square wave to sine wave

[The74namesofgod]

i dont understand anthing so please make it simple. i have a square wave and i need to make it to a sine wave. its 250 v 20000 watts


20 min later

o yeah its 1-100 KHz

 

[Sadegh.j]

100 khz sine-wave oscillator

How about a band pass filter

 

[echo47]

round off square wave

A twenty kilowatt bandpass filter with 100:1 tuning range? Wow, show us how!

 

[The74namesofgod]

change square wave to sine wave power filter

maybe i should say the 1-100 kHz is what the square wave will be and can i even change the square to a sine at 20000 watts or am i dreaming

 

[GrandAlf]

Very tall order, especially at this sort of power. What are you driving?. If it is inductive, it may just round off the edges a bit, enough to be usable. I would personally be a bit wary of experimenting at such high power levels, such a wide range would be very difficult.

 

[echo47]

A low-pass or bandpass filter could remove the high-frequency components of a quare-wave and give you a sinewave, and a high-power filter is achievable, but the widely variable frequency part seems like a major complication.

Just a wild guess . . . Instead of that approach, maybe you could modify the square wave oscillator into a pulse-width modulator like in a high-power class-D (switching) audio amplifier. The output filter would be a lot simpler (but still nasty at 20kW).

Either way, that's a lot of dangerous power to be fooling around with!

 

[The74namesofgod]

it is inductive but this is the olny part that is a circuit. i dont know if it will work but i have to find out and i wont take anyones word for it. i have to try it my self. i can can change the frequency down .0000001 Hz and up to 100 KHz. i could just pick a frequency and stay with it but i want some range. maybe something like 32 KHz i think i could live with that if i have to pick a frequency. also i must use this crazy amount of power.

 

[rudane]

At a power as such, maybe an expert opinion would be desired. There seem to be too many unknowns, like is there a chance this could inject harmonics into the grid? If so, fines may be incurred. The obvious answer is a filter, but I doubt your average engineer or hobbyist has ever designed a practical filter for such a power.

 

[The74namesofgod]

the square wave modulator is not like a modulator its mechanical



2 min later


o yes your thing about the grid i am using a generator

 

[FvM]

the square wave modulator is not like a modulator its mechanical

You mean the 100 kHz square wave is sourced from a mechanical modulator?

 

[webdog]

Now this is a real mean problem
Square wave is fundamental tone plus odd overtones, i.e. 3, 5 7... These will contain several kW and if you go for a simple lowpass (or bandpass filter) the power will reflect back into your generator. Maybe you need to have a di-plexer, i.e. a highpass/lowpass filter and then terminate unwanted energy. If you just want to filter the signal, do look at a bandpass filter, as it may give better supression vs. lpf with same order.

 

[FvM]

In case of LF and without km of power distribution cable, one would normally analyse the harmonics as apparent rather than reflected power. It's not like a RF power generator that possibly needs a circulator and a absorber to protect against reflected power. E. g. in a FET inverter, the power would simply flows back to the bus, causing minor losses in the power switches.

 

[The74namesofgod]

what if there is transformers and diodes right before the squre wave modulator dose that make a differents with the feed back or if a dobble up my transformers to make more separation from the generator. thanks for all your input you guys really gave it your best shot and sorrie about the hard question. i might have to think of a dirrerent way maybe change the squre wave modulator


like i said befor i dont understand anything so this might sound like a stupid question but do i even need to change it to a sine wave its for an inducter

thanks again

 

[johnyaya]

How 'bout converting your 250V sq-wave into DC, then generate the sine wave. I've seen HUGE UPS systems that do something similar, although at a fixed frequency.

 

[kboltz]

What if we did away with the power requirement, and shorten the bandwidth to 20 kHz. Any suggestions?

For example, using a relaxtion oscillator with a pot in the feedback. Then feeding this signal into an amplifier?