A step toward completely off grid!
 

Hello! We just purchased 5.6kW of solar panels for our place in Arizona. They are B quality panels, but still test at full power. It's 20 of them at 280watts each. I'm also basically done with the inverter. I just need to bolt it all together and try it out. I'm also going to get grid tie working. gasp... Oh no, not that. LOL. Sorry, the solar power forums are very "the proletariate shant discuss such things. Far to complicated! You are banned!!!" I'm not sure a good way to test it out yet, but I don't think it'll be very hard.

So, here's the plan! First, get them mounted on the ground. I've got lots of space in my back yard, and no obstructions for the sun. I could do it on the roof, but I don't want to alter it at all. I'm going to hook them up to power my shop at first, and then run a 110v extension cord to power a few isolated things at the house. Then, once I figure out the dang laws, which are not at all easy to understand in my area, We'll try to go off grid completely. My electric company charges like $30/month just to have it turned on, even if I don't use it. I'll be sure to take lots of pictures of the process.

Awesome! Looking forward to seeing your progress!

Sometimes I also get the willing to get off-grid, but living in an apartment makes the whole thing harder.

The 20 panels are arriving today!! Also, I'm picking up the unistrut and some treated 4x4 lumber and cement bags tomorrow from the home crack ho (depot). haha.

I'm doing sort of a hybrid of these 2 approaches:

4KW grid-tied battery backup system - Page 2 - EcoRenovator

http://www.builditsolar.com/Projects...ideViewDim.jpg

Grid tie testing
 

If you have a 120V generator you can set up your own 'grid' and do some testing with your inverter syncing to that. Some of the small generators don't regulate very well, particularly with no load ... so you may have to put a load onto the generator, like a hot plate, so that the frequency stabilizes and your inverter can synchronize to it.

Using a 220V generator would be a better test, but it is harder to keep the loads relatively balanced on both 120V 'phases'.

AND you can test that the output of the inverter shuts itself off when the 'grid' goes down (turn off the generator or trip the output breaker) ... so your inverter does not try to power up your whole neighborhood when a power company breaker trips.

One step further would test the load regulation on the inverter by adding a hair dryer load to your isolated little grid and see if the load on the generator gets higher (easy to tell by the sound of the motor) or if your inverter picks up all of the additional load. Turn off the dryer ... again, which one dropped the load? Generator or inverter?

Another INSPIRING project. I'm EAGER to read all about it!

Thanks, Thingstodo!!!! That's a very good idea. I don't have a generator, but I'll be sure to get one for testing. Here's some progress pictures:

Unistrut from Home Depot

One of 4 of the wood parts that the unistrut will be connected to.

My son was a huge help.

It's getting there! But officially too dark to see what the heck I'm doing.

MPPT charger
 

I was thinking that MPPT was so complicated, until I realized, it's just a regular old buck charger! All you do (during the bulk charging phase) is adjust the PWM duty to maximize the output current! You don't need to maximize the product of the input current and input voltage (input power)! Just the output current! LEM Hass to the rescue! I think I'll sample current at a particular pwm duty.... Say, 0%. haha. I bet that will have a small output current. Then I'll just chase the higher output current by incrementing or decrementing the pwm duty to see if I need to hunt to the left or right for the global maximum. One problem with this is that I heard that there can be local maximums if panels aren't matched well. It's probably good enough to sweep through the whole range of 0-100% duty, and then assume that the global max you found can be kept in your sights by chasing it with the "increment/decrement pwm duty, and go in the direction of larger output current" method. haha that's a long name.

Most people recommend not using wood at all in the structure to support the panels, unless it's well protected, like in the form of a house, but you are in a really dry area, right? reason for this is that your solar panels are going to outlive you, when your son is old there is a good chance that he will have to figure out who he wants to leave them to in his Will, you don't want to support them with something that will only last 20 years, but then again in Arizona you might be fine.

I hadn't thought of that. We won't be here for 20 years though. Also, It rains around 0.1" per year around here. I'm already baking in the fires of heck and darn-nation, and it's 8:02am. Plus, before I die, I want to shoot the panels with a shotgun, and throw some off a big cliff to see them shatter, since that would look awesome. Also, in a couple years I'll get bored, tear the whole thing down, and make a solar tracker.

