Any properly rated FET or IGBJT (standoff potential, current) should be able to do this through a full-wave rectifier bridge.

thanks for your help. i dont have much knowlage about transistors. i am trying to develop a new product to market, your help is appreciated. anthony7

Nice answer. IXYS certainly has semiconductor hardware capable of steering this much power. Which does answer the specific question anthony7 posed. But a block diagram schematic would clear up what you mean by use a FET or IGBJT through a full-wave bridge rectifier. The circuit topology that comes to my mind from your description would produce a chopped DC signal, not a chopped AC signal. What you have in mind might produce anthony7's request of chopped AC, but it doesn't seem an obvious topology to me. (I may just be a little dense this morning.)

Your circuit topology just might answer my concern about what type of load is expected, and what frequency power is being used. For quickly switching on and off a one Henry inductor that has 60 amperes running through it will be quite tricky, particularly since as we all know the current in an inductor lags behind the voltage.

I drew it all up but am remote and using Safari browser and can't figure out how to insert the pic I drew. Will respond tonight after getting back to home-base. Resultant waveform is not clipped. Switched on waveform output is two diode drops lower in potential than input power waveform, plus IR drop across FET "ON" resistance.

This is emc_c.

from anthony7; my thanks to everyone who sent advice. the product i am trying to develop will be used to power a communication machine which is mostely used by the government. it will run on power provided through a common wall outlet 120vac 30amp, the current must pass through my device and into the machine carrying the 120Vac, i do not need it to convert into DC current at any time, as the 120vac will carry on to another machine that the first machine will supply power to, and that next machine runs on 120vac also. so the entire process must run on Ac current the entire time, without needing DC current at any time. i hope this helps to explain what i am trying to do better. thanks anthony7

So from what you are saying it is a 120V ac power supply switch, but why do you need to switch at 1000 times a second (this creates all sorts of problems)?

Is the purpose of this switch for data communications over the ac grid power lines?

jack; the ac grid power lines will not be used to transmit or receive. regular phone lines, satelite, sw radio, and other dedicated ground lines will be used. the purpose of this product i am trying to develop, will add a layer of difficulty, along with the other protocals involved with the opperation of certain communication machines. all wires carrying ac current are sheilded, and so will be the product being developed, to surpress any unwanted frequency problem. when the times comes to assemble my product, i will find a qualified elect.engineer to assemble and make sure i don't fry my %*?. thanks again to everyone for your help. let me know any other questions. tony

Okay, as requested, here's that block diagram of a rectifier bridge, FET & driver that is an actual ac switch, not a pulsed dc converter. Let me know if this makes sense or if it requires further explanation.

EMC C....just as I thought, a steering diode bridge, definitely AC output as promised.

I am a bit curious as Tony hasn't replied (or did he) to any of our latest postings, I would really like to at least know a very basic part of his intended application as I think your solution gives him the basic idea but he is going to need a lot more information to get the job done or the fourth of July is coming early this year!

Yep, that will certainly chop an AC signal! Classic example of a picture worth a thousand words. A little bit of care will be needed to let the FET and FET driver circuit ride the supply voltage. Been a long time since I've played with this approach. Sometimes when a wrong circuit gets in your head... I thank you for the clarification.

This approach won't solve my concern of turning off a reactive load. I too wonder why anthony7 has not answered this concern. I hope he's OK.

redfred,

(This is emc_c again).

Thank you for pointing out one thing I had forgotten to mention. I used an NTE 3090 opto-isolator between the circuit shown in my drawing post and the circuitry that controls the switching. That way the only dc supply that is referenced to the ac mains is the FET driver and FET; the rest is relative to the grounded chassis of the box. As you and the Wiz have pointed out, such a circuit should be put together by someone with experience; not someone building his first or second circuit.

7.2 kilowatts of power is a lot of juice. You'll need to be a little more specific about what you're doing. What is the frequency of the of the alternating current? Will you be driving a reactive load? But I can almost guarantee this won't be a small part to survive switching 60 amperes of current.

While I agree with EMC C that what you are asking for can be done, we really need more information on the load you are switching this high current into. As RedFred asked, is your load strictly AC or can it tolerate the rectified DC that EMC C referred to? I guarantee that your circuit will require more than just an FET and driver to work correctly.

This much power can be very dangerous to mess with, even with the correct fusing, a circuit failure can be really nasty. While 1KHz sounds like a low frequency, switching that much power off and on at that rate entails a lot of parasitics which can bite you. This is not a particularly good project for a novice to tackle. Despite our best efforts to give you good answers, I would strongly suggest you defer to someone with experience with this kind of high power circuitry for your own safety.

emc_c again. I totally agree with the Wiz that you are working with hazardous power levels, and switching this much current on and off will cause nasty switching transients.

But my circuit will produce switched ac just as the OP asked, not switched dc. You will have to take that on faith until I am back in the (home) office this evening and insert the pic.

Okay. EMC C, no problem, I think I know where you are going with this......

I'm just kind of nervous about what this OP is wanting to do and his unknown level of experience. I've done some work with this kind of juice and it isn't for the novice level techies. Fuses and circuit breakers can't work fast enough to prevent fireworks!

No argument from me (emc_c). FETs generally sacrifice themselves to protect fuses. :-)

However, these IXYS big block FETs are the real thing; handled properly, they will do the job.

A minor byproduct of switching 60 Amps on and off rapidly (sub-100 ns rise/fall-times) will be rfi that you won't believe. Your AM radio will squeal like a banshee.

The time-domain manifestation of that (the source of the rfi) will be nanosecond duration transients into the kilovolt range, unless there are low impedances on the bus to load them. I don't see these as dangerous to other loads on the bus in that they are relatively high impedance and easily loaded by a capacitive front end, but your power quality will be horrendous. Now this could all be solved with some serious line filtering just before the switch. That way the switched power will of course still be noisy, but the rest of the bus won't be trashed. Also, if the switched bus is relatively short (in physical length) relative to the unswitched bus, the amount of radiation will be reduced, and the effect on broadcast band AM reception will be more localized to the switched bus.

I quite agree, the key words being "handled properly!"

I can just hear the neighbor's complaints to the power company about all the hash on the line and maybe the FCC....that might get a quick response from them to the cause of it all. I am rather doubtful our OP really knows what he is getting into....like I said, those parasitics can really BITE!. That filter is going to take some serious cash to build an effective one, inductors at that current (likely have to be rated somewhat higher than 60 A to avoid saturation at the much higher peak currents) and high voltage, high dV/dT capacitors, ouch.

I concur with others Tony, what you are actually asking is not simple (you cannot just connect a big transistor and it will work). Can you provide more information on what you are trying to do and your level of electrical experience?

I have this bad feeling that you are a backyard inventor trying to build a energy-saving device by pulse-width modulating the house 120V mains supply.

We are all here to help, but without more information from you we cannot safely supply more information than "you need a FET or IGBJT", especially when your level of knowledge of these basic electrical parts is too low for the specific application you describe, and trial and error on your part could be potentially fatal.

I look forward to seeing emc c circuit diagram. Producing chopped AC directly sounds interesting. I suspect I'm going to look at it once and say, "Of course that will work!"

I hope that anthony7 is soaking up our safety comments, 7.2 kilowatts will bite people and things very hard. I agree this should not be a first project for anyone.

I would like to know what the objective of switching this much power a thousand times a second though. That would give us an insight as to how reactive of a load he has.

Look for 1kHz frequency inverter. You may easily get for 400Hz version but 1kHz version can be custom built. I did get one custom built for 1kHz.