Old 06-18-2018, 10:45 AM   #1
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default Computer noise over ThunderBolt cable

NOISE issues - frustration mounts...

I use Reaper on a i7 6700K with a ASUS main board and Thunderbolt 3 PCIe card.
The card connects to a TB 3 -> 2 adapter, then a TB2 cable takes it up to a Presonus Quantum.

I had quite a time initially getting the Thunderbolt "stuff" working, (virtually no info available on BIOS settings, drivers etc) but the machine has been working fine with the Quantum now for roughly 6 months. Very low latency, etc etc - except for one thing: computer noise.

I've been designing and (occasionally) using audio products for 30 years, and I have NEVER heard as much computer noise as with my current lashup. If I lift grounds and hold my head right I can get the computer noise level down to -almost- acceptable, but not completely gone. Move the mouse, you hear it. More/less/different effects loaded, you hear that too.

The Quantum is in the same rack (thus the same ground potential) with a 4-channel more-me headphone monitor, an ADAT mic preamp and an active DI box. That rack is connected to the PC by the TB2 cable above, and all the gear is on two filtered power strips both plugged into the same grounded outlet.

I have fiddled with signal grounds, mains grounds, ferrite beads you name it. It's obvious the computer is the source of the noise, (wiggle the mouse) but the path it takes to get into the audio isn't so obvious. I can hear it everywhere - the Quantum headphone jacks, mains, individual channel outputs - everywhere.

I have a session coming up next Sunday, and I'm considering getting a Mac shipped in here overnight... Not that I want a Mac or to throw money away, but I'm just about at the end of my rope with this. My guess would normally be that the noise is conducted common-mode over the T2 cable, but no amount of any ferrite mix has had any effect on it. Lifting signal grounds has had the most effect, but not 100% . I must be missing something... ?

Any ideas that don't involve firearms or explosives?

Thanks!
--jim
audiomath is offline   Reply With Quote
Old 06-19-2018, 06:23 AM   #2
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
NOISE issues - frustration mounts...


I can get the computer noise level down to -almost- acceptable, but not completely gone. Move the mouse, you hear it. More/less/different effects loaded, you hear that too.

ey- on windoze i had a problem of noise on 1 interface which did not exist before some time ago--the onboard realtek mic input was the culprit for that (disabled @ windoze mixer) -very little bg noise after that..
what kind of noise levels are you actually seeing db wise?
what db input level is 'acceptable' for yourself?
mouse and programmes still have a direct effect here,barely [just passable] for actual recordings..do not think there is a comfortable solution to currents passing thru devices..until the way we use electric changes..>?

oh+ both monitors in use here make mega noise-- when switched off while music plays,all is much much betterererr.

Last edited by Bri1; 06-19-2018 at 06:37 AM. Reason: 1Xtra.
Bri1 is offline   Reply With Quote
Old 06-19-2018, 08:23 AM   #3
serr
Human being with feelings
 
Join Date: Sep 2010
Posts: 12,559
Default

That the TB data connection is functioning - the audio interface is connected and audio data is going back and forth - suggests this is a noisy ground issue and not a data corruption issue. FWIW and all that. But that means it might be possible to isolate! (If it was data corruption from the pci card straight up not working properly, you would have garbled audio at best and probably simply no connection to the interface.)

Having said that...
The part you don't want to hear. Every time I've gone looking for a PCIe thunderbolt expansion card, I read in all forums and reviews I see that these cards flat out do not work. They give you the connectivity but not the performance. You can plug in a thunderbolt dock and run USB devices off that or maybe get away with an external monitor. But anything requiring thunderbolt performance and speed with stability just won't happen. I'm told that if thunderbolt is not native to your logic board that I should "give up now". And that's the deal even with native Mac hardware!

Maybe this is finally a thing of the past? I haven't looked anything up in 6 months or so.


Thunderbolt is a heavy line in the sand as it were. Like PCI-X vs. PCIe. You can have a more than capable machine with PCI slots available and everything but as soon as you paint yourself into a corner with the first TB-only device, you have a mountain to move that doesn't like to be moved. Going back to a firewire or higher performing USB interface may be easier and cheaper in the long run if the computer is still current.

Last edited by serr; 06-19-2018 at 08:30 AM.
serr is online now   Reply With Quote
Old 06-19-2018, 03:40 PM   #4
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by serr View Post
That the TB data connection is functioning - the audio interface is connected and audio data is going back and forth - suggests this is a noisy ground issue and not a data corruption issue. FWIW and all that. But that means it might be possible to isolate! (If it was data corruption from the pci card straight up not working properly, you would have garbled audio at best and probably simply no connection to the interface.)

Having said that...
The part you don't want to hear. Every time I've gone looking for a PCIe thunderbolt expansion card, I read in all forums and reviews I see that these cards flat out do not work. They give you the connectivity but not the performance. You can plug in a thunderbolt dock and run USB devices off that or maybe get away with an external monitor. But anything requiring thunderbolt performance and speed with stability just won't happen. I'm told that if thunderbolt is not native to your logic board that I should "give up now". And that's the deal even with native Mac hardware!

Maybe this is finally a thing of the past? I haven't looked anything up in 6 months or so.


