USB-powered amp/ps
Jan 8, 2007 at 12:32 AM Thread Starter Post #1 of 29

Yen

100+ Head-Fier
Joined
Apr 18, 2006
Posts
461
Likes
10
Hi!

I am currently using a battery powered cmoy with my laptop and wondering could I make an USB PS for cmoy. Walwart is unconvenient and rechargable 9Vs (and the charger) are rather expensive and I though: having a Aliendac running from USB -> why cant you power the amp from usb also!

Is there reasons why it is stupid idea? USB is rated at 5V, which is quite low, but if one 9V (+/-4.5V) is enough for cmoy then why not?

Is usb unstable or the current output not enough? Noise and distortions? Somewhere even a usb-ps design?
smily_headphones1.gif
Or can I take the power from Aliendac?

Or should I simply go with hpdac?
evil_smiley.gif
Apparently there are no boards left... but something similar would be perfect for my laptop use.
 
Jan 8, 2007 at 1:17 AM Post #2 of 29
the hpdac uses a voltage booster to increase the voltage. I would go with it, speak with dsatvisk to see if he has some boards left or if he is planing to make more.

Manuel
 
Jan 8, 2007 at 1:48 AM Post #3 of 29
Hi,

I think now powering amps from computer isn't so bad. (Although there is a big contraversy going on building amp INSIDE the computer.)

I would try to build an amp using chips that are designed to work at that voltage. I would use 74HCU04 and operate it in analogue mode. This makes a neat amp that operates in ... Class-A.

Only problem is that 74HCU04 amp may have RF-Pickup depending on your building skill. I don't have much pickup but operating mine near vacuum cleaners or refrigerators can noise up my 74HCU04. ... Computers don't do this, thank gawd.

I think there are 5V rail-to-rail opamps too. I have never used them ...

Tomo
 
Jan 8, 2007 at 3:16 AM Post #4 of 29
Quote:

Is there reasons why it is stupid idea? USB is rated at 5V, which is quite low, but if one 9V (+/-4.5V) is enough for cmoy then why not?


Because the minimum voltage of operation for most of the normally "usable" audio op-amps is +/- 2.5 volts which puts you right at the margins and if you add any supply decoupling will fall below the minimum voltage required for operation.

Part ll of the potential for a problem is voltage drive.The less voltage you have at the supply rails the less overall voltage drive you will have at the output which depending on your cans may or may not be a problem.if low sensitivity/med-high impedance headphones you would likely find the headphone amp did not better than striaght out of the headphone jack of the source.

Options that work ?

Special dedicated headphone chips designed to work on a low power buss,using a DC-DC converter to bounce the voltage up from the single +5 volts to a respectable +/- 12 volts (but means proper filtering at the opamp pins or you will have a crap sounding amp) and /or CMOS Op-amps,most of which suck in SQ though I have had some amount of success using the AD8655/AD8656 in portable DAC/HPA and Mic Preamp/ADC designs.

Usable with either a single polarity +5VDC or a split supply of +/- 3VDC,low noise,low offset voltage,unity gain stable,operates rail to rail,exellant slew rate and settling time,not too shabby for use in a tough voltage environ.Oh yeah,gets +/- 220mA output drive current with +5V reducing the need for a buffer following the op amp to zero.

not perfect and maybe not THE best in SQ but one of the few single part options for really low voltage use

http://www.analog.com/UploadedFiles/...D8655_8656.pdf
 
Jan 8, 2007 at 4:10 AM Post #6 of 29
Quote:

USB will supply 500ma enough for a DC/DC converter. Just be sure to filter and bypass properly and maybe even sheild the converter separately from the op amp



but that is the trick here that is hard to pull off.Current delivery from the USB port never has been a problem.Clean voltage is and in the case of analog circuits enough of it to gain usable voltage output with dynamic headroom.Mate a DC converter to an Op-amp and 9 out of ten times the sound of the converter will dominate the sonics and the higher the speed/bandwidth of the op amp even more so placing all the emphasis on any filtering and that would be entirely chip dependant so toss op amp rolling right out the window.

Can DC conveters sound good ? With bandwidth limiting and "slow" analog audio devices yes but no way with high speed devices when compared head-to-head to the very same part fed from a linear power supply and that is the meaure of SQ.

