Whether it can generate IR signals is not the issue as I've not been talking about remote controlled amps.
Excuse me, but it has always been the issue for me. Just because its not an issue for you, you think that automatically makes it not an issue for me? Think again pal
ISTM your communication skills are lacking here. Let me quote what I actually said :
Just as one example, an AP generates a near perfect sinewave stimulus but its differential mode only, by design.
Notice the words 'just as one example' there? I'm not attempting to catalog the full capabilities of the AP, just highlighting one of those capabilities. So your claim that I think 'the AP only does sines' turns out to be imagination only. As I've already said, I did use an AP1 in my former life, I did notice and use its capabilities beyond merely generating sines.
As far as I'm aware (and here I must admit I never tried to reverse engineer the AP) the oscillator output from the AP is isolated via a transformer. So in normal generating mode, there's by design no common mode output. That's a feature, not a bug for the purpose of the AP1 which is to be able to inject a low distortion sinewave into practically any position in a circuit. In a word, its almost perfectly floaty, by design.
So no, I did not say 'sines are only output in diff mode' either if you care to re-read what I actually wrote. But here I admit that I could have phrased my meaning a little clearer than I did. What I meant was there's no facility to add common mode noise onto the diff mode signal. I had hoped that the context would help clarify the meaning. I'm aware that it has CMRR testing abilities too, just as one example. But CMRR mode won't help in testing the unbalanced input of an amplifier - its for checking balanced inputs.
So, moving on to your next claim...
Your conclusion is a false one, I last used an AP I think in 2001 when I was an employee, I needed to use it in my job as amplifier designer, I first used one in 1989 or perhaps 1990. So given that, please re-check your premises.
So to you, an observation that you don't live up to your own ideals is rubbishing you on a personal level? If so, then guilty as charged your honourThank you. I am going to test as noted by Mooly and others, and I will add the LM4562 to the growing short-list!
A question for all... almost all op-amp datasheets show distortion rising with frequency, so when I make tests what can I do to minimise this problem, to keep distortion level with frequency?
I presume the answer includes: match impedances at the inputs, decouple close to the supply pins, use the lowest possible gain for the application, and use the highest possible load for the application, but what else to keep distortion level with frequency?
Well there's always the resistor to load the VAS
but there are no sensible reasons for doing that, even if the pin is brought out of the package. Why do you want to keep the distortion level with frequency? To keep it level we can either increase LF distortion (fairly easy, that's what's done by the VAS resistor as it decreases LF OL gain or you could try loading the output with an inductor, its impedance falls at LF, increasing the output stage distortion) or try to reduce HF distortion (much more difficult). Running the opamp in inverting mode minimises common mode effects of the input LTP and associated non-linear capacitances, so its one way to keep HF distortion down, compared to the non-inverting config. Another trick might be bootstrapping the supplies, I've seen that used in an app note somewhere - useful if you really must have a non-inverting topology. Using a decompensated amp and other than single pole compensation might do the trick too.My experience of the LM4562 is very favourable. I have used three of these dual op amps in the ESP project 97 preamplifier (with tone controls). Compared to the NE5532s, which are not quite as bright, the 4562s appear very transparent and offer a superbly detailed sound stage. It was necessary, however, to add some low-value paralleled feedback capacitors to roll-off the HF response at around 200KHz or so. Without these capacitors, there was mild sibilance suggesting possible instability.
It might be prudent for anyone wishing to compare similar devices to limit the HF response as a matter of course. I was about to return to the 5532s (which didn't need compensation in this circuit) due to the harshness of the 4562s, but now they sound as sweet as I could have hoped for.
Regards,
Steve
No problem Cliff. Generally, I would agree with your statement, but ears are also capable of measurement and on having previously experienced this effect, I was in little doubt as to the cause.
Regards,
Steve
Sounds more like RF susceptibility than instability to me. Unless you've got too much capacitance already on your -ve input, adding capacitance across the feedback resistor will generally increase the amount of feedback at HF, leading to slightly greater instability, But since the LM4562 is already unity gain stable, why would it be close to instability?
Multiple registrations are not allowed. I have merged "Cliff" and "Clifforrest." In the future, please just use the Cliffforrest account.There is little measured RF in the vicinity of the amplifier (using a "Zapchecker"), no wi-fi in the house and an RF filter is soldered to the rear of the phono connector (1K & 470pF). The whole preamplifier is fully screened. It seems unlikely that the problem was RF susceptibility.
From http://focus.ti.com/general/docs/lit/getliterature.tsp?baseLiteratureNumber=SLYT087&track=no:
"Op amp stability and input capacitance"
Internally compensated op amps can be made unstable in several ways: by driving capacitive loads, by adding capacitance to the inverting input lead, and by adding in-phase feedback with external components. Input capacitance is hard to avoid because the op amp leads have stray capacitance and the printed circuit board contributes some stray capacitance, so many internally compensated op amp circuits require external compensation to restore stability.
The article later states that adding CF increases stability.
