Grid-untied

System · November 2011

I am wondering about a solar configuration I will call grid-untied.

In many states, such as mine, the regulators and power companies wont buy back power at a decent price. Furthermore, the regulation and certification requirements make it impossible to experiment anyway. So Grid tie is out.

Furthermore, many people are prepping for scenarios of EMP attacks, solar flares, arab revolutions, Canada selling their oil to China, etc in which a grid tie system would be useless.

A generator backup is of no help in a moderately long term disaster, without a few thousand gallons of stored fuel.

And off grid is too expensive because of the batteries, I figure $0.20 a watt to store and retrieve at the very least.

I want to build something between these two polar opposites. I am surprised no one has thought about it.

What I would like (I have about 7500 watts of cheap panels) is a system which acts like a UPS, only backwards. My house would work on public power, untill the mid part of sunny days, where it would switch over to the solar->inverter if that was producing. At night, I will run on utility power.

Technically, I visualise a MPPT tracker delivering about 170V DC to 13 highly isolated small agm batteries or a capacitor bank, just for starting loads. That will feed a bridge inverter to run my house during the day, including a minisplit heat pump. I hope to get efficiency up to that of microinverters.

I am still learning about batteries, I think I could float some gently for emergency use.

Does the solar community totally trash quasi-sine inverters? It seems like a quasi-sine would be more efficient and less expensive, and I can't think of anything that cares about waveforms except light dimmers.

Any thoughts on implementing this? Am I reinventing the wheel?

niel · November 2011

Re: Grid-untied

"Any thoughts on implementing this? Am I reinventing the wheel?"

well, yes. you are reinventing the wheel. many off grid sine wave inverters have a transfer relay built into them and these relays are available separately too for those inverters that don't have one built in (see below link). my magnum will run items with the grid or whatever other ac power you may have and when that is gone it can switch to the batteries and inverter acting like an ups. pvs can keep the batteries charged, but with mine it also has a built in ac charger.
http://www.solar-electric.com/autrswandgea.html

ggunn · November 2011

Re: Grid-untied

okie wrote: »

I am wondering about a solar configuration I will call grid-untied.

In many states, such as mine, the regulators and power companies wont buy back power at a decent price. Furthermore, the regulation and certification requirements make it impossible to experiment anyway. So Grid tie is out.

Furthermore, many people are prepping for scenarios of EMP attacks, solar flares, arab revolutions, Canada selling their oil to China, etc in which a grid tie system would be useless.

A generator backup is of no help in a moderately long term disaster, without a few thousand gallons of stored fuel.

And off grid is too expensive because of the batteries, I figure $0.20 a watt to store and retrieve at the very least.

I want to build something between these two polar opposites. I am surprised no one has thought about it.

What I would like (I have about 7500 watts of cheap panels) is a system which acts like a UPS, only backwards. My house would work on public power, untill the mid part of sunny days, where it would switch over to the solar->inverter if that was producing. At night, I will run on utility power.

Technically, I visualise a MPPT tracker delivering about 170V DC to 13 highly isolated small agm batteries or a capacitor bank, just for starting loads. That will feed a bridge inverter to run my house during the day, including a minisplit heat pump. I hope to get efficiency up to that of microinverters.

I am still learning about batteries, I think I could float some gently for emergency use.

Does the solar community totally trash quasi-sine inverters? It seems like a quasi-sine would be more efficient and less expensive, and I can't think of anything that cares about waveforms except light dimmers.

Any thoughts on implementing this? Am I reinventing the wheel?

Are you kidding me? Of course someone has thought about it. Folks come in here asking how they can run their house on PV without batteries virtually every day. The short answer is that to run off grid you must have batteries, and batteries aren't cheap.

The reasons why this is true and the various methods of implementing backup/offgrid power are involved enough to fill several good sized books, but it really boils down to that there ain't no free lunch.

Cariboocoot · November 2011

Re: Grid-untied

"I am surprised no one has thought about it."

Actually lots of people have thought about it. A few have even done it. Usually in locations where grid power is unreliable.

The thing is, it's no different from a typical off-grid set-up - except for being more expensive per Watt hour because it doesn't get used as much. $0.20 per? No, more like $2.00 per.

You still have to have batteries enough to handle the loads and panels enough to recharge the batteries. The only difference is using the grid as the "back-up generator". For the most part it's not a very practical situation. Off-grid becomes more economical the more you use it.

