USB DAC project

Status
Not open for further replies.
Jump to Latest
#21
very interesting , Ukram!!

I have read the datasheets of the 290X chips, but found no information about duplex operation, but it might be so .

In that case, these little chips could be úsed with Speakerworkshop and similar loudspeaker programs with very good results, 83 dB S/N is ok, much better than any crappy Soundblaster card. Plus you can use it with a laptop making portable!

/rickard
 
#22
I made RMAA loopback test with the PCM2902. The input has some noise, so i got 80 dB S/N with 44.1 kHz and 82 with 48 kHz.

I used the REG102.

Still fiddling with ground connections etc. to get the promised 89 dB...

RMAA results http://hepso.dna.fi/misc/pcm2902/PCM2902.htm

I have not tried to use the SPDIF, because i do not have anything to connect it to...

The device must work in full duplex, because i could use RMAA program to simultaneously play and record the test signal. USB transfers the data in packets, so it is actually half duplex, but the USB "full speed" bandwidth is enough to move the signal both ways...
 
#25
Hmmm... im gonna try using a PCM2902 spdif out to a DIR1703 to a PCM1730.
The result should be interesting.

BTW... anyone know if the SPDIF out on the 2902 does any resampling?
Im hoping its simple enough that it doesnt.
The PDF kinda lacks on these kinds of details 🙁
 
#26
PCM2902 document: http://www-s.ti.com/sc/ds/pcm2902.pdf

DAC can do 32, 44.1 or 48 kHz. As long as your sound sources use any of these sampling rates, there is no need for resampling.

Can somebody explain what SpAct means?

About bandwidth: I looked the USB data lines with scope while playing at 44.1 kHz. There was packets of data at 1 ms intervals. Packets lasted about 1/5 of ms, rest of the time the USB bus was idle. When recording at the same time, there was still more than 50% free capacity.

In the RMAA noise spectrum the 1 kHz packet rate can be seen as small noise spike (also at 2 kHz, 3kHz and so on).
 
#27
SpAct is used to clock a circuit, its used in the DIR1703 to recover the clock from a S/PDIF input.
Do u think the digital out could be fed straight to a DIr1703 for use with a digital amp (TDAA's)?
I'm guessing a similar thing was used for making a the SB Extigy.... 5.1 sound would be nice 🙂
 
#28
ukram said:
DAC can do 32, 44.1 or 48 kHz. As long as your sound sources use any of these sampling rates, there is no need for resampling.
Click to expand...

Soundcard DACs also support many sampling rates, but most still resample SPDIF out to 48Khz.
What you said above doesnt mean that the 2902 doesnt resample.
 
#31
ukram said:
What puzzles me about SpAct is the way it works with multiple 1 ms packets? Does it change the sample rate for each 1 ms packet as needed, or does it let the fifo buffer grow or shrink for a while before changing rate? I have not found any details about SpAct in TI web site.
Click to expand...

I don't think 1ms packets have anything to do with it. The PLL is spec'd to lock within 1ms, that is all.

The fifo does just what you say; look at the PLL filter response on the DIR1701 datasheet. It tracks for low-frequency changes and rejects high frequency ones.
 
#32
tiroth said:


I don't think 1ms packets have anything to do with it. The PLL is spec'd to lock within 1ms, that is all.

The fifo does just what you say; look at the PLL filter response on the DIR1701 datasheet. It tracks for low-frequency changes and rejects high frequency ones.
Click to expand...

PCM2902 gets samples from USB in packets that arrive with 1 ms intervals. One packet is sent in about 1/4 ms, so first change to lock into incoming sample rate is when the second packet arrives. However there is no guarantee that PC can send data in exactly 1 ms intervals, so it might be a good idea to adjust the sample rate rather slowly...
 
#34
ukram's USB-DAC

Just finished & tested!

Arne K
 

Attachments

  • usb-dacxx.jpg
    usb-dacxx.jpg
    51.4 KB · Views: 1,193
#35
Hello,

i have built the USB-DAC based on the Ukram project in order only to use it as an interface for USB to S/PDIF. Now that i have finished the board and it is not working :xeye: i have a few questions.

I get no signal out of my Parasound DAC when plugging in USB, what might be wrong? The supply voltages are there (4,8 V USB and 3,39 V from regulator) and i checked the whole board three times for mistakes.

Is it right that i can plug it in right between USB out and digital IN of my DAC?

Do i have to change something in my PC setup?

Thanks in advance

Frank
 
#39
An Ethernet ADC/DAC would be interesting, as USB does not allow dc isolation. The Ethernet transformers would isolate the audio interface nicely from the PC psu noise. It could be built using a Micrel Ethernet chip, microcontroller and small linear psu. Software driver could be a headache.
 
#40
A transformer does only galvanic separation, it has no way to determine which is noise and which is usefull signal.

An USB dac should have it's own psu with separate sections for digital and analogue parts to minimize interference. Not that I'm sure somebody could actually point out blindly a usb-powered from a standalone-powered.
 
Status
Not open for further replies.
Top Bottom