[yejun]

Can I bias 50mA on buf634?

 

[tangent]

Theoretically, yes, you can. The trick is coming up with a good 50mA current source. A pair of 2N5484s isn't going to cut it.

 

[yejun]

http://www.maxim-ic.com/appnotes.cfm...mber/159/ln/en

a programmable current source?
any idea? since I am thinking add digital pot and LCD to my headamp, maybe adding this circuit isn't a problem.

 

[yejun]

Another idea is use LM377.

 

[jeffreyj]

A very simple but high performance "two terminal" current regulator that can source 50mA and has a compliance of 1.2v can be made most simply by connecting a 24 ohm resistor to the output of a LM317T regulator and tying its adjust pin to the output after said resistor. The current output then becomes the adjust terminal reference voltage, 1.2v, divided by the resistor value, 24R, which equals... 50mA.

(edited to remove reference to an image I was going to insert, not attach, here)

 

[jeffreyj]

Quote:

Originally posted by yejun http://www.maxim-ic.com/appnotes.cfm...mber/159/ln/en a programmable current source? any idea? since I am thinking add digital pot and LCD to my headamp, maybe adding this circuit isn't a problem.

Holy Switching Noise, Batman! That circuit would be a poor choice for the application at hand on account of complexity, noise, transient response and bandwidth issues! Egads, man, there are much simpler solutions to be found!

 

[rickcr42]

a better choice would be to just build a discrete class-A voltage follower/buffer

Why use a monolithic buffer IC for class A output when a discrete stage would do the job better and which much more simplicity

 

[yejun]

is discrete really better than op? why?
i don't have any experience about amp, just wondering some simple tweak can make it sound better.

 

[rickcr42]

you can find an example of an op-amp gain stage folled by a discrete class-A current stage (buffer) at the headwize projects library - see sheldon stokes

and since we are talking about monolithic buffers-

they sound just fine for the intended use but are optimized for high current gain with little current draw so by definition are class AB or class B devices : they have to be

Class a operation means (if full class A) there is no or little differnce between the idle current of the stage and the same stage at full output

and while it is probably possible to make the buf634 into a class a follower it is better IMHO to use as simple a stage as needed to perform the result you are looking for

remember , if you look at the actual internal structure of the buffer chip it is far from simple-unlike first generation buffers such as the LH0033/63