How to measure slew rate of Op-Amp Practically ?

[Aswinth]

Hi, I am under a internship and my supervisor wants me to practically measure the slew of AD8602ARMZ op-amp.

I am aware that this value is available in the datasheet along with a well plotted graph, but my task here is to recreate the graph on the oscilloscope. So far I have read the basics of slew rate and learnt how to calculate it if I get a graph. But I am not sure if I am in the right path and if this will be possible.

My knowledge and idea on Op-amp slew rate and how to measure it come from this article here: https://components101.com/article/u...in-op-amp-and-how-to-measure-and-calculate-it

Here is what I have done so far,

1. As discussed in article, I have created a voltage follower circuit using AD8602ARMZ op-amp.
2. I have verified that the op-amp circuit is working my measuring the input and output voltage
4. Used an Arduino to generate PWM signals for the input signal of voltage follower
5. Mounted one channel of my scope to input and other channel to the output of my voltage follower circuit

With that I am stuck. The input wave from arduino and output wave from op-amp both look identical I am not able to measure any change in time or change in voltage.

I have tried changing the freqency of input pulse but it did not help. would appreciate if anyone can suggest a way forward.

[jbb]

Sounds like the slew rate of the opamp is as fast as, or faster than, the Arduino signal source. What slew rate did you measure, and how does it compare to what the data sheet says?

[MathWizard]

How do they really measure it, feed in a big, steep, voltage step, with the op-amp as diff. amp, then measure how fast the output swings from swing rail to rail ?

[Picuino]

https://www.analog.com/media/en/technical-documentation/data-sheets/ad8601_8602_8604.pdf
Slew rate: 5v/us

You need to inject a 1MHz square wave signal. 0.5us high and 0.5us low.
In that time the operational must show a triangle wave output with 2.5Volts peak to peak.

[Kleinstein]

The square wave from the source should have a much higher slew rate than the OP-amp. 5 V/µs is way to slow for the logic signal.  Also make sure the scope and probe is fast enough to actually allow a faster slew rate. A bad 1:1 proble could be an issue here as they can be quite limited in the BW and the grounding could also matter.  So use a 1:10 probe or alternatively  Z0 probing with a series resistor at the source and 50 ohm termination at the scope.

The OP-amp may want some series resistor (e.g. 50 ohm) at the output  to isolate it from cable capacitance.

There is not need to have a very fast input signal to get all the way to a triangle, just enough to get a good scope picture.

[Terry Bites]

Apply a fast ristetime squarewave  and take a measurment at the amp output on a scope.
(S) = dVo/dt => Vo((90%)-Vo(10%)/ t(90%)-t(10%))
Apply ac and measure the -3dB bandwidth.
Slew rate (S) = 2πfmVtest
A test Vtest say 1V.

[iMo]

..
With that I am stuck. The input wave from arduino and output wave from op-amp both look identical I am not able to measure any change in time or change in voltage..

What is your oscilloscope there?

[jonpaul]

Pulse or Function Gen>>50 Ohm term>>DUT>>scope of sufficient BW and RT.

We avoid cheap Chines digital junk, and use Tektronix 2465B

j

[Picuino]

You need a fast switch like this:

Code: [Select]


const int ledPin = 2;

void setup() {
    DDRD |= (1 << ledPin);
}


void loop() {
    for(;;) {
        PORTD |= (1 << ledPin); // ON
        asm volatile("nop");
        asm volatile("nop");
        asm volatile("nop");
        asm volatile("nop");
        asm volatile("nop");
        asm volatile("nop");
        asm volatile("nop");
        PORTD &= ~(1 << ledPin); // OFF
        asm volatile("nop");
        asm volatile("nop");
        asm volatile("nop");
        asm volatile("nop");
        asm volatile("nop");
   }
}


[Terry Bites]

Make sure capacitance at the opamp inputs is a low as possible. Keep the distance between the GPIO or other generator and the opamp as short as possible.

This little circuit will give bags of current drive (to overcome stray capcitance) with rise times <10nS, running on 3.3V it will outputs about 1Vpp. I hooked this up just now and I got a tr and tf of about 5nS into 50ohms. It may be better than that, I've got a junk scope.

Make sure that the test signal is not pushing the opamp into saturation. Chucking 3.3V or 5V from a GPIO into a 5V opamp is not a great idea.
You will need a scope with significantly higher bandwith than the opamp you're trying to test, say 10x. For a 5MHz opamp a 50MHz scope is needed for accurate results.

[HalFoster]

https://www.analog.com/media/en/technical-documentation/application-notes/an94f.pdf

[MarkT]

You will need a scope with significantly higher bandwith than the opamp you're trying to test

Not necessarily, you need time resolution to match the slewing period.  If you have 5V opamp slewing at 5V/µs, you need time resolution say of 1/10 of that 1µs period, ~100ns, but if you have a +/-18V opamp slewing at the same rate you have perhaps 7µs for the slewing and a much relaxed requirement on the 'scope's time resolution of 500ns or so.

Be aware that some opamp's have input protection that prevents the inputs being at different voltages more than a diode drop or so, so a series resistor on the input is wise protection to add.

Another approach uses a spectrum analyser and a signal generator.  You use the opamp as a follower, apply a known amplitude signal and increase its frequency till the harmonics suddenly jump in magnitude due to slew-limited distortion.  That will measure the slowest of the two slew rates (positive slewing and negative slewing are often different).  The maximum slew rate of a sinusoid is 2*pi*f*A, where A is the amplitude and f the frequency, so for instance a 5V amplitude (10V peak-to-peak) signal at 100kHz slews at 3.14V/µs

[DimitriP]

Slew rate measurement ; short and sweet version:

https://youtu.be/BKA5ESd0IJs?si=KB4qkQaxKt4WrYIp&t=785

[jonpaul]

see spec sheet of opamp mentioned/

Most will have both SR specs (eg vs Vcc or load) and some have application notes on SR, fast followers, etc.

Measurement requires proper equipment.
perhaps your setup is incapable  give any good SR info, slow gen eges and RT, low output voltage of gen, poor scope.

j

[kimballa]

Rather than rely on the raw speed of the gpio, you could add a very fast comparator like TLV1811 (the push-pull version, not the open drain TLV1811) or some other logic gate with Schmidt trigger inputs to drive a super fast input slew to the opamp.

[TimFox]

One thing to watch for when measuring slew rate on an actual circuit (op amp with feedback network).
For small signals (where the output amplitude is less than that limited by slew rate), the bandwidth is fixed by the op amp's frequency response and the feedback network in the usual way.
If you drive such a circuit from a good (fast rise time) square wave within the small signal condition, the output's apparent slew rate (seen from the rise time) will depend on the bandwidth and not the op amp's slew rate.
When you increase the input drive voltage from there, you may see the output condition change from constant rise time to constant slew rate, where the rise time increases with increased amplitude.
From that, you can calculate the op amp's slew rate limit.