Ancient_Hacker said:
elrefaie wrote:
Problem is, these high-power circuits can be touchy. And one little
wiring mistake or short circuit and all those MOSFETS go Poof!.
I'd start with a much simpler inverter, such as your basic Royer
inverter.
One transformer, two transistors, two resistors, two capacitors, and
voila, a usable inverter. For a basic schematic, see the bottom
schematic in:
http://www.butlerwinding.com/elelectronic-transformer/switch_mode/inverter.html
The circuit you asked about looks like trouble with a capital T, that
ryhmes with pee, and that's what you'll be when it goes up in smoke.
motorola actually makes a really good set of documents on inverter
design, however their impossable to find, especially because they
restructure their documentation authorities every so often, I only have
a printed copy.
but, may I suggest throwing some newer technology at it?
I'v been mulling over a 4kW UPS design for running my house (the hot
water, stove dryer etc is on gas, so no huge power demand)
what I'm thinking so far is this
48v bank -> dc-dc converter -> modulator -> filter ->
here is why..
48V battery bank:
this voltage is still somewhat sane, and yet high enough to keep
currents down to something sane at 4kW
approximitly:
4000/48 = 83A
if it were 12V:
4000/12 = 333A
which is getting quite scarry. also if the drive to keep it as efficient
as plausable, you want to keep currents down, so your I^2R losses are lower.
,001 ohms of wire * 83 amps ^2 = 6.9W (now imagine at 333A
)
DC-DC converter:
this is to boost the 48V to 170V (my output goal is 120V, not 240)
As a switching converter, these can be quite efficient and are still
somwwhat easy to build, I'd be going for a boost converter,
non-isolated. Because there is no way its practical to do this in 1
converter, it would be done with a set of converters (say 10) which
makes the core of each one obtainable, and keeps for current for each
converter reasoanable. The other advantages of splitting it up are that
I can offset their timings, in effect a 10 phase converter, this eases
life on capacitors etc etc. splitting it up also allows for some
redundancy, if a converter decides to bite the buscut, it can be
isolated and life can go on (which may be where the system goes into a
chain reaction failure, but thats what microcontrollers are good for
detecting) it also means I only have to design 1 converter and make it
10 times. This would be FET based, possibly IRF1404 or soemthing
modulator:
the 170V is perfect for bring modulated into a sine wave to form
120Vac, this would be done via pwm, making the process nicly efficient,
probably PWM at a freq of 25 to 50Khz. This would probably be fet based
also, although at these voltages, bipolars usaully prevail efficiency
wise (high voltage fets tend to have a high Ron, a bipolar transistor
has a fixed saturation voltage, there is a point that the I^2R losses
exceed the IV losses)
filter:
this would be a low pass so that the pwm freq dosn't get through, just
a nice true sine 60Hz.
Another nice thing is that this circuit is easily made on a small scale
for testing with a 40W lamp
dan