Thunderbolt is a heavy line in the sand as it were. Like PCI-X vs. PCIe. You can have a more than capable machine with PCI slots available and everything but as soon as you paint yourself into a corner with the first TB-only device, you have a mountain to move that doesn't like to be moved. Going back to a firewire or higher performing USB interface may be easier and cheaper in the long run if the computer is still current.
As you say, the TB data isn't corrupted. I spent another hour on it last night and found that if I lift ALL the input-side grounds going into the 4-ch more-me headphone monitor, (a cheap Behringer, I should know better) the noise is down to *almost* acceptable. Bonding the PC chassis to the rack frame with a 24" #8 stranded wire had no effect.

This will sound unlikely, but since I also do RF design I have two modest spectrum analyzers, one a portable 9 kHz to 3 gHz. I put a H-plane EMI probe on the portable and spent some time sniffing about. What I found is that I had to go to the largest H-plane loop and crank up the gain to even see noise from the PC. There's more noise from lighting and other electronics in the room than that PC. If there's noise being conducted to the rack it's either RF *in* the TB cable or an incredibly bad ground between the PC main board and the chassis. Of course having to use a TB3 to TB2 adapter isn't the best thing either - adds to the variables.

You may be right about a Mac. I'm a pragmatist, so I get hung up on paying 2X for the hardware just to get a decent Thunderbolt implementation. MSI has a 2nd-gen (Z370) motherboard with an integrated TB3 port that goes right to the PCIe bus. So the gamble appears to be "spend $1500 on a PC that might be better" or "spend $3K on a Mac that will probably be better" . OR - (again as you suggest) give up on the Quantum and find something else. I had a tantrum when M-AUDIO stopped driver support on the ProFire 2626, I had two of them chained up that worked well over FireWire with Windows 7. Not so much with Windows 10.

I need a beer. ( I don't even drink beer)

And for the earlier gent that responded - What I can tell you about noise is that if I can hear it, it's too much.

[edit] Oh and BTW: Thunderbolt is an Intel invention, and requires support at the CPU level. Apple, thus far is the only MFG that's taken it's seriously. Apple uses the same CPU and PCH everyone else does but they do a better job supporting it in the OS, and the integrated connector is *probably* far better than the generic add-in PCIe card. Those were built by ASUS and GigaByte directly from the Intel reference sheets; not exactly first-cabin implementations.

Thanks for the responses.
--jim

Last edited by audiomath; 06-19-2018 at 03:54 PM. Reason: Forgot to add...
audiomath is offline   Reply With Quote
Old 06-19-2018, 06:59 PM   #5
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
What I can tell you about noise is that if I can hear it, it's too much.
heh-audiomath__ if you can get rid of noise,then your doing extremely well coz it's everywhere...well..pretty much everywhere..
there's noise,and there's noisey though-- why not post some levels your seeing itb so others can see db's? >> coz>> if it's pre cabling,then yea-it won't be easy to rid that.. circuitry happens itb,then otb after cabling.
all the time, radio+telecommunications interferances are beyond our personal 'controls' all sorts of waves are happening that can be picked up by devices or cablings..glhf.
Bri1 is offline   Reply With Quote
Old 06-20-2018, 09:24 AM   #6
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by Bri1 View Post
heh-audiomath__ if you can get rid of noise,then your doing extremely well coz it's everywhere...well..pretty much everywhere..
there's noise,and there's noisey though-- why not post some levels your seeing itb so others can see db's? >> coz>> if it's pre cabling,then yea-it won't be easy to rid that.. circuitry happens itb,then otb after cabling.
all the time, radio+telecommunications interferances are beyond our personal 'controls' all sorts of waves are happening that can be picked up by devices or cablings..glhf.
The opportunity to measure it objectively has passed, but 80 dB SNR would be exaggerating. Using a 16-bit stream I can still see it as a burble on the baseline, so you could say I've been able to improve the SNR to *almost* 96 dB. Where I come from, that isn't acceptable. I had planned to record at 96/24 this coming Sunday, but unless I can make significant improvements before then I'll do the session at 44.1/16 and hand out CD's.

Last night I fiddled with the mains grounds, to no effect. I'm not sure how the RF would be conducted across the Thunderbolt cable except as common-mode, but since it doesn't respond to common-mode treatment I have to assume it isn't.

--jim
audiomath is offline   Reply With Quote
Old 06-20-2018, 10:47 AM   #7
cyrano
Human being with feelings
 
cyrano's Avatar
 
Join Date: Jun 2011
Location: Belgium
Posts: 5,246
Default

Just a thought.

I believe all TB adapters contain a processor, but I'm not sure about TB to TB cables/adapters. What if the noise stems from within the cable?

Maybe it's just a faulty cable?
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell
cyrano is offline   Reply With Quote
Old 06-20-2018, 11:46 AM   #8
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by cyrano View Post
Just a thought.

I believe all TB adapters contain a processor, but I'm not sure about TB to TB cables/adapters. What if the noise stems from within the cable?

Maybe it's just a faulty cable?
The TB3 to TB2 adapter certainly could be a problem, so could the TB2 cable from the adapter to the Quantum. When purchased, both were what was available at the time.