Pictures
 

All I get is an error message from Dropbox ...

MPPT Charger
 

I'm not quite following ... do you have a fixed output voltage, and that's why the output current is the only thing that you need to maximize?

DISCLAIMER: This made sense to me last night, after being in the sun all day, so I could be wrong. haha:

Yes, since it's feeding a 48v battery bank, the voltage is approximately constant. Well, higher current into the batteries causes the battery bank voltage to go up slightly because of the internal resistance. Am I thinking about this wrong?

MPPT Charging
 

I guess charging a battery bank is close enough to constant output voltage. It will be a narrow range of output voltage ...

So maximizing output current ... will maximize the input V * I for the buck converter ... that follows!

I don't know how you have the PV arrays wired. Do you need a buck/boost circuit for cloudy days ... wait ... you said Arizona. Never mind!

My wife had made the pictures set to private. dang it. But she has some pictures of her nursing on there too, lol, so I guess I understand. Here's one picture anyway:
http://i419.photobucket.com/albums/p...5-19190634.jpg

I was up at 3:40am, and worked until around 8pm. And it was like 95 degrees outside today. But it's getting close to done:
http://i419.photobucket.com/albums/p.../solarPic1.jpg

http://i419.photobucket.com/albums/p.../solarPic2.jpg

Two 3/8" x 3" lag bolts to connect the "superstrut" from home depot to the 4x4:
http://i419.photobucket.com/albums/p.../solarpic3.jpg

Still rather cool there, eh? Don't you have another 20 - 30 degrees to go? :p

hahaha. Yes. We moved here in December, and I was dying from the heat by the time February hit. It just keeps getting hotter!! And it's MAY!

I was thinking that if you're up to it - and you certainly appear to be - you could mitigate a lot of the heat load your house experiences by shading the house with the solar panels.

That would sharply reduce how much air conditioning you need. Sure, you give up a lot of power generation by foregoing tracking, but I think you get it back in reduced weathering on your roof, reduced cooling load, etc.

I would think the clay tiles would reject more solar heat than the PV panels since they are lighter in color and do not conduct heat as well, but I'm no expert on the subject. I'd use landscaping to provide shade on the house, as it appears there is none at the moment.

Tracking is absolutely crucial to getting the most performance per M^2 out of a PV array. However, the pictures of the rack don't suggest a method of adjusting the vertical tilt seasonally, and they certainly don't pan east to west.

Still, I'm sure the setup is more efficient than if I were to install the same M^2 PV on a tilt and pan tracker in the Pacific NW. It seems insane to me that people install PV up here, when there would be a better return by installing it in AZ. Essentially leasing out the PV system to those that can better take advantage of it.

A solar tracker would be much better. It's actually really simple. You can have a photo diode that only has a small slit for light, and each time it gets dark, a motor turns on until the peephole gets sunlight again. One hard thing was making the mounting strong. We get CRAZY winds around here. My neighbor used to have 21 chickens, and after the monsoon season in July/August, he only had 1 chicken left. haha. So, I was just afraid it would break if I tried to get tricky on the first draft. :p (My son wanted to click the face with the tongue sticking out)

OK I went with photobucket. I think they'll show up now...

http://i419.photobucket.com/albums/p...s/CAM00728.jpg

http://i419.photobucket.com/albums/p...s/CAM00725.jpg

I just used some square washers that are 1/4" thick to squash each panel down at 4 different points. I should get a picture of that part.

There were about 15-20 mph winds today and there's not even a hint of wobble. I think they will be fine in really really fast winds, at least I hope.

Pictures of mounting...
 

