Solar for Sailboats -- microMPPT

[nimblemotors]

I've finished up the design of the microMPPT controller.
Now just waiting for the boards to arrive to confirm it works as documented,
and then the kickstarter for the $25 boards will open.

The writeup of how and why this is the best way to setup solar for a sailboat is on the Electric Boat Forum website.

I am also looking small scale manufacturing of solar panels specifically to deal with shading.
Using the lightweight and flexible and highly efficient sunpower cells.
So unlike say the Renogy (chinese) panels that only have 2 bypass diodes,
these panels will be designed for multiple microMPPT controllers.

In addition, to maximize output for a given space, an additional 5% can be gained by not having the 'holes' in the panels from the cut corners.
Ever wonder why the super expensive Solbian panels still leave you short changed with 'solar holes' in the panels?


In addition, like all good things on a boat, the solar panels can double duty as solar water heaters. I will investigate water cooling so one can get hot water while water cooling the panels to maximize their output. It's a win-win. Better solar efficienty AND hot water.

Yes these will cost more than cheapest panels.

[Dsanduril]

Jack,

I don't want to rain on the parade, because I think this is a noble and valuable effort, so the following is offered in a constructive vein:

1. Solar 'holes' - the ingots from which mono cells are cut are generally round. There are lots of reasons for squaring them up, but getting them fully square is costly in terms of wasted silicon. Putting circular cells on the panels is the most material efficient, but wastes a lot of space. Square cells are the most space efficient, but waste a lot of silicon. What you see is the market is the most cost-efficient middle ground. You can certainly cut down cells to square, but it won't be cheap.
2. MPPT - What are you doing with voltage? Are you planning buck, boost, or both? Most of the existing MPPT use buck (least expensive) but that then requires a minimum number of active cells to reach a voltage above the charging point. If you're going to boost then it doesn't really matter, but at $25/unit seems like this could get really costly really quickly if you are putting a controller on only a few cells.
3. Water cooling - We've played with this a fair amount, on paper it sounds like a good idea, but you actually need a huge quantity of water. Most panels operate ~20-25C above ambient (and your Sunpower cells will be at the lower end of that range). In order to get a useful electrical boost you want to reduce that operating temperature by ~10C, which leaves you with a maximum heat increase on the water side of ~10C (you can't really get a counter-flow heat exchanger going here - otherwise you have uneven temperatures in your cells). 10C only gets you lukewarm water. With no feasible way to amplify the heat, you need to move a lot of cool water across the panels (the ocean is handy, but doesn't suit your stated purpose of heating water for use). What you have is ~300W/m2 heater (so plenty of heat really) that can only raise temperatures by 10C.

Just some food for thought as you put together your project, I look forward to seeing your test results.

[nimblemotors]

It isn't costly to square them, if you can use the cutoffs in other smaller panels and applications that don't have the space premium of a boat.
Like my shop rooftop, plenty of space there, no worry about holes.

The MPPT board for the kickstarter boosts the voltage from a 12v panel to 24v, so an off-the-shelf 12v panel can be used. They don't make any 4-cell panels that I know of. This is why I'm looking into producing them.
The MPPT for this will cost much less than $25.

You are probably right regarding heat from the panels, however, even that small amount of heat is useful as input water for the watermaker making it more efficient. Will have to research it certainly, as it would increase the cost of the panels.

Originally Posted by Dsanduril

Jack,

I don't want to rain on the parade, because I think this is a noble and valuable effort, so the following is offered in a constructive vein:

