# PWM frequency divider

Discussion in 'General Electronics Chat' started by DominusDRR, Sep 3, 2009.

1. ### DominusDRR Thread Starter Member

May 25, 2009
26
0
Greetings.

Is there any way to divide the frequency of a PWM signal and simultaneously also the duty cycle with flips flops and logic gates?

Thanks.

Fabián

2. ### bertus Administrator

Apr 5, 2008
15,799
2,385
Hello,

When you put a PWM signal into a divider, you will end up with a signal of half the frequency and 50 % duty cycle.

A possibilty is to take an AND port take the input signal and the divided signal.
You will get a signal with half the frequency and half teh duty cycle.

Greetings,
Bertus

3. ### DominusDRR Thread Starter Member

May 25, 2009
26
0

Is it correct?

4. ### Ron H AAC Fanatic!

Apr 14, 2005
7,050
657
As bertus said, it will give you half the original duty cycle. That's not what you asked for.

5. ### DominusDRR Thread Starter Member

May 25, 2009
26
0
Hello. Thanks for responding. I do not want half, I want double the time in duty cycle

Last edited: Sep 3, 2009
6. ### DominusDRR Thread Starter Member

May 25, 2009
26
0

I have this circuit. In IN when the duty cycle is 10% in OUT duty cycle is 95%. When IN the duty cycle is 90%, the duty cycle at OUT is 55%. Responds to the equation 1-t / 2 (where t is the duty cycle in IN in%)

I need to divide the PWM signal several times (I need to measure the duty cycle of a high frequency signal), but using this circuit the duty cycle changes according to:

Period __________Duty cycle (%)
T ______________ t
2T _____________ 1-t / 2
4T _____________ 1/2-t/4
8T _____________ 3/4-t/8
etc
Ultimately, the duty cycle variation is so slim that you can not measure.

Thanks

Last edited: Sep 3, 2009
7. ### hgmjr Moderator

Jan 28, 2005
9,030
214
It seems to me that one solution would involve converting the PWM signal to its equivalent DC level using an RC filter. You would also need to divide the squarewave by whatever factor you are looking for. You would then need to use the lower frequency squarewave to generate a new ramp the repetition rate of which would be the same as the new lower frequency squarewave. You would then use the recovered DC and the new ramp as inputs to a comparator. The new PWM would be the same as the old PWM only at the slower rate.

hgmjr

8. ### DominusDRR Thread Starter Member

May 25, 2009
26
0
It seems a solution. But it doesn't work in a SPWM signal.

buscar

9. ### hgmjr Moderator

Jan 28, 2005
9,030
214
So do I take this to mean that you are dealing with a Sinusoidal PWM signal rather than a conventional PWM signal?

hgmjr

10. ### Ron H AAC Fanatic!

Apr 14, 2005
7,050
657
Can you describe in detail what you are trying to do, and why?

11. ### eblc1388 Senior Member

Nov 28, 2008
1,542
102
DominusDRR,

Are you asking the question if a given PWM frequency can be reduced to(1/2, 1/3, 1/4...) by logic gates while keeping the same duty cycle?

Interesting question albeit I don't know the answer.

12. ### DominusDRR Thread Starter Member

May 25, 2009
26
0
Yes. It's a sinusoidal signal that is compared with a triangular wave.
The sinusoidal signal is 1 MHz and 100 MHz triangular

13. ### DominusDRR Thread Starter Member

May 25, 2009
26
0
I'm building a kind of AD converter. I don't want to use a 100MSPS ADC for the following reasons:

1. The ADC is very expensive.
2. I use a 40 MIPS microcontroller, so it is very difficult for my microcontroller reaches 100 MSPS ADC.

14. ### Ron H AAC Fanatic!

Apr 14, 2005
7,050
657
I don't understand how your 100MHz PWM relates to an ADC. Please provide a block diagram of what you are trying to do.

15. ### DominusDRR Thread Starter Member

May 25, 2009
26
0
The variation of duty cycle is proportional to the analog signal. When the duty cycle is zero, implies that the analog signal voltage is zero. When the duty cycle is equal to 100%, the analog signal is at its maximum level. The amplitude of both signals is the same.

I get different and proportional duty cycles per 10 ns (100MHz).

:S

16. ### Ron H AAC Fanatic!

Apr 14, 2005
7,050
657
How about a block diagram of the hardware you propose to use?

17. ### DominusDRR Thread Starter Member

May 25, 2009
26
0
It's not complicated I want to do. Simply divide the SPWM signal for example up to 32 times, (3.125 MHz). Assuming the duty cicle somehow remains in proportion to the original, I would use the module comparison of my microcontroller to measure the duty cycle time. This time is proportional to the amplitude of my 100 MHz analog signal.

18. ### Ron H AAC Fanatic!

Apr 14, 2005
7,050
657
If you are modulating the duty cycle at 1MHz, how can you get useful information about that modulation if you divide the PWM frequency by 32?

19. ### DominusDRR Thread Starter Member

May 25, 2009
26
0
For example assuming that the duty cycle at some point is 10%, this implies that it's 1 ns, by dividing the frequency of some special arrangement of logic gates will be the new frequency of 3,125 MHz or a period of 320 ns. If the duty cycle is kept in proportion, also will be 10%, meaning that it's 32 ns. This time can be measured by my microcontroller.

20. ### blueroomelectronics AAC Fanatic!

Jul 22, 2007
1,758
98
Ok I'll ask, what is this for? Are you trying to measure the speed of light?