I want to build an USB soundcard with an headphone amplifier in it. The soundcard will be based on a PCM2702, the headamp will be based on Tangent's MINT. I need the following voltages:
+5V for analog part of PCM2702
+3.3V for digital part of PCM2702
+9V for positive rail of opamps
-9V for negative rail of opamps
I can get the +3.3V directly with a LDO regulator from the USB port, but I would need three separate DC-DC converters to get the other voltages. That would cost me a lot of board space, parts, errors and time.
Are there any low noise DC-DC converters that could generate +9V and -9V from USB 5V with a single chip?
Do you know any other good solutions to get these voltages? An external PSU is not a solution since it has to be portable.
+5V for analog part of PCM2702
+3.3V for digital part of PCM2702
+9V for positive rail of opamps
-9V for negative rail of opamps
I can get the +3.3V directly with a LDO regulator from the USB port, but I would need three separate DC-DC converters to get the other voltages. That would cost me a lot of board space, parts, errors and time.
Are there any low noise DC-DC converters that could generate +9V and -9V from USB 5V with a single chip?
Do you know any other good solutions to get these voltages? An external PSU is not a solution since it has to be portable.
you can save some time and effort by defining your load, 32 Ohm headphones require 2.5 V/80mA peak for 100 mW/ea drive, this just fits in 5 V with modern R_R output op amps - but 160 mA peak headphone current draw cuts into the current for the USB intf + DAC (actually taking RMS current, and using a big AC ground Cap cuts the averaged DC draw from the 5 V down 2.8x, but large filter caps may require smart inrush current control to meet USB current limits)
another option is to use sw-C converters, again with low Ohm headphones simply inverting the raw USB V gives plently of voltage swing - but now you only get 1/2 the current
slightly better systems-wise would be to double raw USB volts and regulate down to get clean 5 V for the DAC, use the doubled V for the op amps
another option is to use sw-C converters, again with low Ohm headphones simply inverting the raw USB V gives plently of voltage swing - but now you only get 1/2 the current
slightly better systems-wise would be to double raw USB volts and regulate down to get clean 5 V for the DAC, use the doubled V for the op amps
You can use a simple flyback transformer topology (and then use linear regulators if you want to reduce the switching noise). The only difficult part is getting hold of the small transformer. There are plenty of standard ones out there to give +/-5V & +/-15V volts for example, but most are custom made.
As for the driver chip, Maxim, LT and TI have plenty of chips that are suitable for a flyback topology.
/Magnus
As for the driver chip, Maxim, LT and TI have plenty of chips that are suitable for a flyback topology.
/Magnus
I want to build the MINT headamp, with at least 12V r-r voltage, because it sounds so good. I really need a negative voltage, because the PCM2702, the USB port and the headphone have the same ground. Altough it is separated.jcx said:
One solution I have is this:
Use a LDO regulator to get +3.3V
Use a step up converter to get +9V from 5V. Use a inverting converter to invert the +9V to -9V. Put a regulator between +9V and GND to get analog 5V.
This solution uses 2 converters and a lot of capacitors, resistors etc.. and I want to keep the opamp rail source separated from the PCM2702 source.
What is sw-C?
The problem is not to get the voltages, but to get them with as less as possible components.Swedish Chef said:
switched-capacitor converters don't use magnetics, just charge up a big cap and stack it on the rail
they are usually limited to 2x doubling and 1x inverting applications below ~ 100 mA but are pretty trouble free compared to inductor/xfmr converters
you really need to look at your current/power budget, you may not have enough current to run low z headphones from +/- 9 V
remember V*A out < V*A in for any converter, so going from ~ 5 V to +/- 9 requires 4x current load on the 5 V side if you can approach 100% efficency, at low powers breaking 80% is hard due to fixed losses
they are usually limited to 2x doubling and 1x inverting applications below ~ 100 mA but are pretty trouble free compared to inductor/xfmr converters
you really need to look at your current/power budget, you may not have enough current to run low z headphones from +/- 9 V
remember V*A out < V*A in for any converter, so going from ~ 5 V to +/- 9 requires 4x current load on the 5 V side if you can approach 100% efficency, at low powers breaking 80% is hard due to fixed losses
If you are to take signals from the USB to the DAC you would need som way of isolating the grounds between the USB and the DAC in all three suggested solutions, optocoulper or transformer.