Good points. I'll check around and see if anything better than BestBuy consumer grade stuff exists now.

Thanks,
--jim
audiomath is offline   Reply With Quote
Old 06-20-2018, 12:06 PM   #9
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Come to think of it - if there are any Mac gurus lurking, do you have a model or date range for a 27" i7 iMac with TB2 port or ports?

Even used I'll probably make a prune face when I start checking prices, but I'd really like to keep the Quantum if I can. I have a rack built around it and (other than the noise) I like using it. Two ADAT mic/line preamps couple seamlessly to it over fiber, and not having a patchbay/mixer in the middle makes it very usable. Add the (very) low practical latencies and I probably sound like a Presonus ad...

I can probably afford to throw one more new PC (Mac or Windows) at this. After that and another couple of hours messing with grounds and ferrite, it will be off to auction with the Quantum if I can't get the noise down far enough to make a nice recording at 96/24 .

Thanks,
--jim
audiomath is offline   Reply With Quote
Old 06-20-2018, 12:16 PM   #10
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

ey-just if you have any other onboard or seperated interfacings--apparently these can cause 'issues' as posted... i had no idea why,an un-attached mic input would have a direct effect on a seperate asio card-but it did until disabled.
try disabling anything <before> any expensive purchasings,or just go for the c4 remote detonator >instead<
the spec is saying about 120db of clean signal which your not getting...have you tried brands support also?
itb_ if set to rec in reaper using that interfacing..are the meters cleanish(full dr) there for input stage?
Bri1 is offline   Reply With Quote
Old 06-20-2018, 01:48 PM   #11
cyrano
Human being with feelings
 
cyrano's Avatar
 
Join Date: Jun 2011
Location: Belgium
Posts: 5,246
Default

Quote:
Originally Posted by audiomath View Post
The TB3 to TB2 adapter certainly could be a problem, so could the TB2 cable from the adapter to the Quantum. When purchased, both were what was available at the time.

Good points. I'll check around and see if anything better than BestBuy consumer grade stuff exists now.
Despite the price, it would be wise to buy Apple adapters. From experiences with RME gear, most cables give problems. The Apple ones are 99% trustworthy.

Quote:
Originally Posted by audiomath View Post
Come to think of it - if there are any Mac gurus lurking, do you have a model or date range for a 27" i7 iMac with TB2 port or ports?
I think 2014 is the latest year before USB-C. Check everymac.com for specs.

https://everymac.com/systems/apple/imac/index-imac.html
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell
cyrano is offline   Reply With Quote
Old 06-20-2018, 02:17 PM   #12
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Thanks for that, will do.

--jim
audiomath is offline   Reply With Quote
Old 06-23-2018, 08:16 PM   #13
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Unfortunately, replacing the T3->T2 adapter with an Apple-branded one didn't change anything. Same noise, no detectable difference in performance.

I've looked at iMacs over the last few days, and decided that:

(1) It makes no sense to support one iMac in a building with 15-20 PC's

(2) It would cost over twice as much to buy an iMac with a i7 7700 than to build a i7 8700K (mini ITX using Intel graphics). Granted, a lot of the cost of an iMac is in that gorgeous display, but I already have two brand-new 34" displays at that position.

So: I'm ordering the parts tonight to build a i7 8700K around a mini-ITX board from ASRock that has integrated TB3. Will it fix the noise or change anything? I have no idea. If it doesn't, I'll have only bought yet another PC sooner than I would have wished, and I can cuss liberally and put the Quantum up for sale. If it works, I'm a hero for a few minutes.
audiomath is offline   Reply With Quote
Old 06-24-2018, 11:40 PM   #14
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

What speakers/amp are you connecting the interface to? A common cause of this sort of noise is unbalanced connection between the interface and speakers/amp.

If you listen through headphones connected directly to the interface is the noise still there? If not, the noise is being caused by the way your interface and monitoring system connect to each other.

edit: Just noticed you said this:

"Quantum headphone jacks"

OK, so noise there too. Have you started from basics with just the interface connected to the computer and headphones plugged into the interface with NOTHING else connected. Is the noise still there? If no, connect things one at a time till the noise appears again.

Last edited by drumphil; 06-24-2018 at 11:46 PM.
drumphil is offline   Reply With Quote
Old 06-25-2018, 09:32 AM   #15
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by drumphil View Post
...
OK, so noise there too. Have you started from basics with just the interface connected to the computer and headphones plugged into the interface with NOTHING else connected. Is the noise still there? If no, connect things one at a time till the noise appears again.
Thanks, drumphil.
Yes - the noise exists with a "bare" system, although at a very low level. At first I thought something was wrong with the Quantum, and then I moved the mouse. It then became obvious it was computer EMI. Adding *any* external monitor to the Quantum mains results in more and more noise. The worst is a Behringer headphone amp. (big surprise) I've also tried several distribution amps I have on the shelf, which all suffer to one degree or another. Connections are ALWAYS balanced, (both TRS and XLR tried) although ground lifts on the monitor amp sometimes improve the situation.

At this point I have to point at the PC. Poor power supply or motherboard design, don't know. There almost has to be a high-impedance ground in the mess. No combination of ground ties in the power fixes it, although using a single 10-outlet strip with an advertised 40dB noise filter did reduce it a small amount.

