Want to purchase Solar Panels from China and import to USA...

[G00SE]

So should I say to heck with the Over wattage concern (5500w vs 10s=6200w on paper in the nonexistent perfect world), and go full on 10s?

Well, that’s over 10% over the rated. I’d let others chime in if that’s excessive or not

 

[42OhmsPA]

OK, not trying to be a wize guy here... but VOC, as I understand it is the rating when the circuit is disconnected., meaning you won't be producing any power or at least consuming any power since you've removed the load.

What about early morning when panels first see light but aren't making power? How about when your batteries are full and loads are minimal?

So I interpret that as long as only the VOC # is above the 420 you recommend, then that should not be limiting my production. What I don't know of is how the real world translates from paper. I suspect voltage drop and connectors and such each are going to rob things, plus clouds and dust on the panel and any other sort of thing is going to mean I'll never see close to the paper numbers anyway.

Below 420VOC, look at the mppt range.
It's my understanding you won't gain much of anything going over 420 and 500VOC is the limit as well as cushion for cold weather spikes.

So should I say to heck with the Over wattage concern (5500w vs 10s=6200w on paper in the nonexistent perfect world), and go full on 10s?

You could with the panels you mentioned above but I'd go 5s2p before I went full on 10s. Or follow Hedges good advice and orient strings different directions.

In my attached screenshot you can see voltage climbing but no power being made, it's incredibly foggy out. This should give you and idea of how voc could spike with panels connected under ideal conditions.

 

[Zwy]

So should I say to heck with the Over wattage concern (5500w vs 10s=6200w on paper in the nonexistent perfect world), and go full on 10s?

Some will argue the point that you have peak SCC output over a longer time frame by overpaneling.

I look at it as wasted power. I'd rather capture all of it at any time as I don't know if it will be sunny 2 hours from now or if tomorrow it is cloudy followed by 5 more days.

I've mentioned it several times, but another standalone SCC would make things so much easier. Maybe if I say it long enough it just might cause that lightbulb to go off in your head. Sure would be easier than trying to justify this or that arrangement.

All my PV goes thru 2 SCC's at this point, I don't use any of the MPPT's in my inverters. It's not hard to install one, 2 cables, 2 wires and a ground. 4 fasteners to the wall. PV wires need metal conduit, EMT or MC works. MC can make it easy, just be sure to put the red bushing in the end.

Watch me discuss how I installed these starting at this point in one of my videos. I did a followup to this video and after the firmware update, I have not had a single problem with these SCC's.

 

[tialiden]

What about early morning when panels first see light but aren't making power? How about when your batteries are full and loads are minimal?

I hadn't looked at it that way. Sorry again for my newbness. Appreciate ya'll tolerating my questions (it's how I learn, before I make the more costlier mistakes and learn the hard way).

And yeah Hedge had another interesting thought I'm now bouncing around. I do like the wisdom of intentionally spreading my output further into the morning and evening, so having some panels aimed SE and SW accordingly makes good sense in accomplishing that. Probably more important considering I'm trying to shoestring into this PV / Battery world, with goal of being energy independent.

And to Zwy's reminders about how I could benefit with Independent MPPT/SCCs added to the mix to increase gains and flexibility is also wisdom I'm trying to let marinate in this thick skull. Probably the biggest factor is juggling keeping intial costs down as I do not believe in debt.

Thanks again!

 

[tialiden]

Maybe if I say it long enough it just might cause that lightbulb to go off in your head.

Lol. Sometimes I need to hear something a few times before I see the vision. Haha. Not trying to ignore ya, but still trying to put it all together.

 

[ksmithaz1]

Parallel doesn't require diodes either. Up to two strings parallel, no fuses/breakers required.
Series only requires the built-in bypass diodes.

What happens when one of two or more parallel panels or strings have a short or fault condition? I would think this could cause you some grief as the good panels start pumping their juice at the Zero Ohm load they are in parallel with. I would put a secondary diode for each tied connection at the + junction in my combiner box. we are talking $1.00 or so per wire.

