Quick review of the AD823 vs OPA2132 vs AD8066

Yesterday, I got some opamps.. amongst them one I could directly put in my cmoy, an Analog AD823. So I was curious how it would sound.

It pretty much blows away the old BB OPA2132. I thought the 2132 was a stereo opamp. 'Stereo' in the sense that there's a little difference in what is coming out of the right and left channel. Well, with the 823, the 2132 nearly sounds like a mono amp... the separation between left and right is much better, and the most important thing: 3D! Finally I can hear where the sounds really are coming from. The 2132 sounds very flat in comparison, the exact sources of the various instruments are pretty diffuse. The 2132 also tends to somehow put one element of the sound in the foreground, blocking what is going on left, right, and in the background.

With some tracks, all I could do was listening in awe and realising for the first time what really was going on on the stage. The 2132 just shoves some part of the music directly in your face, like a nagging salesman in some sort of cheap tourist establishement.
The down-to-3-volt operation is also very interesting. According to tangents website, it starts to clip at 4.3v. 4.3/9= 0.47 volt per cell. With this you can even run your cmoy with supposedly 'dead' batteries (ok, output impedance might pose a problem with such deeply discharged batteries).

Ok, no chip without drawbacks. It seems to miss a certain slam I had with the opa2132, might be because of the lower output current it can supply, 16 vs 40mA. Also with some Keller Williams live recordings (Live at the Vic Theatre, Chigaco, 2002), the sound seemed to be quite far away, as if you were listening to small speaker box 2m in front of you. I liked the presentation of the 2132 with this album more, it had a more intimate feeling to it. Might be the recording, the in-your-face properties of the 2132 could have just made it sound perfectly (an opamp correcting sonical flaws )

I really wonder why the opa2132 is often proposed as the default chip in a cmoy. It doesn't sound better than most headphone outs. The AD823, on the other side, makes you anxious to get even more of this better sound.

i quite liked the sound of the AD823 as well in the limited devices i could roll it into (z-audio mu, zhaolu d2.0, heed canamp) nice punchy presentation and tight fast bass

i quite like the AD house sound

Which headphones were you using? it seems to have a bit troubles driving 32 ohm grados...

edit:
bpribadi on sgheadphones.net about the opa2134 (sonically identic to the 2132):
Quote:



link

I like the "head in a bucket" comparison

Quote:



I've built several portable amps using the AD823 and i like it a lot.
Personally i think its a big step-up from the OPA2132. Ironically the amps were built originally for my Senn PX200 didn't suit them but sounded great with my Senn HD600s.
Not sure if the OPA2132 bottom end is better, it may be sort of wooly/boomy, whereas the AD823 i think has as much bass its just tighter/punchier maybe?

Are the OPA2132 & OPA2134 pretty much the same sounding op-amp?

Welcome to the world of opamp rolling! I agree when it comes to OPA2134. It's congested, muddy, mono-sounding and undynamic. I find the AD823 very grainy and harsh, even when used in front of a buffer.

If you like the "AD-sound" (AD823/8620/8397). 8620 is smoother and leaner, and except for the coarse treble it's very natural and musical. 8610/20 are used in many high end amps. A problem is it's high cost. If you want more slam, there's the 8397. I've never heard anything like it. It's very forward sounding, very detailed, presents a very good soundstage, and gives a full and kicking bass. When you roll back to whatever opamp, they all sound dull. They say it's hard to tame, but M3-mini is soon to be released, developed around AD8397. The amp is built to fit a small Hammond aluminium case and looks absolutely great. I think this will be a killer amp if you want a small portable with lots of slam, spaciousnes and details.

If you want the best, the opamp of choice seems to be AD744 using comp pin as output. You have to use a BrownDog adapter (2 single to 1 dual). LM4562 is also a contender if you want a clearer treble, but with the cost of a more "sterile" sound.