I would have swapped in a different PC long ago to test if there was another one within reach that had a Thunderbolt port.

Yesterday's session was a little lame because of the noise in the monitors, but the recording quality was great, the guys were more than happy. There was one acoustic guitar with a really high impedance soundhole pickup that picked up the same noise, but swapping it out for a different guitar with a LR Baggs mic/pickup and preamp took care of that. The mics are all 1" condensers on 20-foot cables direct into the Quantum, bypassing the patch bay- no noise heard there. I didn't use either of the ADAT mic pre's, so don't have a current report, although I would have been floored if I'd heard RF noise transmitted over fiber... ;-)

Summary: I really like the Quantum, perhaps irrationally so. Maybe I can find a laptop to borrow that has a Thunderbolt port, anything to change the computer end of the equation to test. In the end, though the noise has to go.

--jim
audiomath is offline   Reply With Quote
Old 06-25-2018, 09:57 AM   #16
serr
Human being with feelings
 
Join Date: Sep 2010
Posts: 12,559
Default

Wait...

Are you saying that you can get cleanly recorded tracks from the interface to the computer? The noise is only in the monitors?

If the noise is only in the monitor feed and you can record clean tracks to computer and hard drive, then you just proved you had uncorrupted data from the interface to the computer over thunderbolt. That leads to the noise problem being in an analog stage somewhere in your monitoring system.

That doesn't automatically mean the thunderbold pci card is innocent here. But it suggests the digital system is functioning error free. If there's noise getting into the analog end of the monitoring chain and it's coming from that pci card, then we have a shielding issue here.


Yeah... not sure who to believe with the TB pci card comments. Some folks are really sure that a TB pci expansion card will NEVER work on any system sold that didn't have native TB to begin with. The language I've read sounds like you wouldn't expect it to function at all digitally. Not that it would function but just have some noise.

I might have to pick one up and try it in the Mac Pro and see for myself I suppose.

Careful with the newer Macs. They went to overpriced disposable lower quality stuff after 2012 IMHO. I would only recommend one of their genuine pro machines made between 2008 and 2012.
serr is online now   Reply With Quote
Old 06-25-2018, 10:19 AM   #17
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Yes, with the current config, the noise apparently exists only in the monitors.

Note I said "apparently". :-)
I did have an issue with a (really) high-impedance acoustic guitar pickup "hearing" the same noise as the monitors, but as below changing it out for different guitar/pickup eliminated the noise, or reduced it to the degree where a quick review of the recordings didn't catch it.

The other caveat I'll throw in is that the recordings made yesterday were 44.1/16 and I haven't examined them critically for noise. I may or may not get to that this week. The gamble of the moment is that swapping out the PC will fix things to the degree that I'll be able to make "honest" 24-bit recordings.

Not that it's directly relevant here, but I do have a fair understanding of the issues at hand - this is, however the worst and most persistent case I've run up against in a small studio. The fact that it's MY project studio this time only makes for dark comedy... :-)

Thanks,
--jim


Quote:
Originally Posted by serr View Post
Wait...

Are you saying that you can get cleanly recorded tracks from the interface to the computer? The noise is only in the monitors?

If the noise is only in the monitor feed and you can record clean tracks to computer and hard drive, then you just proved you had uncorrupted data from the interface to the computer over thunderbolt. That leads to the noise problem being in an analog stage somewhere in your monitoring system.

That doesn't automatically mean the thunderbold pci card is innocent here. But it suggests the digital system is functioning error free. If there's noise getting into the analog end of the monitoring chain and it's coming from that pci card, then we have a shielding issue here.


Yeah... not sure who to believe with the TB pci card comments. Some folks are really sure that a TB pci expansion card will NEVER work on any system sold that didn't have native TB to begin with. The language I've read sounds like you wouldn't expect it to function at all digitally. Not that it would function but just have some noise.

I might have to pick one up and try it in the Mac Pro and see for myself I suppose.

Careful with the newer Macs. They went to overpriced disposable lower quality stuff after 2012 IMHO. I would only recommend one of their genuine pro machines made between 2008 and 2012.
audiomath is offline   Reply With Quote
Old 06-25-2018, 06:06 PM   #18
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

So, just to be clear, if you plug your headphones into the interface directly, with nothing else connected, and play back a file that wasn't recorded through the interface, is there any noise?

I'm slightly confused. Does the noise exist to some degree in a "bare" system (being just headphones and NOTHING else connected), or does it only exist in the monitors? Both can't be true.
drumphil is offline   Reply With Quote
Old 06-25-2018, 06:53 PM   #19
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by drumphil View Post
So, just to be clear, if you plug your headphones into the interface directly, with nothing else connected, and play back a file that wasn't recorded through the interface, is there any noise?

I'm slightly confused. Does the noise exist to some degree in a "bare" system (being just headphones and NOTHING else connected), or does it only exist in the monitors? Both can't be true.
Sorry - I know it sounds inconsistent.
When I first put the Quantum in the rack and got the Thunderbolt working, yes - there was some low-level computer noise in the direct headphone outputs. That was around the end of November 2017. Since then, I've rewired the rack with low impedance grounds, ground-bonded all the gear and loaded the permanent and semi-permanent cabling with mix 31 ferrite beads. After doing those things, if there is any noise in the direct headphone outputs I can't hear it.

