NAD PP3 via USB - Loud noise in Win7 but not in XP

[Paul Sanders (AlpineSoft)]

OK, thanks.  Seems like a lot of money for such a simple thing.

I'm, not sure why NAD elected to bus-power their USB chip.  Seems to me it could be powered from the PP-3's own power supply.  Next time I talk to them, I will ask.

[corboid]

Isolating analogue and digital stages is normally a very GOOD idea because it minimizes risk for getting the digital noise into the analogue path.

The problem is that USB allows very unclean voltage because is meant for powering digital computer peripherals.  Using such a voltage quality to power ADC is a BAD idea.
Thus they should do some cleansing of the voltage - capacitors and voltage regulators etc, *or* use an own power supply of known quality (like Aqvox does).

[corboid]

I just fetched the replacement unit. This time it has S/N H9XPP3Gxxxx, ie made 2009/October.

And now to the anti-climax.
The noise is still there, and the news is that when USB is connected to the HP mini-PC the noise also gets to the amp+speaker. This as I know I did not get earlier with my previous box (S/N H7ZPP3Gxxxx (2007/december)).

Luckily I cannot hear the noise via amp+speakers when connecting to Lenovo - it comes just a few seconds when the USB-cable is plugged in and then it becomes silent.

So to sum up - the PP3-design is still not suitable for use with computers.
To fail to fix a "known problem" is not really a trade mark of a master. I did not expect this from NAD and their QA. Apparently someone forgot to test the product on a range of computers.

So it's time to go for the plan B and replace  PC's voltage feed to the PP3 by few parallel connected  4.5V batteries (4.5V is almost within the allowed voltage range for USB) and if it works, use a stabilized and low-pass filtered power supply. A DIY-version of Aqvox unit. Must get an oscilloscope to take a look at the signals.

[Paul Sanders (AlpineSoft)]

OK, good luck.  I'm a bit disappointed, I have to say, although I think your case is a bit unusual.  There are many thousand happy PP-3 owners out there.

4 x NiMh = 5V, more or less.  It probably doesn't matter if you creep a fraction over.  Anyway, please keep me posted.  I don't have a scope either, otherwise I would take a look myself, but my PP-3 is a pre-production unit so I might not learn much anyway.

[corboid]

I just soldered a experiment board with USB-sockets so I can manipulate the USB/VCC-voltage to PP3.

The first test was to add some capacitors between VCC and GRD (pin 1 and 4) - I don't want to overload USB so I used modest 100nF parallel with 470pF hoping to suppress some of the disturbance frequencies. But it did not help. Now I try to replace the PC's USB-voltage by injecting some other stable 5V voltage.


Another finding - I also compared with my Behringer UCA200 ADC/DAC that only supports the line level signals (I use it for interfacing an audio-mixer to computer over USB).
And this is SILENT, as expected even with my noisiest USB-socket, when I connected UCA200 to the PP3's output.
On the other side when PP3 is switched to line level the noise level is 40-50dB lower (over USB).
Maybe it's not the ripple in the USB voltage feed to the ADC is the main problem (as I assumed)?

I.e. apparently NAD screwed up seriously the phono-level when interfacing the ADC... Doh!...

[Paul Sanders (AlpineSoft)]

Thanks for the update.  I think the problem is that the high sensitivity needed for a phono input (just a few mV) makes it particularly sensitive to noise pickup.  I shall be interested to hear how the PP-3 performs with a clean USB power source.

[corboid]

Bad news.
I made a setup where I replaced the USB-VCC/5V voltage coming from the computer by a VERY clean voltage from an AC/DC-unit and unfortunately  the noise is still there on quite same level and spectrum.
I used a 5V power supply specified for medical usage - Mascot 9619/5V. It has low ripple < 30 mV p-p and switching freq of 40kHz.

My setup was:

[Computer-USB]----<standard USB A/B cable>--- [injection device with fresh 5V]---<standard USB A/B cable>---[PP3]
Nothing is earthed, and I run my PC (HP mini-PC 210-1023) on battery.

I do not like that the 1st cable still contains the dirty voltage that can contaminate the other three USB wires, but I do not think that it is the main problem.
(It seems all the USB-wires are already contaminated when leaving the PC - unfortunately difficult to prove w/o an oscilloscope).
Anyhow I will make a custom cable that cuts of computer's VCC at the socket and injects the fresh 5V at the PP3-socket. But I have low believe in this.

