MOTM

At 2:31 PM -0500 07/25/02, Paul Schreiber wrote:
>
>You need the output caps. Just bypass the electrolytics
> with a 0.1uf film cap.

Thanks, I've seen that trick used for crossovers, but I guess I forgot.

If I use a non-inverting amp, should I still AC-couple the inputs??  Or am
I better off using a DC-coupled inverting amp??  I've read an inverting
headphone amp made from the same op-amp will be noisier, but it will use
much fewer parts.

>Use a dual gang, 25K or 50K log pot on the *input*.

OK :)  Mouser sells 50K dual ganged log pots.

At 12:36 PM -0700 07/25/02, Tony Karavidas wrote:
>
>Yes, those caps are there to keep DC from frying your headphones or
>>anything else downstream. You cannot simply AC couple the input and
>expect there to be no DC on the output. Look at the data sheet functional
>>diagram and you'll see regardless of what comes in, the output is still
>at >some DC level.

That makes perfect sense, but you know what is odd??  I found this site:
http://headwize.com/projects/opamp_prj.htm
and _none_ of the op-amp circuits have output caps.  According to that
site, offsets can be sufficiently minimized by using the right feedback
resistors.

Yet, all of the commercial applications of the 2135 that I found used
output caps.  So did all of the examples in the AD's application notes.
Two out of two professional EE's on this list agree.  So I guess, what is
possible in theory isn't safe enough in practice.

>This device can be powered bipolar, and that might allow you to circumvent
>>the DC blocking caps.

It can, but I chose it because it can be run single-sided.  That, and
because I already have a few.