while that is an excellent achievement, the addition of the fourth harmonic with a 1k load is odd as this is was designed to have a better output section according to JohnC124 in his introduction to the chip here on the forum.
OPA1656: High-Performance CMOS Audio Op Amp
Cheers
Alan
I can only read the scales at max size. more contrast would be nice.
OPA1656: High-Performance CMOS Audio Op Amp
Cheers
Alan
I can only read the scales at max size. more contrast would be nice.
Last edited:
Next time I'll try. Now for the best view you can click on the picture, then mouse right click and choose "view image". Then click "+" (mouse left click).
The quiescent current of the OPA1656 is 3.9 mA per channel. The 1kohm load at 2.7V RMS means 3,8mA output peak current. Probably the output stage of the opamp leaves the linear region, and AB class starts to run. So higher order harmonics may appear.
Victor- I know how disorienting it can be to change software but Diana may be better suited to these tasks. The latest version will work with a notch filter and can isolate just the harmonics. DiAna The residual analysis can be really helpful as well. With your notch filter you should be able to see well into the noise.
And you get a lot of info in table form as well.
And you get a lot of info in table form as well.
DiAna is great, it measures correctly (the other software i've seen have trouble dealing with harmonics close to noise floor, and lack in ability to de-noise the measurement in presence of phase noise/wander. Ah, and the scalloping errors too... I haven't verified whether DiAna could deal with scalloping thing, TBD.).
Unlikely since AFAIK windowed FFT's are not involved. I still have not tried phase locking one of Victor's oscillators.
What you can do if the pin looks "fine" with a 10x probe but has problems without the probe connected is turn up the the vertical gain and put the probe tip NEAR the pin (like 1/4 inch or a mm away). If it's oscillating or generating some funky RF, the tip will pick it up and you can see something on the screen increasing and decreasing in amplitude as the tip moves toward and away from the pin.
The menus and features on the Rigols are just ... I'll be nice and say not intuitive. I need to read through the whole manual someday (it's the hackerspace's scope I don't use often, if it were mine I'd have more incentive to read the manual). Maybe it's not too bad, but I can get a voltage measurement value to show up on the screen, but then I don't know how to make it go away...
Looking at TI site, is the OPA1612 better than the OPA1656? THD 0.000015% vs. THD 0.000050% http://www.ti.com/assets/js/compare...78typ,p1498,p1192,p2954,p1130,p233typ&lang=en Somehow this confirms my knowledge that BJT distort less than FET.
Last edited:
I see the OPA1652 there, not the OPA1656.
The open loop gain is one of the most important things when the opamp runs in the inverted mode:
OPA1612 ... 130 dB
LME49720/LM4562... 140 dB
OPA1656... 150 dB
By the way, seems strange that the TI sells the LM4562 for around 1$ and less now:
https://www.ti.com/store/ti/en/p/product/?p=LM4562MAX/NOPB
They moved it from the old obsoleted 6" wafer process to the new 8" wafer process. That move was what saved the chip from extinction. Perhaps the better economics of a larger wafer make it possible to lower the price?
In any case, while my tests are nothing more than anecdotal, I think the new 8" process results in a chip with less (i.e. no) popcorn, a behavior that would plague about 10-20% of the old LM4562 chips. I'm hoping this is a side benefit of the new 8" fab line. Again, I have tested only a few chips so far, but it seems that they are uniformly "clean" and free of the random popcorn that would plague the old product. If anyone else finds a current datecode LM4562 with popcorn problems, please post here - so far, I have found zero such chips.
With modern op amps, THD+N, which is actually what is being specified and not THD, is completely dominated by noise. The actual distortion products of most of the op amp ICs that people consider to be "clean" are 30-40dB into the test circuit's noise floor.
So, THD+N specs are really just comparisons of the total noise of an amplifier in a specific impedance test jig. And, as you note, bipolar input amplifiers will measure better if the resistors surrounding the amplifier are low valued, since the entire amplifier will have less voltage noise, and thus result in a better THD+N number.
Actual distortion values as measured using an FFT to measure specific harmonics can paint a completely different story: noise and distortion are often unrelated, but THD+N with a modern clean amplifier is basically a noise measurement. Direct harmonics are the best way to compare distortion.
At the same time the price for the LME49720 is more higher:
https://www.ti.com/store/ti/en/p/product/?p=LME49720MAX/NOPB
In future I am planning to use the OPA1656 in my oscillators, but also ordered 10 pieces of "new" LM4562. Probably next week they will be received...
https://www.ti.com/store/ti/en/p/product/?p=LME49720MAX/NOPB
In future I am planning to use the OPA1656 in my oscillators, but also ordered 10 pieces of "new" LM4562. Probably next week they will be received...
Notchfilter too?
But keep in mind that the open loop gain can deviate downwards according to data sheet. And that would be worse than the LM4562 --> min: 134dB, typ:154dB
LME49720/LM4562... min 125 dB
OPA1656 ... min 134 dB
Of course, the OPA1656 is very good candidate for the buffer amp after the passive notch, needs to try. The OPA627 runs well in this place. I can see only very small distortion contribution at 10 kHz. Now I have also the discrete buffer which is better than OPA627 in terms of noise and distortion.