[HAL]

Had the opportunity to setup the CLIOwin 8.51 Standard measurement system on my Netbook to get ready for some speaker measurements.   Dave had asked me about trying to measure the noise floor from the BatteryBUSS so had an idea.  The CLIOwin system uses two USB bus cables to interface to the PC.  One is for the SC-02 I/O box and the other is for the M-Audio Transit USB soundcard.  The system has self calibration for accurate dBV measurements.  I decided to check the shorted input noisefloor of the system.  Here is a screen capture of the results:



The measurement is done with a 262144 point FFT with a 48KHz sample rate.  This gives a bin frequency resolution of 0.183Hz.  The Transit soundcard is 24bit/48KHz full duplex to be used with CLIOwin.  The data was averaged 100 times to stabilize the noisefloor.  There are artifacts in the 1KHz-5KHz range that are above the noise floor.  Source seems to be the 5VDC USB power supply.

I used the BatteryBUSS and a powered DISCREET USB cable that Dave built to isolate the USB Transit soundcard from the PC power supply.  Here is the same setup with just the one cable swapped out and BB running the soundcard:



The power supply artifacts have dropped 6dB.  I cannot change out the SC-02 USB cable as I only have the one external powered USB cable from Dave.

I tried it with another PC and got very similar results to the above. 

Along with this test, tried the same cable and BB in a friends system with an MS-1 and Musiland Monitor 01 Mini USB DAC.  This is an asynchronous USB device with native up to 24bit/192KHz DAC output powered by the USB.  We heard switching power supply artifacts from the system at our listening positions when the Musiland was powered directly from the MS-1.  When the DISCREET USB cable and BB were swapped, the noise floor of the system dropped below audibility very close to the drivers.

 Now time to build some speakers and make measurements!

[Speedskater]

I wonder what the spikes at 1 kHz are all about?

[dBe]

I wonder what the spikes at 1 kHz are all about?

I wonder about that, too.  Now that I'm back in town I'll be in touch with Rich and see if we can figure that out.  I'm wondering if those are anomolies related to his test set up.

I need to dust off my Virtins system and get some other measurements... this time from a Mac Mini to a DAC.  I'll have to borrow a standard Mini for AC testing since my gear is all battery powered.

Dave

[HAL]

Dave,
Since it has been consistent with multiple PC's with the same measurement suite, I can only speculate that it is switching artifacts on the USB power lines getting to the SC-02 and the Transit.  The USB bus connections are the only ones from CLIOwin to the PC's. 

The Transit sound card has a single clock driving both the A/D and D/A's for full duplex use.

All of the PC's have multiple internal switching regulators for the different voltages needed for the CPU, memory and peripherals.  Not sure how synchronus they are to not cause intermodulation products.

My only idea would be to try a second powered DISCREET USB cable to break the PC 5VDC supply to the SC-02 to see if the rest of the artifacts drop.  The SC-02 drives the A/D inputs on the Transit.   

[dBe]

Dave,

My only idea would be to try a second powered DISCREET USB cable to break the PC 5VDC supply to the SC-02 to see if the rest of the artifacts drop.  The SC-02 drives the A/D inputs on the Transit.   

  Then it wouldn't be "DISCREET"!   

I'm with you on the artifacts.  I'm wondering if it is a beat frequency between clocks.  Other than that I have no idea.

Dave