1. Solar 'holes' - the ingots from which mono cells are cut are generally round. There are lots of reasons for squaring them up, but getting them fully square is costly in terms of wasted silicon. Putting circular cells on the panels is the most material efficient, but wastes a lot of space. Square cells are the most space efficient, but waste a lot of silicon. What you see is the market is the most cost-efficient middle ground. You can certainly cut down cells to square, but it won't be cheap.
2. MPPT - What are you doing with voltage? Are you planning buck, boost, or both? Most of the existing MPPT use buck (least expensive) but that then requires a minimum number of active cells to reach a voltage above the charging point. If you're going to boost then it doesn't really matter, but at $25/unit seems like this could get really costly really quickly if you are putting a controller on only a few cells.
3. Water cooling - We've played with this a fair amount, on paper it sounds like a good idea, but you actually need a huge quantity of water. Most panels operate ~20-25C above ambient (and your Sunpower cells will be at the lower end of that range). In order to get a useful electrical boost you want to reduce that operating temperature by ~10C, which leaves you with a maximum heat increase on the water side of ~10C (you can't really get a counter-flow heat exchanger going here - otherwise you have uneven temperatures in your cells). 10C only gets you lukewarm water. With no feasible way to amplify the heat, you need to move a lot of cool water across the panels (the ocean is handy, but doesn't suit your stated purpose of heating water for use). What you have is ~300W/m2 heater (so plenty of heat really) that can only raise temperatures by 10C.

Just some food for thought as you put together your project, I look forward to seeing your test results.

[jbpatents]

Jack, I see that you are in Sacramento. If you need some 4 cells panels for testing, let me know. I probably have some in my lab that we used for testing the cells reliability. These are ugly panels with no junction box, but for testing you may be able to use them. I should also have single-cell panels. All of these were reliability testing panels and are glass-glass with tabs coming out - no junction box. They are in Sunnyvale - a short drive for you.

As for water heating - unless you going to do concentrated solar, forget about it.

As for MPPT, check out what Tigo is doing instead (as opposed to SolarEdge). I have Tigo at my home, although I personally know the founders of both companies. Maybe you can think of something like that.

As you may have seen in my other posts, if you'd like bi-facial panels, I have a few MxB290SL available in Sunnyvale. These will give you 290W from the front, with probably about 20% boost from the back.

[nimblemotors]

jb,
I already have 5kw of individual thin cells for my solar boat, so that is not a problem. They are not high efficient cells, but were cheap.
I don't think sailors are too interested in making their own panels.

Certainly I believe the huge grid panels are not useful for sailboats,
but the bifacial cells themselves might be interesting, and would like to check them out. Will PM you..

[leftbrainstuff]

Hot water from waste heat from panel cooling would certainly be a compelling point of difference. The numbers arent great but I like your thinking.

We 're currently designing a hard dodger for our Liberty 458. We are considering incorporating a solar hot water concentrator into the dodger. Mainly to augment our other hot water heating options. I havent run the numbers yet but I mentored several solar projects in a graduate engineering program a few years back. Solar concentrators easily meet our needs and fit within our available space claim.

We're relocating our solar panels to the transom arch to address the shading issues from the current dodger mounts. These will be in a readily stackable and deployable group.

Sent from my SM-N900T using Cruisers Sailing Forum mobile app

[barnakiel]

I know there is an extra gained from back linking. Look up black panels available from SVB. Now the other guy mentioned backlit panels too - some extra gained from the reflected light hitting the belly (underside) of the panel.

Maybe a backlinked translucent panel would beat them all?

With twin glass, we could possibly shave the alloy frame a bit to save weight ...

b.

[OceanSeaSpray]

Multijunction GaAs cells already reach >40% efficiency, so twice that of the old monocrystalline technology we are using. If we were prepared to spend a whole lot more money on solar power, it could be better spent on better panels rather than what would be marginal improvements in comparison.
MPPT, cooling etc all attempt to reclaim very small margins that are not even always available.

The annoying thing at the moment is that the solar cell industry is driven by the cost-per-watt figure: space in land installations is usually plentiful. For us on yachts, higher efficiency would make a big difference, but that market is rather small in comparison...

Another area where money could be better spent is in energy efficiency on board. Forget about always trying to generate more: reduce the energy footprint. These days, it is normally fairly easy to be self-sufficient on board with well-designed, sensible systems and without resorting to some kind of floating solar farm either in all regions that have reasonable clear sunlight hours.

In short, I think it would pay to have a good hard look at the economics and alternatives before heading all the way down one track.

[nimblemotors]

Did you read the article I linked to on electric-boat-forum?
The more efficient the cells, the more it matters to lose 50% from shading. That is not just a 'marginal improvement'.
MPPT might be marginal, but to connect cells in parallel you need a dc-dc anyway, so having it do MPPT is just that much better at no additional cost.