You have 7500 Watts of panels? That's a huge amount for off-grid. The #1 rule applies to "no grid", "grid tie", and "emergency power" systems: reduce the loads first. You will get a better return on investment, including investment in any kind of solar power, by reducing the amount of power you use to begin with. After that the system needs become smaller = less expensive.

"Technically, I visualise a MPPT tracker delivering about 170V DC to 13 highly isolated small agm batteries or a capacitor bank, just for starting loads."

You need to read up a bit on how systems work. There are no 170 VDC inverters, and most of your household stuff won't run on DC. Super-capacitors have limited storage/lifetime. Loads need to be started, and run; solar panels alone will not work without the battery capacity to take the fluctuations in supply and demand.

If by quasi-sine wave you mean the MSW (Multiple Square Wave, Multiple Step Wave, Modified Sine Wave, Really Not Sine Wave) inverters you have to understand that they simply do not work on some loads. Unless you know the loads will work on MSW, you're better off with true sine wave. Refrigerators, for instance, don't like the non-sine waveform and that's one of the things you probably want to keep going in an outage.

Here's a very long and involved thread full of info and links that can get you started on understanding the whole solar electric power situation: http://forum.solar-electric.com/showthread.php?t=5556

System2 · November 2011

Re: Grid-untied

I am not looking at something for nothing.

I am going to go off grid during the day using a small battery/capacitor bank just to hold up loads for 5 min as a cloud passes by. As I said, the opposite of a ups.

170V is the peak voltage of the 120V cycle. If you can design a Mosfet inverter using a 170V rail, you know what I am asking. I am going to design a MPPT charge controller that puts out 170V (since none exist) with an integral inverter. Basically, an Enphase without a grid reference.

Actually, motors start well with quasi-sine if it is very low impedance and a narrow duty cycle. Of course, a Chebeyshev filter is required to get rid of RF.

I'm sure the Enphase enginners were told they were nuts, also .

An off grid system is too inefficient, as is spending $10,000 for batteries that I may never use.

Thank you for your thoughtful replies.

Cariboocoot · November 2011

Re: Grid-untied

Here's one more reply for you;

Solar panel output doesn't vary just for a few seconds while a cloud passes by. It varies quite a lot over the course of a day. With the sun at less-than-perfect angles to the panel, output is low. When it's square-on to the panel, output is at its maximum potential. By then the panels have heated up, causing the output to be less than maximum. If there's no place for the power to go, they don't put out anything.

GT inverters work quite differently from OG inverters. They don't have to supply steady Voltage or current because the grid can take anything they do supply, from 0/0 on up to their maximum capacity. That is no good for trying to run things off grid. Many people here have looked for work-around to utilize existing OG installs in a no-power situation. The only ones that work are the inverters with batteries designed to be synched with GT inverters (Xantrex XW, SMA Sunny Island).

UPS units, btw, have batteries in them to supply power for a finite length of time. If you connect one to a solar panel, the panel will be able to recharge the battery but you can not count on the panel having enough power to supply a load at any given time. With both the panel output and load demand varying, you need the battery capacity in-between to mitigate the discrepancy. There is no way around this; it's physics.

I'd hate to think I wasted all those years learning this stuff, but maybe I did. :roll:

ggunn · November 2011

Re: Grid-untied

okie wrote: »

I am not looking at something for nothing.

I am going to go off grid during the day using a small battery/capacitor bank just to hold up loads for 5 min as a cloud passes by. As I said, the opposite of a ups.

If you can design a Mosfet inverter using a 170V rail, you know what I am asking. There are no building blocks in your store for this. I'm sure the Enphase enginners were told they were nuts, also .

An off grid system is too inefficient, as is spending $10,000 for batteries that I may never use.

Thank you for your replies.

Um, yes you are. The output of a PV inverter is not a constant thing, and the variation is not just when a cloud passes over. Furthermore it's very unlikely that your loads will *ever* exactly match the output of your PV inverter. You can't look at a PV inverter as a demand delivery system like the grid. If you have a 4000W PV inverter running at capacity driving 2000W of load the inverter will raise the output voltage trying to push out 4000W until it hits its voltage limit and shuts down. Likewise, if you are trying to drive more load than it is producing at the moment, it will shut down. Batteries provide a buffer to cover the deficiencies and absorb the excesses between production and demand. You cannot run sustainably off grid without them.

TnAndy · November 2011

Re: Grid-untied

okie wrote: »

Technically, I visualise a MPPT tracker delivering about 170V DC to 13 highly isolated small agm batteries or a capacitor bank, just for starting loads. That will feed a bridge inverter to run my house during the day, including a minisplit heat pump. I hope to get efficiency up to that of microinverters.