What hasn't changed is that if I attach any line-level gear to the Quantum outputs, (regardless of pedigree) I get RF noise. The subjective amount varies with the type and amount of equipment and connection method. For the most part, lifting the grounds on the monitor/amp end of the cables makes an improvement. The more outputs I use, the higher the subjective noise level becomes. (my goal is to fully implement "more me" monitoring, which requires using both the main outputs and individual line outputs)

Hopefully that will fill in the gaps a bit.

Thanks,
--jim
audiomath is offline   Reply With Quote
Old 06-25-2018, 09:28 PM   #20
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

Quote:
When I first put the Quantum in the rack and got the Thunderbolt working, yes - there was some low-level computer noise in the direct headphone outputs.
Was that with NOTHING else connected other than the headphones (well, and the thunderbolt cable to the computer)?

Because if NOTHING else was connected, in a setup with no possible grounding issues with other gear, and you still hear noise in the headphones, then root cause isn't interactions with your other gear, although they may exacerbate the problem when connected. That would point back at the quantum or your computer as the source of the problem. At that point, you know there is going to be a problem regardless of what you do with the rest of the setup.

Just a thought though, does it make any difference, with just the computer and headphones, if the quantum isn't touching the rack at all?

Last edited by drumphil; 06-25-2018 at 09:39 PM.
drumphil is offline   Reply With Quote
Old 06-26-2018, 01:18 AM   #21
cyrano
Human being with feelings
 
cyrano's Avatar
 
Join Date: Jun 2011
Location: Belgium
Posts: 5,246
Default

I reckon the grounding scheme on the PC mainboard isn't really well adapted to audio. Sometimes, just tightening the screws is enough to make a difference.

It's not that there's one point that has DC resistance to the ground, it's that there are multiple different resistances. And these tend to add up.

Adding locking washers to the screws might also make a difference.
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell
cyrano is offline   Reply With Quote
Old 06-26-2018, 01:44 AM   #22
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

Oh yeah, should we assume you've re-seated the thunderbolt card? I've had sound cards that had weird noise problems that went away when I re-seated them.
drumphil is offline   Reply With Quote
Old 06-26-2018, 10:38 AM   #23
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by drumphil View Post
Was that with NOTHING else connected other than the headphones (well, and the thunderbolt cable to the computer)?

Because if NOTHING else was connected, in a setup with no possible grounding issues with other gear, and you still hear noise in the headphones, then root cause isn't interactions with your other gear, although they may exacerbate the problem when connected. That would point back at the quantum or your computer as the source of the problem. At that point, you know there is going to be a problem regardless of what you do with the rest of the setup.

Just a thought though, does it make any difference, with just the computer and headphones, if the quantum isn't touching the rack at all?
Well - it's my recollection that I did, indeed hear the noise when I first attached the Quantum to the PC - without anything else attached. Bear in mind this was many months and miles ago.

I haven't tried "floating" the Quantum... I'll try that if the new PC with the integrated TB3 port doesn't prove to be the ultimate solution. All the parts should be here by Thursday evening. Assuming I don't get streamrolled, the plan is to have it assembled, tested and in place before the session this coming Sunday.

Thanks,
--jim
audiomath is offline   Reply With Quote
Old 06-26-2018, 10:47 AM   #24
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by drumphil View Post
Oh yeah, should we assume you've re-seated the thunderbolt card? I've had sound cards that had weird noise problems that went away when I re-seated them.
No I haven't, but I'll give that a shot - as well as re-attaching the PSU and MB with star washers, and De-Oxit everywhere. Back in the ISA-bus days I used De-Oxit on card fingers as well, but I haven't tried it on PCIe cards yet. It's supposed to be electrically inert, but it doesn't take much capacitance to screw things up at 40Gbps - or whatever the bus speed actually is these days.

--jim
audiomath is offline   Reply With Quote
Old 06-28-2018, 02:59 AM   #25
Philbo King
Human being with feelings
 
Philbo King's Avatar
 
Join Date: May 2017
Posts: 3,201
Default

It sounds to me that your PC is generating RF artifacts. These could be conducted or radiated. Given all the checking you've done so far,I'd suspect radiated emissions from the PC. Try isolating it, either by distance, or by putting it in a faraday cage (aluminum foil works for a quick test, just make sure the foil is grounded to earth ground (not the PC chassis). If this helps, you need a better designed computer case.
__________________
Tangent Studio - Philbo King
www.soundclick.com/philboking - Audio streams
Philbo King is online now   Reply With Quote
Old 06-28-2018, 09:34 AM   #26
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by Philbo King View Post
It sounds to me that your PC is generating RF artifacts. These could be conducted or radiated. Given all the checking you've done so far,I'd suspect radiated emissions from the PC. Try isolating it, either by distance, or by putting it in a faraday cage (aluminum foil works for a quick test, just make sure the foil is grounded to earth ground (not the PC chassis). If this helps, you need a better designed computer case.
An EMI check with a spectrum analyzer didn't find any undue direct emission from the PC, so my favorite theory is simply noisy grounds. I'm building a new PC with a motherboard that has an integrated TB3 port and I selected a well-reviewed, brand-name power supply.

