Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › AD797 now sounds like it should
New Posts  All Forums:Forum Nav:

AD797 now sounds like it should

post #1 of 80
Thread Starter 
EDIT: This thread has become a bit longer than expected. In short. I implemented AD797 in a bad way and found the sound to be colored. Now it's better implemented and it sounds very neutral, and it has become my favourite opamp. Very smooth and unfatiguing, maybe there's a lack of aggressiveness.

I bought some AD797AR (SOIC) from eBay, and they don't sound as I expected from the reviews by others. I thought they would sound dark and bassy. I've tried them in 3 different amps. Buffered and unbuffered, Jung multiloop and normal multiloop. 9, 12 and 18 V supply. I've tried to add small caps in the feedback loop as suggested in the data sheet. I've burnt them in over night. There's no excessive current draw, noise, hum or clicks to indicate malfunction.

To my ears they have a coloration in the lower treble that make them sound bright. This coloration is similar to AD823, but AD797 sounds much cleaner and more sofisticated. They're not the least bassy. I find them very neutral in bass and mid.

I've only read one review of AD797 that correlates to my findings and that's from Majkel: "AD797 - is like more analytical version of AD825, less impactful and with some glaze over highs." Are the eBay AD797 from another batch or even fakes?

As I understand AD797 seems to go through a revival, used in Zero DAC and D1. Please comment your impressions of AD797.
post #2 of 80
wow, you managed to tame a BJT AD797 even with multiloop?

please share your resistor values!!!


AD797 is used in RSA's top end amps IIRC. They shouldnt sound too bad . I think, from observation, the Apache is AD797 ---> Push Pull Output stage driven at centre point of two diodes, of course biased to Class A.
post #3 of 80
Thread Starter 
Shopper, you should try AD797 since you like AD823. Regarding the Linear amps, I just haven't got around to order any samples yet. If you have "too many" LT1364 or 1363 lying around, I'll be glad to trade for AD797.

I haven't got that many opamps left, since many of them have been fried in various experiments and placing them the wrong way on sockets, but the best balance is with LM6171 in ground channel. OPA2134/AD797 is too dark and dull and AD825/AD797 or AD829/797 is a bit bright. I'm awaiting a batch of opamps bought from a fellow Head-fi'er (only dual, I'm afraid).

AD797 can be unstable. It oscillated in the LISAIII-clone until I added 10 pF caps in the local loop. In the PINT-style 18 V amp it's totally stable with or without such caps. In the 9 V three ch amp with LISAIII buffers it's also totally stable, Jung multiloop or "normal" closed loop, with or without capacitors.

I can't remember the exact resistor values. In the LISAIII-clone, I think it's more or less the same values proposed by Walt Jung in his original article. In the other buffered amp, the values are a bit lower, but I think it's about the same proportions. In the PINT-style amp, I think it's about the same values as in PINT, with something like a 6 or 8 Ohm "output resistor" in the loop.

Have you heard AD797? What do you think of it? I don't think it will become a favourite for me, if I can't find a perfect match for ground channel (should be dark sounding yet powerful and detailed). AD797 is a fine sounding opamp in every aspect, but it's far to colored for my taste.
post #4 of 80
OPA211 is what I'd call "dark AD797". After another bunch of listening sessions I think OPA211 is good, similar to OPA637, the latter being a bit laid back and "confined" in direct comparison.
post #5 of 80
Thread Starter 
I think it's funny to compare the "PINT" amp to "LISA". All this talk about capacitors and "the right" resistor values. The coloration of AD797 is exactly the same in these two amps. The LISA-clone with it's Jung multiloop and isolating JFETs on the rails and discrete buffers, expensive OS-CONs. The PINT-style amp with no electrolytics at all right now, only small ceramics like in RA-1. LISA is slightly better, the soundstage is very special, and it's a bit more detailed, but the major difference is the opamp. It's not easy to be sure either. It could be placebo since I only have one AD829 and one LM6171, and I have to roll opamps while comparing them. I have two NE5534. I think I'm going to compare the amps with the same opamps.

Majkel, you should do a new list of opamps, like an updated version of this: http://www.head-fi.org/forums/f6/bes...2/#post3032335
Sound is very subjective, but I think your comments are very accurate. I'm eagerly awaiting two more AD825 to be able to use it in all channels and ADA4841-2 amongst some others.
post #6 of 80
Thread Starter 
I did the comparison AD797/NE5534 in the two amps mentioned above. The LISA-clone is better, no doubt about it. I guess AD797 isn't good driving headphones unbuffered. The tonality was quite allright, since NE5534 is dull and dark. The combination with LM6171 is still better, and it's a better opamp than NE5534 in every aspect. I could live with this combination. I have two PIMETA boards and a lot of discrete buffers. I think I'll make an amp of it.
post #7 of 80
I personally like the AD797, but my experience has been (and I'm told my knowledgeable engineers) that it is a somewhat difficult chip to use properly (for reference, I don't consider the AD8397 "difficult"). It cannot just be casually inserted into a circuit; you have to build around it (which means no using SOIC-to-DIP adaptors, especially Browndog!). The topology makes it very sensitive to a variety of environmental factors that you can be much more casual towards with many other chips. If it isn't used properly, you will get substantially rising harmonic distortion with frequency at the least, along with probably other distortion products. My experience has been that it is not suitable for driving headphones directly.

What resistor values did you use for jung multiloop? The bias current requirement is several thousand times that of, say, the AD744 that ppl seems to like.

