• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Measurements of RME ADI-2 DAC and Headphone Amp

Music1969

Addicted to Fun and Learning
Joined
Feb 19, 2018
Messages
522
Likes
142
I just ordered the Delock 62982 galvanic isolator (3kV optical isolation, 480mbps), it costs only half as much as the Intona device that RME recommends. Should arrive Wednesday, then I'll report back.
Would be interested to know how ground is isolated.

I would guess the data lines are optically isolated and power (when not using optional external PSU) is using DC to DC converters.
 
Joined
Jul 14, 2019
Messages
18
Likes
14
Location
Europe
I'd assume it's not providing power on the USB device side, and the power input on the isolator is only there to drive the isolator itself when USB host power is not sufficient (the isolator requires 320 mA). Not a problem with the ADI-2 DAC, though.

The Intona devices do provide power on the device side using linear regulators.

EDIT:

Nevermind, I just found this thread:

https://www.eevblog.com/forum/beginners/129322-help-with-crappy-usb-isolator/

I'm going to cancel the order. I guess the Intona devices are really the only proper solutions on the market.
 
Last edited:

KSTR

Active Member
Joined
Sep 6, 2018
Messages
216
Likes
518
Location
Berlin, Germany
Short summary:

- so-called balancing currents along the cable shield of an audio interconnect are the *only* place where the error signal is formed, nowhere else (at least not in the RME devices which are immune to even huge cross-currents). The shield has some impedance and that allows for voltage drops that appear as error signal at the receiving end.

- the root cause for the balancing currents is the power supplies, more excactly the mains coupling which forms a secondary path between the audio grounds of the devices. The path with the lower impedance wins and when it connects to noisy points in a distributed net this noise is forced to pass through the audio cables. Advice: use cables with lowest possible shield resistance and/or make additional beefy connections between device grounds of the analog gear to reduce the shield current. And, of course, use balanced interconnects whenever possible.

- there is no way to "break the loop" with simple means, all we can hope for is empirically find a connection setup that produces the lowest balancing current along the interconnect cable.

- the only way to break the loop is true galvanic isolation, preferably on the digital side (for USB I can recommend the Intona). This solution is 100% guaranteed to always work perfectly and reliably.

- the stock supply of the RME is *not* an issue here. It produces some leakage current but that is several orders of magnitude lower than what is discussed here. To make audible impact the interconnect cables must be of very poor shield quality and with significant length.

- I've built an supply isolator using a TRACO THM 30-1213 medical DC/DC convertor inserted after the stock supply that lowers the mains coupling to totally irrelevant levels. Together with the Intona thus I have a "truly floating" measurement setup for some specialist investigation. For critical listening, the better isolation of this supply is irrelevant whereas the Intona USB isolated really can make a difference if one is forced to use a setup which is especially prone to suffer from balancing noise current on the audio interconnects. The typical case is both the source (PC) and the amps being earth-grounded, together with unbalanced cables. That never works, never did, never will.
 

Nango

Senior Member
Patreon Donor
Joined
Aug 6, 2018
Messages
317
Likes
162
Location
Rheinhessen, DE

KSTR

Active Member
Joined
Sep 6, 2018
Messages
216
Likes
518
Location
Berlin, Germany
I have those Olimex Isolators, they don't seem work with many USB2.0 audio devices including the RME but work well for isolating test equipment, programming adapters and the like.
 
Joined
Jul 14, 2019
Messages
18
Likes
14
Location
Europe
Advice: use cables with lowest possible shield resistance and/or make additional beefy connections between device grounds of the analog gear to reduce the shield current. And, of course, use balanced interconnects whenever possible. The typical case is both the source (PC) and the amps being earth-grounded, together with unbalanced cables. That never works, never did, never will.
Thanks for the technical insight.

The active monitors are connected to the ADI-2 DAC using a pair of 1.5 meters of Cordial CRM FM XLR connections, plugs are Neutrik NC3 XX, cable is CMK ROAD 250. Datasheet of cable:

https://www.cordial-cables.com/product-pictures/bulk-cable/cmkroad250/cmk-road-250.pdf

Datasheet of Neutrik connectors:

https://www.neutrik.us/en-us/product/nc3mxx.pdf

I'd say that's high-end enough for a balanced connection at 1.5 meters. ;) At least it shouldn't cause audible interference. So I don't think that this is a case of poor, unbalanced connection shields transporting currents. Unless you meant the power chord cables or USB cables or DisplayPort cables ...
 
Last edited:

Purité Audio

Major Contributor
Industry Insider
Barrowmaster
Joined
Feb 29, 2016
Messages
3,201
Likes
1,049
Location
London
We stocked the Intona isolators in fact my ‘avatar’ is an Intona created USB eye pattern, measured I believe by Mattias of RME.
I believe they ‘worked’ but with properly designed equipment I couldn’t discern any audible difference, one did solve an issue with a ANUK dac though!
Keith
 
Joined
Jul 14, 2019
Messages
18
Likes
14
Location
Europe
The one issue I have with Intona is that they have started dabbling in the cable voodoo business:

https://pro.intona.eu/en/stories/cable

Considering that the USB cables RME bundles with their high-end interfaces are just bleak, black standard cables that also happen to conform to USB 2.0 specs just fine.
 
