boxes or similar off-the-shelf products, but they definitely do exist.

[nctnico]

So how exactly do you think the DI box would solve this?

A pair with "earth lift" will break the ground connection between the speakers and the Mac.  The audio signals pass through the symmetric transformers, typically 600Ω:600Ω.  The Eris 3.5 has two balanced inputs as 1/4" TRS connectors (sleeve is for ground/shield), so it will act as an unbalanced-to-balanced transformer also.

Since there is no leakage current path to ground from the active speakers' ground to Mac chassis to human to physical ground, there would be no tingle either.

I agree. Getting a signal tranformer / isolation transfer in between is an easy solution to break ground loops at the cost of frequency response flatness. Back in my audio gear design days, we always made it a point to have a ground on the amplifier and design the rest of the equipment to be double isolated (and powered using a transformer). That ensured no mains hum travelling along the ground of unbalanced audio lines. In cases that was not possible, we'd use a signal transformer to break a ground loop.

[Nominal Animal]

at the cost of frequency response flatness.

The Triad Magnetics' TY-250P (from Mouser, 5.1€ each in singles; need two) I would DIY per reply #10 is -0.6dB to +0.6dB within 20Hz .. 20kHz; if the active speaker input impedance is 10kΩ, it increases from +0.6dB to +1.2dB in 20kHz..30kHz.

Other than "it would suffice for me", I don't know how good that is in practice, though, especially when used with active studio monitors like the Eris 3.5.

[SiliconWizard]

There are tons of now cheap commercial DI boxes. If you're in Europe, just have a look at Thomann - if you're in the US, one equivalent shop would probably be Sweetwater.

One of their cheapest: https://www.thomann.de/intl/millenium_die_dibox_passiv.htm

[Someone]

Easier would be to buy a power adaptor that already has the bonding built in. But that can/will cause ground loops as soon as you introduce another protectively earthed devices into the system.

Thanks, if I understand correctly, using balanced audio cables usually stops any ground loop hums?

Reduces hum from different ground potentials, but doesnt eliminate it. Engineering rarely has true zero or infinity. You're in the middle of the balance:
high impedance earth for improved interconnection performance
low impedance earth for improved RFI and touch current

It is not a group loop that is your problem, but the reverse, a too little current passing to ground.

[themadhippy]

One of their cheapest

pity about the £10 p+p ,maybe https://www.studiospares.com/dual-passive-di-box-by-lambden-audio-449300.htm however if you want a known quantity your looking at 10x the price for something like the  emo dual.

[nctnico]

at the cost of frequency response flatness.

The Triad Magnetics' TY-250P (from Mouser, 5.1€ each in singles; need two) I would DIY per reply #10 is -0.6dB to +0.6dB within 20Hz .. 20kHz; if the active speaker input impedance is 10kΩ, it increases from +0.6dB to +1.2dB in 20kHz..30kHz.

Other than "it would suffice for me", I don't know how good that is in practice, though, especially when used with active studio monitors like the Eris 3.5.

My old boss had a rule of thumb to either keep the response flat to approx. 25kHz or filter below 19kHz. So it would be a good idea to source and terminate the transformer close to 1 kOhm for optimal sound quality. This would mean a loss of 6dB though.

[HobGoblyn]

One of their cheapest

pity about the £10 p+p ,maybe https://www.studiospares.com/dual-passive-di-box-by-lambden-audio-449300.htm however if you want a known quantity your looking at 10x the price for something like the  emo dual.

True,  I’d need two of the Thomann ones, so with postage would come to a total of £33.40

The one you linked to with postage is. £26.98

I presume they’re all as good as each other at that price range?

Am I correct in thinking the Jack output is just a through for the input (it’s being sent straight back out) and I’d need to use the Xlr outputs?  I’d also have  to get a couple of Xlr to 6.35 cables as well

Thanks

[themadhippy]

Am I correct in thinking the Jack output is just a through for the input (it’s being sent straight back out) and I’d need to use the Xlr outputs?

yep

I’d also have  to get a couple of Xlr to 6.35 cables as well

or just cut off the jacks at one end of your existing leads  and chuck  on an xlr.

[HobGoblyn]

Just ordered the one themadhippy linked to from studiospares

Was thinking of cutting existing leads and putting on Xlr but that would leave me short of leads, I need the existing leads to go from the Audio interface to the Di box, then need additional leads to go from Xlr on Di box to my Eris monitors

Thanks

[HobGoblyn]

Was talking to Presonus support about this and told them what both”tooki” said in reply 1 and what “Whales” said in reply 14.

