Dual op amp experiments.

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
I've known for a while op amps in general have problems with inputs near or at the power supply. I decided to run some formal experiments to see what these limits are. I've also run into problems with earlier designs. So I am going to run a series of experiments using a 1458, LM385, and 4565 dual op amps. I will be using a 555 circuit to input a square wave. I finally settled on a 555C timer For all the experiments , reason being it was a little bit more stable on my oscilloscope. I found out using an old 555 FD bypass cap it's not just important but the location of the cap is too. The inductance on the line there protoboard was enough to cause some apparent jitter which was an artifact. This is why C4 a 220µF capacitor was put in such an odd location. For the moment I will just post the data . I will speculate why the outputs were the way they were in the following posts.Feel free to add your 2¢. I would be glad to rerun any experiment if asked. I will also include a picture of the circuit if asked.

555 / 555C signals
555 output.png 555CN output.png
.............555....................................555C

Experiment 1
Schematic/Layout
[/COLOR]45651 experiment.png
Waveforms
1458 e1.png 4565 e1.png LM358 e1.png
…...........1458…...........………….LM385…...........………….
 
Last edited:

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
Experiment 2
Schematic/Layout
experiment 2.jpg
Waveforms
LM358 e2.png 4565 e2.png 1548 e2.png
…...........1458…...........………….LM385…...........………….4565
Experiment 3
Schematic/Layout
3 experiment.png
Waveforms
3 1458.png 3 4565.png 3 LM358.png
…...........1458…...........………….LM385…...........………….4565
 
Last edited:

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
Experiment 4
Schematic/Layout
4 experiment.png
Waveforms
4 1458.png 4 4565.png 4 LM358.png
…...........1458…...........…....…….4565.....………….LM385LM358
 
Last edited:

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
I held off on analysis until I finish this thread until I finish this experiment post. Some of the waveforms are just plain strange. It is getting near my bedtime so I am going to finish for now.
 

MisterBill2

Joined Jan 23, 2018
9,792
Interesting experiment indeed, Wendy, but it is not clear to me where the three wave forms are from, nor what the voltages are. And one more device to run would be the LM324, the common quad op amp.
It may also be interesting to see the results at different supply voltages, some devices perform differently at higher voltages, at least they used to.
Good luck with the continuation of this project.
 

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
The waveforms are on the output of U2b. I marked the protoboard probe points in red on each illustration. Because I didn't wanna spend all your doing this I just used a12 volts power supply. A quad opamp can be useful for many applications but generally I use a dual. The lowly 1548 performed surprisingly well something to keep in mind in future designs. The illustrations are what I had in mind when I bought that oscilloscope. Again I am willing to rerun any experiment with changes. Experiment 3 using the 4565 was a puzzler. I expected better from that little op amp.
 

dl324

Joined Mar 30, 2015
13,505
Do you mean LM358? LM385 is voltage reference.

1634400193565.png
That's an odd way to tie off an unused opamp. You've grounded the output and inverting terminals, so there's no way for the opamp to function as a voltage follower. It'd make more sense if the output wasn't grounded.

LM1458 and NJM4565 don't include the negative rail in the the common mode input voltage range.
 

Audioguru again

Joined Oct 21, 2019
3,828
Hi Wendy, your problems with the old 555 were probably caused by the contacts resistance on the breadboard and the ESR of the main filter capacitor, since the 555 draws a "shoot-through" supply current pulse of 400mA (!) each time its output switches.
Therefore I recommend using a more modern Cmos 555 that does not produce the current pulse and its output is rail-to-rail.

In experiment #3, the slow rise and fall times of U2b look like a LM358 opamp, and its inverting input is not biased properly.

In experiment #4, the LM358 opamp U2b has no bias resistor on its inverting input so the PNP input transistor base floats high causing the output to go too low.
 

Attachments

Audioguru again

Joined Oct 21, 2019
3,828
@dl324, the grounds are different. I have never, ever used an opamp as a "virtual ground" like U1a because the output usually biases the extremely low input current of an opamp when it has a single supply voltage.
 

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
Do you mean LM358? LM385 is voltage reference.

View attachment 250471
That's an odd way to tie off an unused opamp. You've grounded the output and inverting terminals, so there's no way for the opamp to function as a voltage follower. It'd make more sense if the output wasn't grounded.

LM1458 and NJM4565 don't include the negative rail in the the common mode input voltage range.
yes i meant an LM358. I fixed the post.
 

Attachments

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
Hi Wendy, your problems with the old 555 were probably caused by the contacts resistance on the breadboard and the ESR of the main filter capacitor, since the 555 draws a "shoot-through" supply current pulse of 400mA (!) each time its output switches.
Therefore I recommend using a more modern Cmos 555 that does not produce the current pulse and its output is rail-to-rail.

In experiment #3, the slow rise and fall times of U2b look like a LM358 opamp, and its inverting input is not biased properly.

In experiment #4, the LM358 opamp U2b has no bias resistor on its inverting input so the PNP input transistor base floats high causing the output to go too low.
Moving C4 directly next to the 555 completely fixed the problem. For me it is an unusual configuration, one I will probably use in the future. I do like the 555 relatively massive output current. But the 555CN works better for many applications. To me the 7555 and 555CN are interchangeable.
 

dl324

Joined Mar 30, 2015
13,505
What is the point of experiments 1, 3, and 4?

Experiment 2 shows the LM358 exceeding it's slew rate spec of 0.5V/μs by a large margin. Are you certain that isn't the input signal?
 

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
Do not forget a 555 has a true digital output. Experiment 1 does show the slew rate of the opamp as well how the the op amp handles a rail to rail input. Experiment 3 shows how the op amp handles a 1/2 Volt Vcc signal. Experiment 4 was a brainfart on my part.
 

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
what can I say except, oops? If I were to rerun this experiment I would take it to the virtual ground. I made a similar mistake with experiment 4.
 

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
i am going to rerun experiments 3 and 4. Since this follows idea of a science experiment repost the brain farts here.
Experiment 3
Schematic/Layout
3 experiment.png
Waveforms
1548 e2.png 4565 e2.png LM 358  2.png
…...........1458…...........………….LM385…...........………….4565
Experiment 4
Schematic/Layout
View attachment 250417
Waveforms
View attachment 250418 View attachment 250419 View attachment 250420
…...........1458…...........…....…….4565.....………….LM385LM358

Experiment 4
Schematic/Layout
Old experiment 4.png

Waveforms
4 1458.png 4 4565.png 4 LM358.png
…...........1458…...........…......…….4565............………….LM385
 

Attachments

Last edited:

Thread Starter

Wendy

Joined Mar 24, 2008
22,646
OK, Finished. The new work is on posts 1, 2 and 3. I just updated them with the correct results. Future experiments will be with a sine wave see what's the lowest voltage each opamp will take without distorting. I am trying to find the practical limits of these opamps.
 
Top