Have you priced those 40% cells?

Originally Posted by OceanSeaSpray

Multijunction GaAs cells already reach >40% efficiency, so twice that of the old monocrystalline technology we are using. If we were prepared to spend a whole lot more money on solar power, it could be better spent on better panels rather than what would be marginal improvements in comparison.
MPPT, cooling etc all attempt to reclaim very small margins that are not even always available.

The annoying thing at the moment is that the solar cell industry is driven by the cost-per-watt figure: space in land installations is usually plentiful. For us on yachts, higher efficiency would make a big difference, but that market is rather small in comparison...

Another area where money could be better spent is in energy efficiency on board. Forget about always trying to generate more: reduce the energy footprint. These days, it is normally fairly easy to be self-sufficient on board with well-designed, sensible systems and without resorting to some kind of floating solar farm either in all regions that have reasonable clear sunlight hours.

In short, I think it would pay to have a good hard look at the economics and alternatives before heading all the way down one track.

[OceanSeaSpray]

Originally Posted by nimblemotors

Did you read the article I linked to on electric-boat-forum?
The more efficient the cells, the more it matters to lose 50% from shading. That is not just a 'marginal improvement'.
MPPT might be marginal, but to connect cells in parallel you need a dc-dc anyway, so having it do MPPT is just that much better at no additional cost.

Have you priced those 40% cells?

What you lose from shading depends on the construction of the panel, not so much the cells themselves. The best panel should have one ideal bypass diode over each cell, so you never lose more than what is actually shaded. You can actually find some with one Schottky diode over each cell, but most panels have a couple of diodes only and shading losses come in huge steps instead of very gradually.

The improvement from MPPT is directly proportional to the mismatch between the panel voltage at its maximum power point and the charging voltage. If you design to have a small mismatch most of the time, MPPT isn't actually of great value. It is a small gain that comes at a price. As long as it is cheap... ok. It often isn't and MPPT is primarily used to connect mismatched panels or create dubious series configurations. Just don't do that for a start.

The price of multijunction GaAs cells is not going to come down until people actually start asking for them and using them.

It is great to have a play and see what can be done, but sometimes if you do the numbers up-front, you just don't bother... here I am not sure which way it would go, but there certainly isn't a lot of money to play with to remain competitive because the gains are too small compared to a good setup using current technology.

[txg]

Originally Posted by OceanSeaSpray

The improvement from MPPT is directly proportional to the mismatch between the panel voltage at its maximum power point and the charging voltage. If you design to have a small mismatch most of the time, MPPT isn't actually of great value.

While that's actually theoretically correct, it is impossible in the real world to get a really good match between panel and battery. Solar Panel output voltage decreases with rising temperature, and battery voltage increases with charge level.

This is a really well done comparison between MPPT and PWM controllers:
MPPT vs. PWM Controllers Photo Gallery by Compass Marine How To at pbase.com

In summary: In a practical installation with a "standard" panel for 12v applications, there will be a gain of around 20% with a MPPT controller, even without any shading. In my opinion, this is a HUGE difference.

And another thing to consider is that MPPT controllers allow you to use higher voltage panels. At least on bigger boats (let's say with a 250W panel and 7m wire from panel to batteries, so 14m wire needed all in all) the possible reduction of wire diameter due to the higher voltage and lower current saves a part of the higher investment for the MPPT controller.

[SailFastTri]

One thing I don't see discussed often is the interaction between multiple charge controllers in parallel on the same battery bank. Surely the voltage rise caused by one affects the sensed voltage of the others, and therefore the output from them?

[nimblemotors]

That is a good question.

The 'microMPPT' controller that will be in the kickstarter, is really a 'miniMPPT' because it works for off-the-shelf 36 cell 12v panels, not at a micro cell level.

It is NOT a battery charger. There is a single 24v output bus all panel outputs are connected to.
So you need only a single battery charger that gets the 24v input.
This can be a cheaper 'PWM charger' for lead acid, or any charger that will take 24v input.