All the stuff came from home depot or a cheaper equivalent part from ebay:

3/8" bolt through square washer, screwed into those unistrut spring things:
http://i419.photobucket.com/albums/p...s/CAM00755.jpg

Push into unistrut channel, then rotate clockwise 90 degrees. It won't let you rotate it counterclockwise:
http://i419.photobucket.com/albums/p...s/CAM00766.jpg

http://i419.photobucket.com/albums/p...s/CAM00767.jpg

Torque that sucker down! The panels can take it!:
http://i419.photobucket.com/albums/p...s/CAM00756.jpg

Two clamping points per side:
http://i419.photobucket.com/albums/p...s/CAM00759.jpg

I had to get sort of ghetto/creative for the ends. haha:
http://i419.photobucket.com/albums/p...s/CAM00760.jpg

U better get som diagonal cross beams on that thing to resist wind etc, or it might tumble over.

Ya, I'm going to add some 2x4's. I got the grounding lugs and #6 wire threaded through all 12 panels and then drove a big grounding pole thing like 8 feet into the ground, so I officially won't get shocked by touching the frames now, not that I would have anyway. I also drove another grounding rod into the ground next to the shop, where the AC box thing is going with all those "20 and 30 amp trip circuits" inside it. I also just got 250 feet of 10/3 underground feeder cable. I'm going to put outdoor plugs near the house so I am not messing with the house wiring.

I got a small protoboard soldered together for a DC voltage monitoring circuit. (seemed like a good time for a new paragraph, but I was never very good at knowing when to do that). I'm just going to feed the output of that to probably the throttle position on one of the DC controller control boards that I had done for the "1000 amp controller". That will be the MPPT charger control section. I've done up a driver section for a 600v 600amp IGBT half bridge. Since I'm just doing a buck converter, I'll only need the single module. The copper sheet is here, the 500uF ring capacitor is on it's way. I probably need another 8 6v batteries since I have enough power to charge the 48v battery bank at over 100 amps when the ambient temperature is over 100 degrees. Not a good combination. Maybe I'll make a little battery box and put a super tiny air conditioner inside it.

I just went ahead and made a whole new board for the MPPT solar charger. It monitors solar voltage, solar current, battery voltage, and battery current, and "temperature" of whatever. It powers 2 contactors, and has 2 precharge relays that are good to 450vDC at 10amps. It also has onboard precharge resistors. The power section should be good to charging 400v at 300amp or so with adequate cooling. The inductor is only rated for 200amp continuous, so if you want to do the whole tesla charging thing, you'll need a slightly bigger inductor. None of this of course is at all relevant, since my array can only charge at around 5.6kW. haha. And my golf cart batteries would be none too happy charging at 300 amps, and I don't have 375v worth of 6v batteries either.

Testing?
 

Oh, I see ... you just CREATED ANOTHER BOARD!!

Someone like ME would make it fit into the existing design to avoid fabrication, testing, debugging, et al.

You amaze me ... and speaking of that ... what news? Any news?

Hello! The boards will be here in about 5 more days I think. I just got an aluminum plate tonight from Ebay. 12" x 24". Not the most space conserving, but that's OK. Tonight I'm also doing the power section layout, and figuring out where all the holes go in the aluminum plate. Then maybe tomorrow I'll write the gcode.

Also, the 250 feet of 10 gauge 3 wire (plus a ground wire) got here. It's the direct bury kind. I'm running that from the shop to the outside of my house, and I'll plug in our 3 window air conditioners to it. It was 110 today I think. You can actually feel your eyelids burning in the shade. It's a very weird feeling. Also, I was racing a road runner today in the car. Those things are pretty quick. Also, my son made a trap, and it actually caught a weird desert squirrel creature today.

Also, I got some of those solar panel connector plug things. I'm going to do 4 parallel strings of 5 panels each, which makes for a voltage at peak power of 175vDC at around 40 amps. I could have done 2 strings of 10 panels, but a peak voltage of 450v was going to be a problem for the off the shelf "universal input" AC-DC for the control voltage that I ordered yesterday. It can take any DC voltage from like 110VDC up to 370vDC, and puts out 15v for the control board and powering the igbt driver. Oh ya! All the parts for populating the control board and the driver board arrived today. And I already have the igbt half bridge that will function as the switch/freewheel diode in the buck converter.

Also, I ordered a combiner box that can take up to 12 inputs and can have a single 1000V 1000amp output I believe. It's a monstrosity, but was really cheap on ebay.