Sometimes it is better to use a single comprimised part and accept the warts(low voltage CMOS Op-amps for instance) than to spend extra loot creating an even more comprimised overly complicated circuit that may in fact never sound good unless you have a lot of experience in that area and even then a strong "maybe"........

At least that has been my experience
cool.gif
 
Jan 8, 2007 at 4:50 AM Post #7 of 29
Couldn't you do like some USB harddrives do and use 2 USB ports? Then you would have +/-5V, correct?
 
Jan 8, 2007 at 5:23 AM Post #8 of 29
Quote:

Couldn't you do like some USB harddrives do and use 2 USB ports? Then you would have +/-5V, correct?


no way to get a +/- voltage of the same value as a sinlge voltage so since a USB biss is set at +5 volts the only way to get there is a DC-DC converter or a pair of linear low noise LDOs running at a slightly lower voltage because even the best of them still have a "drop out" even if in the milivolt range.

This leaves the choices still being :

1-DC to DC converter for a dual polarity higher voltage
2-A pair of LDO regulators running of the same +5VDC feed
3-Use an Op-amp specced for low voltage operation
4-Combination of #1 and #2 : DC-DC to bump up into the +/-12 VDC range for good voltage drive followed by the low pass filter then on to the LDO for the final voltage (helps clean up the signal even more) then a final shunt regulator stage for a stable output impedance that does not fluctuate with the dynamics of the music hence noise levels will remain a constant-not for the meek-

That last is likely the only way to real power over and above what is already available on a portable player output with a true medium impedance headphone while a split LDO supply feeding a chip that can operate down in the +/- 2.5 volt range (like the OP2134,some buffers for instance) more than good enough for any low impedance headphones or IEMs as long as the supply is designed with audio in mind as the end use.

No free rides on these things unless just an excuse to do it.What works for "non-audio" work WILL give you the voltages you seek but in the end will be a waste of time unless you design for SOUND which places a whole new set of rules on the final design,all having to do with power supply performance in the audio band and at any frequencies within the range of the actual audio part.

The power supply is not an afterthought just to get the circuit to operate but part of the signal path and needs to be as good or even better than the audio circuitry quality or the entire SQ suffers whatever shortcomings the powering source has.Not trying to dissuade anyone from experiemnting just trying to point out how much of a b*tch it can be if you want to do it right
cool.gif
 
Jan 8, 2007 at 3:10 PM Post #9 of 29
Thanks every1 for the help, apparently bad/complicated idea.

I'll have to aproach dsatvisk if he has any of the hpdac boards left.
 
Jan 8, 2007 at 3:15 PM Post #10 of 29
How about a battery powered amp that can be recharged from a USB port? You can use a DC-DC step up to recharge a 9 to 24 volt pack.

Turn the charger off when you are listening to music.
 
Jan 8, 2007 at 4:19 PM Post #12 of 29
Quote:

Originally Posted by Yen /img/forum/go_quote.gif
I'll have to aproach dsatvisk if he has any of the hpdac boards left.


It's dsavitsk

Unfortunately, there are no more boards for now. But, if you are going to do this on breadboard, I'd look at either DCP020509 or DCP020507. They are DIP sized DC-DC converters, and can be used in series to produce +/- supplies. The datasheets should help you get started.

Otherwise, the schematic from the hpdac is online so you can copy the dc-dc converter from there. PCB layout is important to keeping it stable, however.
 
Jan 8, 2007 at 4:58 PM Post #13 of 29
Quote:

Thanks every1 for the help, apparently bad/complicated idea.


not a bad idea at all,a good thread.

Complicated ? Can be if you are trying to get beyong just powering the device from the buss to creating something that sounds as good as a non USB-buss powered amp.

You can get there by using a "slow" op-amp with a DC-DC converter but that converter needs to be good at audio frequencies,needs proper decoupling plus would be best served by having a shunt reg following it or you can use one of the new single supply low power op-amps,easily sounding as good as a better chip done with an improper power source,or if all you need is current drive for low impedance cans one of the drivers that operate down in the low voltage region.

comes down to what you need to drive and then you look at what voltages you have available.If a single +5VDC meant for driving voltage hungry headphones you are limiting yourself if you try to split it into two voltages unless you increase the voltage so it is better to go with a singe polarity op-amp or the DC converter.You may get volume but no way you get any headroom for peaks so you kill the dynamics or you use the "trick" of Fake Dynamics which is an improperly filtered DC-DC converter tossing artifacts into the sound that when the amp runs out out headroom sounds like upper octave detail hence the appearance the music is "exciting" so considered dynamic.

