Is this a safe way to build a solar inverter?

[SunEagle]

Here's a link to a comparable grid tie inverter http://www.amazon.com/SUN-600G-Micro.../dp/B0089UA9KO, the only difference is the one I use is designed for 220V AC.

Unfortunately that type of inverter is not allowed to be used anywhere in the US and Canada due to the lack of a UL listing.

Second issue is that most of the "plug in" grid tie inverters do not go through a rigorous inspection program and are built substandard which have resulted in failures causing fires. I would think twice before you use one of them.

[rritoch]

It turns out that I made a critical mistake in this design. The UPS batteries weren't connected parallel, but in series so it's designed to handle 24V internally, and not 12 volts. I corrected for this by adding 6V (2x 3V) more of zener diodes, for a total of ~ 19.4V , this was enough to drive the circuit and the UPS activated and was able to power a small load without a problem. The 3v zener diodes can only handle 1/2 W and were hot enough to burn while the system was operating, so I'll need to get some more zener diodes, or higher-wattage diodes, but the important thing is that the circuit functioned without a problem.

[rritoch]

Unfortunately that type of inverter is not allowed to be used anywhere in the US and Canada due to the lack of a UL listing.

Second issue is that most of the "plug in" grid tie inverters do not go through a rigorous inspection program and are built substandard which have resulted in failures causing fires. I would think twice before you use one of them.

I'm not located in the U.S. or Canada, but I know people in the U.S. operate these without a problem. These can handle 600W, per the specs, but I have 3 100W panels connected to it, and they only produce about 120W output (A/C) and I've been operating it for at least 6 months without a problem. The first one I had was 300W and it broke in the first week so you may be right about the QC being sub-par, but the 600W grid tie we have now is running without any problems. I have no plans of adding more panels at this time, but if I did I would probably add another grid tie. Reading the reviews it seems these grid ties can safely handle about 1/2 the power they're rated for, but since the solar panels only produce about half the power they're rated for it probably equals out. I honestly think this is a well-constructed "scam" where the panels produce 1/2 the power they claim to produce, and the equipment can only handle half the power they're rated for. This may work for people that don't know anything about electricity, but anyone that does try to push these things to their specs are burning up their equipment.

[SunEagle]

I'm not located in the U.S. or Canada, but I know people in the U.S. operate these without a problem. These can handle 600W, I have 3 100W panels connected to it, and they only produce about 120W output (A/C) and I've been operating it for at least 6 months without a problem. The first one I had was 300W and it broke in the first week so you may be right about the QC being sub-par, but the 600W grid tie we have now is running without any problems. I have no plans of adding more panels at this time, but if I did I would probably add another grid tie. Reading the reviews it seems these grid ties can safely handle about 1/2 the power they're rated for, but since the solar panels only produce about half the power they're rated for it probably equals out. I honestly think this is a well-constructed "scam" where the panels produce 1/2 the power they claim to produce, and the equipment can only handle half the power they're rated for. This may work for people that don't know anything about electricity, but anyone that does try to push these things to their specs are burning up their equipment.

I don't doubt the people in the US use those plug in inverters. A lot of people do crazy things because they think they know better or don't care if the "rules" are broken. People use to make their own booze which was against the law but they didn't care.

The issue is the high chance of the inverter failing, causing the circuit that the unit is plugged into to overload resulting in a fire. Also without the UL listing, if the item fails and causes a fire, your insurance company will not cover your losses. Seems like a poor decision to save a few bucks only to lose your home.

If you want a grid tie system then get a UL listed grid tie inverter and have it blessed by your POCO. Doing anything else is like playing Russian roulette.

[bcroe]

It turns out that I made a critical mistake in this design. The UPS batteries weren't connected parallel, but in series so it's designed to handle 24V internally, and not 12 volts. I corrected for this by adding 6V (2x 3V) more of zener diodes, for a total of ~ 19.4V , this was enough to drive the circuit and the UPS activated and was able to power a small load without a problem. The 3v zener diodes can only handle 1/2 W and were hot enough to burn while the system was operating, so I'll need to get some more zener diodes, or higher-wattage diodes, but the important thing is that the circuit functioned without a problem.

