How best to isolate computer's noise from DAC?

[barrows]

PS Audio Duet (the Duet features much more effective filters than the old Ultimate Outlet) is a great choice for isolating the computer gear AC from the audio gear. This is an affordable power conditioner, with very powerful series filters. Both Panamax and Monster make similar products as well.

Lifting or blocking the power connection on Firewire probably will help, but remember there is still a ground connection to the computer-it would be interesting to hear how firewire DAC designs deal with (or not) noise isolation from the computer in their internal design. I suspect that well designed firewire DACs like Metric Halo and Weiss address the noise issue internally somehow.

 

[Skip Pack]

In about 2006, I heard a fellow who had extensive Boulder

mods to his Squeezebox SB3. At a meeting it was compared

to a stock SB3, both feeding a good DAC via SPDIF. Since

there were multiple mods, you could not, of course, isolate

the effect of the Bybee, but the total difference was

startling, and for the good. (I have heard several SB3 power

supply only comparisons, and those are not nearly as

obvious.)

 

A year and a half later, I was assembling a Twisted Pear

Opus DAC, and started off stock taking a single output from

the DAC board. Next I added a pair of Lundahl 1544A transformers

for balanced to single-ended conversion plus isolation. That was

a nice improvement. Later (6 months of heavy use), I added

a Bybee silver quantum purifier to the SPDIF input (fed by

an stock SB3) That was a very remarkable improvement.) I share

everyone's discomfort with the Bybee mystery, but heard an

improvement in the first instance and replicated it in a more

controlled, stepwise process.

 

Skip

 

[hifitubes]

It is more than possible for noise to travel down any cable from the computer to the DAC. Noise can travel through the air to the DAC (true story - we were once doing tests on an 8 transistor CAM chip that fabricated - only at roughly 3pm every day the chip lost its data - we traced it to a Bus driving past the window on its daily routine! BUS engines emit unbelievable RFI pollution ).

 

This has been my direct experience as well. I reported my experiences on this forum as I was confounded as to how PSU noise was getting into my DAC.

 

I haven't tried floating the ground my AES cable yet, but the quick fix was to float the ground on my PC's power supply. I have a generic linear on order to try next.

 

PSU noise is not just relegated to ultra-high frequencies. I use 94db/w speakers and I could hear the PSU hash (much like RFI) from 3 different mainstream dc-dc converters: PicoPSU, WinMate, and the stock OrigenAE M10. I think Chris even had a noise issue with his. I tried various systems (4 motherboards, 4 PSUs, and 4 amps, 2 locations) using Lynx PCI and Lynx PCI-E solutions. Ironically, my neighbor's Dell had no issue. The common thread was the small dc-dc converters touted as the most relevant to computer audio; they may simply be in dire need of linear power.

 

I believe there is a grounding issue with this new breed of DC-DC converters, at least when used with Lynx cards; that is the only AES card with which I have experience. The noise is audible and with certain single-ended cables cables, pretty bad. I even had EMI getting in the signal chain with the PC out of the signal chain. I used to own 98db/w speakers and had a larger HTPC with a classic ATX PSU and no noise issues.

 

Mind you, the quietest amplifier/speakers doesn't imply the best sound; I always have some tube rush but this EMI/PSU noise was unnerving and pretty loud. Others here have spoken to phase issues which I think is an important conversation; you have the system's performance compromised by this UHF activity as well as possible audible noise issues (which actually may be less detrimental in the long run).

 

I'm guessing those with zero noise issues from the get go, based on my experience, may be using very quiet SS gear or expensive tube gear, and classic ATX power supplies on their PCs.

 

Paul Hynes replied to me regarding these issues and I will be adding one of his linear supplies to my Zodiac+:

 

The problem with switchers is that they generate HF noise that is difficult to eradicate. The more switchers in the system the more complex the interference. Also the way all the connected equipment is grounded can cause problems, as you have found out, by passing interference along the ground system. I currently use an custom SR7 heavy duty 12v 8 amp linear power supply to feed my Intel Atom dual core motherboard via a Pico. The chassis is grounded but the regulated output voltage is floating with respect to ground and takes it's ground reference from the equipment to be powered. This should alleviate your ground issues when powering the motherboard but it will also provide a very quiet clean supply to the Pico. The SR7 regulator error amplifier has a noise spec of 0.5 nanovolts root Hz with a noise corner of 1 Hz. The voltage reference is very quiet and has a wideband 40 dB noise filter before the error amp input. This regulator is so quiet that I have been able to use it for providing the HT for my valve MC head amp.

 

 

 

 

 

[earflappin]

Interesting Hifitubes. I have had similar experiences.

