Creating a MicroInverter

I was thinking, these microinverters are expensive. Could it be done that you could make one from an arduino nano/mini (or equiv) and some mosfets, i2c connection, voltage/ampere meter, and a little bit of programming?

My thinking is that, AC is AC. It bounces back and forth at a certain frequency by a certain amplitude. The only reason why it's a "wave" form from the power company, is because the power is generated by a rotating electrical field.

A simple DC-to-AC converter would be two PWM's inverted to each other. We could use a similar set up that Julian used in his setup, but using a nano instead of a mega or full sized board.

We would have a master in the shed (or other central location) that would send a sychronizing command to the micro's so they oscillate at the proper frequency. Back in the shed, we would use a bridge rectifer to convert back to DC voltage.

Each micro would also report back to the master what's going on with it; voltage/ampere, temperature, performance, even angle of panels if equipped with a tracker.

If this can be done, it would drastically reduce the cost of installed solar in bulk and getting every drop out of them w/o worry of 1 panel/cell shorting out the others.

Whadda think?

This looks sort of interesting.

https://www.amazon.com/uxcell-Inver...80&sr=1-26-spons&keywords=mppt+inverter&psc=1

Hard to compete with that price if it actually does MPPT and you can spoof it into thinking a pure sine inverter + load is actually the grid. That little gadget does about 75% of what you want to do. I'm thinking it's not terribly bright and would be relatively easy to fool once you figure out how it detects the grid.

Ive actually seen that inverter on a youtube channel. its not as good as it claims. about 1/3 less than rated. plus, thats putting to much into one unit. what i was talking about was for those who have panels and wanted to microinverters on the panels. i currently dont have any, but plan to in the future.

I've found it's better energy spent for me accommodating old technology with new on my own ..

Right now I'm hacking a 5,000 BTU 4.5A 120VAC AC to allow Arduino control... The AC I'm hacking doesn't have so much as a diode in it, although it does have a capacitor for motor starting purposes.

Three wires plus a hot control everything, two speed fan and the compressor. I'm running this thing on a 600W pure sine wave inverter so keeping the load to a minimum is important, I'm going to start the compressor first, let it run about 20 seconds and then start the fans after the compressor is well started and drawing less current. After the compressor cuts off I want the fan to run for a minute or so to get the last of the cool out of the evaporator coil and proceed with a 10-ish second fan run every few minutes to sample the air temperature in the room.

I reckon I can save 15% on an already halfway decent little AC energy use by putting a bit of intelligence with it.

I'm finding not only do I need battery management, I need load management too. I have enough inverter to run everything, just not more than one big load at a time. I'd like a system that prioritizes my heavy loads, microwave before AC, AC before refrigerator unless critical temp on refrigerator and so on so I don't have to think about it so much.

"I'm sorry, Dave. I can't turn on that receptacle right now."

"Would you like to hear me sing a song while you wait?"

A watt saved is a watt harvested.

LOL nice

Yes, I hadn't thought of device management like that. Especially the control of AC units in such a way. That makes a lot of sense; definitely the getting the last bit of cool out of the coils part. I will make sure to take note of this and be prepared for the adjustment of my equipment as well

Wonder what else would need timed delays? Hrmm....


I guess what you could do, is put programming interrupts in each of your high load devices and the communicate with each other. If one is starting, or has started within x seconds, then the others will wait a few seconds and re-query to see if it's safe to start yet.

Korishan said:

Ive actually seen that inverter on a youtube channel. its not as good as it claims. about 1/3 less than rated. plus, thats putting to much into one unit. what i was talking about was for those who have panels and wanted to microinverters on the panels. i currently dont have any, but plan to in the future.

Click to expand...

It should be pointed out for new comers that micro inverters ( theyare fixed on the back of panels ) are Grid tie ...they convert the DC straight away into AC 240 or 120 , and the wire from them goesinto the grid .... the advantage of these is that you can monitor the output from each panel on your computer , and ,resistive losses in cables are kept to a minimum ...