AD825 is a great chip. One of my favourites to use if I need a JFET-input op-amp, especially if it needs to go over 12V.
post #8 of 80
Thread Starter 
Quote:
Originally Posted by Filburt View Post
I personally like the AD797, but my experience has been (and I'm told my knowledgeable engineers) that it is a somewhat difficult chip to use properly (for reference, I don't consider the AD8397 "difficult"). It cannot just be casually inserted into a circuit; you have to build around it (which means no using SOIC-to-DIP adaptors, especially Browndog!). The topology makes it very sensitive to a variety of environmental factors that you can be much more casual towards with many other chips. If it isn't used properly, you will get substantially rising harmonic distortion with frequency at the least, along with probably other distortion products. My experience has been that it is not suitable for driving headphones directly.

What resistor values did you use for jung multiloop? The bias current requirement is several thousand times that of, say, the AD744 that ppl seems to like.

AD825 is a great chip. One of my favourites to use if I need a JFET-input op-amp, especially if it needs to go over 12V.
As far as I can remember, I used basically the same values as suggested by Walt Jung except for a 10k pot and 100k (not 10M) from input to ground, and of course since the buffers are unity gain, there's no local feedback loop for them. Rs = 1k. I don't use output resistors outside the loop.


Filburt, how does AD797 sound to you? Do you find it "perfect" without any flaws at all? Do you find it colored, and if so in what way?

And Filburt, are you one of the few who think there is a JFET and a bipolar sound? I do.
post #9 of 80
Thread Starter 
And about BrownDogs, what's so bad about them? I often solder the caps and some of the resistors directly on them. If you take "good" amplifier boards, there are often small ceramics close to the opamps, but the electrolytics are far away. I've seen some measurements of doing this (read it on Diyaudio, not measured myself), and the result is worse than using just electrolytics close to the opamps.
post #10 of 80
Thread Starter 
After some more listening time, I've been more habituated to the coloration, and I can appreciate the very dynimic, detailed but most important musical sound of AD797. It's very engaging. I think the coloration I hear is what people report as tube-sounding, more pentode than triode. Like Majkel said, a glaze over the treble. If you like AD823 you must love AD797. It's like an AD823 but in every single aspect better and "bipolar-sounding" (= non grainy). I'm still using it with LM6171 and LISAIII buffers, multi loop and with 10 pF caps in the local loop.
post #11 of 80
Thread Starter 
Quote:
Originally Posted by Shopper View Post
Pity for the stability concerns, especially mounted on adapters, as mentioned by Fliburt.
I have no oscilloscope. I don't know if it's unstable. There's no indication of it with regards to current draw, offset, pops, clicks, interference, noise etc.
post #12 of 80
Quote:
Originally Posted by NelsonVandal View Post
As far as I can remember, I used basically the same values as suggested by Walt Jung except for a 10k pot and 100k (not 10M) from input to ground, and of course since the buffers are unity gain, there's no local feedback loop for them. Rs = 1k. I don't use output resistors outside the loop.
Hey there! Sorry about the delay in replying. You know, holiday festivities and all

Anyhow, be careful about the multiloop topology. You need to balance the bias currents. I haven't checked those values to see if they match, but the AD823 is JFET input so it's not as big a deal; the 797 has several thousand times the Ib.

Quote:
Filburt, how does AD797 sound to you? Do you find it "perfect" without any flaws at all? Do you find it colored, and if so in what way?
It should sound a lot like the 8599, if that's any help to you. The 8599 is two re-compensated 797s, much like how the 5532 is a recomp'd 5534. The 797 isn't unity stable without external comp; the 8599 is (likewise 5534/5532). I don't find it particularly coloured, and it has a touch of what people call warmth, although in reality I think people are used to output stages which have some third or fifth harmonic distortion problems and so 'neutral' has shifted such that what is actually very low distortion is now termed "somewhat warm" or something of that sort. I think the 797 is better overall, though it is the more difficult of the two to use.

Quote:
And Filburt, are you one of the few who think there is a JFET and a bipolar sound? I do.
Not in a specific sense (or "house sound" of sorts as seems to be the popular kind of term on head-fi), but the physical characteristics (and respective performance parametrics) of each type of device presents different challenges for a designer and may lead to different distortion spectra in a particular application.
post #13 of 80
Thread Starter 
AD797 doesn't sound a bit like AD8599. AD8599 has recessed mid and a warmer than neutral tonality.

Filburt, or anybody else, how should I make AD797 sound it's best? Are there any "known to work" schematics out there?
post #14 of 80
Quote:
Originally Posted by NelsonVandal View Post
AD797 doesn't sound a bit like AD8599. AD8599 has recessed mid and a warmer than neutral tonality.

Filburt, or anybody else, how should I make AD797 sound it's best? Are there any "known to work" schematics out there?
Honestly, since I don't know what your circuits look like, it's pretty difficult for me to help you on your troubles. I'll reiterate, however, the topology is such that these are very temperamental devices that require care in application. Performance can degrade in these chips, relatively speaking, especially rapidly absent proper application.

AD's own application notes should give you some insight into what you need to do. In order to obtain specified performance, you will probably have to use professionally made PCBs with proper layout, in addition to the schematic requirements.
post #15 of 80
Thread Starter 
I've read the data sheet, and I found an article by Walt Jung http://waltjung.org/PDFs/ADI_2002_Se..._Drivers_I.pdf and now this opamp sounds like it should. I just crumpled the paper and stuffed it in the amp.

No, I added a capacitor from pin 8 to pin 6 for "distortion cancellation". I had no 50 pF plastic, so I used 100 pF ceramic. Now the heavy coloration in the lower treble is gone, and it sounds just like in most reports - smooth and even a tiny bit on the dark side. It even sounds OK driving headphones without buffers, and even used unity gain in the ground channel. Why wasn't this cap built in since it seems to be mandatory?

I won't show pictures of it since the cap I use, a SMD1206, is soldered directly on the opamp pins. People allergic to hardwiring and BrownDogs would get a rash.
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › AD797 now sounds like it should