Ideas for circuit to clean USB power.

[pho_boi]

Hey everyone,

I'm about to build another USB DAC based on the PCM2702 and using this following design

The site states that it uses about 45ma, and can be powered by the USB 5v or from 2xAA supplying roughly 3v. In my previos builds, I built a dedicated power supply for it but this time I would like to build something portable and will run off the USB 5v supply.

I did a little reading and found that the USB supply isn't very clean at all, so I need to build some kind of circuit that will be able to clean it nicely. I was thinking of taking the 5v supply then regulate it to a nice 3v.

But I'm not sure how to.

[pho_boi]

ok well I think I found something that may work.....

I was thinking of using this regulator from TI: link

It has an input voltage range of 4v - 5.5v (thats what is says in the datasheet but differently on the site) and output of 3v. And noise is stated to be 40micro-volts.

So here is my idea, it's just basically replacing the battery source:

-Use the TPS79630 to regulate the USB 5v into a nice clean 3v.
-Then the 3v will be fed to the MAX1722 DC-DC converter which takes it upto 5.3v
-The 5.3v from the MAX1722 will then be fed into the REG102-3.3 and REG102-5 which will output the specified voltages for the PSM2702.

[jackinnj]

if you are considering the MAX1722 which is SOT23-5 you might want a simper solution use a Linear LT1763 AND a 22uH choke ahead of the regulator. The LT1763 is an ultra-low noise, low dropout linear regulator in Thin-SOT package and will handle 500ma. My own experience with the LT1763 and the companion LT1964 negative regulator has been good.

Maxim states that the MAX1722 is "low noise" but they don't characterize the noise in their PDF. They do explain the "damping switch which dumps the ringing current, but I would appreciate seeing a graph of the noise spectra. Also you are putting one switching regulator in series with a prior set of switching regulator induced problems.

[pho_boi]

Originally posted by jackinnj
if you are considering the MAX1722 which is SOT23-5 you might want a simper solution use a Linear LT1763 AND a 22uH choke ahead of the regulator. The LT1763 is an ultra-low noise, low dropout linear regulator in Thin-SOT package and will handle 500ma. My own experience with the LT1763 and the companion LT1964 negative regulator has been good.

Maxim states that the MAX1722 is "low noise" but they don't characterize the noise in their PDF. They do explain the "damping switch which dumps the ringing current, but I would appreciate seeing a graph of the noise spectra. Also you are putting one switching regulator in series with a prior set of switching regulator induced problems.

The reason I am using the MAX1722 is because the PCB has the pads already for it.

[jackinnj]

the TI TPS7960X is SOT223-6, not SO8 btw

[pho_boi]

Originally posted by jackinnj
the TI TPS7960X is SOT223-6, not SO8 btw

That should be fine, because I will be building the USB regulator exteral to the DAC baord

[tiroth]

Listen to jackinnj, you definitely need a nice choke in there to keep all the noise out.

[00940]

you could measure first your usb voltage output. With some luck, on such light loads, it could be up to 5.2V. Which would let enough voltage drop to just feed the usb power to a choke then directly to the REG102, which is a fairly good beast.

Another option is to use a 4.75V regulator, fed from the USB supply, with a choke first. Then don't use the reg102 and connect the 4.75V to the 5V pad on the board.

The pcm2702 should be fine with a 4.75V analog supply (supply must be between 4.5 and 5.5)

[pho_boi]

Originally posted by 00940
you could measure first your usb voltage output. With some luck, on such light loads, it could be up to 5.2V. Which would let enough voltage drop to just feed the usb power to a choke then directly to the REG102, which is a fairly good beast.

Another option is to use a 4.75V regulator, fed from the USB supply, with a choke first. Then don't use the reg102 and connect the 4.75V to the 5V pad on the board.

The pcm2702 should be fine with a 4.75V analog supply (supply must be between 4.5 and 5.5)

Guzzler has also told me a similar thing, the DAC board already has a 10uH radial inductor in an RC network configuration. But the noise level is still very detectable.

The following graph compares the noise level from using the USB power and batteries, it shows that their is still a 10db difference which is very audible.