Partial shading - two panels - two controllers

Maine Sail · Jan 1, 2014

walt said:
I did a little experiment recently that might be interesting regarding multiple solar panels and partial shading.

I did this because two people who’s opinion I respect use a separate MPPT controller for each panel in some configurations and the reason is that this should work better if you have a situation where you are likely to have shade on one of the panels (like on a sailboat).

I was surprised at the results… maybe just the old panel I used for the test?

Anyhow, you hear that partial shading on a solar panel significantly reduces its output – and it does. What I wanted to look into a little more was how much benefit you get if you have two panels and one has partial shading and you use one MPPT controller for both panels or a single MPPT controller on each panel.

I hooked up a 5 watt solar panel that I picked up with a used boat purchase so don’t know much about it and measured current vs. voltage for this panel for both full sun and partial shade (see the picture for partial shade – its just some electrical tape). The setup is simple, just a power supply that can either source or sink current with a 1 ohm resistor in series to the solar panel. A hand held DVM is used to measure the voltage across the 1 ohm resistor (current is the same as the measured voltage) and also measure the actual voltage on the panel.

What I measured somewhat surprised me as I expected to see more benefit from using a separate MPPT controller on each panel than I did but this may be due to the particular panel I used for this test??

I took this single panel and put in full sun with a way to “force” the voltage (with a power supply) and measure the panel current. I then put a piece of electrical tape on the very same panel right after I did the first test (so the sunlight was the same) and repeated the measurement.

This is the panel with the partial shade:

The image below shows the voltage and currents I measured (put in a spreadsheet to calculate power).

Volts = forced panel voltage
I full sun = panel current with NO shading
Partial = panel current with the tape across it (partial shading)
Pfull = power generated by the full sun panel at the Voltage (i.e., Volts * I full sun)
Ppartical = power generated by the partial sun panel at the voltage (Volts * partial)
Ptotal = Pfull + Ppartial at the forced voltage

What you see from this actual data is that this particular panel has a peak power 4.81 watts at 12.75 volts for full sun and this is where the MPPT controller would also operate the single panel. The peak power points are shaded yellow in the figure above.

What you see when the tape is put on the panel is that the current is reduced considerably (from .378 amps down to .228 amps at 12.75 volts). However, the current output does not change as much with voltage as for the full sun case.

Now the interesting thing is that this panel had a peak power at 12.75 volts for the full sun. But the peak power shifted to 13.75 volts for partial shading. This would imply that an MPPT controller per panel (when there is more than one panel) is going to give maximum power, and this is true.

But.. look at the total power you get from using two solar panels (one shaded) with either one or two MPPT controllers.

One MPPT Controller for two panels (one shaded):
Peak power is 7.7265 watts @ 12.75 volts (see Ptotal in the spread sheet)

Two MPPT Controllers on two panels (one shaded):
Peak power full sun panel 4.8195 W @ 12.75
Partial shade panel 3.025 W @ 13.75
Total Peak for both panels 7.8445

I.e., if we have two identical panels and one has partial shading, using an MPPT controller on each panel only gets you about 1.5% extra power out of the panels – and you may actually get worse since you now have the operating current of two controllers vs. one.

Not what I expected.. I thought I would see a lot more benefit using an MPPT controller on each panel. Maybe its just the old panel I tested?

FYI, I also looked at the open circuit voltage for this panel for different shade conditions. The tape I put across the panel only reduced the open circuit voltage from 17.83 volts to 17.69.

All good in a static test. Now put the controllers on a boat with varying & changing shade and conditions and the boosts for a controller per panel grow even more. Shade out a string and things change differently too. As the panels get larger with higher voltages and bypass diodes etc. this also changes the game. There are lots of variables...

I have personally done side by side comparisons with Kyocera 140W panels and two vs. one controller as has Bruce Schwab and Nigel Calder and we all saw the same thing, better results with one controller per panel..

The gains however may not be worth YOUR financial expense but when doing a system on say a dodger, where every amp counts, the gains can be worth it for some.

Alex MeVay the owner of Genasun (an MIT guy) was the one who really discovered this and built his controllers to be affordable so as to be able to do this and see the most gains where shading and a dynamic environment are in play as they are on a sailboat.. Boats are not a static environment and most of Alex's design work and testing has been around and mostly for boats...

I believe Nigel & Bruce's work can be seen in Sail Magazine but I can't recall the issue...