Joined
Jul 14, 2019
Messages
18
Likes
14
Location
Europe
I have now found why I was assuming for a while that the USB input of the ADI-2 DAC would be galvanically isolated. It was in a review of the unit at Thomann (Europe's largest retailer for music gear):

https://www.thomann.de/gb/rme_adi_2_dac.htm

I am extremely pleased with the sound quality, the soundstage improved quite a lot, not to mention the depth and realism I get.. it's like buying a new set of speakers in a much higher category. I enjoy the fact that the USB input is galvanically isolated and the power supply and every part of the power stage is done in a way that it does not need fancy stuff before the DAC to "filter" noise out. There is NO noise whatever you put before this.
Not true at all, it seems.
 

KSTR

Active Member
Joined
Sep 6, 2018
Messages
216
Likes
518
Location
Berlin, Germany
Intona seemed to add noise to RME ADI-2 DAC measurement... even though level of noise added was small, RME ADI-2 DAC measured better without it?
I would assume this (the low frequency noise -- mains hum?) is more likely a measurement problem than a real phenomen. This kind of measurement needs special care and expertise and lots of double-checking to validate the results. You need a lot of time and patience to really dig into matters to the required level (read: several hours, at least).
 

DDF

Active Member
Joined
Dec 31, 2018
Messages
237
Likes
219
I have those Olimex Isolators, they don't seem work with many USB2.0 audio devices including the RME but work well for isolating test equipment, programming adapters and the like.
The adum4160 based olimex worked for me to break usb2 ground loops but is limited to 24/96 and injects jitter (which a good dac cleans up). Like you, i ended using a permanent solution on the psu side (of the pc)
 

Music1969

Addicted to Fun and Learning
Joined
Feb 19, 2018
Messages
522
Likes
142
I would assume this (the low frequency noise -- mains hum?) is more likely a measurement problem than a real phenomen. This kind of measurement needs special care and expertise and lots of double-checking to validate the results. You need a lot of time and patience to really dig into matters to the required level (read: several hours, at least).
Noted. Maybe something for @amirm
 
Joined
Jul 14, 2019
Messages
18
Likes
14
Location
Europe
The adum4160 based olimex worked for me to break usb2 ground loops but is limited to 24/96 and injects jitter (which a good dac cleans up). Like you, i ended using a permanent solution on the psu side (of the pc)
Which PSU did you end up using?
 

Tks

Senior Member
Joined
Apr 1, 2019
Messages
379
Likes
324
Location
NYC
@entrophy

I read about your groundloop ordeal and had a similar experience, but when I tested the issue, it had nothing to do with ground loops or anything of the sort. The whole section about "Coil whine" transitioning to the DAC even with USB input is the situation I was referring to. I also have a dedicated workstation(barely)/gaming PC. I've unbound the GPU bottleneck as much as I could and ran synthetic benchmarks to induce as much coil whine as possible with approximately 900+ frames per second easily make the GPU (GTX 980 and 1080Ti) scream. I have never experienced anything coming out of the RME with respect to the coil whine nor groundloop issues (took an extension cable to make sure the coil whine wasn't leaking through my open back headphones and fooling my into thinking it was leaking through the DAC).

Also I think it's important you use a USB port that doesn't share a bus with the rest of your connected devices. I tried all my USB ports (2.0 and 3.0) and they were all fine, so I simply believe I'm not having ground loop issues since eliminating some idiotic florescent desk lamp I had that was completely flickering my monitor, and making G-Sync go haywire.

Oh and my PSU is the Seasonic 750W Prime Ultra Titanium.
 
Joined
Jul 14, 2019
Messages
18
Likes
14
Location
Europe
@Tks Thanks for the report. My setup is a bit particular in that the coil whine is not transitioning to the ADI-2 DAC through the source PC. Actually, the source PC with the GPU whining and the ADI-2 DAC are not directly connected, but there are multiple paths:

An HDMI cable is connected to the offending GPU and a display, and at the same display is a DisplayPort cable running to a MBP, which is connected through USB to the ADI-2 DAC. Also, the computer with the whining GPU is connected through USB to the same display's builtin USB hub, and at this USB hub there is another USB cable running to the MBP (so that I can use a single keyboard/mouse on two computers, where the displays automatically switches USB inputs when changing the video input from one computer to the other). So there are at least two paths that could pick up the coil whine and ultimately end in the ADI-2 DAC through the USB connection from the MBP.

I'll probably pick up a 12V battery for the ADI-2 DAC to double-verify that its floating PSU is not picking up the coil whine from the power socket.

I have used Seasonic PSUs exclusively in the past, but at one point basically every Seasonic PSU I purchased had coil whine itself (even with a bare mainboard mounted, no cards, no peripherals), so I gave up and switched to BeQuiet. Might be time to come back to Seasonic, as spec-wise the latest models look very interesting.
 
Last edited:
Top Bottom