This is really something of a speculative long-shot but is this stage one where the opamp's inverting input goes to the volume pot's wiper? If so the wire between this pin and the wiper might be acting as antenna. You might try some RF low-pass filtering into the -ve input pin.
This illustrates the fallacy of assuming what owdeo's hearing is THD. He's not listening to static test tones when he says the design doesn't deliver aural satisfaction. Perhaps R A Belcher's 1978 Wireless World article would be a good read here?
There are many reports that the LM4562/LME49720 sounds ultra detailed, yet is ultimately fatigueing for longer term listening.
Some who replaced the OPA2134 in the Silicon Chip Studio Series Stereo Preamplifier with the LM4562, ran into stability problems as it didn't like the added capacitance of the short shielded cable to the volume control potentiometer.
Fitting a 100 ohm resistor at it's output rectified that problem.
Personally, I find that the LM4562HA/LME49720HA in the metal can sounds better than the DIP8 version. I don't know whether it comes down to better thermal management or the additional screening of the metal can.
Many also fit a TO5 heatsink to the HA version as it can get quite warm, despite having the identical chip internally. I agree that many opamps do have a "family" sound about them, just as many Audiophile electrolytic capacitors do.
For that reason, in my modified version of the S.C. designed DAC I mix opamp types. I have had quite a few reports from members of another forum that they heard similar improvements by doing this too.
Alex
Some who replaced the OPA2134 in the Silicon Chip Studio Series Stereo Preamplifier with the LM4562, ran into stability problems as it didn't like the added capacitance of the short shielded cable to the volume control potentiometer.
Fitting a 100 ohm resistor at it's output rectified that problem.
Personally, I find that the LM4562HA/LME49720HA in the metal can sounds better than the DIP8 version. I don't know whether it comes down to better thermal management or the additional screening of the metal can.
Many also fit a TO5 heatsink to the HA version as it can get quite warm, despite having the identical chip internally. I agree that many opamps do have a "family" sound about them, just as many Audiophile electrolytic capacitors do.
For that reason, in my modified version of the S.C. designed DAC I mix opamp types. I have had quite a few reports from members of another forum that they heard similar improvements by doing this too.
Alex
Hi Bonsai,
The differences in absolute SQ are certainly not night and day as you correctly surmise. They are subtle when flicking between it and another preamp while listening. But I've found that longer listening is where it counts, and with the LM4562s in circuit the enjoyment factor of the music is severely diminished. The 5534/2 at least has a "relaxed" presentation with a nice euphonic bloom in the lower mids that is musically pleasing to me, and makes its slight graininess in the top end easy to forgive. The LM4562 just sounds sterile and slightly hard, with a very slight edge that makes it uncomfortable to listen to for long periods.
In terms of trying to characterise the sound of the Self preamp, as an example I was listening to John Lee Hooker's "The Healer" album the other night and it showed the problem very clearly. On track 4 (Chillin' Out with Canned Heat), with my other opamp preamp there is a clear soundstage with great depth behind the speakers despite the complexity and volume of the musical group. Switching to the Self preamp, the depth is greatly reduced to the point where is really just sounds 2D and some of the instruments are harder to follow. On track 8 (Rockin' Chair) Hooker is playing solo in some sort of hall with a very reverberant acoustic. On the other opamp pre the acoustic is huge and the sound utterly realistic. Switching to the Self pre and the acoustic is only a vague impression and it's harder to tell what sort of environment he's playing in.
These observations apply regardless of whether 5532s or LM4562s are installed, so I tend to think it comes down to the circuit configuration.
In terms of instrumental timbre sometimes it sounds cleaner and with other recordings it's less clean - the recording quality is obviously a factor. But with LM4562s in the Self preamp I found brass sections to sound a little too bright and string sections to loose their creamy smoothness to sound a bit thinner with orchestral works.
