# Phase Shifted PWM Signals

Discussion in 'General Electronics Chat' started by george020106, May 29, 2013.

1. ### george020106 Thread Starter New Member

May 29, 2013
6
0

I am trying to build a circuit using op-amp (LM7171A, TI) to delay 180° of a PWM signal (0v-3.3v, 500kHz).

I barely know anything about circuit, since I am a civil student. That's why I am here looking for help.

Any input will be appreciate. Thanks all.

2. ### tshuck Well-Known Member

Oct 18, 2012
3,531
675
If all you need is a 180° shift, an inverter will do that...

3. ### GopherT AAC Fanatic!

Nov 23, 2012
6,315
4,041
if you already have the LM7171a, just build an inverting amplifier with gain of 1.

Start with both resistors at 10k ohms (output voltage will equal input voltage).

4. ### Ramussons Active Member

May 3, 2013
562
92
Is Inverting the same as a phase Shift? The op was the need for a delay by half clock cycle.

Ramesh

5. ### crutschow Expert

Mar 14, 2008
13,502
3,376
An inverted signal is the same as a 180° phase shift.

Oct 18, 2012
3,531
675

7. ### george020106 Thread Starter New Member

May 29, 2013
6
0
Thanks, but my advisor preferred to use a op-amp instead of a hex inverter. That't the problem.

8. ### george020106 Thread Starter New Member

May 29, 2013
6
0
Thanks, I was thinking of inverting amplifier circuit, but the initial amplitude is 0v, not 1.65v (half of 3.3v). So after inverting, the PWM signal is not right.

9. ### Stuntman Active Member

Mar 28, 2011
189
50
If I am reading the goal of this circuit correctly, george, the inverting amp will not work.

An inverting amp will only shift the signal 180° (and retain PWM characteristics) when the master signal is 50% Duty Cycle. (figure 1 and figure 2)

Unfortunately, to actually produce a PWM phase shift of 180° regardless of duty cycle, you'll need a propagation delay of 1us (period of 500kHz=2us). (figure 3)

This can be done digitally without too much trouble. But analog will get substantially more complex.

• ###### Signal Delay.png
File size:
121.7 KB
Views:
56
Last edited: May 30, 2013
10. ### Ramussons Active Member

May 3, 2013
562
92

Sorry, I don't think so. Its only when the stream is a square wave can one take the inversion as 180°phase SHIFT. The delay can be considered as +180° (leading) or as -180° (lagging).

In the case of a PWM, the two are NOT the same.

File size:
22.5 KB
Views:
38
11. ### GopherT AAC Fanatic!

Nov 23, 2012
6,315
4,041
George,
Could you please post a hand drawing of your desired output based on some input of varying pwm duty cycle. Your currnet question can be interpreted several ways, we need to know what you really want.

-Mark

12. ### george020106 Thread Starter New Member

May 29, 2013
6
0
Thanks for your reminding. I should have done it earlier, so there will be no misunderstanding. I have attached the input and output signals I want.

File size:
59.7 KB
Views:
46
13. ### GopherT AAC Fanatic!

Nov 23, 2012
6,315
4,041
George,

this is an interesting problem but very challenging unless we move to a microcontroller. Can you explain "why" you need the PWM delayed by a half of a pulse? Better understanding of the problem will help us design a better solution.

14. ### george020106 Thread Starter New Member

May 29, 2013
6
0
Because my advisor wanted me to generate a PWM signal from -3.3V~3.3V, and kept the initial amplitude at 0V.

I was wondering I could make that out with delaying 180° of the signal, inverting it and combining it with the original one. That is why I want to delay that PWM signal.

File size:
57.4 KB
Views:
24
15. ### Stuntman Active Member

Mar 28, 2011
189
50

I will say, this sounds like kind of an "odd" project. The zero initial amplitude has me puzzled as to what that is achieving, but lets move on.

Have you considered using your op amp to generate the -3.3 to 3.3V wave (perhaps something like this: http://hyperphysics.phy-astr.gsu.edu/hbase/electronic/square.html)

But instead of powering the device right away, use a 555 or even another 741 to put a time delay to turn the power on and off to the device, giving you the 0 voltage at the beginning and end of the signal?

This would be my suggestion.

16. ### george020106 Thread Starter New Member

May 29, 2013
6
0
Really appreciate your suggestion, I will think about that very carefully. Thanks.

17. ### GopherT AAC Fanatic!

Nov 23, 2012
6,315
4,041
Sometimes an advisor has to be taken by the hand and guided to a workable solution. In other words, an advsor's first impression of how to solve a problem is rarely the best one.

I am guessing you are trying to maintain a system around a center point. If that is the case, a center point Pwm is not an easy solution.

If you need some alternatives to solve this problem, let us know some details.