Dirty USB shenanigans

[Noodles]

So, i thought i'd start this to see people's opinions on dirty usb signal issues.

I had a few run-ins with several dacs over the past years where my PC hardware has not really changed. I've had a few SMSL dacs, some topping dacs and now i got the USB converter (d10s) and an RME ADI-2 DAC.
I've had the same cables and i do not recall having changed something in the mean time that could affect what i am about to describe.
I got a dirty USB
--Not that dirty, stop what you are thinking.

So i've suspected this for a while but i can confirm it now - the USB signal through all the motherboard's controllers/ports is quite dirty. I have either some coil whine bleeding into the power or unshielded traces, idk what specifically.
Plugged into my RME i have absolutely no issues whatsoever. The previous Topping Dx7 pro also had 0 problems.
I got the d10s today and what do you know - there it goes - the whine is back.

Anyone have any suggestions on how i can alleviate this (not going through a powered hub) ?

 

[gvl]

If all USB ports on your PC are "dirty" I doubt you can solve this w/out extra hardware and a powered hub ain't it although it is worth a shot. The Intona Isolator should help but it is not cheap. iFi makes several USB isolators but I found them useless in a similar situation but ymmv. Some have been successful by powering USB DACs from an external power source using a custom made USB cable but you can also find them pre-made. Does your amp has a grounded power plug? If so make sure it is plugged in into the same outlet as the PC.

Out of curiosity, were you using single ended or balanced interconnects with the RME/Dx7 ?

 

[Killingbeans]

Ground loop maybe? Like glv points out, I'd try some galvanic isolation or to break the loop in some other way.

 

[Noodles]

Out of curiosity, were you using single ended or balanced interconnects with the RME/Dx7 ?

Balanced, but i did just try single ended and i can hear the noise, it's just MUCH further on the volume dial, goes to say how good the filtering on the RME is...
Speaking of filtering, the Topping A90... XLR in from the RME, RCA in from the D10s: input set to XLR - noise is present. It appears the switch does not properly isolate the SE/Balanced circuit.
I have a powered hub and the noise is minimal when i plug the USB in it, however trying to avoid getting another powered hub altogether...sooo yeah.

 

[Matias]

Why do you use the Topping D10s into the RME ADI-2 DAC? Just use the USB directly into the RME and problem solved?

 

[Noodles]

Why do you use the Topping D10s into the RME ADI-2 DAC? Just use the USB directly into the RME and problem solved?

Trying out some stuff - got powered speakers with SPDF in, which have their own remote/volume control - so don't need to use the preamp, but do want to judge the differences between the dacs, also - same thing on my other two amps.
Just playing around, but thought that it might be worth solving this issue once and for all...

 

[gvl]

Balanced, but i did just try single ended and i can hear the noise, it's just MUCH further on the volume dial, goes to say how good the filtering on the RME is...
Speaking of filtering, the Topping A90... XLR in from the RME, RCA in from the D10s: input set to XLR - noise is present. It appears the switch does not properly isolate the SE/Balanced circuit.
I have a powered hub and the noise is minimal when i plug the USB in it, however trying to avoid getting another powered hub altogether...sooo yeah.

These inexpensive DACs skimp on USB isolation measures and you get these sticky issues with some hardware forcing you to buy additional hardware to fix it, so in the end not so cost effective solution... I don't endorse it as it is dangerous (as in it can kill you) but if you somehow lift the ground pin from the A90's power connector the noise will likely disappear. In the US we have "cheater plugs"... Combined with a GFCI-protected outlet they can be "reasonably safe" to operate and help break ground loops. Still I would only do it as a quick experiment and not a permanent solution.

 

[Noodles]

These inexpensive DACs skimp on USB isolation measures and you get these sticky issues with some hardware.

Yeah, that much is obvious
Intona looks interesting but does not have enough reviews so i can judge it... anyone have any experience with a powered hub that has a decent enough power supply to remove this noise ?

 

[gvl]

Yeah, that much is obvious
Intona looks interesting but does not have enough reviews so i can judge it... anyone have any experience with a powered hub that has a decent enough power supply to remove this noise ?

The noise likely leaks through the USB ground conductor, so a clean power supply isn't going to help. You need to galvanically isolate your PC from the DAC. The Intona should work. Paying for it more than for the DAC is of course nonsense.

 

[Noodles]

Eh, i guess i'll just use the D10s as the usb interface i intended it for and i will have clean nice optical/coax signals - literally dead silent.
Thanks for your input guys.

 

[Blumlein 88]

A new motherboard perhaps?