Overall I would not complain about the timbre or balance of the Self preamp with (JRC) 5532s installed. The problem is the slightly over-processed or compressed character that also diminishes the subtle details of acoustics etc, and that's where a lot of the magic in many recordings that allows the suspension of disbelief lies.
I suspect this is not going to help you speculate because you are assuming that the problem comes down to one of the easy-to-measure (if you have equipment
) parameters like THD or noise.
This is illustrative of the subjectivity of listening reports. One man's 'ultra-detailed' is another man's 'digititus'
In my experience 'detail' is actually intermodulation distortion induced by RF, something I'm keen to eliminate. 'Detail' is what went away when I started listening to TDA1543 DACs rather than AD1955 ones. At first, I missed all this 'detail' but eventually I realized it was an artifact. It takes time spent listening to adjust to electronics which lacks 'detail' but the rewards in longer-term satisfaction are worth it, IME.I'm expressing my opinions and don't need to prove anything. You're only demonstrating that you're out of your depth. Higher harmonics are much more objectionable and clearly audible though speakers that have higher levels of lower harmonics. And if you'd bothered to read my earlier posts my speakers are B&W801, arguably one of the best ever produced. Quoting from the spec sheet:
Distortion at 95dB at 1m from 100Hz - 20kHz
2nd: <0.6%
3rd: <1%
4th: <0.1%
5th: <0.15%
No there's a unity gain non-inverting buffer attached to the pot wiper as this need to see a high Z to maintain the gain law. This feeds the main inverting gain stage and the pot wraps (variable) feedback around the whole thing.
What do you suggest? Ferrite bead and/or cap to ground at +ve input? What sort of f3 do you think is needed for RF immunity?
I've had success with using small (1206) common-mode chokes as inductors. They measure about 6uH (from memory) with DCR under an ohm when both windings are paralleled. Ferrite beads have lower inductance but get more lossy at higher frequencies - they might work well. I use strings of beads to get higher inductance - the TDK ones designed for power supplies (not signal, these have too high DCR to run strings of them).
Try with just series (inductor or string of beads) first without any cap to ground. That's because the local ground might also be contaminated at RF. Probably cap to -ve rail (as direct as possible to pin4 of the opamp) would be superior to any ground. Go for an f3 in the moderate hundreds of kHz if you can but I'm not sure how critical this is.
Try with just series (inductor or string of beads) first without any cap to ground. That's because the local ground might also be contaminated at RF. Probably cap to -ve rail (as direct as possible to pin4 of the opamp) would be superior to any ground. Go for an f3 in the moderate hundreds of kHz if you can but I'm not sure how critical this is.
Glad to hear we are still agreed
I tried the LME49710 to replace 5534s in my other opamp preamp and I quite liked the result, though overall the 5534s are more musically satisfying. This is supposed to be the exact same device in a single package so I don't understand why the LM4562 was so much more irritating, guess it is as I speculated earlier that different circuit configurations tend to affect how much the opamp's character comes into play. But it seems from my experience with other devices too that singles tend to sound better than duals, even when they are exactly the same device internally (unlike 5534 and 5532).
Ok thanks for the suggestions. I'll see what I can do. Unfortunately the opamps are socketed which means they won't be as close to the device pins as ideal.
I must admit I'm not entirely convinced this will make much difference but I'm interested to try it.
I haven't said such thing
... you are asuming that I am talking about THD.But... I guess you don't believe that distortion only shows under static test tones, don't you? Anyway... it seems to be a long way from noise to music/voice signals (I don't know the correlation between them), and this doesn't help to recover the pre-amp distortion from the loudspeaker one.
I've had various active speakers (NE5532 based in the XOs) with relatively fatiguing sound that have been rendered enjoyable to listen to with this trick of protecting inputs from RF. But ferrites do weird things over time - they somehow need to 'break-in'. So my modded speakers would sound fine for a day or so but then revert back to the old fatiguing sound after that, for a day more. Eventually they settled down to the relaxing presentation.
