Build 3-Phase Sine Wave 2.5-3 MHz Power Supply

[biff]

I am trying to build a 3 phase sine wave 2.5-3 mhz power supply.
To do that with electronic's is beyound me. I have a 3 phase 600hz
100amp generator. I was thinking that I could use frequency
multiplier (12x) to get the frequency to 2.5 mhz. Any ideas on a
circuit that would work?

 

[berkeman]

Welcome to the Physics and Math Forums (PF), biff. Sorry, when it's apparent that any advice we could give you would result in your immediate and serious injury, we hold off giving advice.

Your question and the answer are well within the forum capabilities. Please post more details about the problem, and your qualifications to deal with the problem.

 

[biff]

Thanks for the reply
I assume you are referring to a lack of shielding? A heavy gage steel box should redirect that stray flux. There shouldn't be any high voltages. Is there something else I'm overlooking?
The coil that I have been working with is 250mv @ 700 amps @ 400hz.
Not a lot of safety required.

Any help would be appreciated!

 

[berkeman]

Thanks for the reply
I assume you are referring to a lack of shielding? A heavy gage steel box should redirect that stray flux. There shouldn't be any high voltages. Is there something else I'm overlooking?
The coil that I have been working with is 250mv @ 700 amps @ 400hz.
Not a lot of safety required.

Any help would be appreciated!

Nope. What is your background? What degrees and professional certs do you hold? If you are a home-brew experimenter with no advanced training and knowledge trying to work with AC Mains power, you could get yourself seriously hurt. And we're not going to help with that. If you have an MSEE and routinely fix TVs for your friends using an ISO transformer and SAMS Photofacts, then that's different, right?

 

[Panda]

It can be done quite easily if you are looking at signal generation, but I agree with Berkman that your post seems to imply that you are trying to use this for power generation which is potentially painful.

400Hz looks like an aircraft power supply what application are you trying to achieve?

There may be more sensible methods to achieve 3Mhz source power than using a frequency multipying circuit, although what you would do with it I haven't a clue, I've never seen anything much above 1200Hz for source power applications.

 

[waht]

You mean 2.5 MHz ?

Your input frequency would be 208 KHz right?With passive multipliers there is always a conversion loss, by the time you get to x12 you would lose so many amps you couldn't even power a light bulb.To achieve this kind of power level your need to utilize a vacuum tube amp. First convert your input source into a high voltage DC, then feed it into a carefully tuned amp at your frequency of interest.

I'm still confused,

Is this your input source "250mv @ 700 amps @ 400hz"

That's about a 200 Watts and solid state amps should handle that without HV.

 

[Averagesupernova]

I'm quite confused. You are starting out with 600 hertz and want to end up with 2.5 to 3 Mhz? This is more than a x12 operation. I know switch mode power supplies go up to around 500 Khz but never heard of anything as high as 2.5 to 3 Mhz. This would essentially be a radio transmitter. Exactly why do you need this kind of frequency?

 

[biff]

The Project:
--3 PHASE-- 2.5 mhz (min.) 500 ma (I will amplify later)

A simple oscillator will make 2.5 mhz. The trick is to get it to 3 phase.

All car alternators are wound 3 phase. I have a 6 pole unit. At 3600rpm it will put out 360 hz. The max. is about 600-650 hz.
I thought it would be easier to work on getting the frequency right rather than the phase.

I understand that power loss is going to be high with a frequency multiplier.

If I could get 10amps out at the end I could skip the amplifier.

Thanks

 

[Panda]

Still can't work out why you need 3p 3MHz POWER supply.

You could build three oscillators and phase lock them.
You could build one oscillator and split it into three phase shifted paths.
Or you could multiply 600Hz 12x using resonant amplifiers.

All these schemes will only give you a 3ph 3MHz signal sources though.

Your scheme with the alternator also seems to imply that you think that by multiplying each phase of a 600Hz supply will result in 3p 3MHz. It won't because the phase angle between the phases will still be the 600Hz phase angle not the required 3MHz phase angle.

If you truly want 200W odd of power then you have two possibly 3 options.

1. Have a rotary convertor built, but I have never seen any running above 100kHz and they were only single phase.

2. Build a switch mode PSU, but again the chop rates are going to huge to get a reasonably smooth 3Mhz output.

3. Use resonant cavities and cross feed the outputs through a phase shifter. I have only seen this done at GHz frequencies though, a 3MHz cavity would be quite a beast.

 

[biff]

I'm a little depressed my alternator idea won't work, Well back to the drawing board.
I like the idea with the 3 oscillators phase locked. Could you recommend an area to look for infomation on phase locked systems.
I also liked the idea of one oscillator with 3 paths. An inductor would shift the current 85 deg.or so, but its still 30+ deg. short!

Thanks for your help

 

[biff]

panda:
Typing without thinking can be dangerous!
I was thinking current not power. Your right! get the current to lag and the voltage to lead (with the right calculation) will be 120 deg.
I'll sharpen the pencil an get with it.

hopefuly I can get this project to work.
thanks again

 

[berkeman]

One other way to generate the 3 sine waves would be to use ROM-based DACs. You could store 3 sine waves in 3 ROMs (calculating the data for the ROMs with the appropriate phase shifts) and have the ROMs drive DACs to make the 3 sine waves. And if you want to get trickier, just use one ROM and have the addresses offset by 2*PI/3, and interleave the fetches for the 3 DACs.

EDIT -- You probably want at least 16 samples for each DAC, so 16 * 2.5 = 40MHz for the clock for the 3 separate ROM circuit. That's very do-able with simple cheap ROMs.

 

[biff]

berkeman:
This is similar to my first idea since my house is powered by an inverter. In the inverter there is 15 different sized windings on the core, with a computer to decide witch is on or off. The transformer for 2.5 mhz would be very hard to make. 40ga. wire or thin foil,ferrite,low permablity core. O my, it would be a hard wind!

but without the transformer hmm.
Thanks for the help.