I will probably remove the feedback capacitors and get the 'scope out after all, even though the modification is effective...
Regards,
Steve
I already took care of it by merging the two personae. I understand that it wasn't deliberate (like some folks who like to waste everyone's time with sock puppets). So now you're just "cliffforrest."
So is this LM4562 being used as the input stage of a phono preamp? In which case it would most likely have at least 20dB, probably up to 40dB gain? In which case the gain margin given that its unity gain stable, will be huge. I'd not be so certain that 1k with 470pF would make a very effective RF filter, though of course I don't know a thing about your layout or the specific types of components so I'm guessing. But in my experience RF gets in via the ground connection. Given that the (few) phono amps I've examined bring the ground connection on to the board (albeit sometimes with a cap to chassis) this is the most likely route. Caps to chassis have self-resonant frequencies, above which they'll be inductive.
Do they explain how that happens? If they're assuming a lot of stray to ground from the -ve input then I'd agree. In other cases, I probably won't.
<edit> I had a quick look through the appnote you cited - have you worked out where CG is in the schematic? They refer to it in the text but no schematic shows it.
Last edited:
This preamplifier is based on ESP project 97 with no phono input. I was just referring an input connector type.
CG definition is missing from the article. It is the stray input capacitance between the inverting input and ground. (Refer article-cited link: TI - Op Amps For Everyone)
Regards,
Steve
Back to the original question. What is wrong with opamps.
In my opinion there is nothing wrong with opamps. It reduce the rate for failure for less experienced designers. The last one isn't degrading towards any less experienced designers. It can be a huge relieve if you can spent your attention and time on other subjects as layout and power supply, if you don't have to care anymore about all the stages in a discrete opamp. Look at how many wonderful sounding amplifiers are out there with a chip amp. There is more air and room to focus on other things once the chip take over a lot of burden to think off.
And if you forget standard opamp applications, you can do surprisingly creative things with those 8 feet IC's. Look at the NE5534. U can add new (FET) input-stages. You can access the VAS output. You can do many things.
A tube, discrete or an opamp. They have their own sound in their own way. Every circuit or piece of electronics have their own sound. So the argument that you can always hear an opamp is a bit lame in my opinion.
Starting at the 80's indeed every piece of music that is recorded went through opamps. First in the mixing consoles, lateron in the digital interfaces from digidesign and Apogee for pro-tools and Logic recording systems. No matter how, your music wen through opamps already. If the "opamp sound" was that worse, every cd you buy will be worse.
Opamps are handy, save space and therefore allows more compact circuit layouts which benefits the sound. and in my opinion doesn't need to be less good then discrete. It is up to the designer how good the end result will be.
I mentioned it before in other topics, but I will mention it again. I replaced the Jfet discrete stages in expensive Marantz SA1, and SA11 SACD players by a far shorter and simpler circuit made by the OPA627 and a DC servo, fully differential. The latter sounded (subjectively) much better. Most people assume discrete is saint and automatically superior. I really disagree with that statement.
With kind regards,
Bas
In my opinion there is nothing wrong with opamps. It reduce the rate for failure for less experienced designers. The last one isn't degrading towards any less experienced designers. It can be a huge relieve if you can spent your attention and time on other subjects as layout and power supply, if you don't have to care anymore about all the stages in a discrete opamp. Look at how many wonderful sounding amplifiers are out there with a chip amp. There is more air and room to focus on other things once the chip take over a lot of burden to think off.
And if you forget standard opamp applications, you can do surprisingly creative things with those 8 feet IC's. Look at the NE5534. U can add new (FET) input-stages. You can access the VAS output. You can do many things.
A tube, discrete or an opamp. They have their own sound in their own way. Every circuit or piece of electronics have their own sound. So the argument that you can always hear an opamp is a bit lame in my opinion.
Starting at the 80's indeed every piece of music that is recorded went through opamps. First in the mixing consoles, lateron in the digital interfaces from digidesign and Apogee for pro-tools and Logic recording systems. No matter how, your music wen through opamps already. If the "opamp sound" was that worse, every cd you buy will be worse.
Opamps are handy, save space and therefore allows more compact circuit layouts which benefits the sound. and in my opinion doesn't need to be less good then discrete. It is up to the designer how good the end result will be.
I mentioned it before in other topics, but I will mention it again. I replaced the Jfet discrete stages in expensive Marantz SA1, and SA11 SACD players by a far shorter and simpler circuit made by the OPA627 and a DC servo, fully differential. The latter sounded (subjectively) much better. Most people assume discrete is saint and automatically superior. I really disagree with that statement.
With kind regards,
Bas
Last edited:
Well there's always the resistor to load the VAS
As I have mentioned before, a musical instrument, including the voice, produces overtones (or harmonics). When the distortion is rising 6dB/octave in an OPAMP, the result is an alteration of the overtone spectrum.
But multiple opamps paralleled don't have that distortion. One should try that out for instance and be blown away about the sound.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.