 

Currently I am A/B'ing a Mac Mini on stock SMPS, a Macbook Pro on battery, and Macbook Pro on its SMPS. With the SMPS's I am using both straight power from the wall outlet and an isolation transformer. All three going asysnc Firewire to a Metric Halo LIO-8 DAC. The sonic signature of all three combinations is audible. While battery has the lowest noise floor, there is a slight, but audible decrease in dynamics. The best sound is the Macbook Pro on SMPS plugged into the iso xfer.

 

I have an optical firewire repeater/isolator coming tomorrow to test. I would like to believe with optical isolation of the Firewire interface then the noise signature of the source computer should be irrelevant. I'll know soon enough. I also have a variable linear power supply coming that will allow use with the Mac Mini and LIO-8.

 

I agree with Paul Hynes that the noise signature of various devices is a complex, dynamic summation of all noise generating components. The issue I have with an isolation xfer only approach is that I don't see how it can eliminate the potential for noise transfer on the Firewire interface as you will still noise being transmitted by the various computer components, albeit at a lower level.

 

I'll be posting back once I have run some more tests.

 

[hifitubes]

My system is now all grounded to an Exactpower EP15 and I have lowered the noise floor to the best possible prior to adding linears and toying with grounding some more (the PC is still ungrounded).

 

So, when I return from vacation I'll post back the results of adding the linear supplies.

 

It is interesting to power on and off the Zodiac+ as I can hear it raise the noise floor but almost like it is using a dithering method to make a "gentle"; I think it's by design. It came with a switching power supply of course.

 

[barrows]

"I have an optical firewire repeater/isolator coming tomorrow to test. I would like to believe with optical isolation of the Firewire interface then the noise signature of the source computer should be irrelevant. I'll know soon enough"

 

That isolating the computer's firewire connection will only treat noise that is delivered to the DAC via firewire. The noise the computer puts back on the AC line is shared with any components plugged into AC. And as noted above, RFI can be airborne as well (radio frequencies). As far as airborne interference is concerned, keeping the computer separated by some distance from sensitive audio components may be an advantage. You can check for the relative strength of airborne RF fields with an small AM radio. Tune the radio low in the AM band, between stations where there is only low level white noise-then hold it close to your components-this will quickly indicate how much relative hash they are putting out. For fun, try taking the radio around to your AC outplets as well as any wall switch dimmers (change the dim setting as you listen) you may be surprised what you find!

I disagree that filtering high frequency noise from the AC line is difficult, as an example, the PS Audio Duet is over 80 dB down at frequencies of 150 kHz and above. I really do not see any advantage to powering a computer or hard drive with an external linear supply, as it is easy to filter the noise it puts back on the AC line, and internally, the computer is going to be producing so much noise by DC/DC converters and mHz range processors regardless of the source of the incoming power.

 

[hifitubes]

Interesting points but the noise isn't just pushed back down AC path to other outlets. My noise rides direct on the AES cable; one it is unplugged from the DAC, I don't monitor the noise even thought the PC is still plugged into the EP15. Maybe the onboard DC-DC converter is the cause? I'll report my findings.

 

[BobH]

A lead-lined subterranean room, completely devoid of electricity, would seem to be the answer!

 

Seriously though, where is the perspective in all of this? My room has a dedicated mains spur, that's it. The rest of it is taken care of by the very clever chaps who designed the kit I use!

 

Last year my son invested very heavily in constructing a completely battery powered system to drive his kit. The gains were modest at very best. So yes, I have heard treated systems. And no, it is not money well spent IMHO. If 'huge' gains can be had by throwing money at the electricity supply of a piece of kit then I respectfully suggest that the same money would be better spent on a better designed piece of kit!

 

Obviously, those with a vested interest in mains conditioning, isolation etc, etc will argue strongly in defence of their corner, rightly so, but a 'because I'm clever and I

said so' attitude is more than common in discussions such as these and in no way addresses the fundamental question I posed right at the beginning.

 

 

Why is it assumed that the manufacturers of high-end audio don't know what they are doing?

 

Mind you, if you really can't hear this alleged noise unless music is actually playing, then perhaps the manufacturers, like me, are labouring in ignorance of their dreadfully noisy designs!

 

[BobH]

more than enough from me. Have fun.

 

[The Computer Audiophile]

"If 'huge' gains can be had by throwing money at the electricity supply of a piece of kit then I respectfully suggest that the same money would be better spent on a better designed piece of kit!"

 

Hi Bob - I respectfully don't want a DAC with a built-in 200 lbs. isolation transformer.

 

[cbarber]

One thing that seems to be being voiced here is that it's better to put a couple hundred dollar filter between your computer and the wall than it is to go for a linear power supply.

 

Does anyone disagree with this?

