[Mad Max]

   
Thanks for the info. The stock op amps that come in my DAC + headphone amp do not have an actual ground connection, but some of the official upgrade op amps do have an actual ground wire and a place to connect them to earth ground. So I assume this means I should have a real ground I can use. The ground is really close to the op amp socket, so I was planning on crimping a ring terminal onto the end of one of the legs of each capacitor to connect to the ground screw and then solder the other leg of each capacitor to the proper power pin on my op amp adapter PCBs.
 

 
I do not recommend crimping terminal rings directly to cap leads, use a short length of wire instead.  I use 24awg or 22awg Mogami wire for ground leads, whichever I happen to have on hand.  These stranded wires will take more punishment than solid capacitor leads, unless you intend to swap in a different opamp only once and never again.
 
If the upgrade opamps in question are audio-gd's HDAMs, some people like those best with ungrounded decoupling and a 1uF Soviet K42-Y2 film capacitor across the power pins in place of the original EVOX MKT capacitors.  I got to try that out earlier this year, and Sun is even more fun and higher fidelity that way, but I sold off Earth and Moon long ago and can't try those with upgrade caps.  The little grey EVOX caps they normally have don't bring out the best sound in Sun at all, I have no doubts that the same applies to Earth and Moon.
 
As nice as audio-gd opamps are, though, they don't beat TI and ADI chips in sound like OPA827, ADA4627-1B, LME49990, or OPA1612, but they are indeed a nice way to flavor things up a bit.  I recommend Vishay MKP1837 decoupling caps if you go with any of those chips I mentioned, best film caps.  Or WIMA MKP4 if you can't source those.
Edit: MKP1839 is fine, too, just as good as 1837.

 

[EniGmA1987]

 
I do not recommend crimping terminal rings directly to cap leads, use a short length of wire instead.  I use 24awg or 22awg Mogami wire for ground leads, whichever I happen to have on hand.  These stranded wires will take more punishment than solid capacitor leads, unless you intend to swap in a different opamp only once and never again.
 
If the upgrade opamps in question are audio-gd's HDAMs, some people like those best with ungrounded decoupling and a 1uF Soviet K42-Y2 film capacitor across the power pins in place of the original EVOX MKT capacitors.  I got to try that out earlier this year, and Sun is even more fun and higher fidelity that way, but I sold off Earth and Moon long ago and can't try those with upgrade caps.  The little grey EVOX caps they normally have don't bring out the best sound in Sun at all, I have no doubts that the same applies to Earth and Moon.
 
As nice as audio-gd opamps are, though, they don't beat TI and ADI chips in sound like OPA827, ADA4627-1B, LME49990, or OPA1612, but they are indeed a nice way to flavor things up a bit.  I recommend Vishay MKP1837 decoupling caps if you go with any of those chips I mentioned, best film caps.  Or WIMA MKP4 if you can't source those.
Edit: MKP1839 is fine, too, just as good as 1837.

 


Good thought on the solid capacitor leads. I don't really plan to change out the op amps much, once I decide which one I like it will probably stay for years, but who knows how many swaps back and forth will go on over a couple weekends while I decide. lol

I have an OPA Earth and Moon, currently using the earth because I wanted a bit more bass. Before that I had used some AD797B's (much better than regular AD797A) that I really liked and they do sound better than the Earth, but don't have quite as much low end. I never even thought much about modding them too but Ill definitely be doing that now too. I had never really heard of the capacitors on the op amps before recently, but I decided to put them on my AD797B's and give them a go again, and I see after looking at the datasheet that Analog Devices actually recommends multiple bypassing using a 1-4.7uF tantalum and a parallel 0.1uf ceramic capacitor on each power leg to ground. I am going to use a 2.2uF Mundorf aluminum+oil film cap as well as a 0.1uf Mundorf silver|gold +oil film cap. Those would be each power pin to ground. Or my other thought is, upgrade both cap's to the silver|gold + oil's and only use two capacitors just between the two power pins and don't bother with the ground wire. Any thoughts on which way might be better?

