AD8620 or AD823 as an output buffer?
Mar 15, 2006 at 5:20 PM Thread Starter Post #1 of 8

Andrea

Banned - aka HeavySoul - aka inconnu - aka Albert - aka layman - aka joe_average - aka altglos - aka Mr boobi - aka mikesand - aka blindbuy - aka The Well - aka yummy-fi
Joined
Dec 5, 2004
Posts
4,891
Likes
11
Hello,


I'm debating whether to use an AD8620 or an AD823 as the output buffer of my Zero Audiocraft amplifier. The op-amp will be configured as a unity gain follower (output connected to inverting input), placed in the feedback loop of the controlling op-amp (another AD823).

My doubt is, will the AD8620 in particular be stable in that condition - and with no series resistor on the output, to lower the output impedance? Or had I better add 47 ohm output resistors, and/or resort to the AD823 which has lower unity gain bandwidth (so perhaps has more chances to be stable) ?



Thanks for any hints
smily_headphones1.gif
 
Mar 15, 2006 at 5:56 PM Post #3 of 8

rickcr42

Are YOU talkin' to me?
Joined
Jun 21, 2001
Posts
13,874
Likes
14
Stright buffer used open loop.

100K in z resistor,1K series resistor to buffer,10 ohms on the output,4700uf/0.1uf to each supply terminal (or 2200uf/10 ohm 1W/2200uf/0.1,better man),bipolar 15VDC shunt reg power source and done !

Get the TO-Package LT1010,run that puppy HOT and heat sink the crap out of it or the old standby TO-Package version of the BUF634 high bandwidth mode,also with a heat sink.
 
Mar 15, 2006 at 10:31 PM Post #4 of 8

Andrea

Banned - aka HeavySoul - aka inconnu - aka Albert - aka layman - aka joe_average - aka altglos - aka Mr boobi - aka mikesand - aka blindbuy - aka The Well - aka yummy-fi
Joined
Dec 5, 2004
Posts
4,891
Likes
11
Thanks for answering.


I'd rather stay with the AD8620 just because I have one mounted on a browndog, while the BUF634's or OPA551's or LT1010's I would have to pay for (which I don't plan to do since I'm short of cash
smily_headphones1.gif
).


So, how about my concern with the stability of the AD8620 in the condition said above ?

Thanks again
smily_headphones1.gif
 
Mar 16, 2006 at 12:40 AM Post #5 of 8

rickcr42

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

Opamp gain/Opamp buffer really is the last choice if there is one but on the "use what you have" mode of DIY operations yes.It will work.

Maybe not as good as a true dedicated buffer would OR as well as lopping out the original opamp and just replacing it with the single opamp performing both gain and buffer duties but still workable if you must.

Standard PS bypass caps etc at the new chip and maybe set the input Z to 50K (47K std,value),1K series resistor between stages plus add a small amount of resistance outside the loop for output load isolation on the last opamp .
If you plan to use Grado cans or other Low-Z headphones then keep this from 10-22 ohms.If only mid impedance to highish impedance then 47 ohms is a good happy medium.
 
Mar 16, 2006 at 7:08 AM Post #7 of 8

jcx

Headphoneus Supremus
Joined
Jul 24, 2002
Posts
2,371
Likes
368
its not clear that the ad8620 is that much better in output drive than the ad823 - its hard to find the # on the data sheet but at +/- 12 V supply the 823 should be competitive - the low sc currents are for 3 or 5 V supplies, @ +/- 15 Vs 823 sc currents are +80,-60 mA compared to +/-65 mA for the 8610 @ +/-13 Vs

if amp gain is > 5 a unity gain ad823 would probably work but the loop stability story is the reverse of what you imply in your 1st post, the faster ad8620 as a unity gain buffer will add less excess phase to the loop and therefore be more stable - slightly complicated by the relative Cload stabilities of the 823 vs 8620 in unity gain, load isolation R or lossy ferrite bead helps
 
Mar 16, 2006 at 5:11 PM Post #8 of 8

Andrea

Banned - aka HeavySoul - aka inconnu - aka Albert - aka layman - aka joe_average - aka altglos - aka Mr boobi - aka mikesand - aka blindbuy - aka The Well - aka yummy-fi
Joined
Dec 5, 2004
Posts
4,891
Likes
11
Thanks for the useful answers.


jcx -- actually the AD823 performs very well at unity gain (at 12V already), showing better grip on the headphones' drivers than the OPA2227 or OPA2134.


Clutz -- PM replied!
 

Users who are viewing this thread

Top