But these systems are not really suited for those who use powerwalls because you have to charge the powerwall by changing mains AC back again into DC ... losses in both micro inverters , and in the mains charger and the expense of these two when they are not needed.... much better to put the DC from panels straight into the battery .. this can literally be done ,panels with open circuit 21v are designed to feed 12V Lead acid (41V open circuit can feed 24V LA)with nothing in between ,but most do not do this , as adding a charge controller increases efficiency ... for our lithium power walls we need to protect against overcharging so need a standard LA regulator..

I only know of two designs of micro inverter .. the ridiculously expensive enphase ( don't even think of these you will never get your money back in earnings from the power company) .... and the chinese knock off (about 1/3 the price of enphase)... I've been running some of these chinese copies for a few years with no complaints they're very well designed , sealed aluminium unit , waterproof, computer monitoring possible... http://www.ebay.co.uk/itm/1200W-gri...456366?hash=item237e28c4ae:g:JNsAAOSw4A5YpxLr ... but even the price of these does not make this a sensible idea ..

I would not recommend any powerwaller to consider micro grid tie... much better build a big battery fed directly from panels , and have a standard inverter to supply your home , cut off from the grid..... if you do this there will be times when you're panels are producing and the powerwall is full ... what to do with this power .... not easy .... electrolysis springs to mind convert it to hydrogen to store and use for cooking ? .. buying grid ties to feed this to the grid will lose you money .

The whole reason for the micro-inverter in my case, is as I mentioned above. If there is a long run between panels and storage, AC power losses are a lot less that DC power over a long run. I'm not sure of the lengths of each to make a determination as I have not done panels yet. I was just interested here.

Also, for those who may want to have a Grid-Tie system DIY setup, and have Lithium as storage backup, I think this would a project worth looking into. As we want to do as much as possible in the DIY arena.

Another reason for this setup would be so you could power your house directly from the panels without first going through the batteries; again, using them as storage for the night or when there is cloudy days.

No matter the setup, your panels will produce power regardless of what type of storage medium you are using or what type of controllers you are using. There are always 2 ways to deal with this.
1) Turn off the panel at the source; just plan disconnect it from the wiring using a relay
2) Dump the excess power into something really useful; a water heater would be a prime example. People in the north would dump into radiant flooring; people in the south could dump it in radiant roofing to cause positive airflow into the home by pulling in fresh cooler air from under the home or other means.

There are plenty of ways to deal with something. Just because YOU don't see a practical application does not mean that someone else might not see one. I'm in it for the learning experience and knowledge. To better understand how things work and don't work. Why is this method better than another. Could it be improved upon? Can the cost be dropped? Etc.

Just because the Chinese make these, does this mean that "WE" can't?!? No, it doesn't.

One slight caveat about grid tie inverters, they are *supposed* to be grid tie, that doesn't mean they have to be operated that way, I suspect the cheaper ones are dumb enough to be tricked into operating with just a regular sine wave inverter in order to share the AC load in an entirely off grid situation.

MPPT controllers are expensive, inverters are expensive, the majority of my load is daytime cooling and if I can combine an MPPT controller with an inverter to help with that daytime load straight off the panels at a reduced equipment cost then I would be remiss in not taking advantage of that situation.

The idea here is not to take AC power, rectify it and charge batteries but rather to take the load off the controller that is charging the batteries by shunting AC production mostly off to another device, leaving the controller free to charge batteries at its full power.

On edit: I'm thinking that the manufacturers have used the same parts used to do the DC/AC conversion to also do the MPPT, some sort of toroidal transformer more than likely, it's quite possible to do more than one thing with a single component saving money and the toroid and power driver to switch it are the costly bits here. In this case the only real cost is in firmware development to make the same parts do two different things at the same time. That is why a micro inverter is cheaper than a controller>charger>battery>inverter combination.

Twenty five years ago I watched a brilliant developer generate NTSC video in real time with a PIC controller to display a weather station app on a (then standard) television, I'm not remotely in that league but I know that some amazing stuff can be done in firmware.

Korishan said:

The whole reason for the micro-inverter in my case, is as I mentioned above. If there is a long run between panels and storage, AC power losses are a lot less that DC power over a long run. I'm not sure of the lengths of each to make a determination as I have not done panels yet. I was just interested here.


Just because the Chinese make these, does this mean that "WE" can't?!? No, it doesn't.

Click to expand...