I also bought some LME49990's after reading about their use from people in this thread, and I have some K42-Y2's coming as well with them. So Looks like I'll be auditioning between 3 op amps. Do you think it would be better to use the MKP1837's instead of the K42-Y2's specifically on the LME49990?

 

[imran27]

For my DIY DAC+AMP I've been reading design and some technical papers a lot (I'm Electronics Engineer by profession)
 
What is the technical/electronic counterpart of soundstage? Is it related to "Phase Response" of the AMP circuit? Or is it related to relative phase response of L & R or crosstalk between them?
 
One thing I truly admit and understand is that op-amps DO NOT have any particular sound! It is the circuit around them which is made for a different op-amp makes it sound "different". Just assuming that soundstage is related to phase response then replacing some op-amp with different ones will very likely change it's PR and hence soundstage.
 
Instead of changing op-amps we could try changing the FR & PR of the circuit to see how that affects.
 
I will likely build a test amp in order to test these things soon.
 
If someone would like to comment or extend upon this then please do, you're most welcome
 
EDIT: And hey @EniGmA1987 nice to see you here

 

[Mad Max]

Well, sticking an opamp in a circuit not designed for it is one way to get some thing different, already mentioned in this thread many times by now.  Distortion, feedback loops now having to work for the wrong opamp resulting in mild FR changes, and so forth.  Sticking instrumentation opamps in an audio circuit, this hobby is riot!  And a damned waste of time, but we still do it anyway.  It is fun despite being silly.  I hardly roll lately, though.
 
Phase has something to do with it, I've read.  The rest is psychoacoustics stuff no one gets, hahaha!
I find that soundstaging usually sounds better with higher-output current opamps, regardless of whether one is using low-impedance or high-impedance headphones.  Other than that, I've read here that output current is more important to low-impedance cans, higher voltage swing for high-impedance cans (preferably the +/-15V swing of desktop amplifiers).
 
PR = "pulse reponse" I assume?  From what I've read, which hasn't been very much, pulse response is adjusted with inductance.  More inductor-capacitor-inductor-capacitor layout going on in the PSU to "slow down" the power, giving equipment more of a "soft-start" trait, good for the lifespan of equipment.  But adding inductors for this and that anywhere seems to start adding coloration to the sound for the most part.  Caps in the signal path tend to have a similar effect of adding coloration, though adding bandwidth-limiting capacitors to opamp feedback resistors seems to do no harm - quite the opposite in fact, they sound even better.

 

[imran27]

  Well, sticking an opamp in a circuit not designed for it is one way to get some thing different, already mentioned in this thread many times by now.  Distortion, feedback loops now having to work for the wrong opamp resulting in mild FR changes, and so forth.  Sticking instrumentation opamps in an audio circuit, this hobby is riot!  And a damned waste of time, but we still do it anyway.  It is fun despite being silly.  I hardly roll lately, though.
 
Phase has something to do with it, I've read.  The rest is psychoacoustics stuff no one gets, hahaha!
I find that soundstaging usually sounds better with higher-output current opamps, regardless of whether one is using low-impedance or high-impedance headphones.  Other than that, I've read here that output current is more important to low-impedance cans, higher voltage swing for high-impedance cans (preferably the +/-15V swing of desktop amplifiers).
 
PR = "pulse reponse" I assume?  From what I've read, which hasn't been very much, pulse response is adjusted with inductance.  More inductor-capacitor-inductor-capacitor layout going on in the PSU to "slow down" the power, giving equipment more of a "soft-start" trait, good for the lifespan of equipment.  But adding inductors for this and that anywhere seems to start adding coloration to the sound for the most part.  Caps in the signal path tend to have a similar effect of adding coloration, though adding bandwidth-limiting capacitors to opamp feedback resistors seems to do no harm - quite the opposite in fact, they sound even better.