Since the only motherboard I could find with an integrated TB3 port also happens to be Mini ITX, the selection of "heavy metal" cases was limited. I went with one that doesn't have a plastic viewing window, but it's anything but a Faraday cage. It's also proving to be a PITA to assemble, which is not a big surprise given the size.

It's an Intel i7 8700K with 16GB of reasonably low-latency RAM and the fastest NVMe M.2 boot drive I could find. If a miracle occurs and the new PC eliminates the problem I'll post a build list in this thread. Let's see if I can link photo URL's... (WARNING - large images)

http://www.barberdsp.com/downloads/pc1.png

http://www.barberdsp.com/downloads/pc2.png

Thanks,
--jim

Last edited by audiomath; 06-28-2018 at 09:35 AM. Reason: Correct photo URL
audiomath is offline   Reply With Quote
Old 06-28-2018, 10:27 PM   #27
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

Lol.... "FATAL1TY"

I actually own the dumbest sounding sound card in history.

The X-Fi Titanium Fatal1ty Professional.


Where exactly does the PSU go in this setup?

Last edited by drumphil; 06-28-2018 at 10:32 PM.
drumphil is offline   Reply With Quote
Old 06-29-2018, 08:38 AM   #28
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by drumphil View Post
Lol.... "FATAL1TY"

I actually own the dumbest sounding sound card in history.

The X-Fi Titanium Fatal1ty Professional.


Where exactly does the PSU go in this setup?
I had to do an eye-roll on "FATAL1TY" too. The motherboard book has a forward from an apparently famous gamer who goes by that handle. It's a Creative Labs thing. The onboard audio is branded Creative Labs "FATAL1TY". None of that kept me from disabling it in the BIOS, though, as well as the onboard WIFI. :-)

The PSU mounts in the rectangular "tube" on the right side of the chassis in the photo - it has a white sticker on it. It then hangs over everything else, which means that the cabling has to be squared away first. I had to fiddle the orientation of the cooling block and lines as well. Normally I pretty much throw PC's together, but this one required some thought to get everything connected and inserted. The M.2 SSD actually mounts on the bottom of the MB, and there is a bracket that straddles the top that will mount two 3.5" drives.

Zippy, I will give it that. If I turn on "fast boot" in the BIOS it goes from cold to a login prompt in less that 10 seconds. (quite a bit less I think, but I didn't time it) Login->launch Reaper->open a project is maybe another 10 seconds. I can live with it. ;-)
audiomath is offline   Reply With Quote
Old 06-29-2018, 09:31 AM   #29
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Prelim findings: (WARNING - long winded)
Just before midnight I took the machine down to the studio and plugged it in. No HDD's yet, covers still off, no ground bonding to the rack. Drivers installed, Quantum firmware updated etc etc.

(1) I pulled eveything except power and TB cable from the Quantum, (floated it) then plugged my favorite old Sony cans into the front HP jack. With the PC and Quantum turned on and driver connected, NO NOISE. I set the Windows audio output to the Quantum mains, NO NOISE. Streamed a few tracks from the NAS box, NO NOISE.

(2) Installed and licensed Reaper, set up a track with a single mic, (condenser, 48V enabled) monitor still only through the Sony cans. Room rumble at high gain, but no computer noise. Quantum is still floating.

(3) Connected two ancient EMU monitors (balanced) to the Quantum mains out TRS-TRS . Rats... Low-level computer noise heard.

(4) Disconnected the EMU's and connected the [expletive deleted] Behringer HA-4700 headphone amp TRS-TRS. Computer NOISE is definitely heard in the HA-4700 output, but at least 20dB better than previous tries. Still no detectable noise in the Quantum HP outputs.

(5) Racked the Quantum and the HA-4700, connected ground straps. Noise in the HA-4700 output increases by a few dB, (??) but still far below prior attempts. I pulled the screws with the ground bond wires, scraped and fiddled the rails a bit. NOISE from the HA-4700 REDUCES a few dB. ( !!! )

I crawled off to bed at that point and it's too early for solid conclusions, but the plan for later tonight is to:

(1) Plan a sledgehammer ceremony for the HA-4700. I can't remove until I find something to replace it, (Mackie?) but its days are numbered.

(2) Find a better, lower-impedance way to bond the Quantum to the other gear in the rack. The Quantum doesn't have a ground post, though... I have a roll of Georgia Copper 1.25" wide thin copper strap left over from a RF job, but I need to figure out how to connect it to the gear. Maybe instead find some brass rack screws and bond point-to-point with #10 wire and lugs?? The Quantum rack ears appear to be painted silver, too. I thought they were some kind of metal finish originally, which is potentially the largest issue.

(3) I'm going to try putting the midget PC on top of the rack. If there's no induction field issues, that will shorten that part of the ground equation by quite a bit.

Overall, the new PC has brought a considerable improvement with it, but monitoring is still an issue. I'm encouraged enough not to retire the Quantum, though, at least not yet.