Like everything a boat is a series of compromises. A single controller will certainly work but a single controller per panel when shading is a factor can lead to better performance.

Quantifying it as a % is tough in a dynamic environment but with my Pentametric I was able to measure fairly solid 8-15% differences (2.8Ah's to 5.25Ah's more per day on this particular boat), in captured Ah's, on a boat with a wind gen causing shading and two bimini mounted panels. This was about 8 days of clear weather testing total on a boat that swings on a mooring. 4 days per set up. These are pretty small differences but when every amp counts, and you are out of space, many owners choose to spring for it...

I plan to do more testing this coming summer on actual boats because I find this is the only really accurate way to quantify the most realistic performance expectations...

I have done plenty of PWM vs. MPPT testing and MPPT has yet to lose but I want to do more single vs. dual controller testing.

Maine Sail · Jan 1, 2014

walt said:
I think if you would have just measured the exact same setup for four days, and them measured the exact same setup again for the next four days, I would not at all be surprised at a 8 to 15% difference. That experiment has a huge extra variable, measured on different days.. The controllers we are talking about here all react quickly, I don’t think that maters either.

Walt,

What controllers were you using? The voltages seem oddly low...... I routinely see 16-17V+ with nominal 12V panels using Genasun controllers, Rogue, Midnite, Outback or Morningstar... More expensive controllers give better performance but most boaters puke at the cost of a good MPPT... I will also very often use panels that push 30v when I can.. Perhaps I am missing something in the translation? What SOC & voltage were the batteries you were connected to? What controllers?

Maine Sail · Jan 1, 2014

walt said:
It would be interesting to do the same thing with a more recent panel, maybe vary also the shading. Is this real.. or just good marketing (like 30% boost for MPPT).

The boosts can actually exceed 30% but in the real world 5-15% gains are what I see with the 12V nominal panels often used.. I have seen some in the 20+% range but that is when using high voltage panels on 12V systems with premium MPPT controllers. With nominal 12V panels 5-15% is all you will realistically see. In my demo above you can see 4A of current at 17.4V boosted to 5A+ which is a very healthy gain at 17.4V.... In the real world panel voltage is constantly changing thus the MPPT gains are constantly changing. Still I have yet to see a PWM beat MPPT on a boat. Most boaters run out of real estate faster than they do wattage so MPPT makes sense for many...

Depending upon shade the whack can be 50-80% or more...... I did this one for one of my customers who wanted to put a BBQ here:

Just one small corner shaded and the output went from .39A to .15A...

KristinKaye · Jan 6, 2014

walt said:
I did a little experiment recently that might be interesting regarding multiple solar panels and partial shading.

The results of your test, walt, are not surprising.

My wife and I have lived off-grid here for 28 years this coming June. We obviously use solar and wind systems and standby and prime diesel generators with battery storage. I also do beta and field testing of off-grid solar equipment for Schneider Electric.

We ran a two year test on arrays receiving unequal insolation with single and multiple MPPT controllers. We have solar arrays aimed in three directions - south facing at optimum tilt angle for the season, east facing at fixed 22.5 degree tilt angle and west facing at 22.5 degree fixed tilt angle. This is a photo of two of the 1.5 kW arrays that I was able to get in one photograph and the controllers being tested (dual MidNite Classic 150's and dual Schneider Conext (formerly Xantrex) XW-MPPT60-150's):

Data on solar insolation (measured in watts/sq meter) was collected with a sensor at each array. Power measurement data on the system (including system load, battery voltage and net current, battery current from each controller, array Voc, Vmp, and current, and 21 other variables) was collected and logged every three seconds for two years.

For testing with single controller on all three arrays (4.5 kW STC input capacity to the controller) all arrays were wired at 92.4 Vmp STC. For testing with individual controllers on each array the arrays were wired at 61.6 Vmp STC to achieve better controller conversion efficiency.

The south facing array is subject to shading from pine trees to the east in the summer and from a 90 foot wind turbine tower in the winter months to the SE of the array. The east facing array is subject to shading from the 90 foot wind turbine tower and shadow flicker from the turbine blades during the summer months.

The east and west facing arrays, when operated on one controller, never achieve equal insolation except at solar noon.

Over the two year test we determined that the optimum configuration was to run one controller on the south facing array wired at 61.6 Vmp, and a second controller on the east/west arrays wired at 92.4 Vmp. There was no advantage to running individual controllers on the east/west arrays - kWh production of the arrays was the same with either configuration, within the acceptable measurement error range of the logging equipment.