Originally Posted by SailFastTri

One thing I don't see discussed often is the interaction between multiple charge controllers in parallel on the same battery bank. Surely the voltage rise caused by one affects the sensed voltage of the others, and therefore the output from them?

[OceanPlanet]

Originally Posted by SailFastTri

One thing I don't see discussed often is the interaction between multiple charge controllers in parallel on the same battery bank. Surely the voltage rise caused by one affects the sensed voltage of the others, and therefore the output from them?

We hear this question a lot....however with low-current sources (like solar) there is virtually no interference or "bouncing". The battery voltage rises as they fill up and the various regulators all regulate to whatever voltage levels they are programmed for. Some controllers with the same programming may taper off a bit before the others based on tiny differences in wiring voltage drop, however this would be at the very last bit of charging and have no effect on the charging of the batteries.

Now if you have two larger charge sources (alternators, for instance) that produce enough current to rapidly change the battery voltage, then they can wind up with uneven current (one charging more than the other). This is when you would use a product like the Balmar "Centerfielder" to have them charge at the same rate.

However with solar have never seen any problems with multiple controllers "fighting" each other....

[OceanSeaSpray]

Originally Posted by txg

While that's actually theoretically correct, it is impossible in the real world to get a really good match between panel and battery. Solar Panel output voltage decreases with rising temperature, and battery voltage increases with charge level.

This is a really well done comparison between MPPT and PWM controllers:
MPPT vs. PWM Controllers Photo Gallery by Compass Marine How To at pbase.com

In summary: In a practical installation with a "standard" panel for 12v applications, there will be a gain of around 20% with a MPPT controller, even without any shading. In my opinion, this is a HUGE difference.

And another thing to consider is that MPPT controllers allow you to use higher voltage panels. At least on bigger boats (let's say with a 250W panel and 7m wire from panel to batteries, so 14m wire needed all in all) the possible reduction of wire diameter due to the higher voltage and lower current saves a part of the higher investment for the MPPT controller.

This is actually the typical sales pitch. No, a perfect match is not going to be achieved all the time by matching the panels to the bank voltage, but a damn good match can a lot of the time. If you can choose between investing in one or two extra panels or a MPPT unit, the extra panel(s) blow the MPPT out of sight every time.
Also, performance without power conversion (MPPT) always varies and can degrade with panel temperature increasing, sure, but let's have a look past the hype here:

If the panels are "hot", it is because they are in intense sunlight. At that point, they normally produce far more power than actually useful because if the system was sized to be functional in perfect conditions only, it would next to useless the rest of the time. What happens at peak output doesn't actually matter that much in many instances. I need to generate enough on a cloudy day, when there are no shadows either incidentally, and I couldn't care less about what happens in full sun.

Next is that the so-called "efficiency" of MPPT controllers is always quoted as good as it gets when not fudged outright for the glossy little brochures (by hiding the unit's own consumption etc). As soon as they have heatsinks and fins, you can bet they are not too good at high output. They can be really bad at low output too, because of their own consumption eating into the panel limited output... the higher the conversion frequency, the worse they get at low output.

When it comes to shading, haha, how are you to reclaim power that hasn't been generated in the first place... It only works if the panel is high voltage and shading doesn't cause its output voltage to drop below the battery voltage. The catch is that if the panel is high-voltage, it is also a lot more sensitive to shading a lot of the time.
If shading really is a concern, spend money on high-quality panels that have one bypass diode per cell, long before MPPT gizmos, because this does make the difference: you lose one cell at a time, not a quarter, third or half of the panel.

A good MPPT converter that runs cold, with ultra-low intrinsic consumption, is a really neat component in a system and it creates more flexibility with panel selection, but it is certainly not the "must have" that fixes everything auto-magically like many like to claim. Good units in well-designed systems always have an edge, but this in itself is not enough - and by far - to make it an automatic value proposition.

The future of MPPT, as far as I can see, is in small, low-cost and highly efficient units. Low-cost opposes highly efficient, so there are always going to be trade-offs there, but small (less than 20A) is fully aligned with the other two.
The best topology would be dissociating the charge controller from the power converter(s) and run a few smaller black box MPPT units feeding into a single programmable multi-bank charge controller. For that matter, the vision of the OP is spot on.