[MWP]

Greetings all,

Im designing a USB powered DAC to replace the very ****** (pfft, sh1tty isnt that bad) soundcard/amp in my Laptop.

Itll be somthing like USB -> PCM2707 -> a DAC (undecided atm, maybe PCM1781) -> OPA627 LPF -> Volume control pot -> Headphone driver amp.

Power will be via a 5V to 15V stepup converter (not the best option i know, but i am aiming at USB powered and small size).

Now my question....

Can i use OPA627's for my headphone output amps?
I have quite a few here, so paralleling them up is an option.
Ill be driving 32ohm Beyer's.

Thanks for any help...

 

[Jazper]

You'd be better off paring them up, mint/pimeta style, with buffers imho

 

[MWP]

Yeh, i know, ive only got one BUF634, and cant get another, so im looking for alternatives.

How much o/p current to i need to drive 32ohm phones comfortbly?
OPA627's will o/p up to 45mA.

 

[Jazper]

Quote:

Originally Posted by MWP Yeh, i know, ive only got one BUF634, and cant get another, so im looking for alternatives. How much o/p current to i need to drive 32ohm phones comfortbly? OPA627's will o/p up to 45mA.

I'd want about double that

you could try the ad 8397 (it's a dual though)

or maybe the opa551/552


here's a thought, how about trying to design a Mosfet output stage like the M3 for it?

 

[MWP]

Quote:

Originally Posted by Jazper I'd want about double that

Ok, well i actually have enough OPA627AP's to have 3 per channel.
Parallel three would give ~135ma per channel.

A rather expensive solution i know, but if it works, ill be happy.
Ill put the OPA634's in sockets anyway, so i can always use them in other projects later.

Quote:

here's a thought, how about trying to design a Mosfet output stage like the M3 for it?

I need this to work first time, and i dont have the best luck designing descrete circuits

I dont really have time to make a second PCB if the first fails (im going OS soon and this will go with me).

 

[amb]

Quote:

Originally Posted by MWP Ok, well i actually have enough OPA627AP's to have 3 per channel. Parallel three would give ~135ma per channel.

While paralleling more opamps does increase the current output capability, you need to add resistors in series with each opamp's output before summing them together. Typically this is around 22 to 47 ohms. Unfortunately, the resistors limit the current so the net result is not scaled up the way you think it might.

For low-Z headphones high current capability is your friend. You should really consider using a buffer stage, either IC buffers like BUF634 or HA3-5002, or a discrete buffer. Not only will this be cheaper than using three OPA627s in parallel, it will simply work better.

 

[MWP]

Quote:

Originally Posted by amb While paralleling more opamps does increase the current output capability, you need to add resistors in series with each opamp's output before summing them together. Typically this is around 22 to 47 ohms. Unfortunately, the resistors limit the current so the net result is not scaled up the way you think it might.

Hmm, ok, i didnt think the current sharing resistors would need to be that high.
I was thinking 10ohms max.

Ok... time to rethink this project

 

[rickcr42]

Quote:

Hmm, ok, I didnt think the current sharing resistors would need to be that high. I was thinking 10ohms max.

10 ohms will work fine for simple opamp current "stealing" prevention/balancing but higher is better.Even as high as 47 ohms though there will be some small absolute crrent output losses.
All you are doing is trying to get opamp "A" from reaching into opamp "B" and screwing with the output transistors.Having a resistor there also helps protect the opamps output stage from RFI reflecting back into the amp from the output cable.Something very often overlooked but a problem even if not recognised for what it is.
It also helps to have a safety valve if somthing goes wrong such as a dead short across the output (unplugging/plugging headphones while the power is on).This rsistor will limit current and protect the output stage.Does it also limit stage output current ?Hell yes it does but no free rides.Bullet proof designs or absolute signal purity but living on the edge.your choice man.

Buffered or doubled ? Nothing is perfect so if you add a thing in audio you also take something away.With a buffer that is adding the tonal character of the buffer to the sound of the OPA627,maybe the best overall audio opamp ever,thus taking from it.With any multistage design it is the gain stage,in this case the opamp,that dominates the overall sonic signature but the buffer will impart some of its own sound to the final design.
Doubling retains the overall sonics of the opamp but by the addition takes away a bit of detail in the microdynamics area.nothing blatant and in most cases nonobjectionable but still there and to be considered so the decision is which is best suited to you.

The other option would be a more simple current pump output stage.A discrete transistor or mosfet stage that you can tailor to your needs.Class A/B biased high into class-A,full class-A or for portable use where battery life is imprtant class-A/B.Class B is off the list as a viable output stage.Never "on" until there is a signal it is always playing catch up.

DIY=Options.Building to suit needs and end use.

BTW-a good read is the pdf "Headphone Power Requirements" from the Rane Library.While we tend to overbuild here there real story is you can get by with far less current than you think

 

[rickcr42]

Headphone power Requirements-
http://www.rane.com/pdf/note100.pdf


Practical Line Driving-
http://www.rane.com/pdf/note126.pdf

 

[rickcr42]

More on opamp doubling and output load resistors

http://www.drachen-audio.com/OpAmps/AD826.pdf

 

[Jazper]

have a look here:

http://www6.head-fi.org/forums/showt...highlight=LISA

for a (simple) discrete output stage idea