They responded with (confidential response but don’t think they would mind me quoting this bit)

“I've no other reports of this issue but I would be interested in getting these speakers into a service centre for an inspection and we will replace these units”

Personally after my learning experience in this thread, I don’t think anything is wrong with the speakers and I’ve told Presonus this.  Still it’s nice to see a company taking issues seriously.

[Someone]

“I've no other reports of this issue but I would be interested in getting these speakers into a service centre for an inspection and we will replace these units”

Personally after my learning experience in this thread, I don’t think anything is wrong with the speakers and I’ve told Presonus this.  Still it’s nice to see a company taking issues seriously.

Since you have not had the speakers tested for leakage, they are rightfully worried as there could be something wrong and indicative of future more serious problems. You do not know the speakers are operating within the safety requirements but its a routine test as already mentioned:

If still an issue get the power adapters replaced under warranty then get a p.a.t. or electrician in to test your devices or wiring.

That is for catching faulty equipment before it becomes life threatening.

[Nominal Animal]

While the class Y capacitor in the supply should fail open, it is possible it was a dud or wrong one, or something like the flux used in manufacturing the board has reacted with the potting compound used to keep the big electrolytic caps from flapping in the wind and the resulting gunk is slightly capacitive and just happens to slightly couple over the isolation barrier.  I think both are exactly the situations a responsible company would love to find out the details of, if they ever hear of or suspect it happening.

Props for Presonus support for caring, methinks.  If I was HobGoblyn, I'd try to have their service center check them out.

[HobGoblyn]

Sadly, that DI box cuts the audio to virtually nothing, turning volume up to hear it (still not very loud) and it distorts.

[Nominal Animal]

Sadly, that DI box cuts the audio to virtually nothing, turning volume up to hear it (still not very loud) and it distorts.

The DI box, being passive, attenuates the signal somewhat; this is called insertion loss.  See if you can find it mentioned in the specifications for the boxes you use; it is typically specified in decibels (dB).

Voltage amplitude is multiplied by 10^-dB/20 ≃ exp(-0.115129 dB) across the isolation transformer when dB is the insertion loss in decibels.
Power or volume is multiplied by 10^-dB/10 ≃ exp(-0.2302585 dB).

Did you have a small mixer at hand?  Try putting the mixer between the Mac and the DI box, or the DI box and the monitors.  Use the line out from the mixer.  The idea is to boost the signal to account for the transformer attenuation, without clipping the output.

In the DIY case, the Triad transformers I mentioned have less than 2.8dB insertion loss, which corresponds to about 28% reduction in voltage swing (×0.7244) and almost halving the output power or volume (×0.5248).  One could amplify the input voltage by gain 1.38, without clipping up to output range of about -2.5V to +2.5V maximum, or -2.8V to +2.8V for pro gear –– far wider than normal audio amplifiers allow! –– to compensate for the losses in the isolation transformer.

That makes it into an "active DI box with ground lift".  The currents are very low, in single milliamps per channel.  I would use for USB power, and a 5V to +12V and -12V isolated supply, perhaps Murata CRV1D0512SC or CMR0512S3C or Mean Well DPU01L-12, with both rails LC-filtered and linearly regulated down to +6V/-6V or so, with 0V referenced to the unbalanced audio ground (Mac ground), leaving about 3V between the rails and the output signal.  That would allow the isolator to work from any supply – either the Mac, or from a class II isolated USB wall wart, or from a class I medical-rated USB wall wart.
I know nothing of analog audio amplifiers or analog audio circuit design, but I assume a single dual opamp would suffice.

[themadhippy]

Sadly, that DI box cuts the audio to virtually nothing,

presume youve tried changing the attenuation switches on the box

[HobGoblyn]

Sadly, that DI box cuts the audio to virtually nothing,

presume youve tried changing the attenuation switches on the box

Yes

[HobGoblyn]

Sadly, that DI box cuts the audio to virtually nothing, turning volume up to hear it (still not very loud) and it distorts.

The DI box, being passive, attenuates the signal somewhat; this is called insertion loss.  See if you can find it mentioned in the specifications for the boxes you use; it is typically specified in decibels (dB).

Sadly there’s little information on their site and no leaflet/manual in the box

Did you have a small mixer at hand?  Try putting the mixer between the Mac and the DI box, or the DI box and the monitors.  Use the line out from the mixer.  The idea is to boost the signal to account for the transformer attenuation, without clipping the output.

I do have one somewhere, I’ll dig it out.

This all started as an easy and quick way for me to carry on my music while in my armchair, something that’s quick to move for my meals etc.

