the cmoy slayer - the AD823
Jul 16, 2001 at 1:10 AM Thread Starter Post #1 of 32
Apheared

Apheared

You call that an amp? THIS is an amp!
Joined
Jun 20, 2001
Posts
377
Likes
10
ppl gets the credit for this, but since he's insane and many people can't understand his writings...
tongue.gif


As many of you know, I've tried alot of opamps. Probably 40. Many of them sound good, some sound dreadful, but some have exceptional traits... there's a certain something to the sound of many of them; the LM6171s, SSM275s, OPA228s, 627s, 825s, etc. That's why I sent them to some of you guys, cause I know you're lazy ducks and won't even try it if you get the chip for free.
mad.gif
Well, no free samples this time... but here's my current ultimate opamp:

The Analog AD823. I LOVE THIS CHIP. More than 132s, 627s, AD825/6, 6171s, more than any high-end chip I have currently!

You cannot use it in a cmoy/hansen pcb... it must be buffered by something to drive headphones... a discrete config, a nice opamp like 6171, or a buffer: 6121, buf634, or my personal favorite buffer for this chip, the Elantec EL200x series. An EL2001 provides moderate drive current. This buffer sounds good anyway, but it's a great match to the 823. Both the 634 and the 6121 (and even 6171 really) will work to buffer it's output and keep the sound the same for the most part. If you want to experiment with buffers - do it on breadboard... none of these chips have the same pinouts. Grr.

I built an 823/2001 amp to drive HD600s (in a nearby thread)... well, this has become my main amp for ANY headphone. Head to head it whoops ass and takes names with everything else I've built in the last year. Maybe my skill has increased, maybe it's just that good. I assume it's the chip.
biggrin.gif


132/4s are still good chips. smooth, soft-ish, almost tube-like in the treble warmth. This chip is not that. Bright and sharp, maybe too much so for some people/headphones. 6171 is a great chip, but it sounds sterile next to this, while accurate they don't have the same "liveliness" effect. I don't know if that makes it more or less accurate, but it certainly makes it more enjoyable to me.

This isn't comparing by swapping chips in a single amp; this is comparing the 12-some-odd amps I have sitting here with the same source; all built to specifically bring out the best of what's there, using great parts and detail to build.

This is my favorite amp, plainly. The only thing that equals it is my 823/2008 massive beast (ppl's beast rev3) I goet here by accident - I fried an AD825 that was in my 2008 beast, and started substituting... 823 stopped me in my tracks.

YMMV of course, but if you've liked my chip suggestions on other chips over the last year; trust me when I say you'll LOVE this one.
 
Jul 16, 2001 at 2:44 AM Post #2 of 32
Great! Another amp to build!
tongue.gif

I JUST finished a plain cmoy in Altoid tin with 2 9V, just because I felt left out and had the extra parts.
mad.gif
 
Jul 16, 2001 at 4:48 AM Post #6 of 32
btw, how does your numbering system work? You have the "47" low impedence (e.g. Grado) mod. Is this different from saying #47, like #42 for the HD600's amp?
 
Jul 16, 2001 at 6:20 AM Post #7 of 32
47 was because all the resistor values ended up being 47 470 and then .47 caps... just a previlant number...

this is the 42, as in the answer to life, the universe, everything... HHGTTG. I guess I coulda called it Slartibartfast...

there's no reasoning to it.
biggrin.gif


Serow, you're randomizing all over the place. Make a new thread and ask your parallel and AD questions together so it'll be easier for others to refer to later...
 
Jul 16, 2001 at 10:23 PM Post #8 of 32
hmmm, trying to decided Active or digikey for my next part order (ie BB stuff or AD stuff), i guess this decides it... AD823 is also available in DIP, much easier to work with than 825....

i'll also try the elantic buffers from insight....

Damnit, DIY is draining all my money!
frown.gif
 
Jul 18, 2001 at 5:15 PM Post #10 of 32
Hmm... see, I never tested them swapping... I blew the 825 and then went thru a couple others and when I hit the 823 I stopped. That sibilance issue that the 825 had went away with the 823. It's low thru mid is more forward than the 825... or maybe it's just the config it's in. Again, this is from memory... and my memory sucks. The 823/2008 is what I carry everywhere, ya know?
tongue.gif


But since I fried my 825 with that CRD failure, I haven't re-tested it in your amp... I had that treble issue with it, and an 823 in there is making me damn happy with the sound.

My next test is to see how I like a single BUF634 paired with the 823, side by side with this 2001 version... I think I know the outcome already but I will test for completeness (and then you and me will split a huge order for Elantecs
tongue.gif
)
 
Jul 19, 2001 at 5:43 AM Post #13 of 32
Apaheared: I have noticed that also with the 825 the Midrange is recessed compared to almost every other OPA i have tried. I think this may be the reason for the Exsessive sibelence with this Opamp that is not as noticeable with other opamps. The BUF-634 is also recessed in the midrange about as mutch as the 825 so an AD-825/BUF-634 combo should be only for phones with a real fowared midrange IE Grados. Another Issue i have with the BUF-634 vs the ELantecs is somewhat Bloated bass on the BUF-634. I think Both the EL-2008 and EL-2001 sound better than the BUF-634 but with certin Cans and or sources and Opamp choice it's sonic signature can be exploited. BTW the OPA-627/OPA637 are a good sonic match to the BUF-634. But an 825 and BUF-634 No not for most uses, Only recomended if you mids are allready way too foward.
 
Jul 26, 2001 at 5:25 PM Post #15 of 32
It is sad that today as it was 20 years ago that Buffers are Few and far between But at least nowdays thay cost less than $30.00. The Wonderfull EL-2008 and EL-2009 that have been around for over 10 years is now going away also. No wonder folks are Quick to Look at CFB Type opamps. In Theroy All Current feedback type Opamps are unity gain stable so one would Look Very Good for Buffer use with the Exception of the fact that thay are Still OPamps with voltage gain. So regardless of what elce you do Your output stage is going to use 100% feedback and this Closed loop type of Buffer when combined with the main voltage gain stage and another Feedback loop or two can result in A Radio transmitter rather than a Headphone amp. I Like Open loop buffers because Thay do not use Feedback. now I am not one of those that Dislike feedback In fact i think Feedback is good as long as it is properly Applied. Asking a system to operate well into the MHz. range While driving a Reactive load and also being enclosed within yet another feedback loop is somewhat scary.

The Alternative is to use just a High Curent CFB as Both the gain stage and output. This is again not an ideal solution because of the thermal problems i have gone into before at the headwize site. In Addition The input Bias currents of CFB opamps require a Buffer on the input, or high DC offsets will result as well as killing your vol. pot. What to Do Find the fastest open loop buffer you can get use your fav. Voltage feedback type opamp with low input bias currents. Then apply your fav feedback method Yes there is life outside of multiloop. Get the cleanest low impedance low noise power-supply you can and Build. Getting back to buffers this leaves just about only a handfull of Buffers the BUF-634, The EL-2001, EL-2002, LM-6321, LT-1010, LH-0033 just to name the most available. Analog devices makes a BUF-03 But it's all NPN output stage turns me right off. Linear tec's LT-1010 also uses All npn output stage, but when combined with the LT-1122 sounds rather good.

I intend to investigate the DSL driver at some future long away date since these are just high current CFB type opamps I am not very motivated because of all the afore mentioned problems. But when i get a topology that sounds good i may impliment sutch a design.
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 2 (members: 0, guests: 2)
Back
Top