I agree ...you have identified a bottleneck in the system , which is the high price of these micro grid tie inverters ... I just checked and see now that more companies are coming out with them ... but the prices are going up!! while panel prices are falling ... there should be a massive demand for these (from people without powerwalls) the price is being artificially kept high by enphase , who realise people don't understand the numbers and are buying solar for the 'feelgood' factor .... the true price should be about 1/10th the price enphase sell for ..

So go for it , you should be able to make them (do we really need computer monitoring , this will be the complex part) .... myunderstanding ofelectronics is not up to buildingone .

The prices will fall, power electronics are being driven hard by electric vehicles. Today you can get sine wave (quiet) ebike controllers for very close to what you could get trapezoidal controllers (noisy) five years ago. Even field oriented controllers (very quiet and torquey) are becoming more reasonable. There are five phase proprietary systems that run almost silently and with very high efficiency and so on...

It takes a lot of technical expertise to make heavy power electronics behave, it's closer to a black art than cookbook type electronics. Inductive and capacitative effects we normally ignore as trivial become critical and so on. You have to let a lot of magic smoke out to get the hang of things, it's much more mentally comforting when it's someone else paying for the equipment you are blowing up. Trial and error, with lots of emphasis on the error part.

"Oh, they'll just write it off on taxes."

---

I'm starting to feel like Captain Nitpick.

There isn't much difference between AC and DC as far as transmission goes, DC is better really but not that significantly.

The single biggest advantage for AC is that it's easy to change voltages with entirely passive components, transformers. Induction motors (another form of transformer come to think of it) and three phase transmission are other advantages but the transformer is the biggest, IMO.

On two wires DC is a bit more efficient, modern power electronics kind of eliminates the real difference between AC and DC that has existed up until now.

Regarding losses. Thats one of the reasons im using a high voltage DC system. (>600VDC) Still using 6mm2 and max around 8-9A.

Running grid-tie inverters against a dumb sine-wave "producer" could work but the problem is to manage the output. The Grid-tie is generally meant to just put out all it can unless its coupled with a meter.

So i got 2 questions:
* How hard is it to rebuild a grid-tie not to just overload the non-connected grid? Is it possible?
* Could you hook a grid-tie up with a meter and a load without grid and still get it to function without starting to self-swing? Is the system fast enough?

Above questions may be non relevant...

Micro inverters are quite complex. I have a roof full of these things but the installer and the manufacturer has gone out of business so I reverse engineered the design and have attached the resulting schematic. I have not written a circuit description but would be happy to answer any questions if I can.

WOW!! Thats awesome! Though WAY over my head right now

Enphase has a whitepaper on AC-Coupled battery backup systems and microinverters.
https://enphase.com/sites/default/f...ion-Note_AC-Coupled-Battery-Based-Systems.pdf

From the whitepaper:

The best practice is to shut off the microinverter system when the batteries are charged. To accomplish this, you can use a voltage controlled relay to shut off the Enphase Microinverter system when the battery reaches a voltage within the
manufacturers limit. For more complicated systems, use a series of relays to taper the charge to the batteries.

For additional battery protection, use a dump load to drain excess power out of the batteries in the case of overcharging. Do this by using some of the stored power to heat water or air with a listed and approved heating
element.

Click to expand...

Enphase Microinverters work with a number of battery-based inverter
manufacturers and component manufacturers, including Magnum Energy, Outback Power Systems, MidNite Solar,
SMA, and Schneider Electric.

Click to expand...

So rather than make the microinverter smart enough to handle anything to do with the battery they recommend disconnecting the inverters once you reach your desired SOC. That is something that can easily be controlled with a small amount of code and a voltage controlled relay on any sufficiently equipped devboard.

I presume the plan would be to synchronize the sine wave to the grid once it become available again. I wonder what the challenges would be there, and whether timeshifting the sinewave to re-sync would cause issues. I know there is some tolerance to the sine wave's frequency, I just don't know what that tolerance is off hand.

I presume that so long as any timeshifting stayed within that tolerance it shouldn't have any adverse effects. But I'm curious how that would go. Would the microinverters need to be disconnected when the grid came back up if the sine wave was to far out of sync? Or does a hard swing just not matter?