Google Offers a Million Bucks For a Better Inverter

Would seem that a switcher is a good start. Concentrate on minimizing ON IR losses, then on dI/dT losses which implies that switching frequency / "rate" should not be too high; lastly on core losses. Can't think on what else is needed.

Maybe someone can build a really fast mechanical relay, does it say how long it has to last for?

George H.

Yes, banks of relays or very low resistant MOSFET, switching to pre-charged capacitors.

With 99% of 2KW, 20W is no big deal. Around 5W for the controller and 15W for relays. Might be doable. Current technology is around 97%.

Long enough to demo and get the money.

Hi,

There is this strange sentence in the LBC-InverterRequirements.pdf document:

".. the inverter must also have an efficiency of > 95% at full 2kVA load. The reason for this is so that a team cannot use a forced convection or otherwise parasitic means of cooling only when operating at or near full load, thus obscuring the parasitic cooling method's effect on efficiency."

Also in the same document: "Temperature measurements of openings on the enclosure for cooling, such as holes or a mesh to allow for air to enter and exit, will be taken using thermocouples ... the thermocouples will be small enough and placed in such a way as to not meaningfully impede airflow in and out of holes in the enclosure."

Not sure what that first quote means, but it looks like the second one says fans are ok to use.

Also for the rectangular enclosure the max dimension allowed is

cheers, Jamie

If the load wants to draw some weird and wonderful current with lots of harmonics, then it can, and there's no way the inverter can prevent it without affecting the voltage. "

I think what this means is that under, say a resistive load, the voltage di storion must stay low. There is no other way they could measure it really. Proving it to have a sinosodial current output capability is, I believe, wh at they're after.

The lack of current linearity would mainly affect devices that chop the pow er at its own frequency, such as with an SCR. For example power tools with variable speed. There may be other things affected, possibly adversely, but that's the only thing I can think of right now. If there are still regulat d power supplies in actual electronic equipment working on that principle t he non-linearity could play tricks on its feedback loop. In the power tools , just the linearity of the sped control would be affected.

tion

They won't look inside the box. So, fans are fair game. Also, not grid-ti ed necessary. However, anybody would be foolish to build a commercial prod uct without grid option.

-----------------------------------------

In brief, the other specifications are :

Must be able to handle up to 2 kVA loads Must achieve a power density of equal to or greater than 50 W/in3 Must be able to handle loads with power factors from 0.7?1, lea ding and lagging in an islanded mode Must be in a rectangular metal enclosure of no more than 40 in3 Will be taking in 450 V DC power in series with a 10 ? resistor Must output 240 V, 60 Hz AC single phase power Must have a total harmonic distortion + noise on both voltage and curre nt of < 5% Must have an input ripple current of < 20% Must have an input ripple voltage of < 3% Must have a DC-AC efficiency of greater than 95%

everywhere on the outside of the device that can be touched. Must conform to Electromagnetic Compliance standards as set out in FCC Part 15 B Can not use any external source of cooling (e.g. water) other than air Does not require galvanic isolation

Our testing philosophy is to not look inside the box. You provide us with a box that has 5 wires coming out of it: two DC inputs, two AC outputs and g rounding connection and we only monitor what goes into and comes out of tho se wires, along with the temperature of the outside of your box, over the c ourse of 100 hours of testing. The inverter will be operating in an islande d more?that is, not tied or synced to an external grid. The loads w ill be dynamically changing throughout the course of the testing, similar t o what you may expect to see in a residential setting.

Not efficiency for efficiency's sake -- efficiency for the sake of not bursting into flame with all that power coursing through that small of a package!

Yes, i know. I have an auto 12V to 110V inverter. It feels like 20W to 30W hot when running less than 100W (laptop charger). I beg it is less than 70% efficiency.

But googling the web indicate that current top inverter efficiency is over 97%. So, this project might be doable. It might be very expensive to build, but cheaper than 1M.

Read the detailed sheet carefully. The load is between 0.7 and 1.0 power factor, lagging: a resistor, or R+L.

Low frequency noise or DC offset could result in "THD+N" (i.e., anything but the pure fundamental, including DC or LF noise, not just harmonics and hiss above the fundamental) of higher magnitude in current than in voltage.

In short, it simply has to be a pretty clean sine.

Tim

Inverter compensates for power fact by driving more voltage/current into peak load. The more it can drive, the closer to the ideal factor of 1.

Power factor can be + or -. Could be R, L or C. They just don't want it too far from 1.

Actually, for grid-tied, we want it closest to the grid, not necessary clean sine. For standalone, yes for sine or cosine wave.

That means the controller needs to have a very tight feedback loop. Namely, fast A2D and very close to the controller, for low latency loop. It can be done with mixed-signal ASIC chip, say for 100k.

A typical household is anything but resistive. CF lights, fridge, washer and garage opener. Lots of inductive loads.

How could the inverter affect the current while maintaining a pure sine wave voltage?

Sylvia.

er

Harmonic distortions are finite impulse responses of the system due to chan ging load. The inverter applies equal and opposite impulses to transform t he output into sinusoid. The inverter does not simply output a pure sine w ave. It needs to sample and detect the characteristic matrix of the system and apply the z-transform excitations to transform the error signal. It i s pure classical digital signal processing with lots of power (50A) and mon ey (1M) thrown in.

We are planning on using power transistors for 50A with 5 mili-ohms resista nt. The corrective impulse needs to be quick and powerful. It will be driv en by z-transformed signals in the proper frequencies. So, we need very fa st processor for DFT and FIR.

"Want"?

They aren't asking for a solution, they're specifying a problem. The load is specified -- RTFM.

Nowhere on the website does it say "grid-tied".

In fact if you had read in detail, it is specifically operated in "island" mode, i.e., completely isolated from the grid, any anything else (except ground, and presumably, whatever connection the inverter itself provides between AC and DC terminals).

Tim

What?

I don't know where to begin telling you what's wrong with that sentence...

Do you just read technical jargon and string it together?!

Tim

From Wiki:

"The total harmonic distortion, or THD, of a signal is a measurement of the harmonic distortion present and is defined as the ratio of the sum of the powers of all harmonic components to the power of the fundamental frequency . THD is used to characterize the linearity of audio systems and the power quality of electric power systems. Distortion factor is a closely related t erm, sometimes used as a synonym."

The distortion of the power output is a factor of the impulse response caus ed by the changing load.

Absolutely. There are still nasty rectifier-capacitor type loads out there and they have huge harmonic current content with real sine wave feed (like existing power distribution).

?-)

It is far worse than that. I have had to deal with working electrical engineers that regularly made the same kind of error.

?-(

pre-charged capacitors.

Much worse, 100 hours of testing, including full load resistive, reactive limits both + and -, no load and many other things.

?-/