[rickcr42]

Quote:

Thanks.I am gonna go ahead with the mod and add 100nF bypass caps for the opamp(as specified in the datasheet).

It really isn't rocket science or voodoo electronics but just sound constructon practices.

The data sheets 99% of the time will tell you all you need to know about how to get a particular part to work and all the rest is just overcomplicating the simple for whatever purpose.If that is not enough both ADI and BB have an extensive app.note library that goes into great detail on layout but guys,it IS just an op-amp !

Has a differential input (what makes it an OP amp and not just amp,feedback !),an output and two power supply pins.All you need do is adress each one in turn and the end result is a usable product be it a 741 opamp or the fastest chip known to mankind.

concentrate on layout and sound design and get the "what is the prettties capacitor" crap and you will in the end have a better sounding project.

I will take a properly designed stage over one with the best (sic) parts but used wrongly every day of the week and that includes leaky ceramics and generic electrolytics over V-caps and Black Gates.

Some files to download.Print them and study them and you will have no fear of the lowly opamp

http://www.analog.com/UploadedFiles/...169AN692_a.pdf

http://www.analog.com/UploadedFiles/...30769AN214.pdf

http://www.analog.com/UploadedFiles/...0828AN_345.pdf

http://www.analog.com/UploadedFiles/...388AN674_a.pdf

http://www.analog.com/library/analog...h6_final_I.pdf

http://www.analog.com/library/analog...h7_final_J.pdf

Enjoy the headache

Rickster

 

[rickcr42]

oops !

http://www.analog.com/library/analog...09/layout.html

 

[rickcr42]

Quote:

My CD player uses good ol' NE5532s

nothing at all wrong with that chip and is used in a crapload of great sounding hi-fi AND pro gear.

what has happened here is we confuse headphone capable with gain stage and confusing "end use" damn near 100% of the time.

The original Chu Moy Pocket amp that was the first online blow by blow on how to build a quality headphone amp for cheapc is partially to blame but only because so many choose to misunderstand the intent of the single opamp as both gain stage and driver stage.

Using only one device meant a chip with enough output current to directly drive headphones was needed but it also had to be battery friendly or it would defeat the purpose which originally was as a portable headphone amp.

so the requirement were :

1-good sonics
2-must have enough current output to drive the majority of headphones
3-the idle current must be low for a long battery life
4-low voltage operation is essential and especially so as the batteries drain down
5-low DC offset if it is to be DC coupled at the output

Well this has been twisted up to where these "headphone amp" opamps have become the ONLY opamps under consideration for every single application using one ! Why ? Because most have no understanding of the "why" certain devices are chosen to perform certain duties and not knowing revert to what they DO know and is visible as the forum choices.

Examples :

Why use a high current drive opamp straight into a high current buffer when a possibly better part with lower output current not only can but in usually WILL better the high current/high current combo ?

Why spec an opamp that rules for battery operated portables for a project that will never see a battery ? Why not just choose the BEST opamp for that,one that idles way higher and is in fact an indication of higher bias into class-A ?

Why spec an opamp with 200V/uS slew rate for say a microphone or phono stage that will never see those voltages and in fact is harmful to the signal path ? That one with 2V/us-10V/uS is not only adequate but better ?

How many here know that each succeeding audio stage needs to be FASTER than the previous one or you could introduce S.I.D. (slewing induced distortions) or that the simple "fix" is to simply band width limit the "too fast" stage so it behaves and plays well with others ?

Folks need to match the part to the application for best results and less headaches rather than take a "plug and play" attitude towrds the simple op-amp gain stage which is in reality anything but.

Portable duty

line operated

line driver/headphone driver

straight gain stage

multiloop gain stage/buffer

high gain/low noise operation.......


all different and guys,just because an opamp is not a highly visible presence here does not mean it is crap.Just means most don't know

Rant over..........


continue please

rickster

 

[rickcr42]

In case you are a fast reader and actually retain what you read

http://www.analog.com/UploadedFiles/...m=AN-202&la=en

http://www.edn.com/article/CA476907.html

http://www.analog.com/UploadedFiles/..._Notes/5008249

 

[tangent]

Quote:

Originally Posted by GorE since the AD8397 is said to have some output dc offset,i was wondering if attaching a capacitor at the output of the opamp would work.

That's solving the symptom, not solving the cause. Because the 4556 is a bipolar chip, the input bias current balancing issue has probably already been taken care of. However, read on.

Quote:

Originally Posted by ericj Since the 4556 and the 8397 are both bipolar amps, they should have similar output offset.

The fact that they are both bipolar tells you nothing about how much offset voltage you will get. I would expect the 4556 to have an offset at least half that of the 8397 on average, and potentially a lot more. To see why, study the input bias current, input offset current, and input offset voltage specs in both datasheets. Just looking at the numbers tells you that these chips have quite different input stages. If you want, you can then take the numbers into my offset voltage calculator and see how differently these chips can behave.

So, while the offset current balancing issue has probably been taken care of, it may not be enough to prevent the offset from being too high. What I would do is measure the current DC offset and multiply it by 10 to get the worst case offset you'd get from dropping in an 8397. (Worst case assuming the chip isn't oscillating, at any rate.)

I would also figure out what gain the current chip is operating in before replacing it. If it isn't a gain of at least 2, I wouldn't bother trying. The 8397 really, really doesn't like to be run at G=1 when being used to drive cables.

 

[JahJahBinks]

In my case, when I swapped the opamp from my M3 with another, I noticed increased in output offset voltage even though from the opamp perspective the input offset voltage and bias current of the latter ones are actually lower than the first ones.

 

[tangent]

Quote:

Originally Posted by JahJahBinks In my case, when I swapped the opamp from my M3 with another, I noticed increased in output offset voltage even though from the opamp perspective the input offset voltage and bias current of the latter ones are actually lower than the first ones.

Check the datasheet again. I'll bet you were just looking at the numbers in the "Typical" column. I'll further bet that you'll find that the second op-amp's maximum offset current numbers are greater than the first op-amp's minimum offset current numbers, which means that what you experienced can indeed happen. Don't rely on the Typical numbers in any datasheet.

 

[rickcr42]

Quote:

The 8397 really, really doesn't like to be run at G=1 when being used to drive cables.

why we have buffers for gain of X1 and gain stages,the opamp for gains greater than X1.

A good amount of opamps CAN be used as buffers but I always took the path of buffers to buffer and gain stages for gain rather than trying to make something be another thing just because.I even go to

The real screwup with the 8397 was not going to the new ADI 4-Pin SMD package has a pin layout better than any other opamp package ever invented for audio/high speed opamp layout.since all new products will only be offered in tiny little packages and the DIP is pretty much obsolete except for old devices there is no reason to retain the old standard pin-out unless there are extra pins available such as a DC balnce control or bandwidth set pin

 

[tangent]

Quote:

Originally Posted by rickcr42 The real screwup with the 8397 was not going to the new ADI 4-Pin SMD package

I don't understand this. First, the 8397 is a dual channel chip, so it needs at least 8 pins. Second, a single channel chip would need at least 5.

 

[rickcr42]

point being the old way is obsolete.SMD is here and to stay so the DIP pin-out has no real place in new designs other than habit.The new SOT-23-5 ADI package layout is clearly superior to the way things have been done previously

 

[rickcr42]