Joshatdot: Yes, they are. I didn't hear a difference between them. Neither did Tangent (http://www.tangentsoft.net/audio/opamps.html). The 2134 has a bit lower non-audio specs, and can thus be made a bit cheaper if you intend to use the chips for audio only (which you, I suppose, do). Tangent noted that the 2132 goes to slightly lower voltages, but this may have been a production variance in his samples. He also said that sometimes the 2134 oscillated, and the 2132 fixed it. But c'mon, how can you make the opa2134 oscillating?

NelsonVandal: Already got the 843, 744, 8066 and 8397, I just have to solder them on some browndogs. But the double 843 will not fit in my cmoy, the 744 is useless without buffer, and the 8397 will also most likely oscillate because I have zero bypassing in my cheap cmoy. I will solder the 8066 to a brown dog today to hear how it sounds. And yes, I know about the mini^3, I'm the one who made "the mini^3 v2 is out" thread here I will most likely build one.
And well, the 8620 is still much cheaper than a dual opa627/637. It also has great power consumption specs. Odd thing... I will order 3 AD8610, because it's cheaper this way I have a spare dual browndog, no idea why I ordered two, but now it comes in handy.

The LM4562 seems a bit hard to source, you now have to pay for it if you sample it from national ("sampling"? I'd call it simply "buying"). I'm also eager to hear the opa2228 with my grados, might be a good choice on the burr brown side. but can't get it anywhere, only it's slower brother the opa2227.

This might get a little offtopic, but has anybody tried the opa551/552 as an opamp? some use it as buffer, but apparently nobody uses it as an opamp. How does it sound?

Sorry, I didn't know you were so experienced with opamps.

I have also wondered about the opamps used as buffers, why not just use one of them and get rid of an amplifying stage. I think the problem with most of them are a high DC-offset. I tried this trick with LMH6654 in Xenos OHA-REP. The sound without an input opamp, just using the "buffers", was a major improvement. That's when my evil eye to opamps begun. If there is a way to tame those high current, high slew rate and high DC offset amps, it might just be a success. I'm very curious to try AD8099 LR and AD8045 or 9631 as ground channel and use no buffers. Unfortunately you must have a high quality PCB. The evaluation boards would be perfect, but Analog doesn't ship any more samples to me, and the boards aren't for sale.

I found the same thing... the 2132 is just not as revealing as the 823. IMHO you can use the 823 in a single OP-amp circuit, and it will sound very good, so long as you keep the volumes down. If you are like me and you listen at very low volume levels, it is a fine choice. I frequently turn my volumes down to just above the channel imbalance point on my amps. My source has a VERY high voltage output, so that is not really all that low of a volume, although it is a LOT lower level than most others.

I am not a fan of buffering the 823. To my ears it sounds shouty and harsh.... but IMHO thats a sonic characterisitc of the buffer chip in my pimeta, more-so than the 823 by itself.

I'd love to hear an 823 in a discrete buffered amp circuit.

Quote:



I'm experienced in having them sit around and doing nothing, the opa2132 and ad823 are the first opamps I ever have compared Ok, I did make some 'dry' research (dunno if that metaphor exists in english... research without touching the actual matter), but not more than that.

kramer: no high volume? that seems rather odd, the ad823 supports nearly rail-to-rail output voltage swing. maybe it's the limited output current.

Quote:



I'm working on it
(hammond 1455C801 case, two pcb boards, single 9v battery, jan meier crossfeed, mostly smd, 3 chan config, discrete buffers, opamps socketed, proper bypassing for 'cranky' opamps, ad8397/lm6172/ad8620/ad8066, that's whats floating in my head )

I'm going to solder the ad8066 in a brown dog, will tell you later how I like the sound (btw... the first album by the gorillaz has some mad 3d imaging. first heard it with the ad823, wonder how it sounds with the ad8066)

Quote:



To my ears, and with Sennheisers, it's still one of the worst sounding opamps. I've tried it opamp-bufferd, discrete buffered and standalone. The sound is the same whatever you do with it. It's worse than the headphone output of my portable, and much worse than the CDP's headphone out.

And Kramer if you don't listen at a crazy volume level, you really should try a no-opamp solution. It's a revelation. You get lower noise, improved soundstage, less coloration, more intimacy, more details etc.

