[AudioCats]

upgrading the op-amps in my Stax SRM-001 portable, I have installed 8-sip sockets and made some dip8->sip8 adaptors so now I can roll op-amps very easily. I will try 8599 and LM4562 for sure, but want to add one more option. should I go for 823 or 825 or 826? packaging style is not a problem, I can make adaptors to fit the sip-8 socket.

The op-amp installed in the baby stax right now is NE5532 (which is the only dual I got at this time). Even the old 5532 show some major improvement over the original NJM062L which is just a low power version of TL082.

I am leaning toward the 823 and 826 since they are dual and have SOIC-8 package, can be soldered directly on the Stax circuit board which has pads for SOIC replacements.

Thanks for any input.

 

[Pars]

I haven't used the 823 or 825 (are both of these dual? thought the 825 was a single), but have used the 826 in my CD player, and liked it. After this, I had rolled in LM6172, LM4562, an LT1091 or something like that (forget which one right now). I hadn't noticed it, but the LM4562 can be rather thin sounding... you need to be careful on bypassing, etc. I hadn't done any regression comparison... usually threw a new one in and if it sounded fine, left it. Upon getting a different preamp, something didn't sound right and I wound up putting the 826s back in, and that was when I noticed it... the 826s sounded much better than the 4562s. I didn't try playing around with the bypassing arrangement, as I switched to discrete I/V so no real recommendations there. Search carlos fm on diyhifi or diyaudio as he has some good recommendations regarding opamp bypassing.

 

[majkel]

AD825 kicks ass of most op-amps. There are just two better: ADA4899-1, being 12V max. supplied, which is uncomfortable, and another is OPA1632, with quite different pinout than a regular op-amp. It's dedicated for balanced applications but it's possible and easy to connect it for single-ended on both sides and then you're in heaven.

All other op-amps will sound like crap, with ADA4899-1 being slightly imperfect, AD744 too harsh and cold, AD825 not much direct sounding but still very nice, AD8045 a bit dull, and all the rest being sonic trash. Yes, I mean OPA637 as well.
For my former op-amp impressions - please do some search. I described lots of chips I have heard.

 

[NelsonVandal]

I've recently found the most neutral sounding opamp combination - Majkel's favourite AD825 in ground channel and AD8599 in left/right. Perfect! I use them without buffers and without power caps.

One day I'm going to get banned for this, but AD823 is the worst opamp I've heard, and I think it's unbelievable that it's used in "good" amps like the very hyped Graham Slee's. I've never listened to it for more than a couple of minutes at a time, so I've decided to let them burn in for a couple of houndred hours, not to throw the baby out with the bath water. I've used it in so many different topologies, and it always sounds harsh and unnatural. I have five of them, and all sound the same. I wonder if I've got pirated samples

, since some people like it, and all my conclusions might be wrong. But why pirate one of the cheapest?

It seems impossible to find a perfect opamp. It seems to me they have to be matched with different types in ground and L/R channel. LM4562 is almost a perfect match with LMH6654 and ADA4899-1 with OPA134. AD744 is acceptable with AD829. We should have a sticky with matching opamps, don't you think? What could be the perfect match with AD744 in my LISAIII-clone? It's a bit on the bright side due to the 744's. Majkel's right, they are a bit nasty like most of Analogs JFET opamps. 829 balances this a bit, but not enough. AD825 isn't a good match.

 

[majkel]

NelsonVandal, try replacing AD744 for OPA1632. Watch its pinout which is a kind of a crazy TI's idea not to repeat the pinout for single-ended.
OPA1632 is like more liquid, detailed and warmer AD744. Whatever you focus on, it's done in a better way. Another good and dark sounding is AD8099. As with the ADA4899-1, short pin 7 to 8 for lower offset current.

 

[NelsonVandal]

Quote:

Originally Posted by majkel /img/forum/go_quote.gif NelsonVandal, try replacing AD744 for OPA1632. Watch its pinout which is a kind of a crazy TI's idea not to repeat the pinout for single-ended. OPA1632 is like more liquid, detailed and warmer AD744. Whatever you focus on, it's done in a better way. Another good and dark sounding is AD8099. As with the ADA4899-1, short pin 7 to 8 for lower offset current.

Thanks for the tip, but what am I supposed to do with OPA1632? It's a differential amp. How do you connect it to be used as a single end opamp? 6 uA of bias current, isn't this a problem?

I never had any luck with AD8099. It was very unstable. How did you tame it?

 

[majkel]

When I use the OPA1632 or AD8099 as the virtual ground reference voltage buffer, I have no stability issues. In addition, AD8099 is the most stable when there is no compensation on dedicated pin or in the feedback loop.
Input currents are no issue when you use DC-compensated feedback loop, like i.e. 2x20k voltage divider and 10k negative feedback resistor.
OPA1632 single-ended goes like this referred to a regular op-amp:
pin 1 = pin 2 (inverting input)
pin 2 = leave floating (common mode voltage, Vs/2 when floating)
pin 3 = pin 7 (positive supply)
pin 4 = pin 6 (non-inverted output)
pin 5 = leave floating (inverted output)
pin 6 = pin 4 (negative supply)
pin 7 = leave floating (recommended) or tied to positive supply
pin 8 = pin 3 (non-inverting input)

 

[Mulo]

May try the LT1364: I think it's a great mixture of the opamps you (the OP) mentioned. It's easy to use, too, though not quite as easy as a FET opamp; you can use it without problems in many existing circuits. It's specified from as low as +/- 2,5V which should be good in your application.

 

[NelsonVandal]

I've had my eyes on LT1363 for a while, next time I order from Farnell or Digikey it'll be on my list. I'll probably use them unbuffered.

OPA1632 could be interesting. If it's cheap, I'll try it.

AD8099 isn't unity gain stable according to the data sheet, so I've never tried it in ground position. If I have any left I'll try it this way.

Thank you Mulo and Majkel for this information.