Running at 24V will double your efficiency. Didn't you have 2 @ 12V zeners? Just put them
in series; the output will still be a couple junction drops low. The power through the resistor
will reduce considerably since the voltage across it is reduced by 12V. Keep that resistor at
a value high enough not to heat the zeners too much under NO LOAD. When you have load,
check that the zener voltage doesn't drop, indicating not enough current gain. If it does drop,
I would add a 3rd stage of current gain to the 2 you already have; maybe stick that 3V zener
in the stack to bring up the output (3 junctions lower). Bruce Roe

[rritoch]

Running at 24V will double your efficiency. Didn't you have 2 @ 12V zeners? Just put them
in series; the output will still be a couple junction drops low. The power through the resistor
will reduce considerably since the voltage across it is reduced by 12V. Keep that resistor at
a value high enough not to heat the zeners too much under NO LOAD. When you have load,
check that the zener voltage doesn't drop, indicating not enough current gain. If it does drop,
I would add a 3rd stage of current gain to the 2 you already have; maybe stick that 3V zener
in the stack to bring up the output (3 junctions lower). Bruce Roe

Bruce,

Thanks for the additional design advice. I need to put this project on hold temporarily as I don't yet have the 0.1 ohm resistors, and one of the insulated wires was touching the hot transistor so it's now covered in melted plastic and should be replaced. Now that it is starting to function and bringing the 54V to the 24V needed by most of the equipment that's available in my area, charge controllers & solar inverters, there's a lot more options.

It is a cloudy day today so It wasn't able to power more than ~ 20W load and the UPS shuts down if I connect even a 75W fan to it. Either way I'm just happy it worked, even if the results were less than I hoped for.

The next step, once I'm able to get the additional parts I need, is the physical construction of a more permanent circuit, so I'll need to find some good heat sinks, and mounting hardware.

One additional issue with this new circuit is the transistor at the first stage is now running hot. This may reduce once I have the small resistors, but either way I'll need some heat sinks.

For now this project is going to remain a hobby and I think I have an answer to my initial question. These transistors are very hot so it probably isn't safe for long-term use without good heat-sinks and stable construction. My evidence is that I accidentally burned the insolation off of one of the wires.

Budget permitting, once the circuit is stable I'd like to add some ultra-capacitors to the circuit in place of batteries. As far as I know ultracapacitors can be fully discharged without breaking so that may be a good solution for operating higher power appliances for short timeframes. The UPS is 1200VA so it can certainly handle a higher load even though the panels don't deliver that much power. There are some significant safety concerns with the ultra-capacitors so before I go in that direction I'll need to find a safe place for them to be stored, that's near this circuitry.

The ultimate goal of this project is just to provide emergency power during daytime for running things like fans, water pump, and most importantly, a coffee maker

Edit: I'm not sure that my calculations were correct but it seems that I'd need about 3000F capacitance and that the cost for 9 2.7v 3000F capacitors would be ~ $350. I believe that when fully charged this could operate a coffee maker, and considering the lifetime of the capacitors may be worth the investment.

What I found for a capacitor is @ http://www.aliexpress.com/item/Free-...rchweb201560_9 so if anyone has any experience with these things I'd appreciate the advice. I'm not expecting miracles, just trying to improve the design to something that's useful.

[Mike90250]

Just a note about ultra capacitors. The do indeed hold an enormous amount of power, but your inverter only works in a small range of that, so your storage will mostly at a low voltage range the inverter cannot use.
example:
12v system 12V inverter work 14-11V
ultra cap. 14 - 1 V The inverter operating range is much smaller than the storage range of the cap.

[rritoch]

Just a note about ultra capacitors. The do indeed hold an enormous amount of power, but your inverter only works in a small range of that, so your storage will mostly at a low voltage range the inverter cannot use.
example:
12v system 12V inverter work 14-11V
ultra cap. 14 - 1 V The inverter operating range is much smaller than the storage range of the cap.