The only reason dual controllers was required was because under perfect solar conditions the output capacity of a single controller on all three arrays was exceeded. So we lost potential production due to the controller amp-limiting at it's maximum output capacity.

Now, this testing was done with high voltage strings which are subject to losing an entire string when only one panel in the string is shaded. And it was done with high-end very expensive controllers with shading tolerant algorithms. The cheap low end controllers that many people use on sailing yachts are not going to have this. But the bottom line in the two year test was - spend your money on a good controller and you do NOT need multiple controllers with partial shading problems on the array(s). With cheap low-end controllers with very basic MPPT algorithms you're probably going to get slightly better energy production using multiple controllers. But the money spent on the second controller is not justified when you can buy another solar panel and add it to the system for the same price as a controller (as long as you don't exceed the output capacity of a single controller).

KristinKaye · Jan 6, 2014

For folks interested in learning more about shading issues on PV strings, there is a white paper prepared by Schneider Electric on the topic
http://www2.schneider-electric.com/documents/support/white-papers/seshadetolerantwp.pdf

Again, this is only going to apply to high end controllers. The cheaper low end controllers mentioned in this thread are VERY slow on tracking and use quite basic dynamic MPPT tracking algorithms.

So my point is, folks, don't get hung up on thinking you need a separate controller for every string or panel to optimize output in partial shading conditions. This is already really old technology in the PV business. The more technologically advanced controllers that can detect shading from something as small as a pencil on a 1.5 kW string, and adjust for it within a few microseconds by not tracking the shaded portion of the string have far outpaced the old dynamic controllers mentioned in the thread.

KristinKaye · Jan 6, 2014

walt said:
This is another reason that on the sailboat, the controller per panel doesn’t buy you much as you don’t see the benefit of higher voltage you get with a home grid tie system.

Hi walt - our power system here is not grid-tie. We are totally off-grid and have not had utility power at our home for 28 years this coming June.

You are talking about the I^2R loss in the wire run. And you are basically correct in that the higher you can run the voltage, the less loss you will have. That is a downfall of the standard ~17 Vmp panel operating standalone with a single controller on a 12V system. I recommend not using those panels at all. They are not a good buy as far as $/watt in the first place. Use the standard grid-tie panels that run at ~30.8 Vmp instead. And if you have two, run them at 61.6 Vmp in series. Most of the cheap controllers can't go to 150V so running three in series at 92.4 Vmp does not allow enough headroom to the max operating voltage of those controllers. But if you can run three in series at 92.4 Vmp you can handle 5.3x the wattage on the same two wires that is required for a single ~17 Vmp panel of the same wattage rating.

The benefit should be clear, and no multiple controller arrangement with a controller per panel or per array is going to match it. Each controller only operates at about 94-95% efficiency, so you automatically lose 5-6% just in conversion losses in each MPPT controller. The more controllers you add, the more losses you introduce into the system.

I am refitting our Legend 37 we recently bought for cruising. I am installing two 160W 30.8 Vmp panels in series with a MidNite Solar KID Marine MPPT controller. This controller will handle partial shading with no issues as it uses the same fast sweep algorithm as it's bigger brother, the Classic 150/200/250. You will find that the cheap controllers that use P&O and dynamic tracking are way too slow and not accurate enough to track partial shading on series strings. But the KID Marine is $500. You can get the cheap controllers for around $175. You get what you pay for. Been testing and putting MPPT controllers through their paces here since MPPT became the hot ticket over PWM. And there is no free lunch with solar power systems - there is only well design systems and hacked together ones.

KristinKaye · Jan 6, 2014

walt said:
If I have two panels in series and one gets shaded, what happens to the overall power output of the two panels?

The output of the series panel string drops. How much it drops, and how fast it is tracked depends on the controller. The newest high-tech controllers can detect partial shading and track the global maximum instead of the shaded part. The older technology controllers that use P&O and dynamic algorithms will tend to detect a lower output and drop the voltage of the string - eventually - but can't track it fast enough in an installation like a sailboat where the panels might be changing pitch as the boat rolls on waves and gets intermittent partial shading from sails, etc.. The more expensive controllers with more advanced hardware and software can track this very precisely and very fast (within a few milliseconds).