PR meant Phase Response. Soundstaging definitely needs to have a trait in the circuit that changes with different op-amps. I highly suspect it's the phase response of the amp section, but not sure until I test it once and for all. Any psychoacoustic effect is a direct result of some particular change in output of the amp. Two amps having exactly similar phase and frequency response (I tend to trust Step Response more) will likely sound the same.
 
One more way of making a good amp could be adjust it's step response to have a very low rise time and overshoot but must be stable. This will ensure that amp is perfectly stable yet it can follow the input and recreate it at the output very well.
 
The questions regarding power supply noise and by-passing have been somewhat answered and covered in whatever @EniGmA1987 asked.

 

[Mad Max]

Good thought on the solid capacitor leads. I don't really plan to change out the op amps much, once I decide which one I like it will probably stay for years, but who knows how many swaps back and forth will go on over a couple weekends while I decide. lol

I have an OPA Earth and Moon, currently using the earth because I wanted a bit more bass. Before that I had used some AD797B's (much better than regular AD797A) that I really liked and they do sound better than the Earth, but don't have quite as much low end. I never even thought much about modding them too but Ill definitely be doing that now too. I had never really heard of the capacitors on the op amps before recently, but I decided to put them on my AD797B's and give them a go again, and I see after looking at the datasheet that Analog Devices actually recommends multiple bypassing using a 1-4.7uF tantalum and a parallel 0.1uf ceramic capacitor on each power leg to ground. I am going to use a 2.2uF Mundorf aluminum+oil film cap as well as a 0.1uf Mundorf silver|gold +oil film cap. Those would be each power pin to ground. Or my other thought is, upgrade both cap's to the silver|gold + oil's and only use two capacitors just between the two power pins and don't bother with the ground wire. Any thoughts on which way might be better?

I also bought some LME49990's after reading about their use from people in this thread, and I have some K42-Y2's coming as well with them. So Looks like I'll be auditioning between 3 op amps. Do you think it would be better to use the MKP1837's instead of the K42-Y2's specifically on the LME49990?

 
Tantalums aren't recommended for audio, at least by more experienced audiophiles with electronics knowledge, usually.  I have spare K42s to try on chip opamps but haven't gotten around to trying them out.
 
Try both just two-cap scheme or all four and see which you like better.  You can find pics of both types being used by different people in this thread as well as the audio-gd FUN and Compass (not the Compass 2) threads.
 
As far as AD797, I haven't yet tried adding a 50pF film cap across pin 8 and pin 6 on each chip, that's supposed to minimize a 50pF internal output capacitor.  I suppose I would have to go with WIMA FKP2 10% 47pF caps (10% and not the 5% ones, so- they will vary individually between ~43pF to ~51pF), buy a bunch and get a pair that both measure 50pF.
Edit: and 797's low end should improve with extra bypass caps, it's really one of the best chips around.

 

[00940]

  Tantalums aren't recommended for audio, at least by more experienced audiophiles with electronics knowledge, usually.

 
I'd rather say that tantalums aren't usually recommended for audio by audiophiles with too little electronics knowledge to implement them properly. They're perfectly fine if you know how to use them. Actually, as long as you don't do anything stupid such as forgetting to derate their voltage ratings and provide for current limiting.

 

[Mad Max]

   
I'd rather say that tantalums aren't usually recommended for audio by audiophiles with too little electronics knowledge to implement them properly. They're perfectly fine if you know how to use them. Actually, as long as you don't do anything stupid such as forgetting to derate their voltage ratings and provide for current limiting.

 
I suppose that would include resistors and other stuff?
Edit: I hope that doesn't require adding a dozen more components, hoho!

 

[00940]

   
I suppose that would include resistors and other stuff?
Edit: I hope that doesn't require adding a dozen more components, hoho!

