Block DC Offset at amp. output.
Mar 14, 2006 at 2:33 PM Thread Starter Post #1 of 26

]|[ GorE

500+ Head-Fier
Joined
Mar 16, 2004
Posts
796
Likes
10
I am planning on replacing the JRC4556 on my sound card with the AD8397.And 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.

And what negative impact would it have on the sound ?
 
Mar 14, 2006 at 2:43 PM Post #2 of 26

civilmonkey

100+ Head-Fier
Joined
Feb 17, 2006
Posts
126
Likes
10
I'm pretty sure you can do that, but from what I've learned so far, basically anything in your signal path will degrade the sound. It's just a matter of how much.

For caps in the signal path, polypropylene, PPS and polyester caps are best (in that order).

Check out http://www.tangentsoft.net/audio/input-cap.html

For more information.

EDIT: I think there was a thread on this if you search these forums. Keep in mind I'm brand new to DIY audio, and one thing I've learned here is take advice with a grain of salt.

Good luck!
 
Mar 14, 2006 at 3:00 PM Post #3 of 26

jar

100+ Head-Fier
Joined
Jan 26, 2005
Posts
186
Likes
0
Be aware that the 8397 is quite finicky. It probably won't just drop into an existing circuit and work properly. Take a look at the MINT v2.0 and Mini^3 threads on that other forum.
 
Mar 14, 2006 at 3:17 PM Post #4 of 26

rickcr42

Are YOU talkin' to me?
Joined
Jun 21, 2001
Posts
13,874
Likes
14
Quote:

I am planning on replacing the JRC4556 on my sound card with the AD8397.And 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.

And what negative impact would it have on the sound ?


You can go that route and would even be a good way if you had to use the chip in a single polarity power supply but the cap will be a BIG one so must be the best quality you can get.

Capacitors in the signal path are audible but that does not mean some very good sounding cap coupled amps can not be made (Szekeres for one,almost all SE Triode amps,even some Class-A speaker amps).

Look at around 470uf if driving low Z cans,220uf for higher Z headphones and around 10uf for line level
 
Mar 14, 2006 at 4:09 PM Post #5 of 26

n_maher

Resistorous Conflagorous
Joined
Nov 20, 2004
Posts
8,405
Likes
31
Quote:

Originally Posted by ]|[ GorE
I am planning on replacing the JRC4556 on my sound card with the AD8397.And 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.

And what negative impact would it have on the sound ?



My first question would be why the 8397? Are you hoping to not have to use the soundcard with an amp? In my opinion you'd be much better of going with a more stable opamp (echoing jar's comment) since it's surface mount application and it'd be nice if you could just drop the chip in and be done. Maybe the 2134 or 2227 would be a better choice?

Also, what soundcard is this? Is it really worth modding it or would you be better of buying a decent inexpesive one like the AV710 and using the line out to an external amp? I find my 710 run through my PIMETA to be quite nice given the cost of each.

Nate
 
Mar 14, 2006 at 4:46 PM Post #6 of 26

ericj

Headphoneus Supremus
Joined
Aug 2, 2005
Posts
8,262
Likes
142
Since the 4556 and the 8397 are both bipolar amps, they should have similar output offset.

Unless your 8397 starts oscillating, and it probably will. It is NOT suited to drop-in replacement uses.
 
Mar 14, 2006 at 5:52 PM Post #7 of 26

]|[ GorE

500+ Head-Fier
Joined
Mar 16, 2004
Posts
796
Likes
10
the card is a modded Audigy 1 24bit Value.Why do you think the 8397 would start oscillating ? Coz the modded card has a 220uF cap and a 0.1uF bypass cap for V+ and V- (meant for the opamp).
 
Mar 14, 2006 at 6:29 PM Post #8 of 26

guzzler

Headphoneus Supremus
Joined
Jan 17, 2003
Posts
1,851
Likes
12
The opamp on the sound card will likely be being used a buffer, ie, a gain of 1. All opamps have something called their "gain-bandwidth" product, which means their bandwidth drops with increasing gain. The AD8397 is extremely fast, over 100MHz, while the JRC is in the 10MHz region (check the datasheets!). This means the 8397 will respond to any stray capacitance/inductance at its input pins, and possibly drive it into oscillation. As has been mentioned, an opamp with a similar bandwidth is a more appropriate choice
 
Mar 14, 2006 at 6:41 PM Post #9 of 26

rickcr42

Are YOU talkin' to me?
Joined
Jun 21, 2001
Posts
13,874
Likes
14
Quote:

The AD8397 is extremely fast, over 100MHz, while the JRC is in the 10MHz region (check the datasheets!). This means the 8397 will respond to any stray capacitance/inductance at its input pins, and possibly drive it into oscillation. As has been mentioned, an opamp with a similar bandwidth is a more appropriate choice


or just slow it down with loop compensation and an input shunt cap.Not a bad idea for most audio use opamp devices which are quite content (and will act less badly) at 1-2mhz and slew rates in the 20's.

