Need some guidance for small project series-parallel 18V 1.25A charger

[sgking2]

My daughter and I are building a project that consists of (15) 6volt 250mA panels. We are wiring these in series-parallel as (5) strings of (3) to get 18V nominal at about 1.25A to charge a 12V AGM battery. The individual strings will have schottky blocking diodes before being paralleled and all fed into 5A MPPT controller/charger (Genasun-GV5). We have 3 questions:

1) would bypass diodes be helpful (even remotely) across each 6V module? I have lots of 9A 45V Schottkys available so price is not a concern. But if one panel is bypassed, then the remaining 2 providing 12-14volts would be insufficient charging voltage and be way off from the other 4 strings of 18-21 volts. Or will it not help provide energy but only serve to protect any one cell from the other 2 dumping into it in a given string?

2) would minor voltage differentials present a problem between the strings if all are within a couple hundred millivolts of eachother, all ranging between 18-21, but balanced to be +/- .3V of one another? Is there a way to better balance them or combine them to negate this effect using just one MPPT controller?

3) some strings are SE and others are SW facing and this is unavoidable due to the strict location they must be placed, there is no South facing opportunity. Given this situation we can exploit the morning sun through a set of panels and the afternoon sun through another set. What is the best way to funnel this back into the same battery? Should we use the same charge controller? If so, even with blocking diodes, it is likely to mess up its MPPT tracking and see the lower voltage of both sets all the time. A second tracker is expensive unless it is PWM. But even then, will 2 controllers going to the same battery affect eachother? Or is there a better way overall to handle this that I am missing?

THANKS!!

[sgking2]

The panel inside/outside

We live in a community governed by an HOA (will never do that again) so we have to be low profile. This is the SE-facing panel from the inside and outside. Fortunately, this semicircle window was a pain in the butt and always put a glare on my monitors across the room so I was glad to block it off (was going to limo tint it otherwise). From the outside it doesn't look like much of anything, and that is a good thing. I know it is not very large and the angle is not ideal but it does harness power for the first 4 hours of each day. We are putting together a 2nd for the SW-facing window for the noon-sunset portion of the day. The AGM Deep Cycle battery and 65W MPPT charger seem to be enough to run her desk, charge her iphone, ipad, laptop and external LED monitor. We may buy the same battery and put it in parallel to reduce the DOD cycling and battery wear/tear. We will never see a cost savings from this but it was a neat project.

With (5) 18v strings on one panel (for AM) and (5) more on the other for PM harnessing, is it enough to simply wire all (10) strings in parallel to the panel input on the MPPT Controller? Of course I will use a Schottky blocking diode on each string but is there a better way to get more power from all 10 strings, possibly a 2nd controller? Any ideas?
I am concerned that at one point in the day I may be getting 18V from one panel and 21V from the other. Not sure how the MPPT will respond to this even with blocking diodes. As I understand it, the panel is still providing power to the battery (albeit much smaller) with ambient light when the direct sunlight had moved to the other corner of the house, so I want to be able to absorb as much as possible from both panels. Sorry if I did not explain this well. Just not sure if one MPPT would suffice for this unique application or if I am better off with two economical PWM controllers to maintain some isolation between the panels, like the SunForce 8.5A Pro from Home Depot.

Thanks.

Panel from inside.JPGPanel from outside.JPG

[Sunking]

Do not put your panels inside the house behind windows. Doing so will cut power more than 50% not mention will be shaded more than half the day cutting you down to some 75 to 90%. At that point is technically useless.

[sgking2]

I understand, it is not solar glass and will reflect some energy away, it is not cooled by outside air and thermal runaway will reduce performance, and it is not at a good angle and that sucks too.... that being said ... how to make the best of a crappy situation. Again, this is a project, not a home improvement. This is costing money, not saving - as most of my hobbies cost lots more I am fine with that. Given the scenario as I described it, which is the lesser of the evils I mentioned?

