This is a visual "off the scope screen" comparison of four different op amps often used in audio circuits. Patreon www.patreon.com/user?u=12254451 Mail: PO box 362, Vandalia, OH 45377 Email: johnaudiotech@gmail.com Thank You!
Hello John, Good to see a new JAT video, you do know that Snickers makes the videos fun. Great comparison of the op-amps and I like the O-scope view of their performance. The forest fire smoke has been gone two days, after over a week of it. Had some welcome rain, more to come. I hope you and yours, especially your Brother are doing good. All the best from Oregon, C.
Me too, for me the LM4562 is a bit better on a highend system. Otherwise NE5532 has became a very general multipurpose audio opamp. Most of consoles, electronics instruments etc use always those opamp. But the 2134 is not bad but two smooth me.
Great video, thanks! I agree that a step response / square wave test often tells you what you want to know, particularly about stability. One stop shopping. Always be careful when driving capacitive loads with any amplifier. Wire (speaker or audio cable) is a common capacitive load. If you guess ~30 pF/ft, you'll be in the ballpark for most audio cables, lamp cord, or coax. So a 3' (1m) cable is about 100pf, 30' (10m) is about 1000pf. Your .01uF load is pretty bad for most opamps. You should never do that in a real circuit without mitigation. With 1000pF or less, most medium speed op-amps are stable. With larger >1uF caps, the output pole becomes dominant and they slowdown but are stable. 1 to 100nF is the worst case. FET gate input loading is also capacitive, and in the bad range. The easiest way to decouple a capacitive load from an opamp is with a 10 to 200 ohm series resistor before the cap. Most audio gear does this on line outputs. It also limits short-circuit current to safe levels. I often use Spice to simulate various capacitive loads, as long as you have a good spice model for the particular amplifier. Cheers!
BTW: the LM4562 is capable to drive stereo headphones with 30 Ohm impedance. You can even operate it in class A mode by use of J113 N-JFETs as constant current sources between its outputs and the negative power rail which forces the npn output transistors of the opamp to deliver 20 mA without signal. It should sound very good and precise.
@@elsaarmstrong-zp6ng The 20mA goes from the opamp output to the negative rail via the J113. It does *not* pass through the headphones, which are connected between the opamp output and ground. Hope that's clearer now.
Very nice, gave you basic idea about the most common chip out there. keep in mind that some of these 'basics' behave differently on single supply (hence why 10:11 looks good), or in low voltage (+5v) range but thanks anyway.
I've used (about 7 years ago), the LM4562 for a phono amp - superb, but highly strung! Definitely needs both +ve and -ve supply bypass caps - as near to the chip pins as possible. I'll do a video of it at some stage and put it up on my channel. Another great video, by the way, John :)
@@MustangBoss1973 Thank's for reminding me. Whilst it's more or less conventional, I did design it mathematically from scratch - paying particular attention to its critical power needs. It's just got moved further up my list :) and thank you John for allowing us to converse on your excellent channel. All the best, Beamer.
Very interesting. Now I want to solder together a rig with an 8-pin socket, and switches, with round-pin headers for the caps, so you can do this test with style. I have some in my meager collection that you did not try: the TL082 and TL2272. Perhaps these are unsuitable for audio fidelity?
Hi John, cool demonstration! Opamps are not intended to drive a capacitor without a decoupling resistor (33-150 Ohm) at the output. If you are looking at the stability aspects of an opamp, you have to test it in a +1/voltagefollower- configuation, were stability is more critical. In your test could be a small improvement, if you add a small C over the feedback-resistor.
Many manufacturers from China don't agree with you that 072 is not a good op amp for audio -) They would use 5532 everywhere, but sometimes laws of physics dictate them to use a fet input stage.
5532 of different brands sound differently but not too much. 072 of different brands sound extremely differently. My favorites were lf353 UA772 A well -known English brand of professional mixers (Trident) was famous for the sound of TL072
Great illustration of step response, just curios, the BW limited step response you showed, is that a stable op amp, or even in the BW limited codition can their be instability and hence oscilations.
I am one of the guys who would never use an op-amp in my main stereo system. The LM4562 sounds quite good though. The 5532 goes right in the trash, it should have been discontinued 20 years ago. The Sparkos discrete op-amps sounds very good. I just build an AK4497EQ DAC, where I tried those 3 op-amps. The TL072 is very good for guitar effects, better than the 4558 I think.
