Of course, I have used a similar circuit for the last 40 years. BUT, are IC's a true and equal replacement for your and my discrete designs? I am all for better IC's, in fact I have to design them in some of products, but, bottom line, I still design discrete designs for my 'best' audio products.
If you add a discrete current buffer to an op amp, you can still have your >20 mA of output bias current.
I don't think that's the the problem. The problem, as I see it, lies in the fact that by the time you've added all the discretes you need for a truly SOTA setup, you discover that the op amp is really only a differential input stage, i.e. that in fact it replaces just 2 discrete devices.
Now, that's hardly practical, isn't it, Pavel?
I don't think that's the the problem. The problem, as I see it, lies in the fact that by the time you've added all the discretes you need for a truly SOTA setup, you discover that the op amp is really only a differential input stage, i.e. that in fact it replaces just 2 discrete devices.
Now, that's hardly practical, isn't it, Pavel?
Now everyone, I would like to say a few words regarding Ron Quan's ongoing research on IC and other audio amps.
Ron worked in both the audio and video world, including Sony, Ampex, and many other advanced Silicon Valley video companies over more than 3 decades. He and I have discussed the 'problems' with IC's and even some discrete products for decades. You see, we actually heard differences over the last 40 years and we would like the situation to improve. It wasn't just our ears, but others as well.
Ron decided to add his expertise in video design and measurement, with audio measurements to find what we have heard over the decades. Now, this 'might' be why Ron often uses older IC's (most of which are used today) as measurable examples of exotic distortions, first denoted by Otala in 1978 or so, when we came to understand that TIM could not be the ONLY problem in audio amps and preamps.
Ron has had to custom build much of his own test equipment to make these tests, and they still could be refined even further over time. Now, he presents the measured problems in typical IC's, many of which are used by people on this website. Please remember that the same IC designers would have stated that even these earlier designs were good enough for audio, using SMPTE IM distortion ( the standard measurement before TIM was shown to exist),and THD as well. Ron is just taking it one step further, showing that our concerns about crossover distortion, etc is more easily measured by his methods, than with standard THD.
Ron worked in both the audio and video world, including Sony, Ampex, and many other advanced Silicon Valley video companies over more than 3 decades. He and I have discussed the 'problems' with IC's and even some discrete products for decades. You see, we actually heard differences over the last 40 years and we would like the situation to improve. It wasn't just our ears, but others as well.
Ron decided to add his expertise in video design and measurement, with audio measurements to find what we have heard over the decades. Now, this 'might' be why Ron often uses older IC's (most of which are used today) as measurable examples of exotic distortions, first denoted by Otala in 1978 or so, when we came to understand that TIM could not be the ONLY problem in audio amps and preamps.
Ron has had to custom build much of his own test equipment to make these tests, and they still could be refined even further over time. Now, he presents the measured problems in typical IC's, many of which are used by people on this website. Please remember that the same IC designers would have stated that even these earlier designs were good enough for audio, using SMPTE IM distortion ( the standard measurement before TIM was shown to exist),and THD as well. Ron is just taking it one step further, showing that our concerns about crossover distortion, etc is more easily measured by his methods, than with standard THD.
Scott I'm not sure I get your drift. As I read the paper, the 'staircase' was just a handy way to use five different offset values. As I understand, Ron would measure the THD and whatever at the plateau of each offset, so the staircase isn't really a 'dynamic'input signal.
Or did I miss something?
Jan
The offset is at the output but the input tanh() transfer function is what is talked about. The tanh() input transfer function of the op-amp (again any modern op-amp) is not exercised by the steps i.e. the input error signal only will be at the higher AC frequency and the phase error will be from other effects like VAS junction capacitance non-linearity or non-linear output stage output impedance referred back to the input. Throwing in results of bad 1970's op-amps only creates a smoke screen.
The concept that the staircase is causing transient offset at the input is wrong. Just as wrong in fact as all that asymmetrical music waveform nonsense that pops up occasionally.
I'm afraid I disagree with both.
I have heard a few Scandinavian products using op amps encircled by discrete devices, and they sounded very good to me. Looking at their schematics, it turns out that the op amp is almost only acting as an input stage, since most of the work is subsequently done by discrete devices. As if the op amp is used as a supermatch transistor only.