This since you will have the USB ground as negative rail (with a regulator in series) in all three suggested solutions.
As i see it if you do not want to do that then you could use different topologies for positive and negative rail.
The cuk converter (cannot remember the other name for it) inverts the voltage. And for positive voltage there is the step up converter.
This since you will have the USB ground as negative rail (with a regulator in series) in all three suggested solutions.
As i see it if you do not want to do that then you could use different topologies for positive and negative rail.
The cuk converter (cannot remember the other name for it) inverts the voltage. And for positive voltage there is the step up converter.
Maxim has quite a few SMPS chips,controllers,etc..
www.maxim-ic.com I think. Something there should be useful..
Personally I would connect the input of the 3.3V regulator to the 'raw' 5V from the USB port,instead of the output of the 5V reg. Just so all the 3.3V stuff isn't drawing all it's power through the 5V reg aswell (which has it's own loads to worry about.)..But it may not matter much in this situation..Also,be aware that the USB port can't supply alot of current. 500ma max or something similar.
I was thinking about a couple 9V rechargeable batteries for the output amps..Maybe figure out a way to re/trickle charge them from the USB port. Maybe even just use one for the -9V supply,or something.
Hmm,if you used two,you cold run the whole thing from the batts,and use the USB to charge them when it's plugged in,also that way you won't have to worry about current restraints as much,because it will primarily be running off the batts. Then again,maybe ~500ma is enough..
www.maxim-ic.com I think. Something there should be useful..
Personally I would connect the input of the 3.3V regulator to the 'raw' 5V from the USB port,instead of the output of the 5V reg. Just so all the 3.3V stuff isn't drawing all it's power through the 5V reg aswell (which has it's own loads to worry about.)..But it may not matter much in this situation..Also,be aware that the USB port can't supply alot of current. 500ma max or something similar.
I was thinking about a couple 9V rechargeable batteries for the output amps..Maybe figure out a way to re/trickle charge them from the USB port. Maybe even just use one for the -9V supply,or something.
Hmm,if you used two,you cold run the whole thing from the batts,and use the USB to charge them when it's plugged in,also that way you won't have to worry about current restraints as much,because it will primarily be running off the batts. Then again,maybe ~500ma is enough..
How about this circuit: http://stiftsbogtrykkeriet.dk/~mcs/DC-DC_Con.gif
You can see my version here (it's the small vertical board):
Best regards,
Mikkel C. Simonsen
You can see my version here (it's the small vertical board):
Best regards,
Mikkel C. Simonsen
The easy solution -- use the LT3439 from Linear Technology -- it will work down to 2.8 V -- and is "ultra-quiet" (at least as far as switching chips go (200uV noise). The secondary of the transformer just has to accomodate the voltages you need. Put an EMI filter on the input to the LT3439 to keep the USB Port noise down to a dull roar.
The LT3439 isn't regulated, so you need post-regulation -- I have used the LT1761 and LT1964 low noise regulators. Linear makes a regulated low noise switcher too, the LT1533 and LT1683.
If you don't want to spend the $ for this chip, use a flyback boost converter to generate enough voltage to drive a sine-wave oscillator at 40kHz -- you are going to have to load the flyback circuit to keep it happy -- the inductor value is inversely proportionate to the current. the sinewave generator drives a transformer with the appropriate windings for the output you need -- you can experiment with a signal generator to make sure things are going right. Some switching noise will migrate through so linear post-regulators are a good idea.
GenRad used this scheme in some of their portable, battery operated audio gear --
The LT3439 isn't regulated, so you need post-regulation -- I have used the LT1761 and LT1964 low noise regulators. Linear makes a regulated low noise switcher too, the LT1533 and LT1683.
If you don't want to spend the $ for this chip, use a flyback boost converter to generate enough voltage to drive a sine-wave oscillator at 40kHz -- you are going to have to load the flyback circuit to keep it happy -- the inductor value is inversely proportionate to the current. the sinewave generator drives a transformer with the appropriate windings for the output you need -- you can experiment with a signal generator to make sure things are going right. Some switching noise will migrate through so linear post-regulators are a good idea.
GenRad used this scheme in some of their portable, battery operated audio gear --
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.