thinking of building an inverter

Thread Starter


Joined Aug 13, 2007

with the plethora of inverters on the market (12volt dc to 120vac units) and
the fact that some are pretty cheap i got to thinking

the companies that make them, have to pay for all the overhead, advertising, and then have to sell them to wholesale distributors who sell them to retailers
and they then sell them to us.

so it would appear to me that for every 100 dollars of retail price the underlieing cost per unit in parts and manufacture is likely under 20 bucks.

so the question then becomes what if one was to put 100 dollars worth of parts instead of the 20,, would he not have a more robust if not a better inverter?

with microcontrollers one could certainly control a multistep waveform if not a pure sine wave,, get something pretty close?

with power electronics being relatively cheap surplus, is it possible to build up a pretty serious and dependable inverter for ones money?

the inverters i have been looking into are around 2 grand retail, or a bunch more depending on who you get it from.

i just got to wondering what one could build with his 2 grand in parts?
certainly a unit that could be repaired by the guy that built it i would think.

any thoughts?

bob g


Joined Jul 17, 2007
Some things you'll find it really tough to compete with:

o Brand recognition
o Distribution chain
o Advertising budget
o Cheap Chinese components and labor
o Service/warranty
o Mature designs

You can't count on getting a constant supply of surplus items. Sure, you might get enough to build a number of units, maybe even a LOT of units. But sooner or later, that supply WILL dry up. Then you'll be left with changing your design, or buying from the original manufacturer - and unless you're buying in bulk, you'll be paying far more than major manufacturers.

Don't forget that retailers put a substantial markup on everything - perhaps two to four times what they paid for a given item; whatever the market will bear. But even if you try to buy the same item from the manufacturer, you'll usually find it's even more expensive to buy it that way - unless you're buying a LOT of them. The manufacturer doesn't really want to deal with small sales; those are costly to them in terms of payroll.

Every business has a staff - unless it's a sole proprietorship - and there are ALWAYS expenses that have to be met; payroll, offices/manufacturing/storefront, etc. ad nauseum.

Not to mention that the big-name brands had their guts designed by teams of engineers to get the most they could out of the least expensive components; perhaps going as far as having custom IC's, rectifiers, enclosures/extrusions and whatnot made to optimize a specific design.

But don't let me discourage you ;)

If you can build a better mousetrap, people will buy it. :)

Thread Starter


Joined Aug 13, 2007

maybe i didn't make myself clear,, or made a misspoke

this is for a single unit for my own consumption,, not as some idea of producing them for sale.

but your points are well taken and illustrate better what i was trying to get across,, the manufactures have to make comprimises in order to meet market pricing and further comprimises to maximize profits for their company.

i don't have either of those considerations

what i am wondering is,, say for example

abc inverter company sells a multistep sinewave inverter that is 24 vdc input
and outputs 120volts at 4 kwatts and it sells for about 2300 bucks on sale on a good day, and retails for close to 3 grand.
but for sake of discussion i will use the 2300 dollar figure,
the way i see it is the cost for them to produce it in parts is probably less than 3 or 4 hundred bucks,,,

now what i wonder is ,,, if i was careful and shopped ebay for some high power parts,, what could i build for my 2300 bucks spent on parts?

i am thinking a guy might be able to build an inverter with perhaps twice if not three times the capacity, maybe not as many steps to form the sinewave, but not sure of that, and a unit that is servicable by me.

anyone ever built a high power inverter,, or thought about doing so?

not as a for sale project but rather for your own use?

it seems like it ought to be possible,, this is not cutting edge technology, requiring unobtanium parts.

i have dissected a triplite inverter, and found it to be basically a pushpull square wave unit,, that one should be able to produce in whatever size he wanted with banks of mosfets instead of the two that the triplite used.
granted that design would necessitate a large centertapped transformer, but one can be found and rewound to suit or suitable size... so a square wave unit is possible.

a stepped wave, such as those that are sold as modified sine wave,, basically is a bit more complex,, but with the use of a micro controller to triger the events it looks doable,, i have also dissected a 1 kwatt pure sine wave and found it to have a pic micro at its heart,,, the cost of which is under 10bucks. now that pc board was loaded with stuff,, but a good part of it was the battery charger function because it is a ups, more board was taken up for switching from ac mains over to inverter, there was a buck/boost converter to take care of sloppy mains, brown outs etc... all of which i don't need... you strip that away and there really isn't much in there.

it used two banks of mosfets 6 on each side, presumably 6 steps on the first half cycle + and 6 on the second half cycle --. each appear to be triggered by the pic chip.

so i am wondering if one built up a dc to dc converter that had multiple outputs or maybe multiple dc to dc converters to feed the mosfets and fire them with a micro to form the steps.

i guess the practical limitation might be how large can one build a dc to dc converter in ampacity? and can one converter be tapped for multiple voltages? maybe 50 amps as a topend limit,,,
so i would like to have at least four steps,,
so the first step could be dc input of 24 volts
the second step 36 volts from a 24 to 36 volt boost converter
the third step 48 from a 24 to 48 volt boost converter
the forth step 60 from a 24 to 60 volt boost converter

then use individual pins off the micro to provide the timeing and fire the
mosfet's to form the steps in the sinewave.

i know that the dc/dc converters of 50amp ampacity would probably be fairly expensive,, but we have 2300 bucks to work with.

anyone got any thoughts on this one?

bob g


Joined Jul 17, 2007
Oh, ok. Yes, I did read in something that wasn't there (I was getting tired) ;)

Sure, you could build something. Right now, it seems to me that the big-ticket items are going to be:
1) Bank of storage batteries
2) Your output transformer(s)

The batteries are your critical item. You're going to have to have a LOT of batteries if you wish to maintain high power output for any extended duration; they'll be expensive to buy and expensive to replace every few years.

Storage for the batteries - you're going to need a secure place that's well ventilated. Charging a bank of lead-acid batteries is going to generate LOTS of oxygen and hydrogen gasses - stacking them up in your garage between your breaker panel and hot water heater is begging for an earth-shattering kaboom. ;)

Charging and monitoring - you'll need to keep all of the batteries "topped off" without overcharging, along with monitoring the health of each battery.

Configuration of the UPS - a "true" UPS, like those used in mainframe computer installations is always on - that is, the power for the computer is supplied by the UPS. The batteries are kept charged by the input power supply. The kind that you buy for PC's are different - they switch on only when there is an AC fault detected. Sure, the mainframe-type UPS is less efficient - but it's extremely reliable.

As far as the current switches - sure, you could use banks of inexpensive power MOSFETS or the like, as long as they were well heat-sinked. You might take a look at these:
IRF9150 - 25A, 100V, 0.150 Ohm P-channel MOSFET
IRF520 - 9.2A, 100V, 0.270 Ohm, N-Channel Power MOSFET
IRFZ14 - 10A 60V 0.20 Ohm Single N-Channel HEXFET Power MOSFET

You can find these datasheets on - no charge or login req'd.

I recently bought a dozen of the IRFZ14's on Ebay for under $0.70/each, including shipping - just to give you an idea what they can be had for.