Do you know of a decent battery monitor that isn't cost prohibitive - that I can leave attached to gain more insight into exactly what is going on on this tiny scale system? (like the clipper or Nasa BM1?) I drained my battery to about 65% just to make sure I would be charging in bulk mode and used my Fluke DVM to take I&V readings going to the charge controller every 15 minutes. It seems it used 63Wh over 4.75 hours charging the battery. Its self-consumption is extremely low so it appears that I am generating, on average, 13.25W from a 22.5W array, or about 59% of the best I could hope for. With the second panel I may double this. I know this is nothing to brag about and won't put the pocos out of business but it is fun and more rewarding than spending the money on restaurants and movie theaters IMO.

[Naptown]

You are having fun and hopefully your daughter is too
Not an ideal setup but you do what you can do. (I don't deal with HOA's thank God)
You should fuse between strings of collectors. The NEC requires this when there are more than 2 parallel strings. Now the reality is finding a fuse small enough to work ( generally I would go 125% of the Imp on the string.
There will be a voltage drop across each diode so maybe just blocking diodes on the different facing arrays?

[FloridaSun]

Do you know of a decent battery monitor that isn't cost prohibitive - that I can leave attached to gain more insight into exactly what is going on on this tiny scale system? (like the clipper or Nasa BM1?) I drained my battery to about 65% just to make sure I would be charging in bulk mode and used my Fluke DVM to take I&V readings going to the charge controller every 15 minutes. It seems it used 63Wh over 4.75 hours charging the battery. Its self-consumption is extremely low so it appears that I am generating, on average, 13.25W from a 22.5W array, or about 59% of the best I could hope for. With the second panel I may double this. I know this is nothing to brag about and won't put the pocos out of business but it is fun and more rewarding than spending the money on restaurants and movie theaters IMO.

Check out the Turnigy power analyzer which is used by the R/C crowd. It will show Vm, Ah, Ap, Wp & Wh, also logs so you won't have to check every 15 min. I think I paid around $60 for mine and it's been a great help in showing what's going on in my system.

[inetdog]

As I understand it, the panel is still providing power to the battery (albeit much smaller) with ambient light when the direct sunlight had moved to the other corner of the house, so I want to be able to absorb as much as possible from both panels.

The voltage from the panels with no load attached (Voc) will not change much with the amount of light hitting the panels, from full sun down to moonlight. But the current they produce certainly will.

The current will fall off rapidly as the sun gets to a shallow angle relative to the panels, and indirect light (except on a day of thin clouds causing a bright sky) will not produce much at all. You could measure it with a meter, but the amount of power you get from them into the batteries is not worth thinking about.

[Sunking]

Check out the Turnigy power analyzer which is used by the R/C crowd.

Got a link to that system? I know it is scaled for toys, but it might be real easy and cheap to scale it up. I got an idea on how to make some money off of it, and get around patent protection on the cheap.

[FloridaSun]

Got a link to that system? I know it is scaled for toys, but it might be real easy and cheap to scale it up. I got an idea on how to make some money off of it, and get around patent protection on the cheap.

Scale it UP? S'posed to take 130 amps as is tho it only comes with 10 gauge leads.
Try this; http://www.hobbyking.com/hobbyking/s...dProduct=10080

[FloridaSun]

Check out the Turnigy power analyzer which is used by the R/C crowd. It will show Vm, Ah, Ap, Wp & Wh, also logs so you won't have to check every 15 min. I think I paid around $60 for mine and it's been a great help in showing what's going on in my system.

price correction. Looked up in my records and see I only paid $47.90, delivered, fleabay

[SunEagle]

Scale it UP? S'posed to take 130 amps as is tho it only comes with 10 gauge leads.
Try this; http://www.hobbyking.com/hobbyking/s...dProduct=10080

Wow. Talk about coming full circle. I am one of them RC guys and I own one of those meters. Surprising it is very accurate getting readings off of my plane's motor to determine how much power it draws at full power.

[FloridaSun]

Wow. Talk about coming full circle. I am one of them RC guys and I own one of those meters. Surprising it is very accurate getting readings off of my plane's motor to determine how much power it draws at full power.