 

[ksmithaz1]

Well, that’s over 10% over the rated. I’d let others chime in if that’s excessive or not

This is not complicated. On the panel side, the only thing that matters is not to exceed the VOC of the input of the MPPT. Period. You should have a breaker on the feed line that matches the max current of the MPPT. Your MPPT should never draw more than it's rated current, if it does then it is bad / has failed and you should pop a breaker. E=I/R P=E*I. Simple math R = Load is controlled by the consumer. P = power, is going to be the voltage provided by the panels (E) times whatever the MPPT is pulling (I) which depends on load (R).

An MPPT is a pretty simple device in theory. It varies the load between 0A and (MAX)A. As the load goes up (R goes down) the source (PV) has a tendency to drop the voltage (internal resistance inside the source, deeper discussion, let's not digress). An MPPT monitors it's output watts, and increases the load (lowers R presented to the panels) until output (P) peaks, then starts to fall. This is the "Maximum Power Point". It resamples continuously and adjusts the load up and down slightly to re-center on that peak. An MPPT is a "current control device".

You cannot provide "too many amps" to an MPPT. You will however waste any available power made available to it if the MPPT cannot use the juice. This is no different from the 100-200A panel on your house. The power company is offering you 200A, but your using zero unless you plug something in that asks for it.

Reality is you will not get anywhere close to the rated output of your panels for any realistic length of time. I live in Phoenix, and it may be hard to believe, but it's REALLY sunny here. I should be able to pull 14560 watts from my 32 panels. I have seen numbers right at 14000. Right now I'm lucky to hit 9000. If you have the money and space I'd over panel in a heartbeat, panels are relatively cheap. The maximum amount would depend on a number of factors, but I'd say 25% would be a starting point. Just like an MPPT (or the power company generator for that matter ) you want to make an optimal amount of power available for your load, to charge your batteries, and run your house. When panel output is 50% you would need to be 100% overpaneled to meet the same average peak output. The usefulness of this is going to diminish, depending on the amount of production time you get that is exceeding your ability to use it.

To summarize, to fully optimize you need track your output, but if you live where your 95th percentile output is 70% of rated you could easily overpanel 40% taking only a minimal hit in the net output of your string. There is some algebra and statistics involved. 95th "Percentile" is the 95th highest sample out of 100. It's used a lot in networking.

 

[tialiden]

So I'm wondering, seems like there should be a simple votage monitoring circuit added to combiner box when over paneling, that would kick some output to a big resistance load like perhaps a DC Shunt breaking resistor?? That would acts as a safety mechanism rated for long duration (unlike the SPDs) only activated in the 500-600vdc range. Why don't I see something like this? Basically it could then allow added safety over volt protection at the DC side. This is EE territory now, which I am not!

 

[G00SE]

This is not complicated.

I’m not going to come out and advocate for someone to exceed specs. I don’t think it’s complicated but I have my own plans as well lol.
As @Zwy keeps alluding to, at some point overpaneling is wasting energy. Idk if that threshold is 5%, 10%, etc.
I will say, it is more complicated than you lead on as for example a low frequency Growatt or sigineer absolutely can NOT be overpaneled by wattage nor voltage.

 

[ksmithaz1]

So I'm wondering, seems like there should be a simple votage monitoring circuit added to combiner box when over paneling, that would kick some output to a big resistance load like perhaps a DC Shunt breaking resistor?? That would acts as a safety mechanism rated for long duration (unlike the SPDs) only activated in the 500-600vdc range. Why don't I see something like this? Basically it could then allow added safety over volt protection at the DC side. This is EE territory now, which I am not!

Adding load, aka lowering resistance, does not necessarily lower voltage, and could simply overload the output frying something. If you want to lower voltage reliably you will need an inline voltage limiter.