 

[Noodles]

It's probably the power supply tbh, but too expensive to swap. But yeah i don't think i will be using straight USB to DAC anymore (unless i'm testing the usb filtering of said dac )

 

[Blumlein 88]

I did fix the issue once with one of the self powered active USB extenders. Worked for that DAC and that PC so I can't say if it works in general.

The one I used was USB 2.0 like this.
https://www.amazon.com/BlueRigger-F...rds=active+usb+extender&qid=1609293023&sr=8-5

I suppose a USB 3.0 would do the same thing. This is USB 3.0 only and not compatible with USB 2.0. So the above one is more universally usable.
https://www.amazon.com/CableCreatio...&keywords=usb+extender&qid=1609292953&sr=8-26

They use the USB power and recreate the signal each step. The voltage is a tiny bit lower after each one as a result. So if you daisy chain too many the device at the other end might not see it or work. I've never had them fail to work using just one however.

Cheap enough to try and if it doesn't work send it back.

 

[solderdude]

I have either some coil whine

Coil whine is the coil itself (in a power supply) making audible noise. You can hear hiss or weird noises coming from a physical device. It does not have to have an effect on any signals.

What you call coil whine is a groundloop and is of electrical origin (common mode currents) and the D10 is notorious for this. It's a ground layout error in the D10 which is responsible. The only way is to break the groundloop.

 

[ZolaIII]

How about getting a usb splitter cable for power/data for use with D10 evantualy use a power bank to power it. In my experience USB power from PC usually hurts around 2~3 dB compared to battery source.

 

[Noodles]

That's way too much work, besides as i mentioned on the ifi zen dac thread - its not our problem to fix an engineering decision to skimp on a 10 bucks worth of a signal filter.
If the product is shit, it's going to either get returned or sold. Same goes for the Topping A90 ... i know they haven't separated the balanced and single ended circuits, but come on ....

 

[AnalogSteph]

IMO any other USB-powered DAC would give pretty much exactly the same problems as the D10s here. Unbalanced analog audio connections between IEC Class I devices such as a PC and a Topping A90 are just fundamentally broken. The USB ground return and PC's power supply ground return invariably become part of the signal path - a classic ground loop issue. The DX7 Pro would make it less obvious because it has a PE connection by itself, but I bet you would still have issues in some circumstances.

You could use the D10s with a traditional floating hi-fi amplifier or headphone amplifier (e.g. a trusty JDS Labs Atom) just fine. Likewise, the A90's RCA inputs would work fine with a phone or DAP on battery, or a traditional CD player or a number of phonopres with attached turntable. And yes, that includes a phone + D10s. You just can't combine the two without additional galvanic isolation.

 

[Noodles]

That's exactly my point - it's an engineering decision not to include galvanic isolation or just filters/chokes to lessen the noise. It is quite an easy fix - an implementation of something like a 10 year old ADUM4160 for a manufacturer will cost no more than 10-20 bucks, but they choose not to do so because the bill for the components needs to be pretty tight.
I have no issues paying the extra 20-50 on top of what it will cost instead now i have to solve this problem for myself.
Yes, i get it - the D10s is not exactly high end... but still.

 

[AnalogSteph]

The D10s is a mere $100 USB DAC, with a metal case and fancy display no less. I have yet to see any in this price class with galvanic isolation included... in fact, you can spend more on a USB - SPDIF converter with it. It seems the cheapest DACs with galvanic isolation included are around $300 or so.

I would be surprised if the BOM on a $100 DAC were more than $20-30 max. Subtract the case and display part, and you might have half of that left (if that). Not very much wiggle room at all.

 

[jeffbook]

After reading through this thread I thought I would post my experience with my Topping D10. I use this as a relatively inexpensive USB bridge from my Acer laptop to two miniDSP 2x4HD units that serve as active crossovers and EQ devices for a Linkwitz LXmini system. The DAC function with the unbalanced RCA outputs was considered a bonus feature. Over time I have done some measurements with a 1 kHz sine wave signal using ARTA and a Scarlett 2i2 3rd generation interface. The D10 was fed by the USB output from my Acer laptop. There were 3 different configurations measured.

Number 1 was simply the analog output of the D10. The output is "dirty" to say the least.
Number 2 is the analog output from the miniDSP 2x4HD. The input to the miniDSP was TOSlink from the D10. Very much a cleaner signal.
Number 3 is, once again, the analog output from the D10. However, in this test a Jensen ISO-MAX CI-2RR transformer was inserted between the D10 output and the Scarlett input. The noise is eliminated except for a 2 dB hump at the 60 Hz line voltage frequency.

So the I can eliminate the noise from D10 analog output through the use of a device that is nearly twice the cost of the D10 itself. Luckily I had the Jensen transformer from prior use that was used to break a ground loop that existed in an earlier system configuration
.