I think this is a comparison of useful low cost audio opamps (with the exception of the TL072, which is very noisy). My current favorit is the NJM2114 which is a low noise, high slew rate and high GBW device with good THD+N at a reasonably low price (around 1.00 €). If you look for a good opamp for phono you may consider the LM4562 from the video above (about 1.50 €) or even the LT1115 (6.00 €). However, these devices have quite high quiescent current of 8 to 10 mA per channel. You can also get a very good selection and comparison of useful audio opamps with their data here: www.cycfi.com/projects/six-pack/op-amp-shootout/
Gunter, THANK YOU SO MUCH for your reply, man!! I'm using two opamps, the LT1113 and the OPA2107, do you know them? Very intering the LT1115, but I have one question: you said that "... however, these devices (LM4562 and LT1115) have quite high quiescent current of 8 to 10 mA per channel." Does this mean they should have less Iq ? I'm just trying to understand this parameter. Thank you. :)
@@montech5647 op amps with high quiescent currents are usually very fast. They have high slew rates and high gain bandwidth products. I don't see any influence on Iq. The Burr Brown op amps are examples of audio op amps with slew rates of more than 20 v/µs and 50 MHz gain bandwidth. These parameters appear to be overshooting for audio applications.
You have a viewpoint of faithful amplification. But in the guitar amp community where we like harmonic distortion and compression. Is it possible a chip like the TL072 May sound superior than these other chips?
I you would want a circuit with a lot of non linearity, perhaps soft clipping and compression of higher signal levels. You could create this with various opamp circuits but a simple amplifier would be too linear. Tubes can do this without a lot of complexity.
Since my Denon AVR is full of 833s maybe I should put bypass caps on the supply rails close to chips. 4562 with caps on supply legs, and it's derivatives, vs plain vanilla JRC4558, 4560. How about old AD797- it supposedly be a very good chip if properly implemented. And how about OPA X132 and X134 JFets, against TL072. Too many op-amps to choose...
John, as long as you keep the opamp out of oscillation, they should be fine! All the other characteristics just add character to the sound, well that's the way that I see it anyway. Some Amplifiers just seem to add something to the sound that makes it sound better than other amplifiers.. So what's most important at the end of the day is how it makes your favourite music sound like! You can chase down the most stable with the least amount of slew rate and find that it just doesn't sound as good as you would have expected it to! Keep up the great work! Most of my opamps are the NE5534 and I'm very happy with their sound... Not sure if they performed different to the NE5532 though...
Hey John, I have a question, that also could be a video idea for you. What is the significance of an op amp, that runs dual rail, Vs an op amp that runs single rail? I'm not a tech, but just mess with amps and thing on a simple hobby level of knowledge, and found this a bit of a curiosity. Basically,, why is this a thing? And with main amp circuits also?
I'm a little late here, but I will attempt an answer. Chip op amps do not require a bipolar (+/-) supply, but can be used with a bipolar supply. The thing to understand is, the op amp doesn't work any better with a single supply or a bipolar supply, and in fact the op amp does not even "know" if you are using a single supply or a bipolar supply. Because of the op amps design, if you have zero input on the amplifier, by shorting the inverting/non-inverting inputs, the output voltage will be at a voltage half way between supply voltage on the op amp. For example, if you are using a single 10V supply, the quiescent output voltage will be 5V. Now, if you are using a bipolar supply and are supplying the op amp with +10V and -10V, the quiescent output will still be half-way between these two voltage, which is 0 V. If you are using a bipolar supply for an op amp in an audio application, an output capacitor will not be required, since the audio signal is already centered around zero volts. If you are using a single supply, you will need an output capacitor, to block the output DC voltage, but pass the varying audio DC voltage, which is riding on the (in our example) +5V
Tom Terrific , thanks for explaining that. I've bean very confused about how to hook these up, and the only examples I find, in videos, are using a split rail supply. Because, a computer speaker amp board I've bean trying to learn from, and figure out, is a single rail. The non inverting inputs run from the power line, through a few components, and it's confusing me as to why. From the power + , it goes to a 100uf cap, the through a resistor voltage decider. From the resistor, the line then taps both non inverting inputs together. (I can't remember the resistor values, but they're in the thousands.). This op amp is used for the headphones, but also has a tap to go through the volume and tone, then to the amp.
as a newbie ... I have a thing / hurdle about how to connect the scope for these demos...I guess trial and error as no one seems to show exactly how to do this.
