# Single op amp PWM generator

#### StijnC

Joined Oct 23, 2021
22
Hello, I have to design a single op amp circuit that has only +15v, -15v and GND as input and outputs a symmetrical PWM signal (+V and -V) with adjustable frequency and duty cycle. I have something that works partly but not completely, I can adjust the duty cycle from ±60% to ±100% but af that the duty cycle potentiometer also adjusts the frequency, the frequency potentiometer seems to work. Also the output is not really symmetrical. The values on my drawing are not really calculated because it doesn't work yet.

Last edited by a moderator:

#### MrAl

Joined Jun 17, 2014
11,559
Hello, I have to design a single op amp circuit that has only +15v, -15v and GND as input and outputs a symmetrical PWM signal (+V and -V) with adjustable frequency and duty cycle. I have something that works partly but not completely, I can adjust the duty cycle from ±60% to ±100% but af that the duty cycle potentiometer also adjusts the frequency, the frequency potentiometer seems to work. Also the output is not really symmetrical. The values on my drawing are not really calculated because it doesn't work yet.View attachment 250891
Hi,

First, what types of components are you allowed to use besides the one op amp? For example, NPN transistors perhaps.

Second, what do you call symmetrical. Do you mean the pulse width for both positive and negative sections of the output wave have to be the same as well as the voltage swing?
For example, a rectangular wave at a frequency of 10Hz where the positive pulse is 10v and 1ms and the negative pulse is -10v and 1ms. I also guess you want the plus and minus pulses to be equally spaced too, so if +10v and 1ms 'on' and 9ms 'off', then -10v and 1ms 'on' then 9ms 'off'.
That would be perfectly symmetrical.
I have to ask this because the word 'symmetrical' can mean a host of different types of symmetricality. It's probably one of the most ambiguous terms in science and math.

I must admit that of all the circuits i ever had to design i never had to design a circuit that uses one op amp and has to deliver a symmetrical output as described above, but my first thought is you could probably achieve the symmetricality by using pulse steering diodes. Just a guess as you may not need them.
At first i thought this might be hard to do, but after a little thought i realized that the circuit should work the same no matter if the output is positive or negative although not both. The trick maybe is to get it to trip for both positive and negative outputs in the same circuit.

With that in mind, try designing two circuits, one that operates with a positive pulse output only and one that operates with a negative pulse output only, then see if you can figure out how to combine both functions into a one op amp circuit.

Last edited:

#### Ian0

Joined Aug 7, 2020
9,988
Why aren't you using a 555?

#### ericgibbs

Joined Jan 29, 2010
18,981
Hello, I have to design a single op amp circuit that has only +15v, -15v and GND
It is a Homework assignment

#### StijnC

Joined Oct 23, 2021
22
Hi,

First, what types of components are you allowed to use besides the one op amp? For example, NPN transistors perhaps.

Second, what do you call symmetrical. Do you mean the pulse width for both positive and negative sections of the output wave have to be the same as well as the voltage swing?
For example, a rectangular wave at a frequency of 10Hz where the positive pulse is 10v and 1ms and the negative pulse is -10v and 1ms. I also guess you want the plus and minus pulses to be equally spaced too, so if +10v and 1ms 'on' and 9ms 'off', then -10v and 1ms 'on' then 9ms 'off'.
That would be perfectly symmetrical.
I have to ask this because the word 'symmetrical' can mean a host of different types of symmetricality. It's probably one of the most ambiguous terms in science and math.

I must admit that of all the circuits i ever had to design i never had to design a circuit that uses one op amp and has to deliver a symmetrical output as described above, but my first thought is you could probably achieve the symmetricality by using pulse steering diodes. Just a guess as you may not need them.
At first i thought this might be hard to do, but after a little thought i realized that the circuit should work the same no matter if the output is positive or negative although not both. The trick maybe is to get it to trip for both positive and negative outputs in the same circuit.

With that in mind, try designing two circuits, one that operates with a positive pulse output only and one that operates with a negative pulse output only, then see if you can figure out how to combine both functions into a one op amp circuit.
Hello, with symmetrical I mean for example -14V for low and +14V for high, I know that you can do this with extra components because for example with an extra opamp it's simple but I was told that it is possible with one without a transistor or something else, just an opamp and something with a diode and ofc the RC and so on, so I want to find out how.

Cheers

#### StijnC

Joined Oct 23, 2021
22
Why aren't you using a 555?
Because we were asked to do it with an opamp.

Cheers

#### StijnC

Joined Oct 23, 2021
22
It is a Homework assignment
No, I made one in class with 3 and 2 opamps but they told us that it's possible with one so I'm trying to find out how.

Cheers

#### Ian0

Joined Aug 7, 2020
9,988
Have a look at figure 4.4.7 in this. I know it's a 555 circuit, but if you delete your zener and the circuitry to the left of it, you have your op-amp wired so it behaves like a 555. Op-amp output = 555 output. Op-amp inverting input = 555 trig.
Then you have to figure out how to vary the frequency - think what happens if you vary one of the other resistors.

#### dcbingaman

Joined Jun 30, 2021
1,065
Yes it can be done and you don't need a diode. Just one op-amp, two potentiometers and resistors and one capacitor.
Can't provide the solution as this is homework. But here is a hint:

Consider how hysteresis (i.e. positive feedback can control the frequency) and a voltage divider can control the duty cycle.

Much success.

#### ericgibbs

Joined Jan 29, 2010
18,981
No, I made one in class with 3 and 2 opamps but they told us that it's possible with one so I'm trying to find out how.
hi dc,
The TS has done the set homework, this version is his personal project

E

#### dcbingaman

Joined Jun 30, 2021
1,065
hi dc,
The TS has done the set homework, this version is his personal project

E
ok, sorry about that, here is one possible solution replace R1 and R2 with a potentiometer wiper going to positive input. Replace R3 with a rheostat. The frequency and duty cycle can be set but it is not a linear relationship.

#### StijnC

Joined Oct 23, 2021
22
ok, sorry about that, here is one possible solution replace R1 and R2 with a potentiometer wiper going to positive input. Replace R3 with a rheostat. The frequency and duty cycle can be set but it is not a linear relationship.

View attachment 250920
Hello, I've tried the circuit but it doesn't seem to work, also I have a few questions, how does the frequency change when you're not changing the RC oscillator values? Why are there 3 pots needed, what does the third one do?

Cheers

#### dcbingaman

Joined Jun 30, 2021
1,065
Hello, I've tried the circuit but it doesn't seem to work, also I have a few questions, how does the frequency change when you're not changing the RC oscillator values? Why are there 3 pots needed, what does the third one do?

CheersView attachment 250922
You only need two pots R1 and R2 are just one pot with value of say 20K. Sorry for the confusion.

#### StijnC

Joined Oct 23, 2021
22
You only need two pots R1 and R2 are just one pot with value of say 20K. Sorry for the confusion.
Hi, still does not work.

#### dcbingaman

Joined Jun 30, 2021
1,065
Hmm, the circuit does work in LTSpice. When you change the feedback rheostat, the voltage swing on the non-inverting op amp input is decreased when the rheostat is increased in value. Thus the capacitor does not have to charge up as much or discharge as much, thus changing the frequency.

#### dcbingaman

Joined Jun 30, 2021
1,065
Hi, still does not work.
What parts are you using? Are you breadboarding it or simulating it with Spice?

#### dcbingaman

Joined Jun 30, 2021
1,065
Here is the LTSpice simulation looking at the output of the op-amp:

Also attached is the LTSpice circuit.

#### Attachments

• 1.3 KB Views: 11