Many "bright" sounding op-amps are actually oscillating or ringing in the mhz range with the artifacts being audible much lower which reinforces the upper octaves.Some call this detailed sound but would be shocked to see the reality of it on a scope trace
eek.gif


To be honest I could live with dead flat and well behaved from 30hz to 15khz over ultrasonic with artifacts any day and most others could also if they never knew there was a rolloff at the ends.It is more the knowing than the listening and why so many get hung up on specs over performance
 
Mar 14, 2006 at 8:16 PM Post #10 of 26

MisterX

Headphoneus Supremus
Joined
Dec 30, 2002
Posts
3,880
Likes
23
As someone that has been there and done that (except I used AD8620s) I would suggest that you not waste your time.

The stock opamps actually do a pretty decent job of masking the horrible top end that the Philips UDA1328 6 channel DAC featured.
 
Mar 14, 2006 at 9:26 PM Post #11 of 26

rickcr42

Are YOU talkin' to me?
Joined
Jun 21, 2001
Posts
13,874
Likes
14
Quote:

Unless your 8397 starts oscillating, and it probably will


that statement is way overblown.There are many opamps being used here every day that are more prone to go into oscillation the LM6171 being one and I see that stops no one from using it.They just pay closer attention is all.

The AD8397 is what it is.A bipolar input opamp no worse than any other bipolar opamps that gets GOBS of output current while still remaining battery freindly and if you follow simple opamp housekeeping techniques (see ADI library) is no worse than any other high speed opamp.

Proper layout,loop compensation,and a proper power source and no more of a "chore" than any other opamp.

The above statement is a scare tactic for whatever reason.Read the app. notes and do a proper layout and you are fine.If you need to do a SMD/DIP adapter look to the AD825 plug in modules or AD8065 modules available from various mod services for hints on how.

No scary monster,just an opamp and a slow one when directly compared with most modern opamps in regular use here for audio
 
Mar 14, 2006 at 10:02 PM Post #12 of 26

rickcr42

Are YOU talkin' to me?
Joined
Jun 21, 2001
Posts
13,874
Likes
14
For "clarity folks.A couple of commonly used opamps and their speeds :

OPA637

Bandwidth 80MmHz,
Slew Rate 135V/uS

http://focus.ti.com/docs/prod/folders/print/opa627.html

------------------------------------------------------


OPA2134
Bandwidth 8mhz
Slew Rate 20V/uS

http://focus.ti.com/docs/prod/folder...t/opa2134.html

------------------------------------------------------

OPA604

Bandwidth 20mhz
Slew Rate 25V/uS

http://focus.ti.com/docs/prod/folders/print/opa604.html

-------------------------------------------------------

LM6171/2

bandwidth 100 mHz
Slew Rate 3600V/uS

http://www.national.com/pf/LM/LM6171.html

-------------------------------------------------------

AD825

Bandwidth 46mHz
Slew Rate 140V/uS

http://www.analog.com/en/prod/0,,AD825,00.html

-------------------------------------------------------

AD8620

Bandwidth 25 mHz
Slew Rate 50V/uS

http://www.analog.com/en/prod/0%2C28...8620%2C00.html

------------------------------------------------------

AD8065

Bandwidth 145 mHz
Slew Rate 180V/uS

http://www.analog.com/en/prod/0,,759_786_AD8065,00.html

-------------------------------------------------------

AD8397

Bandwidth 69 mHz
Slew Rate 53V/uS

http://www.analog.com/en/prod/0,2877,AD8397,00.html


looks pretty much "middle of the pack" to me.......
 
Mar 15, 2006 at 8:34 AM Post #13 of 26

fewtch

Headphoneus Supremus
Joined
Jul 23, 2003
Posts
9,559
Likes
31
This thread really makes me ponder if the opamps we commonly use around here are not only insane overkill, but actually counterproductive in some ways. My CD player uses good ol' NE5532s which not only sound pretty good to my ears, but are more logical for audio use than some of these others that would fit better in radar guns and space weaponry
tongue.gif
.
 
Mar 15, 2006 at 3:06 PM Post #15 of 26

NeilR

500+ Head-Fier
Joined
Dec 2, 2005
Posts
929
Likes
12
Quote:

Originally Posted by fewtch
This thread really makes me ponder if the opamps we commonly use around here are not only insane overkill, but actually counterproductive in some ways. My CD player uses good ol' NE5532s which not only sound pretty good to my ears, but are more logical for audio use than some of these others that would fit better in radar guns and space weaponry
tongue.gif
.




I thought the whole point of this forum was to apply insane overkill to what otherwise would be $30 consumer electronics devices
k1000smile.gif
 

Users who are viewing this thread

Top