--jim
audiomath is offline   Reply With Quote
Old 06-30-2018, 11:49 PM   #30
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

Quote:
(3) Connected two ancient EMU monitors (balanced) to the Quantum mains out TRS-TRS . Rats... Low-level computer noise heard.
The next thing to try there is disconnecting the shield at the speaker end of the TRS cables. This page has some very useful information (which you probably already know) and diagrams. The Rane Notes are an extremely valuable resource.


https://www.rane.com/note151.html

Quote:
The Audio Engineering Society recommendation will address a simple issue, the absurdity that one can not buy several pieces of pro audio equipment from different manufacturers, buy off-the-shelf cables, hook it all up, and have it work hum and buzz-free. Almost never is this the case. Transformer isolation and other interface solutions are the best solutions for balanced/unbalanced interconnections, though they are too costly for many systems. Even fully balanced systems can require isolation transformers to achieve acceptable performance. Some consider isolation transformers the only solution. These superior solutions are not covered here.

Another common solution to hum and buzz problems involves disconnecting one end of the shield, even though one can not buy off-the-shelf cables with the shield disconnected at one end. The best end to disconnect is unimportant in this discussion. A one-end-only shield connection increases the possibility of radio frequency (RF) interference since the shield may act as an antenna. The fact that many modern day installers still follow the one-end-only rule with consistent success indicates that acceptable solutions to RF issues exist, though the increasing use of digital technology increases the possibility of future RF problems. Many successfully and consistently reduce RF interference by providing an RF path through a small capacitor connected from the lifted end of the shield to the chassis.

The details of noise-free interconnections and proper grounding and shielding are well covered in other literature. They are not revisited here. Readers are encouraged to review the References listed for further information. Most of these materials have been applicable in the audio industry for well over 60 years, though they have not been implemented or embraced by many.
Quote:
Fully Balanced

Fully balanced systems (left column in Figure 5) provide the best performance when both ends of the shield connect to units with chassis-grounded shields (Figure 5a). When units with signal-grounded shields are encountered, disconnect the shield at the signal-grounded end (Figures 5b & 5c). This keeps the induced shield currents out of the audio signal ground. If both units involved have signal-grounded shields, you have entered the twilight zone (Figure 5d). This is perhaps the most common scheme. Most disconnect one end of the shield, specifically which end is disconnected creates strong political debates and is left for the individual user to decide [6]. Never disconnect both ends of a shield.
So, shield disconnected at one end, or connected with a capacitor may be the solution in your situation. Damnit. This should be easy!

I wonder if all your problems would go away if all your balanced cables had the shields disconnected at one end.

I guess that would also mean floating everything so that the rack doesn't keep all the chassis electrically connected after you've disconnected all the shield wires (although, something seems wrong about that idea. I'm no EE, you may be able to explain what goes on there better than me).

My experience with these issues suggests to me that grounding issues are much more common, and cause greater audio quality problems than compromised shielding. I don't think I've ever had any real problems with the shield disconnected at one end due to RF noise. Grounding issues with the shields connected however are very common, and degrade audio quality in a very obvious way.

Anyway, I'm a computer and studio techie.. Certainly no form of actual electrical engineer, so I hope I'm not just adding more confusion.

Another Rane Note on the situation:


https://www.rane.com/note110.html

Quote:
A balanced line requires three separate conductors, two of which are signal (+ and -) and one shield (see Figure 1a). The shield serves to guard the sensitive audio lines from interference. Only by using balanced line interconnects can you guarantee (yes, guarantee) hum-free results. Always use twisted pair cable. Chassis tying the shield at each end also guarantees the best possible protection from RFI [radio frequency interference] and other noises [neon signs, lighting dimmers].

Neil Muncy, an electroacoustic consultant and seasoned veteran of years of successful system design, chairs the AES Standards Committee (SC-05-05) working on this subject. He tirelessly tours the world giving seminars and dispensing information on how to successfully hook-up pro audio equipment2. He makes the simple point that it is absurd that you cannot go out and buy pro audio equipment from several different manufacturers, buy standard off-the-shelf cable assemblies, come home, hook it all up and have it work hum and noise free. Plug and play. Sadly, almost never is this the case, despite the science and rules of noise-free interconnect known and documented for over 60 years (see References for complete information).

It all boils down to using balanced lines, only balanced lines, and nothing but balanced lines. This is why they were developed. Further, that you tie the shield to the chassis, at the point it enters the chassis, and at both ends of the cable (more on `both ends' later).

Since standard XLR cables come with their shields tied to pin 1 at each end (the shells are not tied, nor need be), this means equipment using 3-pin, XLR-type connectors must tie pin 1 to the chassis (usually called chassis ground) -- not the audio signal ground as is most common.

Not using signal ground is the most radical departure from common pro-audio practice. Not that there is any argument about its validity. There isn't. This is the right way to do it. So why doesn't audio equipment come wired this way? Well, some does, and since 1993, more of it does. That's when Rane started manufacturing some of its products with balanced inputs and outputs tying pin 1 to chassis. So why doesn't everyone do it this way? Because life is messy, some things are hard to change, and there will always be equipment in use that was made before proper grounding practices were in effect.