Quote:

i mean, headroom bitheads are USB powered amps AND dacs


DACs are another audio device that respond well to "feeding" but being already low voltage devices having their own internal single polarity CMOS Op-amp stages I sometimes wonder if it wouldn't just be better to tap the signal at the DAC output itself and couple it directly to the output jacks through a DC blocking cap and call it a day rather than screw around adding ANOTHER op-amp (op-amp to op-amp) unless you need the current drive and then why not shoot right to a low power buffer ? If you need more voltage then yes,you are stuck with it but if not then just another useless addition mucking things up.

I have zero clue what Headroom uses or how they go about it but would be very surprised if the buss powered portables sounded as good as the battery operated versions.

if there are other ways to get from A-to-B I don't think I know them and would be REAL intersted in someone explaining the "this is how you do it rick" part
wink.gif



Quote:

Otherwise, the schematic from the hpdac is online so you can copy the dc-dc converter from there. PCB layout is important to keeping it stable, however.


anything that operates in the Mhz range (digital,DC Converters,High Speed Analog) is a pain in the ass because you are no longer dealing with audio frequencies but RF so to get up to speed better reading material would be antenna theory and radio theory not audio primers and definately NOT the manufacturer data sheets which are mostly meant to tell a manufacturer using the parts how to best save a buck rather than how best to do it......
 
Jan 8, 2007 at 10:33 PM Post #14 of 29
Quote:

Originally Posted by rickcr42 /img/forum/go_quote.gif
...I have zero clue what Headroom uses or how they go about it but would be very surprised if the buss powered portables sounded as good as the battery operated versions.

if there are other ways to get from A-to-B I don't think I know them and would be REAL intersted in someone explaining the "this is how you do it rick" part
wink.gif
...



Not sure either but according to the headroom website, they say that using battery power is superior:

Quote:

...If, however, you are of an audiophile mind, turn the power switch on and the Total BitHead’s digital-to-analog converter (DAC) and power amp section will run off the internal batteries for ultra-clean performance and lower power drain on your computer...


 
Jan 9, 2007 at 3:45 AM Post #15 of 29
4AA batteries in the BitHead so regulating down to the 3.3 VDC of the USB DAC chip a cakewalk leaving the headphone amp powering method the question.Is it

A-one of the new generation of single supply CMOS Op-amps
B-A dedicated "HEADPHONE AMP" chip that are cropping up from Maxim,National,TI and ADI
C-Traditional op amp,one that can operate at low voltage levels such as the OPA2134 (specs to +/- 2.5 VDC) powered by a split supply with a sythesized ground
D-A "real" op-amp running at +/- 12VDC supplied by a DC-DC converter so it has proper voltage drive and meaning all the battery is doing for the "audiophile" performance is not having any supply glitches due to the computer power supply limitations.Not built for strictly audio use means any mains line abberations are usally part of the power stream to whatever device is on the motherboard,a thing that causes more freeze up and dropout than most realise...........................

My loot is on #3 if only because Headroom already has a track record in using DC converters in their amps which warts and all is a better solution than an underpowered amp or for strictly road use.

The Cosmic and the Supreme used a Cosell DC converter to get +/-12 VDC from a battery and were far better amps than the strictly portable "Airhead" amps even though using a straight battery power supply.

reason ?

Oyput voltage and power supply voltage are related so there was more pure voltage drive in tha Cosmic,an amp that was more "home system" related than it resembled any portable amp (big mamma hunk of portable amp !) there was a tiny little switch on the back that if you purchased the home docking station and connected it would 100% bypass the converter and power the amp from a "proper" +/-15VDC analog power supply.

The dock also added RCA line inputs that in combination with a switch bypassed the wimp Mini-TRS jack used for portable duty making the amp a proper domestic system amplifier.

With that history pedigree I'm guessing a DC-DC converter powers the analog stage...........................unless they are using a low power chip..............or.............wait.............
evil_smiley.gif
 

Users who are viewing this thread

Back
Top