Now I’ve got two monitors, a Di box, an Audio interface and now possibly a mixer too

In the DIY case, the Triad transformers I mentioned have less than 2.8dB insertion loss, which corresponds to about 28% reduction in voltage swing (×0.7244) and almost halving the output power or volume (×0.5248).  One could amplify the input voltage by gain 1.38, without clipping up to output range of about -2.5V to +2.5V maximum, or -2.8V to +2.8V for pro gear –– far wider than normal audio amplifiers allow! –– to compensate for the losses in the isolation transformer.

That makes it into an "active DI box with ground lift".  The currents are very low, in single milliamps per channel.  I would use for USB power, and a 5V to +12V and -12V isolated supply, perhaps Murata CRV1D0512SC or CMR0512S3C or Mean Well DPU01L-12, with both rails LC-filtered and linearly regulated down to +6V/-6V or so, with 0V referenced to the unbalanced audio ground (Mac ground), leaving about 3V between the rails and the output signal.  That would allow the isolator to work from any supply – either the Mac, or from a class II isolated USB wall wart, or from a class I medical-rated USB wall wart.
I know nothing of analog audio amplifiers or analog audio circuit design, but I assume a single dual opamp would suffice.

Thanks, I’m seriously thinking of going this route now

[themadhippy]

Did you try cutting the earth on the signal cable?both channels of course.

[HobGoblyn]

Things I've tried this afternoon (not using Di box), all with MacBook on battery.

1) I unsoldered the earth in both jacks where they go into the monitors, exactly the same problem. (resoldered them before I continued)

I then  made up a lead, just the earth connected to UK 3 pin plug, and plugged into the same double socket as the monitors are in.

2) Other end of earth connected to one of the jack inputs in the back of the monitor.  Same problem, I unplugged the monitors and just left earth cable plugged in, still had same problem.

3) Other end of earth connected to the Anker hub (Anker hub connects to MacBook via USB C, Audio Interface connects to Anker hub via USB 2), Same problem, I unplugged the monitors and just left earth cable plugged in, still had same problem.

4) Other end of earth connected to MacBook case, problem still there but about twice as bad.

That's all I can try  today as have been summoned by my loverly wife.

What's confusing me now is I THINK the above does what has been suggested such as attaching the output GND to mains ground. I thought one of these would have worked.

Tomorrow I can insert my small mixer into the chain to see what happens and I might still go the DIY route onceI 100% understand what's going on.  As I said somewhere above in an earlier post, previously when I've had the tingle with my right hand, touching metal on the monitors with my left hand such as one of the jacks instantly stopped it, hence I was convinced running earth from there to mains earth would solve my problem, seems I am still misunderstanding.

I really appreciate all this help

thanks

[HobGoblyn]

Did you have a small mixer at hand?  Try putting the mixer between the Mac and the DI box, or the DI box and the monitors.  Use the line out from the mixer.  The idea is to boost the signal to account for the transformer attenuation, without clipping the output.
.

Today I returned the DI box, but thought I would see what happens if I use the mixer.

Macbook to Anker Hub via USB C
Anker Hub to Audiobox Go (Audio Interface) via USB 2
Audiobox  to Behringer Xenyx 502 Mixer, Line in 2/3 via balanced 6.35 (1/4") cables
Behringer Mixer to  Eris monitors via unbalanced 6.35 (1/4") cables

Perfect sound

Been using it for an hour and not one tingle, yay

Mixer specs here if anyones interested (exactly the same mixer as mine, but my mixer has Tracks on it where this manual says  CD/Tape)

https://mediadl.musictribe.com/media/sys_master/hdd/h5b/8850067685406.pdf

Many thanks everyone

[HobGoblyn]

It gets weirder

Today I plugged everything into the same places (cables labelled, hence in exactly the same sockets) and the tingle was back.

However, if I turn the fig 8 (c7) mains cable for the speakers 180 degrees, it completely goes away, if I turn it back again, its back.

Tried two different cables, exactly the same

[tooki]

Since it’s been suggested multiple times and (unless I overlooked it somewhere) you still haven’t tried it: what happens when you power the MacBook with a grounded mains cable? If you don’t have one, you can order it from Apple. (Not sure if the brick and mortar stores stock them.)

[HobGoblyn]

I said in post 11, “ There sadly doesn’t appear to be any continuity from the psu ground stud to the usb-c.  As Ian.M says, Apple do an extension cable that also connects to the ground stud and doing a lot of research today, a fair few people have said it solves their problem, but no continuity from mine so not worth buying”

The MacBook also has a very long battery time and a very short fast charge time, I use it for hours on just the battery, all of the above experiences has been on battery although when I do charge, it makes no difference, tingling is still there.

I  also said in post 45 that I tried grounding the speakers to mains ground and the tingling still continued