Please refer to page 8 on the whitepaper produced by Schneider Electric Solar that I posted the link to. It explains how the I-V curve of a PV module or PV string changes with shading. It also explains how the traditional MPPT controllers handle it vs the more high-tech ones that use Fast Auto Timed I-V Sweep technology that precisely adjusts a string's output to the global maximum instead of slowly tracking what it detects as a change of more than 10% in output (dynamic).

To the human eye your solar MPPT controller might look fast as it tracks. But in electronics 1 second is like a century. So do not base your assumptions on the likes of a Morningstar Sunsaver on 17.5 Vmp panels.

Bill Roosa · Jan 6, 2014

If all you are concerned about is getting max output current in the dynamical sailboat environs then go for lots of bypass diodes. I don't think this has been addressed by industry as most solar installations are not troubled by shading like sailboat panels are.

KristinKaye · Jan 6, 2014

Bill Roosa said:
If all you are concerned about is getting max output current in the dynamical sailboat environs then go for lots of bypass diodes.

Well, the thing is that maximum output current is not what you're looking for. Maximum kWh harvested is what gets the job done. You can put out 10 amps @ 14 volts or 1 amp @ 140V and the system operating at 1 amp is more efficient than the 10 amp system, and consequently will harvest more kWh.

The reality is that for most people out cruising a few lost watt-hours from shading is not a big issue because you're going to run your propulsion engine once a day anyway for hot water. The alternator on the engine can produce more amp-hours in 15 minutes than a couple small solar panels can all day in perfect conditions.

The solar panels are most important when you're dangling on the hook and don't want to run the engine to meet onboard power needs.

I'm just pointing out for the people who like details that a properly designed MPPT solar power system does not suffer the losses from partial shading that necessitates the use of multiple controllers. And I have years of experience to back it up.

Maine Sail · Jan 7, 2014

This is all good info and will lead me to do more testing on the panels and controllers I would normally use for a dodger mount. A dodger is really the only time I use one controller per port/stbd panel.

Some questions I can't seem to get answered by manufacturers..

#1 How does the MidNite compare RFI/Noise wise to Tri-Star, Outback, Rogue etc.? I have had trouble with an Outback (Flex 60) with regard to noise. The Rogues have been bullet proof and quiet as have been the Genasun and TriStar MPPT's.

#2 Why doesn't the solar industry have a standard rating for MPPT refresh rate/perturbation frequency /MPPT frequency? Perhaps a rating for shade tolerance? Why doesn't the solar industry tell you what type of algorithm they are using? They all fall into a category even if it is a custom algorithm. Why all the secrecy?

#3 What controllers today still use dynamic tracking (pause charging, check voltage, charge)?

#4 Do ANY of these controllers that claim to have Li charge settings actually discontinue charging when the bank is full and not proceed to a float charge?

#5 Which is the fastest and most reasonably priced controller that allows for full custom programing of charge parameters?

#6 The MidNite KID seems great but it was supposed to ship in January and now my wholesaler is telling me late first quarter, early second quarter. I am quoting projects now for spring and hope they don't run late as I would like to try one.... Do you know any different or have a supplier who is getting them sooner..?

#7 The MidNite KID is awfully bulky for the limited panel real estate we have on boats. Can the 150 remote be used with the KID? My wholesaler was shown a prototype that looks nothing like the KID they are now showing....?

#8 Is the MidNite KID going to be faster than the Rogue 3048 when dealing with shade? I do know from talks with Alex at Genasun that the little GV controllers are "sweeping" 20X per second. How does that compare to a MidNite, Outback etc..?

#9 If these controllers deal so well with shading can different brands/sizes and voltages of panels be combined? Generally that is not suggested with MPPT. Do these new shade tolerant controllers change this? Due to space constraints we often have to mix panels on boats to get the most array possible.

KristinKaye · Jan 7, 2014

Maine Sail said:
Some questions I can't seem to get answered by manufacturers..

Hi Maine Sail - lots of questions here and I'll try to answer as best I can.

#1 How does the MidNite compare RFI/Noise wise to Tri-Star, Outback, Rogue etc.?

I don't know how they compare because I have never compared side by side. I have used a MidNite Classic 150 on our motor yacht and never had any problems with RFI noise with it.

#2 Why doesn't the solar industry have a standard rating for MPPT refresh rate/perturbation frequency /MPPT frequency? Perhaps a rating for shade tolerance? Why doesn't the solar industry tell you what type of algorithm they are using? They all fall into a category even if it is a custom algorithm. Why all the secrecy?