On the single ended vs. fully complementary input stage, I own two pieces which use single diff. amps for their input stages, one only a BJT discrete pair, the other a FET and BJT cascode, which sound to me a hell of a lot better than most fully complementary amps I have heard. That's "most", not all. Which to me suggests that we're back to square one - it's not the topology, it's how you use it.
Which is why the simple single diff amp mentioned above comes from the same source which provided me with a fully complemetary power amp, which also sounds good.
OK, let me explain what Ron's forced offset using a staircase does in practical reality: It shows that almost all IC's have crossover distortion AND this should be addressed, either by the manufacturer, or by deliberately current offsetting the IC output with an external current source. We have been doing this for years, knowing subjectively that it almost always sounds better. This includes the 4562, etc, parts that are popular in new designs today.
This is well known and easily quantified as I showed almost 25yr. ago in a test that separates the crossover from all other sources of distortion. Published in the AES also in fact, it showed also the relative contributions of the major distortion components (in a simplified manner).
A while back I suggested 36V locomotive starting batteries, maybe 200mV sag at 0-1000A transient (assuming the wire has enough 0's in gauge).This is over the top DIY after all.
Wonder how quiet they are. Will need 4 of them. But, now I can use a Prius battery pack and get dealer warentee with it. I wont even take them out of the car. Plug-N-Play. Yet, another use for my welding cables. Sounds like a really good idea to me. Wonder why no one else thought to do it long before now. [Can I get a patent?]
=RNM
Last edited:
Lead acid is documented somewhere, but I would challenge the raw noise vs any active regulator.
We've known for a long time the issues of the IC's OPS and its limitations and sometimes what can be done about it or improve on it. Its major failings have been lightly covered here.
Something I think would be new and interesting is more in-depth on Common-Mode distortion.
Scott - Would you say somethings about some ways unintended cm distortion (beyond the usual descriptions) manifests itself?
THx-RNMarsh
Something I think would be new and interesting is more in-depth on Common-Mode distortion.
Scott - Would you say somethings about some ways unintended cm distortion (beyond the usual descriptions) manifests itself?
THx-RNMarsh
Last edited:
Maybe start with the parameters which could be degraded by different methods of externally increasing bias to the OPS of an IC opamp?
What conditions in a 'system' or in a topology which cause CM to be converted to DM and visa-versa? Which conditions cannot be reduced by nfb?
-RNM
What conditions in a 'system' or in a topology which cause CM to be converted to DM and visa-versa? Which conditions cannot be reduced by nfb?
-RNM
Last edited:
Richard, while common mode distortion is significant, it is not what Ron is trying to show. He is showing 'graphically' how crossover distortion exists in almost every IC, when loaded down to 500 ohms (not 100 ohms that would be worse) and this is caused mostly by a weak output stage, with low Iq. We usually don't MEASURE it with thd measurements, because of the needed instrument bandwidth that allows noise to cover any distortion artifacts, but the crossover distortion region is where we normally listen to the IC, as it is at audio working levels, not anywhere near full output.
Nothing impeaches to use active regulator after a battery source of power.
OK. The point is further made re OPS distortion under low Z loads... and showing that it is cross-over distortion. BTW - wouldnt FFT show this also? It is what most people use today as the cost and accuracy has improved greatly over time. And, to increase the Iq in an IC package may make some things better but others worse. So, we go out-board with the OPS if we need to supply more current than IC can do cleanly. I would never use an opamp at less than 10K load. A few are spec'ed for 1K.
I am just rushing the story along.
Thx for making it even more clear about Ron's paper. Now I dont have to read it (?)
What else about OPS of opamps?
Check. Next? CM distortion?
THx-RNMarsh
Last edited:
But with distortion levels at -120 dB worst case into 600 Ohms at 1V out (see AD797, LME49990 et al) I would argue it's absolutely inaudible.
Separately, bootstrapping the OPS to 600 uA and then using an in loop class A buffer will get you another 10 to 20 dB lower than this.
But, who can hear distortion at these levels on modern opamps? They are superb by any measure.
Separately, bootstrapping the OPS to 600 uA and then using an in loop class A buffer will get you another 10 to 20 dB lower than this.
But, who can hear distortion at these levels on modern opamps? They are superb by any measure.
- Status
- Not open for further replies.