 
Not really. It would include considering what capacitors you put in //, their respective esr, capacitance, etc. It could include using tantalum mostly for RC filters and avoid places which could confront the caps to extremely fast charge-discharge.
 
To back up the tantalum's audiophile credentials, Ray Samuels' products are full of them.

 

[imran27]

Does anyone know if it makes a difference in using an Instrumentation Amplifier after I/V stage instead of a simple Differential Amplifier, soundwise of course since technically IA is always better than the latter?

 

[00940]

  Does anyone know if it makes a difference in using an Instrumentation Amplifier after I/V stage instead of a simple Differential Amplifier, soundwise of course since technically IA is always better than the latter?

 
disclaimer: answering from a technical point of view.
 
There is little point in going for a instrumentation amplifier at this point. The output impedance of the I/V stages are usually well matched and quite low on one hand and, on the other hand, the extra load the differential amplifier puts on the I/V opamps isn't a big deal considering how strongly loaded the I/V opamps already are.
 
If I was trying to improve the typical opamp based I/V stage, I'd rather put buffers inside the feedback loop of the I/V opamps. jcx (among others) has posted about it in the past (an example here).

 

[imran27]

   
disclaimer: answering from a technical point of view.
 
There is little point in going for a instrumentation amplifier at this point. The output impedance of the I/V stages are usually well matched and quite low on one hand and, on the other hand, the extra load the differential amplifier puts on the I/V opamps isn't a big deal considering how strongly loaded the I/V opamps already are.
 
If I was trying to improve the typical opamp based I/V stage, I'd rather put buffers inside the feedback loop of the I/V opamps. jcx (among others) has posted about it in the past (an example here).

I was actually looking at it in terms of CMRR. IA's typically have a very good CMRR that can decrease the crosstalk between L & R. That in turn will likely improve imaging/positioning by quite a bit since L & R are more isolated now. You might think how is CMRR related to L/R isolation? They will have much better silence (probably coupled with lower noise/higher SNR).
 
Just trying to think other benefits it might have, or disadvantages otherwise
 
EDIT: I'm asking and writing in context of designing a completely DIY DAC+AMP for portable use, probably the size of Fiio A17, may be just a little bulkier will also be OK.

 

[00940]

  I was actually looking at it in terms of CMRR. IA's typically have a very good CMRR that can decrease the crosstalk between L & R. That in turn will likely improve imaging/positioning by quite a bit since L & R are more isolated now. You might think how is CMRR related to L/R isolation? They will have much better silence (probably coupled with lower noise/higher SNR).
 
Just trying to think other benefits it might have, or disadvantages otherwise
 
EDIT: I'm asking and writing in context of designing a completely DIY DAC+AMP for portable use, probably the size of Fiio A17, may be just a little bulkier will also be OK.

 
In the context of a I/V, the CMRR is pretty much irrelevant to crosstalk (have a look at the numbers achieved by the pcm1794 datasheet circuit). The only benefit you gain is wrt lower noise as you cancel the noise coming from the DAC analog supply (and some H2 cancellation). In the case of the pcm1794 (to take an example) however, the single-ended performances of the IC are already so good that you don't need much, a few dB are all you need to push you into the crazy low numbers.
 
And, honnestly, for a portable context, I'd look at voltage out DAC such as AK4399 or WM8741 (if you want top performances, otherwise pcm5102 is really easy to use). You can strongly reduce your power consumption by making the output stage more simple without performance loss.

 

[imran27]

Great suggestion bro.

I'll take that, I guess DragonFly uses WM8741 too.

I was thinking of LME49722 for AMP stage or OPA1612

 

[00940]

Great suggestion bro.

I'll take that, I guess DragonFly uses WM8741 too.

I was thinking of LME49722 for AMP stage or OPA1612

 
The Dragonfly uses the ES9023 actually. The wm8740 is used by Amb in his gamma 1.5 for example. 
 
Those opamps are good audio opamps. The opa1612 swings closer to the rails at first sight, that might be a factor for a portable design.