 

[Hedges]

What happens when one of two or more parallel panels or strings have a short or fault condition? I would think this could cause you some grief as the good panels start pumping their juice at the Zero Ohm load they are in parallel with. I would put a secondary diode for each tied connection at the + junction in my combiner box. we are talking $1.00 or so per wire.

If you had several strings in parallel, each with a fuse, then if there was a short and current exceeded fuse rating sufficiently, it would blow.

With as few as two strings in parallel, or with more, there was a short bridging some PV panels, or shorted bypass diodes, then current would flow back through it perhaps without blowing fuses (if any). With two string, current would be limited to 1x Isc, so wires don't overheat. But failed bypass diodes would. This could overheat the diode and cause damage same as they do with forward current.

Anti-backfeed diodes would protect against overheating bypass diodes that way. Nothing prevents them overheating with forward current (you're supposed to avoid hard shadows when in full sun.)

 

[ksmithaz1]

I’m not going to come out and advocate for someone to exceed specs. I don’t think it’s complicated but I have my own plans as well lol.
As @Zwy keeps alluding to, at some point overpaneling is wasting energy. Idk if that threshold is 5%, 10%, etc.
I will say, it is more complicated than you lead on as for example a low frequency Growatt or sigineer absolutely can NOT be overpaneled by wattage nor voltage.

Current specifications on an MPPT (or any device) represent DRAW. The problems occur because as your supply voltage goes up and your current draw at the device remains the same, your wattage goes up. In the case of a static load it goes up expotentially:

Plug a 120V/120W light bulb into a socket
P=E*I so 120=120*I algebra 120/120=I thus I =1Amp
Now the load:
E=I*R so 120=1*R algebra R=120 Ohms

Now lets make the supply voltage for that same bulb 240V. Note: R has not changed it's still 120Ohms
E=I*R 240=I * 120 (alge's bra) 240/120 = I I=2A
P=E*I P=240*2 P=480W

and we get the magic smoke as long as the 240v supply can give you the desired current. Note here we did not limit the current. To limit the current to 1A we would have to double the resistance to 240, say two light bulbs in series 120 ohms each.

Thus what you say about Growatt makes practical sense since watts and volts are directly releated. However, it seems beyond a little odd to design a product capable of cooking itself by requesting more wattage than it is capable of handling within the bounds of the advertised supply voltage. That puts a current limiting onus on the suppy side, which I would consider a serious design flaw. None the less an inline circuit breaker or fuse should be able to prevent that.

 

[ksmithaz1]

If you had several strings in parallel, each with a fuse, then if there was a short and current exceeded fuse rating sufficiently, it would blow.

With as few as two strings in parallel, or with more, there was a short bridging some PV panels, or shorted bypass diodes, then current would flow back through it perhaps without blowing fuses (if any). With two string, current would be limited to 1x Isc, so wires don't overheat. But failed bypass diodes would. This could overheat the diode and cause damage same as they do with forward current.

Anti-backfeed diodes would protect against overheating bypass diodes that way. Nothing prevents them overheating with forward current (you're supposed to avoid hard shadows when in full sun.)

I agree... But earlier you said...

Parallel doesn't require diodes either. Up to two strings parallel, no fuses/breakers required.
Series only requires the built-in bypass diodes.

I must have a context problem.

 

[Zwy]

Reality is you will not get anywhere close to the rated output of your panels for any realistic length of time. I live in Phoenix, and it may be hard to believe, but it's REALLY sunny here. I should be able to pull 14560 watts from my 32 panels. I have seen numbers right at 14000. Right now I'm lucky to hit 9000.

I consistently pull full output from my strings. Of course, I optimize array angle for the sun horizon per season, no shade. I also see by 9 AM about 1/3 of my array power coming in during winter season if there is sun.

Of course, I have to ask, what panels, series and parallel strings, any shade, panel angle and quite a few other questions.