Here's the 200amp continuous 300uH inductor:
http://i419.photobucket.com/albums/p...s/CAM00941.jpg

Here's how I'm getting the control voltage for it. It will do around 50-400vDC (you read that right! DC! Even though it says AC. ya!) input, 15v output. Bootleg ghetto Chinese. Cheap, but evidently reliable. I've heard good things about MeanWell. Once I use them for a while, I can say "I know you, Mean Well." Or I could say "I know you mean well". It's basically the same thing.
http://i419.photobucket.com/albums/p...s/CAM00945.jpg
Here's the copper poop. Oops, I mean copper sheet. LOL. Get it? sheet? hahaha:
http://i419.photobucket.com/albums/p...s/CAM00946.jpg

Here's (here're??) the remaining cables necessary to turn the array into 4 strings of 5 panels per string:
http://i419.photobucket.com/albums/p...s/CAM00948.jpg

A close-up of the connectors. They just snap together! Rated for wet or dry.
http://i419.photobucket.com/albums/p...s/CAM00949.jpg

12" x 24" x 3/8" aluminum plate:
http://i419.photobucket.com/albums/p...CAM00950-1.jpg

5 Mil Kapton tape. And ONLY like $50. barf. Good to 18,000 Volts of isolation!:
http://i419.photobucket.com/albums/p...s/CAM00951.jpg

Input fiter capacitor. 600v 500uF. I threw it up in the air against a white wall, and snapped a picture really fast. It's not a stock photo. I'm not lying.
http://i419.photobucket.com/albums/p...0uFfilmcap.jpg

I finished building the MPPT charger. I'd go install it and test it out if it wasn't 119 outside right now. The voltage monitoring is working well. It was pretty noisy and had an offset of a few volts. Now, it's accurate to +/- 0.05 volts. So, since my output needs to be around 55.2v when done charging, it will stop around 55.15 or 55.25. That's good enough. There's just a tiny bit more wiring to do. I'll start with 1 string of 5 panels, for a total power of 5*280 watts. First, I'll charge the batteries with the inverter off, and then I'll turn the inverter on at the same time.

Eagerly anticipating ..
 

I eagerly await pictures and data!

We don't have the cell phone at the moment. T Mobile has it, but soon they'll say we can get a new phone. It's a long story. Soooo, no pictures until then! haha. Yesterday I installed 3 outlets outside next to the windows. Today I just hooked up the inverter to the house. We have a 12500btu or something like that window air conditioner that I plugged in, and the stupid inverter shut off due to overcurrent protection. But... Since I'm in charge of every stinking line of code on that thing, I just changed it so the hardware overcurrent protection clears itself automatically if it trips. That's the short term fix. I also ordered some higher amp current sensors to replace the LEM 50's I'm using at the moment. That will be safer, but I didn't want to wait for them to ship here.

Tomorrow I hook up the MPPT charger. So, today I'm running the batteries down a bit so the MPPT charger has something useful to do. I"m just waiting on some 2 gauge lugs that should arrive tomorrow!

Starting induction motors
 

If it's similar to my window air conditioner, there's a single phase induction motor in there that draws a very short peak of 6 times rated current (6X). It's connected to your compressor, so it starts up under load. It will trip generators .. that you would EXPECT to be big enough to handle starting up a 0.75 - 1.0 HP motor ... as well as inverters.

Instead of measuring the very high spike of current (which your power electronics will have to supply without melting) it may be more reasonable to put in a 'pre-charge resistor' or 'soft start' that limits the current on a few AC cycles when a SERIOUS overload is detected. Assume that the load is faulty - like a short circuit - and size the resistor - perhaps 6 ohms and 20W - to take a surge of 20 amps for 6 cycles (0.1 seconds). If you energize power electroincs - SCR, MOSFET - to 'bypass' the resistor at that point and the current is still pinned at 20A at that point, let the hardware over-current take it down. I realize that the resistor should be 120W, but that would be continuous. 20W should be a reasonable size and give you a decent safety factor if you limit the time.

