Quick Measurement: Creative Labs Soundblaster 5.1 X-Fi USB interface SB1090

[pkane]

Through an 'accidental' purchase, I wound up with a used Soundblaster X-Fi, surround sound external card/interface. This has a 5.1 channel DAC with up to 24/96k, and a 24-bit 48K/96K two channel ADC. Model # SB1090. I don't think these are still manufactured, but are available on the used market and are really inexpensive ($25 is what I paid for this one). I believe Amir measured the replacement version of this, model SB1240.

While I have it, decided to run some quick measurements. This is on the DAC portion, need to make up new connector to test the ADC.

1kHz, 24/96k, 0dBFS @ 1.1v:

EDIT: Modified the above test by powering X-Fi from a battery to eliminate ground loops and reducing level to produce a much cleaner looking result (-13dBFS @ 555mV):

Frequency response and phase:

Impulse response (updated after redoing the capture with a battery power supply to isolate grounds):

Reconstruction filter, 48k sampling rate (updated after running the test using a battery to isolate grounds):

And at 96kHz sampling rate:

Distortion vs level:

Distortion vs frequency:

Adding jitter J-Test at 44.1kHz:

 

[bigx5murf]

I'm using one of these in my bedroom HTPC setup. I found mine at goodwill for $5. I'm surprised it performs so well. Mine sometimes has driver gremlins and requires a restart before it works properly.

 

[pkane]

I'm using one of these in my bedroom HTPC setup. I found mine at goodwill for $5. I'm surprised it performs so well. Mine sometimes has driver gremlins and requires a restart before it works properly.

I was impressed with what this little box could do, considering the number of channels and the built-in ADC. The headphone output, I understand, is a bit underpowered (didn’t measure it).

I didn’t have any driver issues, but then I didn’t need a driver — I was using this with a Mac. Worked without a hitch as soon as I plugged it in.

 

[mansr]

1kHz, 24/96k:

Looks like slight clipping. What signal level was that, and what happens if you lower it a little?

 

[pma]

Looks definitely like measurement error. I get this from (slightly modified) Creative USB X-Fi HD.

 

[pkane]

Looks like slight clipping. What signal level was that, and what happens if you lower it a little?

yes, strange. I adjusted the level up and down a few dB but didn’t find any improvement. Now that you made me look more closely, THD rises right after -12dBFS, so I can try at around that point.

 

[pkane]

Looks definitely like measurement error. I get this from (slightly modified) Creative USB X-Fi HD.

View attachment 47622

The measurement shows a rise in THD right after -12dBFS. Let me try to see what I can get below that point.

 

[pkane]

Well, lowering the signal level didn't seem to help a whole lot, although it did change some of the relative harmonic levels. Here's the comparison between -15dBFS @ 200mV (black) and 0dBFS @ 1.1v (red):

I'll see if I can find a way to better isolate grounds, may be by powering the X-Fi off of a battery through a split USB cable.

 

[pkane]

Looks definitely like measurement error. I get this from (slightly modified) Creative USB X-Fi HD.

View attachment 47622

With battery power, the results are much cleaner when level is below -12dbFS (here at -13dBFS/250mV):

Here's -1dBFS @ 1v:

And -6dBFS @ 555mv:

 

[Krunok]

Impulse response:
View attachment 47564

Just for curiosity can you please show step response derived from this impulse?

 

[pkane]

Just for curiosity can you please show step response derived from this impulse?

Out of curiosity, what does this tell you?

 

[Krunok]

Out of curiosity, what does this tell you?

View attachment 47642

They both look weird because of these numerous up's and down's.

Here is for example IR of Topping D10:

I remember seeing the same "shaky" IR when my friend measured his speakers. How comes it is like that?

 

[pkane]

They both look weird because of these numerous up's and down's.

Here is for example IR of Topping D10:

View attachment 47643

I remember seeing the same "shaky" IR when my friend measured his speakers. How comes it is like that?

I suspect this is due to a fairly shallow filter with a lot of stopband ripple, but I could be wrong

 

[Krunok]

I suspect this is due to a fairly shallow filter with a lot of stopband ripple, but I could be wrong

I'm sorry but I don't understand: what does "shallow" mean in context of filter and what is "stopband" ripple?

 

[pkane]

I'm sorry but I don't understand: what does "shallow" mean in context of filter and what is "stopband" ripple?

Shallow = the difference between pass-band and stop-band. In this case, you can see it's only about 50dB. And you can see very large ripples in the stop-band, after the filter cut-off frequency:

For comparison, here's a filter I use in DeltaWave for low-pass filtering:

 

[Krunok]

Shallow = the difference between pass-band and stop-band. In this case, you can see it's only about 50dB. And you can see very large ripples in the stop-band, after the filter cut-off frequency:

For comparison, here's a filter I use in DeltaWave for low-pass filtering:

View attachment 47649

Ok, I see, thank you!

I am not sure that is the reason as what I saw with my friend's speakers IR was a room EQ lodspeaker filter so it probably didn't have any stopband at all.

 

[mansr]

Impulse response:

What did you use to record that impulse response? We're likely seeing the combination of the DAC and ADC responses, which can easily get weird. You might want to check it with an oscilloscope.

 

[pkane]

What did you use to record that impulse response? We're likely seeing the combination of the DAC and ADC responses, which can easily get weird. You might want to check it with an oscilloscope.

Here's the impulse response of the DAC/ADC loopback of my 'measurement' device (Apogee Element24):

 

[Krunok]

Here's the impulse response of the DAC/ADC loopback of my 'measurement' device (Apogee Element24):

View attachment 47658

That one looks ok.

 

[mansr]

Here's the impulse response of the DAC/ADC loopback of my 'measurement' device (Apogee Element24):

View attachment 47658

Fine, but what sample rates did you use for the impulse and the ADC? You need to record at a much higher rate to avoid seeing a combined response.