Unbalanced equipment is another problem: it is everwhere, easily available and inexpensive. All those RCA and 1/4" TS (Tip-Sleeve) connectors found on consumer equipment; effect-loops and insert-points on consoles; signal processing boxes; semi-pro digital and analog tape recorders; computer cards; mixing consoles; et cetera.

The rest of this note gives tips on how to successfully address hooking up unbalanced equipment. Unbalanced equipment when "blindly" connected with fully balanced units starts a pattern of hum and undesirable operation, requiring extra measures to correct the situation.

Last edited by drumphil; 07-01-2018 at 12:49 AM.
drumphil is offline   Reply With Quote
Old 07-01-2018, 08:19 PM   #31
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Thanks, drumphil. I appreciate the effort!

And... (drum roll please) we finally have complete success. As briefly as possible:

. New PC with integrated TB3 port (big difference)
. DIY rack, front and back rails bonded with #10 stranded wire with lugs
. Quantum racked per normal, grounded only through rack ears. (no paint removed)
. 4-channel active DI box (grounded through ears only + mains safety gnd)
. Was 2, now 1 8-channel ADAT mic pre (grounded through ears only + mains safety gnd)
. Much-cursed HA-4700 4-channel headphone amp with "more me" aux inputs (more below)

(1) Lifted grounds to monitor/subwoofer system by looping through DI box with ground lift switches engaged. Still had a little RF noise in the monitors. Far, far down from the worst case, but still a bit there.

(2) Disconnected HA-4700 safety ground, (clipped pin) RF noise 100% GONE!

So... What this *appears* to have boiled down to is:
(1) Going to an integrated TB3 port was the largest single factor
(2) There was a small ground differential between the HA-4700 signal ground and the mains safety ground, which allowed just that little bit of RF to flow - and the HA-4700 again appears to be sensitive to it in the extreme.

Last: Smite me if you must for clipping the safety ground on the HA-4700, but please have your ducks lined up first. :-)

Thanks for all the suggestions and for helping me see this through.

--jim
audiomath is offline   Reply With Quote
Old 07-02-2018, 12:47 AM   #32
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

Quote:
Last: Smite me if you must for clipping the safety ground on the HA-4700, but please have your ducks lined up first. :-)
Does the connection between the quantum and the HA-4700 still have the shield connected at both ends? Is low noise operation possible with the safety ground in place if you snip the shields?

Also, can't you just snip the shield on the connection to the speakers so you don't need the DI box in the middle?
drumphil is offline   Reply With Quote
Old 07-02-2018, 09:07 AM   #33
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by drumphil View Post
Does the connection between the quantum and the HA-4700 still have the shield connected at both ends? Is low noise operation possible with the safety ground in place if you snip the shields?
I haven't tried that yet, but I will when I get the new headphone amp in place - if necessary. (Even if I liked the HA-4700, 4 ports isn't enough) I have 2 "clipped" cables to test with, but I need to clip the AUX "more me" cables coming from the Quantum line outputs as well.

Quote:
Also, can't you just snip the shield on the connection to the speakers so you don't need the DI box in the middle?
Yes - I need those 2 DI ports back. I did it that way as a "quick fix" because it was Saturday night and I ran out of time.

Thanks!
--jim
audiomath is offline   Reply With Quote
Old 07-04-2018, 01:41 AM   #34
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

Glad to know you've got everything sorted! It's a breath of fresh air working through issues like this with someone who is methodical enough to break thing down into small enough chunks to determine exactly where the source of the problem is.
drumphil is offline   Reply With Quote
Old 07-05-2018, 10:55 AM   #35
audiomath
Human being with feelings
 
Join Date: Jul 2016
Posts: 35
Default

Quote:
Originally Posted by drumphil View Post
Glad to know you've got everything sorted! It's a breath of fresh air working through issues like this with someone who is methodical enough to break thing down into small enough chunks to determine exactly where the source of the problem is.
Thanks!
It helps a lot to have feedback. Otherwise it's too easy to end up in spirals at 2AM... :-)

I ordered a Mackie HM-800 headphone amp to replace the Behringer HA-4700. It should be here tomorrow (Friday) afternoon. With any luck, the fixes will hold or if I'm astoundingly lucky the HM-800 will turn out to be less sensitive to "these kinds" of issues.

So for anyone that may have suffered through this thread because of an interest in the Quantum - yes, it was worth it. I'm currently using a 64-sample ASIO buffer, (which could be smaller) and loading each track with pretty much the same effects I would use for mastering. A typical session runs two stereo and 4-6 mono tracks in Reaper, with additional hardware outputs for "more me" monitoring. It all runs flawlessly; the only issue of note has been the computer noise.

If I had to do it all over again, I would have waited a few months until I could get a motherboard with an integrated Thunderbolt 3 port, or (if I were someone else) bought a Mac. In fact, if I were a Mac user I probably could have avoided a good part of trouble I've had to slog through. Oh well... My forehead may be bloody but my heart is still pure. ;-)

--jim
audiomath is offline   Reply With Quote
Old 07-05-2018, 06:49 PM   #36
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

I'm very curious to know what happens if you plug your old computer in now that the grounding issues with the headphone amp have been sorted out.
drumphil is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 10:24 AM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.