I haven't said such thing
Yes - when you speak of loudspeaker distortion (1% was the figure you mentioned), what else could you be speaking of? Do enlighten me if I'm mistaken, I'm here to learn.
Indeed I have no such belief.
May I enquire what your point is here?
Excuseme, but you asked for a prove, not me.
OK, so.. you can discern between 0.1% and 0.104%. Nice!
ezavalla,
I think that sometimes it is not harmonic distortions that we are hearing at a low level but injections of noise, whether this is harmonically related or just random noise. This can be that small graininess that people speak of or even an injection of very high frequency rf that can drive you crazy even if you have a hard time perceiving why you are unhappy with the sound. It seems that the higher the resolution of the speakers the more noticeable this becomes as it is not covered by other speaker distortions that hide this added noise. I can't say that this is what owdeo is talking about, but I know that I have had a hard time finding this type of irritation when it has been present but not obvious. In my case it was high frequency hash caused by a titanium tweeter that could only be seen in a waterfall plot doing individual component testing.
I think that sometimes it is not harmonic distortions that we are hearing at a low level but injections of noise, whether this is harmonically related or just random noise. This can be that small graininess that people speak of or even an injection of very high frequency rf that can drive you crazy even if you have a hard time perceiving why you are unhappy with the sound. It seems that the higher the resolution of the speakers the more noticeable this becomes as it is not covered by other speaker distortions that hide this added noise. I can't say that this is what owdeo is talking about, but I know that I have had a hard time finding this type of irritation when it has been present but not obvious. In my case it was high frequency hash caused by a titanium tweeter that could only be seen in a waterfall plot doing individual component testing.
@owdeo - I know this is not adding anything technical to the discussion, but since the 2012 Pre was released over a year ago, have you actually built this version of the pre-amp? I know you said in a previous post, and I quote, "I've been doing some blind testing of my (D.S. '96) preamp against a couple of others". I don't know what the similarities are between the 2012 pre and 96 pre, but isn't it fair to say you are comparing apples to oranges, when comparing the 96 pre to your 'other' preamps? But have you actually used and listened to the 2012 Pre? Isn't it also fair to say that between 1996 and 2012, Doug Self was able to incorporate new techniques that make comparing the 96 pre to other preamps an unfair evaluation - at least in this thread where I thought we were discussing the 2012 Pre?
I'll go back and take my seat with the audience now.
Rick
I'll go back and take my seat with the audience now.
Rick
That alone is going to kill sound quality: I wouldn't take any notice of what a component sounded like until that severe "weakness" had been resolved ...
Glad to hear we are still agreed
I tried the LME49710 to replace 5534s in my other opamp preamp and I quite liked the result, though overall the 5534s are more musically satisfying. This is supposed to be the exact same device in a single package so I don't understand why the LM4562 was so much more irritating, guess it is as I speculated earlier that different circuit configurations tend to affect how much the opamp's character comes into play. But it seems from my experience with other devices too that singles tend to sound better than duals, even when they are exactly the same device internally (unlike 5534 and 5532).
Absolutely right! I also find single opamps to sound better than the equivalent duals and it is very consistent experience. Now I use simple dual mono preamp (no tone controls) with 5534 and overall it's the most satisfying sound. But if you can find them, NJM 2114 are like two 5534 in dual package. It seems that they are a kind of decompensated 5532. When you compare schematics (published for both types) you shall see that one compensation cap is removed compared to 5532 circuit. They sound excellent, have ample current output (60mA), which helps. They are used even in expensive Marantz CD players. That's what I like most for duals. I bought 100pcs from Profusion plc.
When necessary (for experimentation) I use turned pin IC sockets to minimize stability problems, but in this simple preamp 5534 are soldered on board.