It is a very low noise, low distortion op amp with high output current (over 100ma) for driving headphones directly. A very nice chip but relatively expensive. If you don't need the drive current, The 4562 will do just as well.
NE 5532 is the best for it's price but LM4562 sounds a bit better. If you have a CD player and NE5532 was installed you can replace them for peanuts. On EBAy you find 10 LM4662 for about 20 $
The voltage-time relationship for a sine wave is v = a.sin(2πft) where a = the amplitude (half the peak-to-peak voltage) and f is the frequency. So dv/dt = 2πf.a.cos(2πft). Since cos() has a maximum value of 1, the fastest rate of change of voltage for a sine wave is 2πf.a. Therefore, a 20kHz sine wave with an amplitude of 3V (i.e. 6V pk-pk) has a maximum dv/dt of 2 x 3.14 x 10,000 x 3 = 190,000V/s = 0.19V/μs. Any opamp with a slew rate of 0.2V/μs or better will not be slew-rate limited when delivering 6V pk-pk. Although why anyone wants a preamp to deliver 6V pk-pk is beyond me. Of course, square waves have higher harmonics, and a 20kHz square wave will have odd harmonics principally at 60Khz. But unless you're a bat, you're not going to hear those, and I defy any human to tell the difference between a 20kHz sine wave and a 20kHz square wave. If you think you can, you're kidding yourself. The slew rate is not a discriminating factor between those opamps for audio applications. Most of the issues will lie with cross-over distortion in the output stage, or with the noise from the input stage. Although the TL072 (a JFET opamp) has an inferior noise _voltage_ to bipolar opamps such as the NE5532, it has a much lower noise _current_ and therefore may actually have better total noise input when used to amplify higher impedance signals. Horses for courses.
6:18 why 10x has to be used? Why not 1x probe. 10x probe may just be able to measure 10 times the 1x probe voltage. Is it correct to assume that the ramp would be faster than 1x probe so that it can measure 10 times the voltage
The results should be the same, the difference between a 1x and 10x is a potential divider in the 10x. It's simply to protect the front end of the 'scope, as far as I know. As long as the probe is compensated and the 'scope has either sensed or been set to the same setting (1x or 10x) it should show the same information. There will be a very, very slightly higher noise floor on the 10x because of the Johnson noise of the extra resistor.
I'm wrong! Higher frequencies are less loaded (capacatively and resistively) by a 10x probe, as the impedance is 10x higher and the capacitance is lower.
From what I've heard, the LME49720 and the LM4562 are the same chip. The LME prefix is for high performance parts and National Semi. wanted the 4562 to have a part number under the LME series.
Back after a break...not just you but your cat too... .., so electronics aside.. please GIVE YOUR CAT IT'S DUE.. so how about a dedicated vid. on your cat.. on your cat..
This test should be done on a properly laid out board with a ground plane and good bypassing. Otherwise the results are not going to be repeatable. It may not make much difference at a few kHz but as you climb in frequency the tests will not be valid. Having worked for a well known IC supplier for 30 years, we never used those plug strips for anything.
Trying to stuff a square wave into a capacitor WILL exceed the maximum current output specs for the chip, which will then cause the inputs to go unbalanced, which throws the chip into an overload recovery situation. Good circuit design avoids overloading chips, e.g. through dv/dt limitation in the signal source, or with a resistor in series with the capacitor. If you were to try this test with various capacitors, you would see there is a threshold of capacitance, at the point where the chip's output becomes overloaded. Don't blame the chip for doing its best to recover from a bad circuit design. If you see one chip is doing better than another in a particular circuit, it is time to review your circuit design for overloading. This is most easily seen by looking at the voltage differential between the chip's two inputs - which will be ~zero, except where the chip is being overloaded.
Something better than the LM4562? - diyAudio www.diyaudio.com › Amplifiers › Solid State An OPA1612 has slightly better specs than an LM4562---less noise and distortion, faster, and more output voltage swing. May 19, 2017 · 10 posts · 7 authors
@@JohnAudioTech -JFET Type and no Through-Hole Mounting which means the purchase of them on an adapter board. I just posted another link of tested comparisons and on Ebay sellers from China carry the NSC Brand LM4562. 10X LM4562NA DIP8 DualHigh Performance High Fidelity Audio Operational Amplifier Price:US $5.50 - Shipping:$2.80 Standard SpeedPAK from China/Hong Kong/Taiwan | See details Delivery:Estimated between Fri. Nov. 20 and Thu. Dec. 10 ,2020.