 

[Tipper]

On the subject of noise this is quite intersting if you're technically minded.

http://www.sengpielaudio.com/calculator-noise.htm

 

[cbarber]

This is great, we're chipping away here. I'm trying to put the issue to bed of whether it's best to get a linear PS for the computer OR put a $200 or so filter between the computer and the wall.

 

Had an idea, if someone is willing to test and post their results (maybe this test has already been done?):

 

Test 1) Laptop with SMPS, no filter between SMPS and wall

 

Test 2) Same laptop running on battery

 

Test 3 (If you have the equipment)) Same laptop, with SMPS, now with filter between SMPS and wall

 

Thinking this would offer some useful info.

 

I'm trying to get my hands on a laptop, but can't right now

 

[barrows]

Understood. As my posts above indicate, both the wired connection from the computer to the DAC, and the wired connection to the AC can bring rise to noise issues; BUT:

I think that you are experiencing something else: high frequency noise interference should not be audible in terms of your noise floor, as the frequencies are too high (way above the 20 kHz or so limit of humans). It sounds like you are experiencing a ground loop, which will give rise usually to 60 Hz hum, and sometimes multiples. Ground loops occur when their are multiple ground paths between components, and these ground paths present differing impedances to ground.

 

 

[hifitubes]

But if the dc-dc converter in the PC is generating the noise and sending it down the AES cable (or similar), is a filter going to help?

 

I don't think this is a ground loop; a grounding issue? Yes.

 

While waiting for my EP15 I had everything hooked up on two different outlets. Since I use my client PC as an HTPC for film and kids' shows, I had a ground loop via HDMI which was caused by the cable box.

 

So I have heard the hum and buzz of a ground loop which was fixed by both grounding everything to the EP15 only, and introducing a coax Ground Loop Isolator between the cable box and the coax wall jack.

 

This noise oscillates, scratches, and bleeps and is very RFI/EMI sounding, and reacts to what I do with the PC (mouse movements, gpu activty, cpu activity). IRONICALLY, the noise lessens if I load down the PC with a cpu stress test a la Prime95, or move a slider on J. River very fast (loading down the GPU). It's almost as if at idle, these converters, or something in the PC, emanate massive amounts of EMI, but at load they don't.

 

I'm the first one to admit I might be acting a little anal, but I am used to tube amps make a little noise; this noise was observable at the listening position at one point: like an AM radio being on.

 

Now that I have the faux stargrounding in place I just hear a little "rush" at the speakers; it is VERY quiet.

 

I'd really like to hear more about what Chris had a problem too. IIIRC, he had to use a different psu from the stock M10. I have a PicoPSU here that I will swap in to test, then add the linear to it.

 

Another common feature of the completely different PCs I built (and moved 30 miles away in the middle of troubleshooting) was onboard graphics capable of 1080P output.

 

[cfmsp]

 

In my opinion, you'd be wise (just) to do your own testing, as every system's AC noise characteristics (and susceptibility thereto) is unique.

 

 

clay

 

[cbarber]

@cfmsp: You're probably right. Going to try to get a macbook laptop in here this weekend. Will share what I come up with.

 

[barrows]

Ok, what you describe definitely does not sound like a ground loop-this sounds like something is seriously wrong in your setup or computer. RFI pollution is not going to make audible sounds as you describe. Sounds like something in your setup is either very poorly designed/implemented, or something is just plain broken/faulty.

I absolutely agree that any computer setup should address three likely sources of computer/hard drive noise pollution:

1. The noise the computer puts back on the AC line

2. The noise that can be transmitted form the computer to the DAC via the data connection (be it Firewire, USB, or Ethernet)

3. Airborne radiated noise

All of the above sources should be considered and addressed for optimum performance. That being said, I believe your problem is something else, and not typical RFI noise contamination. I do not expect that the noise you are hearing will be eliminated by adressing the three issues above, you have a different problem.

 

[weib]

I've tried using an old usb DAC to clean the noise out.

Maybe you can try using a low end USB only DAC as a noise filter but you can't do high resolution music.

 

The other thing would be using a audio interface like m2tech/RME/Lynx etc.

 

[Spelunker]

Just added an Austere VII to my setup.  Whlly COW, what a difference...but, still getting quite a bit of noise from my Yamaha DAC.  USB cable by itself makes no noise.  Tried some different ports, and a hub as well, but no luck.  Switched to my old Focusrite and Voila!  Zero noise!  The FR also uses the 5V usb to power itself, and I've heard some of the components are hand made, soooo, Quality matters alot.  I do with, I could isolate the voltage from my USB in my Yammy, as it has a few more options built in.  Probably, just going to upgrade my FR.  

 

Has anyone had hiss problems with USB C DACs?

Cheers

M