I have a problem shifting a three PWM signal (10KHz Saw-tooth wave form) with the second signal shifted form the first by Time delay = 0.00005 and the third shifted by 0.000025 from the first signal, is it possible to get this and extract this three control signals (C1, C2, C3) to control the gates of the converter ? any tips or suggestions. PLS take a look at the picture included.
Hi Okba Salah ,
I don't understand your problem completely:
Are the three blocks with a "greater or equal" symbol in your picture comparators?
If so, are d1, d2, and d3 analog control voltages?
Is it really necessary to delay the analog signal (saw-tooth) instead of the digital signals behind two comparators?
Regarding the delay, I assume that the unit of the given delay times are seconds? Then the delays are 25 and 50 microseconds.
A few decades ago, there were two different kind of devices for delaying analog signals on the market, based on the bucket chain principle (the signal is propagated through a chain of capacitors) and based on the propagation of acoustic waves. I guess most of the bucket chains are obsolet today. Besides, the spectrum of a good saw-tooth signal contains many harmonics; with a base frequency of 10 kHz, the spectrum might reach well beyond 300 kHz, and I feel that bucket chain devices were not suitable for frequencies above, say, 20 kHz, while the delay time of acoustic wave devices hardly reaches 50 microseconds.
A possible solution could be to use a microcontroller with ADC and DAC: The analog signal is fed into the ADC, the digital result is written into the RAM, forming a ring buffer, is read from the ring buffer after the required delay, and fed into the DAC which delivers the delayed saw-tooth signal.
Problem: If your application requires a PWM resolution of 256 different values, for example, the sampling frequency of ADC and DAC had to be at least 2.56 MHz, better more. Most on-chip ADCs and DACs in microcontrollers are slower.
I feel that your problem could be solved much easier by replacing the whole circuit in your picture by a suitable microcontroller with three analog inputs for d1, d2, and d3, and three PWM outputs, using internally three synchronized timers.
It might help if you would post a diagram of the three PWM signals, depending on d1, d2, and d3.
PS I just realized that your dSPACE board contains a TMS320F240 as slave DSP with three timers capable of generating PWM signals. Please have a look at page 6-3 (and the whole chapter 6) of the attached Reference Guide.
https://www.physicsforums.com/threads/trying-to-make-a-30-microsecond-delay-circuit.623639/
This all-pass filter concept might work - but you need a wide bandwidth if you want sharp points on the sawtooth.
Hello dear professor Joerg Fricke I really appreciate your time insuring my question which is really helpful and very clear...especially highlighting the existence of the TMS320F240 in my Dspace board also your other propositions using a microcontroller ...I am going to take a look at Chapter 06 (Event Manager Module) ... and if want to replacing the my circuit ( in the picture) by a suitable microcontroller do you sir suggest any one because I think about Arduino board playing the role of shifting the signals. at the end thank you sir very much again and i hope you all the best.
Hello, if you are trying to control a converter using a carrier based PWM generator, so you have to use the slave DSP of you 1104 board instead of the master DSP.
The dspace board contains two processors: a master and a slave. While the master processor is used to execute the RTI application, the slave processor is used to execute all the secondary operations which require very high frequency like generating converter signals. So the master processor is unable to generate high frequency signals like saw tooth signal and execute the RTI application at the same time.
In the Simulink library, you'll find two different libraries, one contains blocks of the master dsp, and the second contains blocks of the slave dsp.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Waves