This will cost you some board space and an additional SCR or MOSFET switching loss in efficiency. But it should save you 4 sizes on your MOSFETs for handling the surge start of an induction motor. It shouldn't mess up your hardware over-current ... too much ... I think ...

Just my opinion ..

I am using some 600v 600amp chassis mount IGBTs that can handle like 2400 amps for a very short period, with a film cap that can handle short bursts of 10,000 amps. So, 6 x 11 amps (continuous is around 11 amps) is still safe. If I wasn't doing such brute force with big fat IGBTs, I would definitely do the resistor soft start.

Drool, drool ...
 

WOW! - that's excessive even for paranoids like me!:D

Did you buy the IGBTs and caps for something else, or did you get a *REALLY* good deal on them? That's the kind of hardware overkill that I'd like to use for an AC drive on an electric truck!

Of course, NOW I want to see the pictures even MORE ... electrical porn at it's best!

haha. They had some IGBTs on Ebay. Tesseract (the guy that makes the Soliton1) was selling a bunch of 600v 600amp igbts in groups of 6 and 24, and I got 6 of them. They were like $65 each or something like that. The ring capacitor was like $150 or so from Future Electronics. I'm going to use 3 of them in the 3 phase controller that I need to just finish. They supposedly will do 400 amps at 400 volts continuously with good cooling.

Hurray it works! It was a super cloudy day... The sun kept getting almost poking through, and then not at all, but I never did see my shadow! But an interesting time to try out the MPPT!!! Here are some graphs. The most interersting thing I found was, the power DIFFERENCE was almost constant between power in and power out. So, the efficiency goes up the brighter the sun is. At least that's what appears to be true at the moment. The x axis is seconds, y axis is whatever the label says! haha. The clouds were racing by, so the cloudiness changed significantly every 0.5 seconds or so.

battery current and solar current:
http://i419.photobucket.com/albums/p...psvstime-1.jpg
Battery and solar voltage:
http://i419.photobucket.com/albums/p...tagevstime.jpg
Battery power and solar power. Note: the batteries were receiving more power than was coming out of the panels! LOL. Just kidding. No need to move me to the unicorn corral...:
http://i419.photobucket.com/albums/p...es/power-1.jpg

Power difference:
http://i419.photobucket.com/albums/p...fference-1.jpg

Efficiency. The efficiency went up as the sun would almost poke through. It peaked at around 90% when the sun was getting through almost. Then went back to stinking it up when the sun disappeared completely:
http://i419.photobucket.com/albums/p...ficiency-1.jpg

Efficiency
 

It makes sense that part of the losses are overhead - switching, powering the electronics, that sort of thing. So the more power you put through, the lower the percentage of the overhead, the better your overall efficiency.

There SHOULD be a component that is a percentage of the power in - resistance losses in the IGBTs, traces, etc - anything to do with current.

Perhaps everything is oversized so much (well, I guess that part is not perhaps ... everything IS oversized so much) that the overhead losses are visible and the percentage losses are miniscule by comparison?

I did some graphs of the power curve from solar panel voltage = open circuit to solar panel voltage = 48v (shorted to the batteries) starting just after the sun peeked over the panels until around 8am. Each period from open circuit to "short circuit" was about 40 seconds.

It's possible that the above curves werent' really the max power. They might be though. I think there may be some local maximums that aren't global power maximums.

Here's a pretty cool graph that I made today by having the PWM duty cycle vary from 0 to 100% at different times throughout the day. The span across the whole duty cycle took 40 seconds each time. The input open circuit voltage was around 200v, and the output voltage was about 49v (48v battery bank that's mostly dead). The lowest curve was when the sun just started peeking over the panels. The highest curve was at solar noon (12:34pm) down here just south of Phoenix in Maricopa. Notice that to the right of the peak power point, there are occasional spots where it's a fake maximum power! Oh that makes me mad. This sort of illustrates the problem with algorithms that blindly search for any local maximum power point. It was 5 280w multicrystaline panels in series. I've got 3 more strings of 5 panels that I'll add to see how the curves change, but I think I need more batteries first. That's a lot of amps at 48v:
http://i419.photobucket.com/albums/p...CycleCurve.jpg