Anybody, use once AD8397 and you will never get back to a classic unbuffered op-amp. 310mA peak current against 20-50mA from normal op-amps is audible even on Sennheiser PX100. It took me a couple of days to completely forget about AD823, it sounds thin and harsh, adds some contrast to the sounds but is not so detailed.
Best "normal" dual op-amp is probably AD8066, AD8620 for darker headphones, and... nothing is able to outperform 2x AD825, 2x OPA637 far behind IMHO.

Quote:



I found the LM4562 @ Digi-Key for $5.50. The opa2227 & 2228...I've heard the even slower cousin, the opa2277. It was a ordering mistake, but it sounded very clean, flat frequency & yet detailed. I liked the 2277 sound better than my 2107. The 2107 is warmer, darker with some bass boost, sound stage is a little small. The 2107 with my KSC35's is way dark, but with KSC35 & 2277 was nice. I can't wait to order more op-amps from Digi-Key to roll them in my CMoy (I know I need a better amp, but I'm a broke college student with very little income)

this took a bit longer than expected, but today I had some time to compare the AD823 to the AD8066.

At a first listen, the sound difference isn't very big. By no means as much as from the OPA2132 to the AD823. I think this has to do with the 'house sound' of AD - it's like Grados, at first the SR80 and RS2 sound the same. I didn't yet spot any differences between the mids and the bass on the two, but the highs are different. You hear them clearer on the AD8066, but curiously, this comes not at the cost of higher sibilance. It may be the same or even less than with the AD823. The highs on the AD8066 have a property you could describe as 'sweetness'. I do like detailed highs, thats also why I like my Grados. If you also like this about grados, then this chip could be something for you. If you hate that aspect of the grados... well, maybe a bit darker chip for you

The soundstage seems to have a bit of difference. I can't comment on it yet, I need to hear through a few albums (I do have to relisten a lot of stuff anyways, to discover the soundstage of my albums. with the opa2132 you heard nothing of this)

p.s.: opa2107 and opa2604 is on the way across the pond, I'll review them also. I wonder how the opa2107 sounds like, hopefully they make up for my bad initial expression of the bb sound

edit: I forgot some stuff...
Majkel: I will be able to test it in the mini^3 or I build myself a 3 channel cmoy with some decoupling

Joshatdot: interesting... looks like an option. have you already tested the opa2227 and 2228? would be interesting how it compares to these two


another thing about subjectivity: I've read reviews about all those opamps. I tried my best to not just going to 'think' that I hear the same as the other reviewers. but to be really sure, I plan to make some blind tests. fortunately I got two cmoys to test the chips in. I just have to find someone who would randomly place two chips in the cmoys and then give it to me to listen. oh and first I'd have to fix my second cmoy, some cable seems to have a bad solder joint


edit2: the ad8066 and ad823 have some differences in their presentation of the soundstage. I'm not able yet to pinpoint it. ad8066 more airy with further apart soundstage maybe?
I also blind-tested myself. Turned on the cmoy to test out the AD8066 sound a bit more. After a few notes I thought that this does not sound like the AD8066... and indeed, the AD823 was still in it. With the AD823, on one song the cymbals seemed to switch position while fading away... also have to listen closer to it.

I must say that atm, I almost prefer the ad823. smaller but more coherent soundstage? will be listening to some more music...

OPA2228 is very nice, has pleasant bass, lots of air and richness in mids, and silky highs. If you like AD8066, you will enjoy OPA2228. It lacks a bit of detail and soundstage deepness. OPA2227 in spite of being a compensated brother of OPA2228 sounds totally different - slow, laid back, dull, awfully boring. The only positive thing is his bass. Overall, one of the worst op-amps ever used in audio IMHO. When I was playing with Cmoy-like designs, OPA2228 was one of my favourites, but I also enjoyed TLE2082 and AD8620. AD8066 alone is for me too mellow and laid back, but sweet, with good space. I went to higher power and more complicated designs so I discontinued using even AD8397, but this one is a real amplifier, not just sound modifier.