I assume he means an MPPT charge controller, not 'tracker', and the high voltage is the incoming side, (which is still too high for most all but that new 1200 buck Xantrex model ) not the output.....but it's not too clear. edit to add: a classic 200 would fit that high voltage classification too. niel

As I take it, he basically wants to run some of his household on solar during the day, sort of directly, using a small battery bank "for starting loads".

Here's a couple problems Okie:

1. You'll have to have a transfer switch on the circuits to use only those you want to run and to be able to flip them back to grid at night.....what time do you make the switch?

2. Assuming you get that done, HOW do you plan to match loads to your solar supply, given it will vary from first sun on the panels to a high around noon, to last sun on the panels ?

You plan to sit there all day and cut things on/off ?

Also, what happens when a cloud moves overhead, and your use is right up there with your PV supply ? You plan to keep a close eye and hand on some circuit breakers to cut that mini-split on and off ? Or you just gonna take a chance on browning out the motor ? Or if the inverter shuts down due to low voltage, you just gonna come home to a cold house ?

See, THAT is the reason for a good battery bank....( or be tied to the grid ).....to handle those variations in situations.

PV is NOT like a generator, in that you don't control the power produced.....your 7500w of panels might give you 2,000w at 9am, 6500w at noon, 4500w at 1pm when those high cirrus clouds start to move that weather front in.....and 500w at 3pm when that cloud cover thickens quite a bit.

BB. · November 2011

Re: Grid-untied

Super Caps (the one or two I have read the specs. on will only last about 2 years of constant power before needing replacement.

You can get a set of forklift/traction batteries that will last 20+ years (not quite as efficient and need a far amount of distilled water to keep filled).

5 minutes on super caps is more or less geared to a computer room or other application that needs non-stop power--and is usually there to allow just enough time to get the genset up and running.

For long term outages--you really need to determine what will be critical to you and how many kWH per month you will need to survive (water, lights, cooking, etc.). Using computers and sat TV may not really be high on the list. A/C may be depending on where you live--but perhaps a mini-split setup in one room vs an entire home, etc.

-Bill

BB. · November 2011

Re: Grid-untied

By the way, there are some loads that work just fine with Solar Panels Only systems... But they are (today) typically water pumping. Connect a solar array to a water pump (typically with electronics to match solar IV curve to motor IV curve--a version of MPPT; one is called a linear current booster).

When there is enough sun, the pump starts pumping. When there is not enough sun, the pump slows down and stops.

-Bill

System2 · November 2011

Re: Grid-untied

Sorry, I meant a MPPT charge controller with 170V output. No supercapacitors will take that, but regular electrolytics will. Since Q=CV^2, it doesn't take as much to help.

I erroneously said tracker since I have to watch PV voltage and current.

If you are not technical, consider computers and everything electronic runs off of 170V dc internally nowadays. The incoming AC is rectified and stored in 2 large capacitors, where it is switched to provide other voltages. By using Shottky diodes or synchronous rectifiers uning mosfets, up to 97% efficiency is possible.

I am thinking an inexpensive string of 13 20ah 12V batteries would help with starting loads, or untill a generator could start. Don't try this at home, everything has to be isolated in a safe metal box.

If you have an internal 170V bus in the box, when your PV voltage droops, your rectified mains power at a slightly lower voltage would automatically power your loads, as long as it is available. If it is not, a DC-DC (48V-168V) converter off of an external battery bank kicks in. Or rectified 170V output from a Generator. Why not sum the sources as DC rather than switching in and out unsynchronised AC sources? We sum the PV and batteries at low voltage now.

It sounds like everyone wants pure sine, which is no problem.

No, I thought about sitting all day by the breaker panel switching loads, did not appeal to me, either.

But if the grid is down or if you are off grid, +/-4 hours a day of reasonably abundant power helps, and the lack of a huge battery bank encourages power thrift at night. I think I personally may use a 12V DC bus at night.

The Fujitsu Minisplits are wonderful, they are 26 seer, pull 400W running, and use DC internally with an inverter to run the compressor.

Cariboocoot · November 2011

Re: Grid-untied

Okie;

Here's a thread you might find interesting: http://forum.solar-electric.com/showthread.php?t=12962
It deals with techntrek's adventure in using his Toyota Prius as a power supply for high Voltage UPS units (Liebert 192 Volts).

