Hi John,
You can't directly compare standing currents equated to performance between discrete and integrated solutions. There are "tricks" that work when devices are on the same die that cannot be duplicated in a discrete implementation.
Besides, you don't like op amps and things like this come down to your perception on sound quality. The best thing for you to do is simply try them out. That will be the only time you will *know* for sure.
Besides, Scott may have also signed and NDA and is prevented from commenting!
-Chris
You can't directly compare standing currents equated to performance between discrete and integrated solutions. There are "tricks" that work when devices are on the same die that cannot be duplicated in a discrete implementation.
Besides, you don't like op amps and things like this come down to your perception on sound quality. The best thing for you to do is simply try them out. That will be the only time you will *know* for sure.
Besides, Scott may have also signed and NDA and is prevented from commenting!
-Chris
I doubt that Scott can't tell me. If so, he can tell me this. I AM designing one of these 2 IC's in an upcoming design. I need to know, but I might already know enough.
I am shocked! Shocked, I tell you, after all that I have given to this website in technical information if I can't get the peak beta current of an output device online.
I am shocked! Shocked, I tell you, after all that I have given to this website in technical information if I can't get the peak beta current of an output device online.
Hi John,
I was mostly poking fun at the situation. The fact remains that you will not know which one to use until you audition both. Could be one likes your layout better than the other.
Given the cost of these, just order a couple and try them out. You can infer the standing current by loading the output stage with resistance or a current source. It's the entire system and its response to your setup that is important. The actual standing current may or may not be important.
-Chris
I was mostly poking fun at the situation. The fact remains that you will not know which one to use until you audition both. Could be one likes your layout better than the other.
Given the cost of these, just order a couple and try them out. You can infer the standing current by loading the output stage with resistance or a current source. It's the entire system and its response to your setup that is important. The actual standing current may or may not be important.
-Chris
Hi John,
In that case, use a variable current source and tune it in. You can easily set up a current mirror for experimentation and play that way. Once you have the magic numbers, you may be using smts on a small daughter board. A more coarse attack may be a standard LED/BJT setup with a 1 K trimmer in the BJT emitter to set current, all sitting on the negative rail.
Also, try it without the current source. You may find that it isn't required for one of your choices. Happy day! BOM just got smaller if that is the case.
-Chris
In that case, use a variable current source and tune it in. You can easily set up a current mirror for experimentation and play that way. Once you have the magic numbers, you may be using smts on a small daughter board. A more coarse attack may be a standard LED/BJT setup with a 1 K trimmer in the BJT emitter to set current, all sitting on the negative rail.
Also, try it without the current source. You may find that it isn't required for one of your choices. Happy day! BOM just got smaller if that is the case.
-Chris
jcx said:
Incindentally I have worked on intensive opamp tests recently, with respect to audio application. The AD797 and LT1028 perform well, the 797 is faster and more suitable for audio. The OPA211 is NOT, its input stage is sensitive to detection and demodulation. The same problem is for the new LM series, LM47910 gave very disappointing result, and is no way a choice for audio for me.
PMA said:
I recently replaced 2 modules on a ML1 for a friend (Phono and Line). I wanted to do something simple, OPA based otherwise it would have been better to redo the entire thing with something closer to the Vendetta. I made the circuit as follows:
An externally hosted image should be here but it was not working when we last tested it.
I tried AD797, AD743, LT1028, OPA134 and latter OPA627 give the best sounding results, all the others have a more pronounced SS or feedback…footprint (what you will). It is important to sink from the output around 5 to 7mA for that (carlosfm trick). It does not work equally with the others devices.
PMA said:
Care to expand a bit? What exactly is "sensitive to detection and demodulation"? And how was LME49710 dissapointing? As far as I know, LME49710 is a single version of LM4562. Is LM4562 "dissapointing" for you as well?
john curl said:
Well if I believe Jim Williams the LT1028 was designed as a breadboard and in his words "several decade boxes" to tweak the poles and zeros. I'm afraid I used only simulation.