I'm pulling 40Kwh easily off 8420W of panels when the sun is out here in late fall/early winter but skies are clear with no dust or haze helps. Only days I run my furnace is when 2 or more days of no sun is in the forecast. I only charged the battery bank 8.5Kwh yesterday and 2.9Kwh today. On November 30th, I not only supplied 28.8Kwh for loads, I also charged the bank 18Kwh. November 29th was 21.7Kwh load and battery 27.2Kwh.

 

[teal95]

I started with a commercially installed system. 8400 rated watts worth of panels on a 7200 watt rated inverter. The mount is at the optimal year round angle. It only clips a few days each spring and fall.

 

[ksmithaz1]

I consistently pull full output from my strings. Of course, I optimize array angle for the sun horizon per season, no shade. I also see by 9 AM about 1/3 of my array power coming in during winter season if there is sun.

Of course, I have to ask, what panels, series and parallel strings, any shade, panel angle and quite a few other questions.

I'm pulling 40Kwh easily off 8420W of panels when the sun is out here in late fall/early winter but skies are clear with no dust or haze helps. Only days I run my furnace is when 2 or more days of no sun is in the forecast. I only charged the battery bank 8.5Kwh yesterday and 2.9Kwh today. On November 30th, I not only supplied 28.8Kwh for loads, I also charged the bank 18Kwh. November 29th was 21.7Kwh load and battery 27.2Kwh.

455's are on the roof, fairly optimal, maybe 5 degrees too flat, due south.
Targets:
8*455 = 3640
8*250 = 2000
7*250 = 1750
32*455 = 14560
8*15 = 3750 (Used, sub-optimal orientation)
T = 18310

On 8/27, it was bright, clear and sunny all day . . . You see A|B-PV1|2 should be able to hit 3640. I do see strange spikes from time to time, discussion below



I see radical spikes, on cloudy days, I think it's due to a capacitive state in the panels while the MPPT ramps up/down. I'm guessing that since the demand was there, this allows a more immediate discharge, The spikey numbers occur as the clouds light the panels back up. So while I've seen numbers that exceed theoretical max (see below) it only happens when I'm getting volatile readings, which are extremely short lived

On 8/14 we had a lot of broken monsoon clouds


And overall production on those days is comparably poor.



I'd be willing to bet if I was overpaneled on the 14th I would have seen a pretty linear benefit from the extra panels, despite the supposed high output numbers. Using an average peak of 3000 from the 27th, a very good day. (Which I actually think is probably high).

100 - ( 3000/3650 * 100)
17.8100

I could over-panel around 20% to generally hit the theoretical spec on a perfect day, and rarely even clip, if at all. That gets back to the 95th percentile stuff. Over the period where ideal production can hit 50% (?) or better what is the sample at the 95th percentile? Is that low enough? Unless you have your panels on a tracker and/or highly optimized, I'm guessing around 25% in a normally sunny environment would create a mostly linear improvement in output. I think the spikes I see in output would mostly just flatten if the MPPT didn't take the current. I think the loss from clipping would be minimal. The larger the percentage of time below whatever threshold you pick the more linear the benefit.

I could be full of it, but in my world, I just don't see it. I'd love someone to show me a graph that shows their panels hitting 100% of rated output for more than a 5 minute interval. Neither do I think the vast majority of panels are 100% optimally facing all the time, but you can only flatten that solar curve so much.

When I have a bit more data after this summer, I plan on doing some percentile calculations.

 

[42OhmsPA]

455's are on the roof, fairly optimal, maybe 5 degrees too flat, due south.
Targets:
8*455 = 3640
8*250 = 2000
7*250 = 1750
32*455 = 14560
8*15 = 3750 (Used, sub-optimal orientation)
T = 18310

On 8/27, it was bright, clear and sunny all day . . . You see A|B-PV1|2 should be able to hit 3640. I do see strange spikes from time to time, discussion below

View attachment 180988

I see radical spikes, on cloudy days, I think it's due to a capacitive state in the panels while the MPPT ramps up/down. I'm guessing that since the demand was there, this allows a more immediate discharge, The spikey numbers occur as the clouds light the panels back up. So while I've seen numbers that exceed theoretical max (see below) it only happens when I'm getting volatile readings, which are extremely short lived

On 8/14 we had a lot of broken monsoon clouds
View attachment 180991

And overall production on those days is comparably poor.