Lots of competition in the PV industry, which is good. But these algorithms that the manufacturers develop are highly proprietary. Especially in the high-end controllers. The MidNite Classic actually has several algorithms in it and you can select the one you like best.

#3 What controllers today still use dynamic tracking (pause charging, check voltage, charge)?

Just about all of them in the lower prices ranges use dynamic tracking. But not all track the same. Some are faster than others. Some are so slow that there's no way they work properly on a sailing yacht. The higher end controllers that do timed I-V sweeps are much faster and more suitable for use on a sailboat.

#4 Do ANY of these controllers that claim to have Li charge settings actually discontinue charging when the bank is full and not proceed to a float charge?

Now, that's a good question. Is there people that use Li-ion batteries on sailing yachts? The fire hazard would seem to be an undesirable thing.

#5 Which is the fastest and most reasonably priced controller that allows for full custom programing of charge parameters?

The MidNite Classic. Hands down winner. And hopefully the KID. I haven't tested my KID yet.

#6 The MidNite KID seems great but it was supposed to ship in January and now my wholesaler is telling me late first quarter, early second quarter. I am quoting projects now for spring and hope they don't run late as I would like to try one.... Do you know any different or have a supplier who is getting them sooner..?

Not at this point yet. I am an established beta tester and I have a KID in my hands that is going in our Legend 37. But to get them thru retail channels is a few weeks out yet. If you would like, I could hook my KID up to our home off-grid system and fire it up and put it through its paces to see what it's got and post my thoughts on it.

#7 The MidNite KID is awfully bulky for the limited panel real estate we have on boats. Can the 150 remote be used with the KID? My wholesaler was shown a prototype that looks nothing like the KID they are now showing....?

No, the MNGP for the Classic is totally different than the KID as far as I know. As per the above quote, I can fire mine up and compare it to a Classic 150 side by side and let you know more about it, if you would like.

#8 Is the MidNite KID going to be faster than the Rogue 3048 when dealing with shade? I do know from talks with Alex at Genasun that the little GV controllers are "sweeping" 20X per second. How does that compare to a MidNite, Outback etc..?

I never got any beta controllers from Mark at Rogue Power Tech. I don't know how fast they are for sure and have not heard from any other users. I think Mark was struggling to get product out the door and get the sales channels established and had a lot on his plate at the time. I have a feeling, that based on my experience with the Classic, that the KID is going to be hard to beat. I have tested a LOT of beta software in the Classic that boB (Gudgel) has sent me and they got the best stuff out there right now. I hate to admit that, running a full Schneider Conext system here, but those Classic controllers are dang fast and precise. They are the only ones on the market fast enough to track a wind turbine.

#9 If these controllers deal so well with shading can different brands/sizes and voltages of panels be combined? Generally that is not suggested with MPPT. Do these new shade tolerant controllers change this? Due to space constraints we often have to mix panels on boats to get the most array possible.

You can, but like you say it's not recommended. You will take a hit on nameplate capacity of the array. You can combine different makes with the same STC specs. But if you use panels with different Vmp's and wattage ratings on one controller you will limit the more (potentially) powerful modules to the output capacity of the less powerful ones. There is no way around that.

Maine Sail · Jan 7, 2014

KristinKaye said:
Hi Maine Sail - lots of questions here and I'll try to answer as best I can.

Thanks for the response. I will get a pre-order in for the KID. They WILL need a smaller remote for that unit though if they plan on selling into the marine market...!!!

KristinKaye said:
Now, that's a good question. Is there people that use Li-ion batteries on sailing yachts?

Lots of people and growing every day.

KristinKaye said:
The fire hazard would seem to be an undesirable thing.

The fire issue is a non issue unless you do what some others do and extrapolate LiFePO4 to a more volatile Li chemistry like LiCoO2..

Damage to the wallet if not properly installed & charged is a potential issue. On boats we are using LiFePO4 which is arguably as safe or safer than FLA.

Our LiFePO4 bank has been truly amazing and I personally will never go back to lead... Not for everyone and you need a good solid understanding but if you are willing to spend the time to build the system and execute it well it is tremendous..

We used to really depend on solar on our boat. Now it sits turned off most of the time because the Li bank charges so fast. We charge to 100% (Full = 13.8V and 5A of accepted current) shut off all charge sources and cycle to 80% DOD then recharge. Simple as it gets.

Partial shading - two panels - two controllers

.

Attachments

.

.

.

.

.

Attachments

.

.

Attachments

.

.

.

.

.

.

.

.

.

.

.

.