Mike90250,

Excluding cost, does this mean that it would be better to put the ultra-capacitors on the input side of the circuit which I believe can operate from 54v - 24v as the latest modified circuit is for 24v not 12v. The components I've used surely can't handle the amount of power the capacitors can deliver, and would constantly be leaking power through the resistor-diodes (first stage) of the circuit, but based on what you said it seems I could get more consistent power from the solar panels by attaching the capacitors directly to the solar panels.

Based on my calculations it seems I would also need to add a 10 ohm 300W resistor between the capacitor-solar grid and the grid tie to keep the circuit < 300W. Since I'd need to import the capacitors, importing power resistors shouldn't be a problem.

[bcroe]

Mike90250,
Excluding cost, does this mean that it would be better to put the ultra-capacitors on the input side of the circuit which I believe can operate from 54v - 24v as the latest modified circuit is for 24v not 12v. The components I've used surely can't handle the amount of power the capacitors can deliver, and would constantly be leaking power through the resistor-diodes (first stage) of the circuit, but based on what you said it seems I could get more consistent power from the solar panels by attaching the capacitors directly to the solar panels.

Those 3000F caps are rated 2.7V; you can't wire one anywhere in your system. If you try to put 9 in
series, the effective capacity will be 3000F/9, or 333F. The energy stored is 1/2 C x V X V. If you try
to start at 28V and run that down to 22V, the energy delivered will be 333/2 [(28 X 28) - (22 X 22)] =
50,000 watt seconds, or 14 watts for an hour. Series connection also requires equalizing circuitry to
avoid failures; I don't see where this is going. Bruce Roe

[rritoch]

Those 3000F caps are rated 2.7V; you can't wire one anywhere in your system. If you try to put 9 in
series, the effective capacity will be 3000F/9, or 333F. The energy stored is 1/2 C x V X V. If you try
to start at 28V and run that down to 22V, the energy delivered will be 333/2 [(28 X 28) - (22 X 22)] =
50,000 watt seconds, or 14 watts for an hour. Series connection also requires equalizing circuitry to
avoid failures; I don't see where this is going. Bruce Roe

bcroe,

Thanks for the correction, I haven't done any electronics in over 5 years so I didn't think about the effect on capacitance. The cost is truly prohibitive than since I'd need like such a large number of capacitors, in series and in parallel (81 caps ~ $3,000). I'll need to throw out the idea of using the ultra-capacitors for now. I suppose if I really wanted to use them I'd need a step-up DC-DC converter to use them and a step-down DC-DC converter to charge them. Lot's of cost for little benefit.

Add in the immense amount of power the 81 caps would store, someone would eventually die if they accidentally touched the capacitors. Considering my inexperience, it would probably be me.

[rritoch]

So I was finally able to get to the city to buy the remaining parts. I have 4 new 2N3055 transistors, 4 new 12V 1W zeners, 4 new 6.2V 1 W zeners, a heat sync and a universal PCB for a permanent (stable) construction. I've also made some final adjustments to the circuit so I can switch between 25v and 19v depending on weather conditions and what the UPS is "willing" to accept.

http://www.falstad.com/circuit/circu...765625+0+-1%0A

It will certainly take some time to fully construct this but I'm working on making some vlogs of the "experience" so when the construction is finalized, and if it works, others can use this if they're facing a similar "problem" of not being able to find a 300W 54V DC to 24V DC converter, which I can't seem to find anywhere, even on Amazon.

[rritoch]

Thank you everyone for your assistance with this project. Now that the construction is complete I'm fairly happy with the end result. It may not be efficient, and may not be completely safe but it's better than nothing. Here's the video I promised of the construction ...

Edit: I tested this today under good weather and it successfully powered a 75W fan. Due to the way the UPS functions it cuts off every time a cloud passes over due to the low-battery cutoff circuit. Now that the circuit is down to 25V I should be able to attach a normal inverter as there are inverters available in the city that can accept the 25V. The addition of the 0.1 ohm resistors and fourth transistor also made a big difference, even at 75W load the transistors didn't get hot.