Output devices are typically run well below peak beta. A general purpose op-amp running class A would have a very narrow appeal. If you exercise the crossover notch (in a normal class A/B opamp) as you walk the DC output current in and out you should be able to make a guess at the standing current.
BTW some of our latest high speed amps are testing our ability to measure distortion as they approach -120dB at 100MHz the low frequency distortion is virtually nonexistant.
Well, yes, but 80 dB noise gain (feedback released of 80dB).
The output stage CCSs are only 400uA and 600uA, and test load is about 1k, so output is NOT in class A.
The output stage CCSs are only 400uA and 600uA, and test load is about 1k, so output is NOT in class A.
I've gone down the path of using both the AD797 and the LT1028/1115 in both phono preamp applications as well as in PLL loop filter applications.
In the PLL, the LT1028 produced worse phase noise from 10-1000 Hz, so I stuck with the AD797. That suggests worse low frequency noise.
In the phono circuit, I couldn't hear or measure much difference in performance. The external components and regulator design had a greater impact. The AD797 had a slight edge, so I kept it.
BUT... there was/is a higher current buffer (either an LT1010 or BUF-03) within the feedback loop. Something unusual that I did try was to connect the AD797 distortion neutralization pin to the output of the LT1010. There was a slight improvement. Using a current source of any kind hanging at the output of the AD797 was not an improvement - it wasn't as good - but in my case the load for the opamp is the very high impedance of the buffer.
One other point that is bit off topic. I tried almost every integrated part available at the time (15 years ago) in the buffer position. The two best sounding and performing parts both happened to be single-ended in design. The push-pull output stages just didn't sound as good, no matter what I did. That could have due to the higher quiescent currents for the single-ended parts. That's only a guess; I have no explanation.
In the PLL, the LT1028 produced worse phase noise from 10-1000 Hz, so I stuck with the AD797. That suggests worse low frequency noise.
In the phono circuit, I couldn't hear or measure much difference in performance. The external components and regulator design had a greater impact. The AD797 had a slight edge, so I kept it.
BUT... there was/is a higher current buffer (either an LT1010 or BUF-03) within the feedback loop. Something unusual that I did try was to connect the AD797 distortion neutralization pin to the output of the LT1010. There was a slight improvement. Using a current source of any kind hanging at the output of the AD797 was not an improvement - it wasn't as good - but in my case the load for the opamp is the very high impedance of the buffer.
One other point that is bit off topic. I tried almost every integrated part available at the time (15 years ago) in the buffer position. The two best sounding and performing parts both happened to be single-ended in design. The push-pull output stages just didn't sound as good, no matter what I did. That could have due to the higher quiescent currents for the single-ended parts. That's only a guess; I have no explanation.
PMA said:
I have to admit, this charades are sometimes tiresome. So, let me summarize and correct me if I'm wrong:
- You are using in your evaluation a metric that you cannot disclose or otherwise specify
- You cannot disclose the test circuit and anything about your measurements setup.
- You are disclosing the results for LT1028 in an -1 gain, inverting configuration, but without disclosing the load and the signal amplitude.
- You are using a concept ("noise gain") which, as long as it's not part of the LT1028 datasheet, needs to be consistently defined for allowing any comparisons.
Based on my understanding of this concept, let me make a wild guess. The noise gain is the reciprocal of the attenuation from the output of (or any feedback loop) to the input. This attenuation is Zi/(Zi + Zf). So the noise gain is (Zf + Zi)/Zi. If I am correct, why do you think such a high noise gain setup is relevant? It certainly doesn't help in your mesurements. And are we are talking here low (80dB noise) vs high frequency (-1 voltage) gains? Sorry for the speculations...
Based on the above, you are claiming that OPA211, LME47910 and LM4562 have poor performance. Set aside my own recent experience with these two opamps (which I am ready to share with anybody interested, to the ultimate detail that I am aware of) and which doesn't match yours, what do you think it's making you statements credible?
Edit: I've noticed you disclosed the load of being 1kohm, thank you.
- Status
- Not open for further replies.