Hi Bonsai,
The differences in absolute SQ are certainly not night and day as you correctly surmise. They are subtle when flicking between it and another preamp while listening. But I've found that longer listening is where it counts, and with the LM4562s in circuit the enjoyment factor of the music is severely diminished. The 5534/2 at least has a "relaxed" presentation with a nice euphonic bloom in the lower mids that is musically pleasing to me, and makes its slight graininess in the top end easy to forgive. The LM4562 just sounds sterile and slightly hard, with a very slight edge that makes it uncomfortable to listen to for long periods.
In terms of trying to characterise the sound of the Self preamp, as an example I was listening to John Lee Hooker's "The Healer" album the other night and it showed the problem very clearly. On track 4 (Chillin' Out with Canned Heat), with my other opamp preamp there is a clear soundstage with great depth behind the speakers despite the complexity and volume of the musical group. Switching to the Self preamp, the depth is greatly reduced to the point where is really just sounds 2D and some of the instruments are harder to follow. On track 8 (Rockin' Chair) Hooker is playing solo in some sort of hall with a very reverberant acoustic. On the other opamp pre the acoustic is huge and the sound utterly realistic. Switching to the Self pre and the acoustic is only a vague impression and it's harder to tell what sort of environment he's playing in.
These observations apply regardless of whether 5532s or LM4562s are installed, so I tend to think it comes down to the circuit configuration.
In terms of instrumental timbre sometimes it sounds cleaner and with other recordings it's less clean - the recording quality is obviously a factor. But with LM4562s in the Self preamp I found brass sections to sound a little too bright and string sections to loose their creamy smoothness to sound a bit thinner with orchestral works.
Overall I would not complain about the timbre or balance of the Self preamp with (JRC) 5532s installed. The problem is the slightly over-processed or compressed character that also diminishes the subtle details of acoustics etc, and that's where a lot of the magic in many recordings that allows the suspension of disbelief lies.
I suspect this is not going to help you speculate because you are assuming that the problem comes down to one of the easy-to-measure (if you have equipment
I always try to listen to those things I listed to characterize the sound of a piece of gear. If you are losing sound stage like that, well that is a bummer, because its one of the most critically important aspects IMV. I have a fantastic Live Chick Corea/Steve Gadd/Christian McBride CD and even though the top end is not particularly extended, the spacey sound stage makes this a superlative CD. When I play it through my Marantz, its lost.
Throw that active volume control out and see if that improves it. I am assembling my new preamp board protos now and I've avoided the active volume control. I have an aversion to summing junctions for audio (although of course the Baxandall tone control is a summing circuit) - anything appearing between the summing junction and the op-amp input gets amplified by the full loop gain - in the case of a 4562, this can be 100dB or more depending upon the closed loop gain. So Abraxalitos advice to screen well, assuming you keep the active gain stage, is good. One other thing you might consider is to force the active gain stage op-amp into class a by loading it to either the + or - rail with a current source.
Agreed, it would at least add some perspective to the comments as to date, we have no details of the SQ of the reference preamp - not even a specification, reference to accessible details or...zip! Just some comments.
I think it would be helpful to introduce a common, widely known design as a reference rather than bandy words loaded with personal subjectivity about these extremely rare devices being compared. I'd be fairly certain only one or two here have heard of, let alone listened to the reference. I suspect it's a local, 1980's Tilbrook DIY design based on hefty discrete opamps that few will remember or know of, but I could be wrong. Actually, few will be familiar with even Self's 1996 design, other than to have sighted extracts of the documentation.
Surely it makes sense to clarify rather than muddy the waters but if we delve into subjectivity, we can only learn about ourselves and project our conclusions onto others. This does nothing for the DIY community other than say an individual is happy or not with what they hear in their audio systems. If preamps make so much of a difference to SQ when operating in their sweet spots, one or both of those devices is in a mess with it's quite measureable parameters. It's time they were defined or the devices simply eliminated from the system to determine what really goes on.
Without any objectivity here, there is nothing to learn in the argument of it.