Well this is interesting. Half sunny half cloudy day, 92 degrees out. Collected this data looking at the turnigy for the maximum panel output (not max amps to battery which was fun but not any real objective yet) for each line:When it is sunny out, my Unisolar 68 watt strip stays between the theoretical Vmppt of 16.5v and a few tenths below that most of the time (16.35v or so average I think).When moderate clouds come out, I saw it drop a hair below 14 volts multiple times (I got full data for one of them). That has interesting implications for serial vs parallel solar setups using 12 V panels on cloudy days. On the flip side, when the battery does not need charging at over 14 volts, that makes PWM around the same efficiency as very good MPPT under those conditions. If the MPPT is not very good, the PWM could actually yield more watts to the battery.It also means that on days with fast moving clouds moving through bright sunshine, a slow MPPT algorithm is going to be WAY off sometimes, such as for an optimization interval where it set Vmppt for sun and then there are clouds for most of the next interval. Output essentially falls of to near 0 because the IV solar output curve shifts radically left.Fun stuff!JimP.S. DSchmidt_2000 you were right on with the idea of using a constant voltage load. Once I created a good approximation of it, it was MUCH easier to adjust the pot to find the max power point.

This makes me want to look at the time and interval on various controllers to scan the voltage range and pick A MPPT voltage.Jim

Wonder if a full scan is even needed. I'm guessing the MPPT 'graph' for a panel would be monotonic. So small scanning could continually occur.

I think if you use all identical photovoltaics mounted facing the same direction it may well be. But picture a mix of panels or mounting angles (say one east and one west), and/ or panels with different IV curves. Or even shade, where one panel is all the way in building shade, another is partial, another full sun. I believe there are multiple local maxes in that case. Mine would have at least a max around 14v and a second one near 16.5v, and possibly a third max for the partial shade unit depending on how the shade hit it.The guys who race solar cars have published some interesting approaches to running MPPT in them.Jim

With the overwhelming majority of affordable panels now being produced for high voltage arrays is not the "debate" of MPPT vs PWM now a moot point (except for very small systems)? Seems like 17-21 voc panels now cost a lot more per watt.

Yes except for those of use who already have 12v. Though I do see people posting about buying them from time to time still due to the cost of MPPT controllers. Mostly smalley systems. But the shift is on and will continue.I figure sooner or later I will have one, and understanding what the real MPPT volts are the controller has to work with helps me filter the various marketing claims of some of the controller makers.If I could get one with a programmable ECU like the PIC in the Eco-Worthy, that would be ideal. Existing hardware platform but all kinds of programmability. Even just changing the upper and lower search voltages to fit my specific panels, and temp compensation values would be very powerful and much less expensive than some other controllers.But the first step for me is doing the work *of* the controller with my system parts, so I really know how it works, rather than speculation from specs of how it should work.JimAlso I really wanted to understand the real world benefit (or not) of MPPT. I have seen real world improvements in charging amps of over 10%, so I did convince myself it is not all fluff. But watts per dollar with my Unisolars I think I would use a working figure right now of a 10% gain vs the cost of going from PWM to MPPT. Which for me would be something between $100-$400 to gain 10% of 136 or somewhere between $7 and $29/watt. Not exactly a bargain if going there for the power gain alone.

Just witnessed MPPT voltage drop below 14V | Good Sam Community

HiTech
Explorer
May 24, 2013
I would do the same once I had enough to retire, just phase back some.

Jim
pianotuna
Nomad III
May 24, 2013
Hi Jim,

I've never worked more than about 40% of full time, since 1984. It was enough to meet my needs. I love my work--so I see no need to stop, though I am finally starting to get "selective" about which clients I rebook.
mena661
Explorer
May 24, 2013
I'm still a little ways off from 50 and it's possible I could retire at that age but I like working so I don't see that happening. At best, I would just find something different to do (ie still work).
HiTech
Explorer
May 24, 2013
I am in awe of you who planed well enough to retire at 50. As I approach that age, I have 3 kids in college. We saved well but college costs considerably more than we saved.

I wish I still had those blank punch cards. There is a company in Conroe Texas who still runs *production* scheduling on an early punched card machine where you hard wire the instructions. With wire jumpers. Apparently blank cards a getting a little hard to find these days...

Won't sell it to the computer museum

Jim
pianotuna
Nomad III
May 24, 2013
Hi Jim,

Your results are why I chose series/parallel for my install. I think it is one reason I do get some charging in the rain. Today it will rain all day--and I have no reason to drive, so I'll survive on the battery bank's reserves. I'll only get enough to cover the parasitic drains today I fear.

I agree--a poor mppt controller can, in some circumstances do less well than pwm for nominal 12 volt panels. (17 volts).
Salvo
Explorer
May 24, 2013
FORTRAN, IBM punch card, been there done that. Retired at 50 to spend more time to mountain bike the best trails out West.
HiTech
Explorer
May 24, 2013
The bimodal operation is very handy.

Jim
DSchmidt_2000
Explorer
May 24, 2013
Was looking around the web and
FOUND THIS article on an electronic load.

This circuit is useful for characterizing photovoltaic modules, which have two source modes.. . . . . At voltages below VMPP, the photovoltaic modules look like a current source. It is normally difficult to characterize this flat region of the curve with a simple current-mode electronic load, because the voltage output is sensitive to small variations in current, and thus a constant-voltage mode load is a better choice.
HiTech
Explorer
May 21, 2013
I was the last engineering class to do punched cards at my college as a freshman in 1982.

So SGI maybe?

Man it would be a miracle for me to retire at 50 LOL.

I have done a lot more chief architect/chief engineer stuff than software engineering for a quite some time now. Code was a lot of fun.

Jim
KJINTF
Explorer
May 21, 2013
Jim

I did programming back in the day with a punch card reader - Damn I must be an old fart. Moved on to embedded systems along with hardware design for extreme dollar video processing equipment, Invested my $$ properly and retired early at 50 young years old, which was a long time ago.