View attachment 180993

I'd be willing to bet if I was overpaneled on the 14th I would have seen a pretty linear benefit from the extra panels, despite the supposed high output numbers. Using an average peak of 3000 from the 27th, a very good day. (Which I actually think is probably high).

100 - ( 3000/3650 * 100)
17.8100

I could over-panel around 20% to generally hit the theoretical spec on a perfect day, and rarely even clip, if at all. That gets back to the 95th percentile stuff. Over the period where ideal production can hit 50% (?) or better what is the sample at the 95th percentile? Is that low enough? Unless you have your panels on a tracker and/or highly optimized, I'm guessing around 25% in a normally sunny environment would create a mostly linear improvement in output. I think the spikes I see in output would mostly just flatten if the MPPT didn't take the current. I think the loss from clipping would be minimal. The larger the percentage of time below whatever threshold you pick the more linear the benefit.

I could be full of it, but in my world, I just don't see it. I'd love someone to show me a graph that shows their panels hitting 100% of rated output for more than a 5 minute interval. Neither do I think the vast majority of panels are 100% optimally facing all the time, but you can only flatten that solar curve so much.

When I have a bit more data after this summer, I plan on doing some percentile calculations.

I assume you're using home assistant?
I really need to dive into that. Really nice graphs.

If it weren't for a damned cloud I would have been over 100% for more than 5 minutes, graph in this link if you're interested.

Edit - working link Post in thread '1Mwh SRNE ASF48100U200-H 10kw.' https://diysolarforum.com/threads/1mwh-srne-asf48100u200-h-10kw.67809/post-929435

 

[Zwy]

@ksmithaz1 I'd give you a graph from SA but my EG4 charge controllers don't communicate with SA. Back when I had EG4 6500EX's it would max out those 4Kw MPPT's at 3.9Kw with 4240W on a string. Installing the EG4 SCC's with 5500W capability, I hit full output under peak power. I've seen as high as 5.1Kw from cloud edge effect but the 4.3Kw is under full sun under peak power those 4 hours mid day.

I think there is something wrong with your PV. Could be you aren't maximizing string voltage for example. I run JA Solar 530W panels, my Vmp is around 360V, this keeps amps low and less power loss. Or the angle is too flat. Something is wrong with it, I would expect to see much higher PV input.

 

[ksmithaz1]

I assume you're using home assistant?
I really need to dive into that. Really nice graphs.

Grafana, custom everything.

 

[ksmithaz1]

I think there is something wrong with your PV. Could be you aren't maximizing string voltage for example. I run JA Solar 530W panels, my Vmp is around 360V, this keeps amps low and less power loss. Or the angle is too flat. Something is wrong with it, I would expect to see much higher PV input.

Angle is not optimal for August, it's on my roof ~ 10 (2/12). I will not be tilting them. I was not recording until August, so we shall see...

• January: 38.4°
• February: 33.4°
• March: 28.4°
• April: 23.4°
• May: 18.4°
• June: 13.4°
• July: 18.4°
• August: 23.4°
• September: 28.4°
• October: 33.4°
• November: 38.4°
• December: 43.4°

But that is kind of my point. Unless you are moving your panels regularly, your panels are sub-optimal the majority of the time, So I may see more clipping in June, having a few extra amps available to the mppt the rest of the time is a win. I have the voltage data, but it isn't going to change anything for me because my panels are mounted flat on the roof, period.

If you have the MPPT over-paneling is dumb, but I probably wouldn't wire up and put in more AIO/MPPT's because I could potentially lose a few watts on a perfect day.