Relaxation Oscillator readings on Analogic Oscilloscope

[wmundstock]

Hi all.

I built a simple relaxation oscillator using a LM358 (circuit below) and I get weird readings in the Tek 2246 oscilloscope. The square wave seems more like a "Slope" square wave as seen in picture below.
If I compare with a digital oscilloscope I do not notice the same behavior. I tested the 2246 with a signal with same frequency and amplitude generated by the signal generator and I do not see this behavior.

Any ideas why?

Here is the circuit.


The sloped square wave.

[Uunoctium]

Is the scopes input feed trough coax-cable and termination is set to 50Ohms?
And the digital one didn't offer this feature?

[Zero999]

What was the sample rate set to on the digital oscilloscope?

I think the slope is correct, due to the slew rate of the LM358, which is about 0.3V/µs.

[wmundstock]

Thanks for the responses.

The difference was caused by wrong assumptions from my part. Because the frequency was too low with the 10k resistor I replaced with 1k when switching to the analogic scope. Turns out that made a difference.

Here is how it look like in the digital scope when I use the 1k resistor.


If I test with the 10k resistor in the digital scope, I get a flat line.


I still don't fully understand why higher frequency causes this slope. I will read about the slew rate, that may be the explanation.

For completenes:
I am using just the regular scope probe for the readings, no coax nor terminator.

Thanks.

[CaptDon]

.5ms per div? Were you a.c. coupled instead of D.C.? Also, as another poster said, maybe you loaded down the circuit with the 50 ohm terminator enabled.

[bdunham7]

Try turning the bandwidth limiter off.

[wmundstock]

.5ms per div? Were you a.c. coupled instead of D.C.? Also, as another poster said, maybe you loaded down the circuit with the 50 ohm terminator enabled.

It was DC Coupled.

Try turning the bandwidth limiter off.

The BW Limit was on in the Analogic, and Off in the Digital. But after the retest using the 1k resistor (794 Hz frequency) I am now seing the slope in the digital too.

[Zero999]

The slope shouldn't change much, when the frequency is changed. You should see similar rise/fall times, irrespective or whether the frequency is set to 800Hz or 80Hz. Zoom in on the parts of the waveform when it's changing from high to low, or vice versa, note the slope. Adjust the frequency and repeat.

[wmundstock]

The slope shouldn't change much, when the frequency is changed. You should see similar rise/fall times, irrespective or whether the frequency is set to 800Hz or 80Hz. Zoom in on the parts of the waveform when it's changing from high to low, or vice versa, note the slope. Adjust the frequency and repeat.

Sorry, slope may have been a poor choice of words. What I mean is the curvature of the signal when the output is high/low as highlighted in the picture below.

The picture now shows 2 signals
- the top one is the output from the Oscilator
- the bottom one is the output from signal generator with same frequency and amplitude.

[Zero999]

The slope shouldn't change much, when the frequency is changed. You should see similar rise/fall times, irrespective or whether the frequency is set to 800Hz or 80Hz. Zoom in on the parts of the waveform when it's changing from high to low, or vice versa, note the slope. Adjust the frequency and repeat.

Sorry, slope may have been a poor choice of words. What I mean is the curvature of the signal when the output is high/low as highlighted in the picture below.

The picture now shows 2 signals
- the top one is the output from the Oscilator
- the bottom one is the output from signal generator with same frequency and amplitude.

(Attachment Link)

That's due to the loading of the output stage as the capacitor charges/discharges. It's insignificant when R is 10k but noticeable when R is reduced to 1k. If you keep R constant and just change C, you'll find it won't change as much.

[wmundstock]

If you keep R constant and just change C, you'll find it won't change as much.

Nice! That is a true statement. Thank you!

This image shows the scope with 1k resistor and 470nF capacitor.


This one is with same 1k resistor and 6.8uF capacitor.


Indeed, both output seem very similar.

I also tried using a smaller capacitor (10nF) but the frequency go up to 10kHz and I think the opamp isnt fast enough. The output is no longer a square wave.


I assume for the Khz frequency range there should be other opamps that are suitable.

[Zero999]

For higher speeds try the TL072 or NE5532.

[Ground_Loop]

Are your probes properly compensated?

[wmundstock]

Are your probes properly compensated?

Yes, they are, but I'm using them in 1X so compensation should not interfere.

[Zero999]

Are your probes properly compensated?

Yes, they are, but I'm using them in 1X so compensation should not interfere.

1X is uncompensated, is low impedance, 1M at DC, much lower at higher frequencies, and very slow, with a bandwidth of under 7MHz. None of that matters in this instance as the 10k or 1k resistor and timing capacitor loads the op-amp much more than the oscilloscope probe.