Slew Rate Limiter.

[Kokoriantz]

Is there any universel slew rate limiter, adjustable in audio range?

[Zero999]

Connect an amplifier with a voltage clamp on the output to an integrator. The integrator limits the rate of change of the output voltage, by charging/discharging a capacitor via a constant current source, which is set by the input voltage. U1 keeps the output voltage equal to the input, via negative feedback.

R1 can be a variable resistor, in series with a fixed resistor for stability's sake. Here's a simulation showing the response to a 2V square wave, with different values of R1.

[shapirus]

Is there any universel slew rate limiter, adjustable in audio range?

Will an LPF do?

[TimFox]

An LPF does not limit the slew rate as a rate limiter would: filters are linear circuits and the slew rate after an LPF driven by a square wave is proportional to tha amplitude.
One possibility for a slew rate limiter might be to use an uncompensated op amp and vary the compensation capacitor, but this would also change the open-loop frequency response (non rate-limited).  The non-linear slew rate limit comes from the limited output current of the input stage charging the compensation capacitor.

[Kokoriantz]

Thank you for the proposition.
I simulated the circuit with single supply using 20khz 1vp and adjusted the resistor just to start limiting. It does limit but gets out of limitation very badly.
I will try to find a remedy for it.

[Whales]

Sorry not familiar with your simulator.  Are you powering the TL072 with 0V to 15V, but then outputting a signal that is -1V to +1V?

[Kokoriantz]

The scope in AC mode.

[Zero999]

The scope in AC mode.

It doesn't matter. The TL072 will not work with a single supply, in the configuration you've shown. And please draw the circuit more clearly. Avoid overlapping text and leave plenty of space between components.

You haven't provided a specification. What's the desired slew rate limit?

Use the following formula, to calculate the maximum slew rate with my circuit.

SR = V/(R*C)

Where:
SR is the slew rate in volt per second.
V is the clamping voltage of D1 & D2, which is a total of around 6.8V. The zener's reverse breakdown voltage is 6.2V and its forward voltage is 0.6V. Add the two together and you get 6.8V.
R = R1's value in ohms.
C = C1's value in farads.

For operation at 20kHz, you probably want to choose a faster slew rate than the values I used in my circuit give. C1 will need to be reduced by a couple of orders of magnitude.

[Kokoriantz]

I am trying out another circuit.
 From your principal idea, I am developing it using the AD844. This is current feedback opamp with integrating capacitor option. I need to  limit the charging current somehow.

[mtwieg]

Texas instruments also suggests a similar circuit: https://www.ti.com/lit/pdf/tidu026
Main difference it that it lacks the zener diodes for clamping. The output range of the first opamp effectively acts as the clamp. That means the slew rate limit will be proportional to supply voltage. If you don't want that, then use the zener diodes like Zero999's circuit.

But this circuit will not work well with all types of opamps. The opamp must be able to withstand large differential input voltages without damage or phase reversal. I'm not an expert on CMFB opamps, but they don't seem like a wise choice for this purpose.

In your simulation, it looks like you're trying to use a single supply with a virtual ground made with a resistor divider. But this will not work well, the virtual ground must be a much lower impedance than the 1K series resistor, otherwise the clamps will be soft and not give a consistent slew rate. Also should use a separate virtual ground for the noninverting input of the second opamp. Or just use bipolar supplies like in the reference circuits.

[Zero999]

Texas instruments also suggests a similar circuit: https://www.ti.com/lit/pdf/tidu026
Main difference it that it lacks the zener diodes for clamping. The output range of the first opamp effectively acts as the clamp. That means the slew rate limit will be proportional to supply voltage. If you don't want that, then use the zener diodes like Zero999's circuit.

The saturation voltages for the high and low sides might not be equal. If you want to omit the zener clamp, then an op-amp with a R2R output would be best, which is probably why that application note uses the OPA192.

But this circuit will not work well with all types of opamps. The opamp must be able to withstand large differential input voltages without damage or phase reversal. I'm not an expert on CMFB opamps, but they don't seem like a wise choice for this purpose.

That's easy to fix. Put clamping diodes between the inputs of U1, if it doesn't have them built-in, and add a series resistor with V1 (the input), to limit the current.

In your simulation, it looks like you're trying to use a single supply with a virtual ground made with a resistor divider. But this will not work well, the virtual ground must be a much lower impedance than the 1K series resistor, otherwise the clamps will be soft and not give a consistent slew rate. Also should use a separate virtual ground for the noninverting input of the second opamp. Or just use bipolar supplies like in the reference circuits.

Yes. Add another op-amp rail splitter circuit, or the TLE2426.

[Kokoriantz]

I could make the AD844 current limiter and adjustable. It saturates the slew rate of 20khz at 200mv. For now, it is not perfect symmetry as the vbe of the transistors are not identical and the -3db is about 40khz which needs to double.
The purpose of this circuit is to emulate the tube amplifiers that became famous ones, the Williamson and the Mullard 5-20. These 15w and 20w amps are sought-after for that sweet high frequency character which is due to slew rate limited. They use low cost 10khz-3db output transformer and 20db NFB to advertise 100khz performance. At 10khz the power is halved and at 20khz is the quarter of it. This why I need an adjustable slew rate limiter, that can be installed at the input of any amplifier.
There has been a Sanyo amplifier that has switched slew rate selector, 5V/us or 20V/us.
The TI circuit is for DC-250mhz only.

[Conrad Hoffman]

Years ago, I built the circuit shown at the top and it worked very well. I think I used parallel diodes, not zeners.

[mtwieg]

I could make the AD844 current limiter and adjustable. It saturates the slew rate of 20khz at 200mv. For now, it is not perfect symmetry as the vbe of the transistors are not identical and the -3db is about 40khz which needs to double.

Sure, connecting a large capacitor to the TZ pin should produce a controlled slew rate. Though the datasheet doesn't give any hint at what sort of variability one should expect between different parts, or temperature dependence, or symmetry...

The TI circuit is for DC-250mhz only.

Not sure what you mean here. It's capable of functioning up to very high frequencies, but will still do what you want at audio frequencies.

[Kokoriantz]

I made the clamped limiter proposed by @zerp999 work clean. For now it is 20khz -3db, needs adjustment.
The AD844 is 2 parts, a voltage to current converter with the negative input resistance ratio and a unity gain diamond follower. All it needs is a means to limit  the current.

[Zero999]

What's the reason for the TL431? It might not have great transient response. I strongly recommend building it.

It should be possible to independently set the positive and negative slew rates independently and implement voltage control. I had a play. Note the polarities for the rise and fall times are reversed because the input stage is inverting.

[TimFox]

When evaluating the frequency response with slew-rate limiting added, you must use very low amplitude so that the sine wave at the highest frequencies is not affected.  When the slew rate affects the output, the apparent frequency response at high frequencies depends on the amplitude.