Hello Gökhan,

do your two MOSFETs form a push-pull stage? If so, you could solve your problem by using one n-channel FET (source connected to GND or the negative supply rail) and one p-channel FET (source connected to the positive supply rail), and connecting both gates to the same PWM signal. With power MOSFETs, the problem remains to delay the signal edge (for each MOSFET separately) which switches the respective MOSFET to the on-state. This can be done by a low pass filter, bypassed by a diode, and followed by a gate with Schmitt-Trigger input, for example.

If, for some reason, using complementary MOSFETs isn't an option, would not inverting your PWM signal yield just the second signal you need?

Thanks for answer. Yes it will form push-pull stage but i can't use n-channel or p-channel mosfet due to different libraries. Also by "inverting" do you mean using "not gate" ?

if so, The inverted signal is not my need. I want, in the second signal pulses are shifted by their duty cycle( or the width when the pulses are "1").

By inverting there are either "0" or "1" as reverse of the first signal's pulses. But in the second signal it is supposed to be "0s" even when the first one is "0".

The problem is the duty cycles of the first and 5th or 10th are different because, as the duty cycle input of PWM generator i am using a sine wave.

If I understand correctly, the second signal should have the same duty cycle as the first. I guess the only way to achieve this is by some kind of "memory" which "records" one pulse at a time of the first signal and "plays back" the pulse after its trailing edge.

In analog technology, a suitable circuit could have some similarity to a two-slope ADC: During the pulse, a capacitor is charged by a constant current. At the end of the pulse, the pulse of the second signal is started and the capacitor is discharged by a constant current of the same amount as the charging one. When the voltage of the capacitor has reached the value it had before charging, the pulse of the second signal is terminated. Advantage: continuous time; disadvantage: minor timing errors possible, maybe dependent on temperature.

In digital technology, for example, the leading egde of the pulse could start a counter counting up from zero; the trailing edge causes the counter to reverse its counting direction, of course clocked with the same frequency. The pulse of the second signal starts at the trailing edge of the original pulse, and ends when the counter reaches zero again. All the advantages of digital circuits; disadvantage: discontinuous time.

In order to amend this disadvantage, the frequency of the clock for the counter has to be several orders (e.g. three) of magnitude higher than your PWM base frequency.

I have no access to Simulink libraries at present, so these solutions are just "generic" ones.

P.S. The principles described above work correctly as long as the duty cycle of the PWM signal is less than 0.5. Since at a higher duty cycle the primary and the secondary pulses are overlapping, two identical circuits have to be triggered by the primary pulses in an alternating way. This is certainly a strong argument for a digital solution. If the PWM frequency is not higher than, say, 20 kHz, one or two microcontrollers might provide an acceptable and inexpensive solution. The PWM pulses could toggle a flipflop, and the signals Q and !Q could (e.g. via AND gates) route the (slightly delayed) PWM pulses to two microcontrollers, alternatingly. The two output signals of the microcontrollers have to be combined, of course.

P.P.S. While your question poses an interesting problem, I do not yet understand the purpose of the second signal ...

Thank you very much.

I am dealing with my final exams these days so I could not return your question in the last. When I get rid of my exams i will use your advice to handle the problem and will explain you better.

P.S The problem is a bit complex even i am telling it in my native language, during translation in English i may have been clearer :)

Hi Gökhan,

good luck with your exams!

Hello again, Thanks for good wish,

All i want for now, to have pulse signal just in simulink, like the second graph in the picture.(the white pulses). I explained better in the pictures. I tried lots of blocks in Simulink but i could not manage it.

For know i should do it in Simulink then i will combine a PV panel and so on. But i am stucked in this shifting matter.

Dear Goekhan,

welcome,

If i understood you correctly you have to generate two pulse trains with one pulse train just delayed by the pulse width of the first pulse train. Is this right? If yes,

You can generate the delayed one by a negative edge triggered monosyllable multi vibrator with adjusting the monostable pulse width to be equal to the original pulse train pulse width. Monosatbles are available as ic chips.

Best wishes