I also tried to regenerate the digital signals D+/D- by passing a hub (powered by the Mascot device). This makes similar or same noise.

Apparently the NAD PP3 is designed in such a way that the USB-signals are received in such a way that they are are radiated internally inside the PP3-box to the sensitive MM/MC-phono stage. The latter has sensitivity of 5mV(MM) or 380 uV(MC). When 16-bit resolution is used it means the phono stage needs to receive radiation of  just 76 uV (or 5,7 nV) to be affected by a single digital level.

My conclusion: This means that PP3 requires that the connected computer MUST provide an excellent internal USB-voltage generation without ripple on any of the USB-wires.
Thus PP3 works satisfactory with the expensive Lenovo T60p but not with the cheap HP mini PC. I will soon get a Lenovo T510 and I am excited to see which noise I'll get. The risk is that newer PC's use more and more the solutions integrated into the main chip-sets and will be equally "crappy" when used for analogue sound

The solution I advice is the one I described earlier:
- do not connect USB-cable to PP3 unless you have noise-free USB-signals (i.e. use PP3 as a PP2, or don't buy PP3 - buy PP2!)
- use another USB A/D-converter to convert line level signals from PP2/PP3 (or your amp's monitor/preamp-out signal) to USB-signals

[Paul Sanders (AlpineSoft)]

OK, thanks for the update.  As you say, a bit disappointing.

It's not completely clear to me whether the noise pickup comes from the PC power supply (which contains various switched-mode components of course) or the USB signal themselves, but the fact that it varies from one PC to another seems to imply that it is the former rather than the latter.  It did occur to me that noise might arise from crosstalk within the USB cable.  The obvious solution there is to inject your clean supply at the PC end of the cable rather than at the PP-3 end, or, as you say, cut the power supply cable at the PC end.

Once your investigations are complete I will send a link to this thread to NAD to see if they would like to comment.

[corboid]

A home made USB does not help (as I feared). The pull-down resistors between D+/D- and Vcc and Ground leak the PC noise to PP3 over D+/D-.

The best solution would be if NAD provided an isolator (or better, built it into the PP3).  It would make PP3 a real usable audio device and not just a half heartily designed device.
Some hint's how it should be done for real:
http://www.ecnmag.com/Articles/2010/03/Simplify-USB-Isolation-in-Medical-Applications/
http://pdfserv.maxim-ic.com/en/an/AN3891.pdf
http://hackaday.com/2009/09/18/usb-isolation/

We are talking about quite a low component cost by < EUR 10 in production volumes i.e. would make a "super-PP3" only 5-10% more expensive if addressed during design.

[Paul Sanders (AlpineSoft)]

OK, I will make NAD aware of this thread.  Thanks very much for your efforts.

One thing though: any phono preamp using the USB chip that the PP-3 uses (which is most of them) is likely to have the same problems.  I'm sure NAD have the shortest possible connections between that chip and the USB socket, and have kept them as far away from the phono stage as they can.

[corboid]

I found another link that addresses power noise propagation (EMI) and design issues/advices for USB from Intel:
http://www.ti.com/sc/docs/apps/msp/intrface/usb/emitest.pdf

(but they do not discuss that USB stage might act as an antenna sending noise and disturbing the low level phono statges)

[Paul Sanders (AlpineSoft)]

OK, thanks.  I'll pass that along too.

[oilburner]

after bouncing a few ideas off Paul I have come to the following conclusions.

To cut a long story short:

the PP3 has always got a very quiet whine when the USB input is active. This is most noticeable on the MM input because this has the most gain.

If I disconnect my cartridge (though leaving it connected doesn't actually make any difference) and connect the arm to the PP3 I get the whine at unacceptable levels.

If I disconnect the earth lead from the pick up arm and earth the arm via the screen on the audio cable instead, the noise goes away.

The conclusion is that the 9 inches of unscreened wire within the arm is causing the problem. Whether this is because it is acting as an aerial or because there is some interaction between the mains earth on the tube and the wires I don't know. All I can say definitively is that changing the earthing stops the noise.

So, unless I have an old batch PP3 (apparently they did a redesign at some point to stop a noise issue?), or a faulty unit, the suspicion is that everyone who has a 'correctly' earthed turntable will get this problem.

Michael.

[Paul Sanders (AlpineSoft)]

Thanks very much for this.  I will give NAD another prod.