Everyone here loves their mini-split A/C's. Except me; I don't need one. It's -10C right now and expected to warm up to zero.

solar_dave · November 2011

Re: Grid-untied

Cariboocoot wrote: »

Okie;

Here's a thread you might find interesting: http://forum.solar-electric.com/showthread.php?t=12962
It deals with techntrek's adventure in using his Toyota Prius as a power supply for high Voltage UPS units (Liebert 192 Volts).

Everyone here loves their mini-split A/C's. Except me; I don't need one. It's -10C right now and expected to warm up to zero.

And it is 80F right here now expected to go to 58F over night! Love the splits by the way.

techntrek · November 2011

Re: Grid-untied

okie wrote: »

If you are not technical, consider computers and everything electronic runs off of 170V dc internally nowadays. The incoming AC is rectified and stored in 2 large capacitors, where it is switched to provide other voltages. By using Shottky diodes or synchronous rectifiers uning mosfets, up to 97% efficiency is possible.

They are only using the ~170 volt peak of each cycle (more like everything above 160 volts), maybe 10% of the total energy fed in. Give it 170 volts dc on a 100% duty cycle and you will likely let the smoke out of your computer supply.

Cariboo led you to something I was going to point out. There are some mid-level UPSs which will operate on roughly 170 to 250 volts dc (192 nominal). Liebert UPStation S is one, but they are hard to find now. APC online UPS in the 3000 to 6000 watt range is another and sometimes can be found very cheap when they are usable but "dented".

BB. · November 2011

Re: Grid-untied

Actually, many of the computer dual input voltage switching type power supplies operate at ~340 DC nominal (voltage doubler at 120 VAC, simple rectifier at 240 VAC).

And yes, they can operate on ~340 VDC input.

Other computer/electronic supplies with Power Factor Corrected input can operate on AC or DC from ~100 to 264 (or higher) VDC.

Problem is you need to understand the input circuitry and the limitations. And most vendors are probably not going to help you with schematics.

There is an initiative to convert data centers to DC distribution by server companies and data centers (to save money/energy use in data centers and probably to provide battery backup more easily too)--But I foresee there are lots of safety and fault isolation issues with this program and it will be hard pressed to replace the typical AC system in use today.

Alliance Boosts 380-Volt DC Power Standard

One place I read that they hope to save ~10% on power costs... I guess it may be worth it with the large data centers out there these days.

-Bill

techntrek · November 2011

Re: Grid-untied

Yes, some power supplies can run from DC unmodified, but not all of them. Some will, but the rectifier's diodes can't handle a 100% duty cycle which is what I was talking about, so you have to remove it or bypass it (rectifier package or 4 individual diodes).

tallgirl · November 2011

Re: Grid-untied

techntrek wrote: »

Yes, some power supplies can run from DC unmodified, but not all of them. Some will, but the rectifier's diodes can't handle a 100% duty cycle which is what I was talking about, so you have to remove it or bypass it (rectifier package or 4 individual diodes).

Huh? Rectifiers aren't rated in duty cycle, they are rated in watts. Problem is, 120VAC is 120 volts RMS and 170VDC is 170 volts RMS. Big difference, and since power is volts times amps, you might just be passing a few more watts through those poor little diodes. Replace them with bigger diodes, add a phat heat sink and be done with it.

techntrek · November 2011

Re: Grid-untied

Heat. A system designed to only utilize the peaks of each half-cycle (which is most of them) will only be passing current through either half of the rectifier for a short period of time. So each side gets half the load and only gets heated for a short pulse. Connect it to a DC supply and now one half handles 100% of the load and there are no breaks in flow to allow cooling.

Why replace them with bigger diodes? If you are converting it to run from DC just eliminate them and the filter capacitor. Supply 160-170 DC downstream of the rectifier stage.

BB. · November 2011

Re: Grid-untied

This getting a little bit into the weeds... The reason that DC power into a "typical" AC power supply should not cause any problems with the diodes is the RMS current of poor power factor + I^2*R heating.

Because you now have an average DC current flow instead of the Voltage Crest current spikes of non-power factor corrected (typical) power supplies, you will end up with less heating (on average). Also this is high voltage DC instead of (possibly) 120 VAC RMS--so the average current is much lower, and I^2*R heating is much less too.

Because only 1/2 the diodes are used on DC power input. So, all of the heating effect is localized to 1/2 the diode array. So, it is possible that you may have diode life problem. But, probably not. There should be more than enough heat sinking for the much reduced overall heat losses.

Anyway, my 2 cents worth.

-Bill

Grid-untied

Comments

Categories