I do know the AD797 it's not recommended for the SR. It said so on the original articles and Jan never seems to get tired of mentioning why Walt and him didn't recommend it. In later articles Walt did recommend the AD817.
This sim uses the AD797 because I was doing a comparison between sims of several regulators with Joe Walton's measured comparison, and one of Joe's tests was using the SR with AD797.
I did ask if the SR simulation setup was correct and if it would output useful results because very recently did I know, on the Denoisator's tread, that those LT regulators that came with LTSpice did not include a model for noise, so the noise figure I had got were not valid. So wouldn't like that to happen on my SR noise tests either, comparing different opamps.
One opamp that was a puzzle was the LT1115, which supposedly is internally a version of the LT1028. To my surprise it provided the best impedance and noise results of all opamps. After knowing about the LT regulators models not being suitable for simulating noise, I wondered if the same thing was happening with other LT opamps, like the LT1115.
The LT1028 has lower noise than the AD797, but slightly worst impedance, even if very low too.
Interesting results provide both OP27 and AD828. But I would prefer if someone else did sim these opamps too.
I have no problem in including the .asc files from my sims, including the models for some parts on the .asc files too. But I don't know how to include opamps models which I imported to LTSpice.
But it might be worth it trying. We can try with one and see what happens. Just the .asc files, not the curves.
This sim uses the AD797 because I was doing a comparison between sims of several regulators with Joe Walton's measured comparison, and one of Joe's tests was using the SR with AD797.
I did ask if the SR simulation setup was correct and if it would output useful results because very recently did I know, on the Denoisator's tread, that those LT regulators that came with LTSpice did not include a model for noise, so the noise figure I had got were not valid. So wouldn't like that to happen on my SR noise tests either, comparing different opamps.
One opamp that was a puzzle was the LT1115, which supposedly is internally a version of the LT1028. To my surprise it provided the best impedance and noise results of all opamps. After knowing about the LT regulators models not being suitable for simulating noise, I wondered if the same thing was happening with other LT opamps, like the LT1115.
The LT1028 has lower noise than the AD797, but slightly worst impedance, even if very low too.
Interesting results provide both OP27 and AD828. But I would prefer if someone else did sim these opamps too.
I have no problem in including the .asc files from my sims, including the models for some parts on the .asc files too. But I don't know how to include opamps models which I imported to LTSpice.
But it might be worth it trying. We can try with one and see what happens. Just the .asc files, not the curves.
These are just a few results from some of the simulations I did with superregulators and others, in this case just swapping the opamps and trying results using a pair of zeners or an LM329 and a zener.
My results on the OPA828 seem quite similar to yours. Not so for the AD825.
My results on the OPA828 seem quite similar to yours. Not so for the AD825.
Attachments
No, from what I could find out, apparently no zener includes noise data in their models.
Here there's a question on how to add the noise data, but you would need the noise data from the manufacturer, which would be hard to find.
https://electronics.stackexchange.com/questions/223271/how-to-model-a-noisy-zener-diode-in-ltspice
The suggestion to add a voltage source with the noise data, like Elvee did with the LM317 on his Denoiser would be hard to apply here on the SR, particularly if you want to use a zener diode as voltage reference instead of LM329.
Here there's a question on how to add the noise data, but you would need the noise data from the manufacturer, which would be hard to find.
https://electronics.stackexchange.com/questions/223271/how-to-model-a-noisy-zener-diode-in-ltspice
The suggestion to add a voltage source with the noise data, like Elvee did with the LM317 on his Denoiser would be hard to apply here on the SR, particularly if you want to use a zener diode as voltage reference instead of LM329.
FWIW, I can confirm that OPA828 spice model from TI includes realistic noise parameter. Noise simulation in LTSpice results with value very close to what was measured on actual circuit. I have a bunch of super regulators (another design) with OPA828 on the bench and in my amplifiers.
There is no much point to evaluate noise using simulation while not all opamp models contain noise parameter. This is apparent from simulation with AD825. Using LTSpice directive ‘"noise V(out) V1 oct 10 1 10e6" will give a correct noise estimate on condition that opamp model contains proper noise parameter.
With OPA828, Jung-Didden regulator will have about 6 nV/rtHz output noise density or less than 1 uV total noise in 20 Hz-22 kHz bandwidth.
Very interesting. I went have a look at several opamps that I'm interested in and on the 3x7 regulators models, opening their .lib files.
All opamps had mod usage comments, including current and voltage noise. None of the 3X7 models I found had this data. So you can only rely in output measurements for regulators noise.
Well, I jumped in the pool and made several simulations using the simple Jung-Didden super regulator.
Please be gentle on the critics, and if something is wrong tell me what or how to correct it.
I simulated noise, impedance and PSRR, using five opamps, some old, some new, some recommended, some not even mentioned.
The not recommended AD797 was not tried.
The idea was to compare the simulation results to what Jack Walton presented on Linear Audio, but the results were generally lower or better on the sims. But as some people say: simulations are always optimistic, measurements is what really matters.
Let's first introduce the three "generic" .asc files I used for each parameter.
Please be gentle on the critics, and if something is wrong tell me what or how to correct it.
I simulated noise, impedance and PSRR, using five opamps, some old, some new, some recommended, some not even mentioned.
The not recommended AD797 was not tried.
The idea was to compare the simulation results to what Jack Walton presented on Linear Audio, but the results were generally lower or better on the sims. But as some people say: simulations are always optimistic, measurements is what really matters.
Let's first introduce the three "generic" .asc files I used for each parameter.
