Using my panels without the grid

[Barrowman]

I have been thinking about this some more and have more ideas as to getting this to safely work.
Assuming the approx. 500 volt no load output of the panel strings is correct several things come to mind which I believe could make this work safely.
I would plan for there to be 3 PC supplies with their primaries wired in series so able to handle 1020 volts input ( peak values of 240 v AC ) which should be safely above what would be present.

First is that if they were originally fitted at night then surely the output of the panels would rise slowly so quite possibly it would give time for the power supply units to start up without damage.

Second I could wire a 250 volt AC varistor and 220 microfarad electrolytic capacitor in parallel across the input of each unit to prevent it being damaged by over voltage. The varistor I looked at reacts in <50 ns and can handle 100 amps.

I could set up each psu to have a default minimum draw from one of it's outputs so it always placed a load on the system if it was running.

Of course I have not been able to find out what happens if the power demand is less than the panel string can deliver. As it is I believe these have an open circuit voltage specification which, as far as I can see isn't much greater than the on load voltage and so should not be a problem
Does this all sound reasonable or not?

[MN Renovator]

I'm not sure wiring up 3 computer power supplies in series will work as well as you expect because you can't be completely sure all 3 supplies will carry the load evenly. That would rely on them a turning on at the same time, not faulting when the voltage goes out of range, their voltage regulation handling the varying voltages identically. Maybe it's worth the science experiment, but I don't think this will work. It all depends on the loads needed, but I'd rather try to find a charger that could handle the input voltage and output to a battery bank sized well enough to buffer the source and load variations. The only things I really care about in a power outage are to run the refrigerator, the furnace, and charge some electronics like USB power banks, phone, maybe a laptop, and AA/AAA batteries for flashlights. For me the cost to buy the equipment for the loads I mentioned above costs way more than a generator would, so I have a generator and keep enough gas to keep the furnace and refrigerator running intermittently as needed. The reality is I'll wear a coat and run long furnace cycles to make sure the pipes won't freeze, theoretically a relatively larger than standard uninterruptable power supply wired to my furnace which only uses 300 watts when the blower is running could keep things more consistent, especially at night when I don't want to run a generator outdoors.

There are periods in the winter where a solar system stands zero chance of powering critical loads. I've had weeks with zero power output due to snow covered panels and even with them clear, thick winter clouds might not even be enough to run a refrigerator for a day even if the energy produced was being stored in a battery as the day went on.

[jeff5may]

Ok, I just read through the thread and I thought I would give you my 2 cents worth...

You're getting a feed in check for power generation. Why would you rob yourself?

You have a power inverter that's doing the risky power conversation for you already. Just use the usual wall outlet devices for the science experiment stuff. Safer, tested for reliability, etc.

If you feel like you just must play with the hvdc, you absolutely must make a custom device that will run reliably at 10 percent or so higher DC voltage than you'll ever be able to generate. Stepping down from 500 to 12 or 17 vdc brings some sort of buck converter to mind.

[Barrowman]

Originally Posted by jeff5may

You're getting a feed in check for power generation. Why would you rob yourself?

I don't intend to rob myself. As I said in an earlier post my home is on higher ground but around it the whole area is a flood plain. In a flood the whole area including my home could be without power for a long time. That is when I would want to use the panels because that is when the inverter would not run

You have a power inverter that's doing the risky power conversation for you already. Just use the usual wall outlet devices for the science experiment stuff. Safer, tested for reliability, etc.

The testing would only be supplied with 230 v AC but I suppose I could rectify it and end up with about 325v unregulated DC and wire 2 psu in series as they will run when only receiving about 160v feed.

If you feel like you just must play with the hvdc, you absolutely must make a custom device that will run reliably at 10 percent or so higher DC voltage than you'll ever be able to generate. Stepping down from 500 to 12 or 17 vdc brings some sort of buck converter to mind.

I guess that would be an alternative, mosfets can now handle a high voltage and control huge amounts of power now.

[jeff5may]

The power supply in series idea will eventually end up in smoke. As with the old series connected vacuum tube heaters, it will work for a little bit. Once your luck runs out, something hogs a little more juice and poof, it's all over.

[Barrowman]

Originally Posted by jeff5may

The power supply in series idea will eventually end up in smoke. As with the old series connected vacuum tube heaters, it will work for a little bit. Once your luck runs out, something hogs a little more juice and poof, it's all over.

Do you think that would be true even though there would be varistors across each input?

[Barrowman]

I know this is an old thread but now I am thinking I could use an IGBT controlled by a pwm signal generated by an arduino to control the rms voltage to a PC psu. my system uses a solis inverter and it's peak voltage spec is 600 volts and its rated voltage is 330 volts. I would set the pwm signal to 60% high and 40% low so it would ensure at even if 600 v is being generated the psu would see 135v rms.