Filtering out USB too...I think I'll use a CLC-RC filter.What would be the best values?
I plan to use
22uF - > 100nF - > 10uH - > 100nF -> 1R - > 100uF
Some advice on that?I think every PC outputs different noise spectrum.but still,I'll be grateful
1R * 1A = 1V
Stick with L's
So...here comes the report
First thing to notice is that my pc won't give me 500 mA out of USB.No-no,not even with dual cable,that evil scamp gives up to 350 mA.The device is recognized as high-power(500 mA).
Initially,I made a dc-dc converter board with 7x09 post-regs.The result was that USB dropped to miserable 3.4 V.XMOS,power supply and the DAC still worked,though regulators were useless as output dropped to +- 7 V.The output was noiseless,but the sound was edgy and grainy - typical "dirty" power supply.
I connected the PCM1794 directly to USB too - it works even with 3.5 V.However,it sounds no good.
To save power I removed the regs,and left the dc-dc converter alone against the whole board.The USB power input is 22 uH/2A coil followed by 100nF and 1000uF.After the convertor we have 10uF on each rail,1 ohm series resistor,and finally 470 uF caps.Each chip has a pair of 100nF also.The DAC is also supplied by the positive rail with 7805.I can't hear any noise,and is sounds good now.
USB still drops to 3.7 V,and therefore the converter barely supports +- 7 V.I should have about 420 mA total current at 5V.So the PC does not meet specifications!
Question:Now,with the dc converter,DAC power is essentialy isolated from PC power.Is it worth to separate the grounds also?They are connected now.
First thing to notice is that my pc won't give me 500 mA out of USB.No-no,not even with dual cable,that evil scamp gives up to 350 mA.The device is recognized as high-power(500 mA).
Initially,I made a dc-dc converter board with 7x09 post-regs.The result was that USB dropped to miserable 3.4 V.XMOS,power supply and the DAC still worked,though regulators were useless as output dropped to +- 7 V.The output was noiseless,but the sound was edgy and grainy - typical "dirty" power supply.
I connected the PCM1794 directly to USB too - it works even with 3.5 V.However,it sounds no good.
To save power I removed the regs,and left the dc-dc converter alone against the whole board.The USB power input is 22 uH/2A coil followed by 100nF and 1000uF.After the convertor we have 10uF on each rail,1 ohm series resistor,and finally 470 uF caps.Each chip has a pair of 100nF also.The DAC is also supplied by the positive rail with 7805.I can't hear any noise,and is sounds good now.
USB still drops to 3.7 V,and therefore the converter barely supports +- 7 V.I should have about 420 mA total current at 5V.So the PC does not meet specifications!
Question:Now,with the dc converter,DAC power is essentialy isolated from PC power.Is it worth to separate the grounds also?They are connected now.
Almost finished.I'm wondering where should I put an LED indicator.So compact and sounds great!Thanks to everyone!
http://s4.postimg.org/5kaqsodtp/IMG_0774.jpg
An externally hosted image should be here but it was not working when we last tested it.
http://s4.postimg.org/5kaqsodtp/IMG_0774.jpg
Nailed it!
I had some problems measuring the device's performance though.Connecting directly to Audiophile 2496 input reveals noise spread over the entire spectrum at about -95dB,which is not very impressive.Combining the FFT,that makes only 75 dB SNR!But the device sounds 10 times better than the M-Audio.I'm hearing details never heard before,and another thing - when I connect my headphone amp to the Audiophile and crank it to max volume,I can hear some very slight hiss.With this DAC,there is absolutely no noise,completely dark image.The AP2496 has about 100dB SNR,so the DAC should be doing better then!So I connected the USB input to my laptop,and the analog output to the AP2496,which is in my PC.The noise floor was now lying at -120dB,which is the AP2496 baseline(so it may be even lower).in conclusion,I think that the M-audio uses some sort of vitual ground or something,and does not like to connect to USB ground.Separating DAC ground from USB ground makes things even worse,but I think the problem has something to to with the AP2496,not the DAC.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
I had some problems measuring the device's performance though.Connecting directly to Audiophile 2496 input reveals noise spread over the entire spectrum at about -95dB,which is not very impressive.Combining the FFT,that makes only 75 dB SNR!But the device sounds 10 times better than the M-Audio.I'm hearing details never heard before,and another thing - when I connect my headphone amp to the Audiophile and crank it to max volume,I can hear some very slight hiss.With this DAC,there is absolutely no noise,completely dark image.The AP2496 has about 100dB SNR,so the DAC should be doing better then!So I connected the USB input to my laptop,and the analog output to the AP2496,which is in my PC.The noise floor was now lying at -120dB,which is the AP2496 baseline(so it may be even lower).in conclusion,I think that the M-audio uses some sort of vitual ground or something,and does not like to connect to USB ground.Separating DAC ground from USB ground makes things even worse,but I think the problem has something to to with the AP2496,not the DAC.
Incredible...the huge voltdrop on USB was caused by the cable...I changed it with some old,ugly and thick USB from a laser printer,and magically regained the power lost -now USB is at 4,78V!With the crappy cable it was at 3.7V.How do the manage to produce a cable with 1V drop at 500mA - I don't know.
Here's a plot for the curious :
An externally hosted image should be here but it was not working when we last tested it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.