Seems like a great tool. Have mine installed between charge controllers and battery for PV input, see exactly what's going into batteries daily and also shows Vm overnight.

[sgking2]

Good info

Thanks guys for the comments. The fuse idea sounds prudent, perhaps a small multimeter type fuse, maybe 300ma per string. That R/C analyzer looks like a cool and economical option to see how things are going. Would you ever use 2 analyzers? Like one between the panel and the MPPT Controller and another between it and the battery so you can see what the panels are delivering and what is either going in or out of the battery on another? Usually the current sense on such devices is super low resistance and shouldn't present too much of a voltage drop to be a concern.

With regard to the super low current at shallow angles, you are absolutely correct I shouldn't worry about even considering that. However, what concerns me is that between 11:30 and 1:30 the sun will actually hit both panels but not equally. Will this be a problem for the MPPT since having unequal strings in parallel will make tracking the MPP difficult at best? This is the part that had me scratching my head trying to find a rationale for one solution over another.... 1) One MMPT and both panels in parallel, or 2) One PWM for each panel. With voltages so close to the charging voltage (18V to 14.2V) I am not certain I am really seeing any gains from the MPPT versus standard PWM. TBH I am not impressed with the Genasun controllers, they do not differentiate between AGM and flooded and will float my AGM at 13.8 instead of 13.2, a good way to shorten the life of a decent batt. They want $50-75 to plug it in and change their microcontroller program to different set points instead of just having a dipswitch or jumper for different battery types. But I guess you get what you pay for, my bad.

[FloridaSun]

Thanks guys for the comments. The fuse idea sounds prudent, perhaps a small multimeter type fuse, maybe 300ma per string. That R/C analyzer looks like a cool and economical option to see how things are going. Would you ever use 2 analyzers? Like one between the panel and the MPPT Controller and another between it and the battery so you can see what the panels are delivering and what is either going in or out of the battery on another? Usually the current sense on such devices is super low resistance and shouldn't present too much of a voltage drop to be a concern.

With regard to the super low current at shallow angles, you are absolutely correct I shouldn't worry about even considering that. However, what concerns me is that between 11:30 and 1:30 the sun will actually hit both panels but not equally. Will this be a problem for the MPPT since having unequal strings in parallel will make tracking the MPP difficult at best? This is the part that had me scratching my head trying to find a rationale for one solution over another.... 1) One MMPT and both panels in parallel, or 2) One PWM for each panel. With voltages so close to the charging voltage (18V to 14.2V) I am not certain I am really seeing any gains from the MPPT versus standard PWM. TBH I am not impressed with the Genasun controllers, they do not differentiate between AGM and flooded and will float my AGM at 13.8 instead of 13.2, a good way to shorten the life of a decent batt. They want $50-75 to plug it in and change their microcontroller program to different set points instead of just having a dipswitch or jumper for different battery types. But I guess you get what you pay for, my bad.

If you're concerned, testing individual panels, using more than one Turnigy might be useful but I'm more concerned with total watts into battery than what's individual panel production. First thing I did with mine was install MC4 connectors to the four leads, thinking I could adapt it in different ways, use alligators or probes in the future. Turned out I left it permanently installed, have even made a case to hold it along with my fuse & ground blocks and other connections. I can still remove it, snap my MC4s together and still use my system but no need so far. My cheapo MPPTs also have digital read out but, no logging.
You will see gains using an MPPT over PWM. On my low volt (18.7v), 145w panels when I switched it went, PWM 7.25amps, MPPT up to 10 amps.... mind you... that was an extremely cheap PWM I was using.

[Sunking]

Scale it UP? S'posed to take 130 amps as is tho it only comes with 10 gauge leads.
Try this; http://www.hobbyking.com/hobbyking/s...dProduct=10080

It does not have to meet NFPA requirements. 10 AWG with high temperature insulation like Teflon in free air can take 130 amps for short periods of time. In its current package I can see some applications on the load side of a charge controller with battery. As it is now completely useless for AC.