Maximizing A Solar Panel

Solar panels seem like simple devices: light in and electricity out, right? If you don’t care about efficiency, it might be that simple, but generally you do care about efficiency. If you are, say, charging a battery, you’d like to get every watt out of the panel. The problem is that the battery may not draw all the available current, basically leaving capacity on the table.

The solution is a technique called MPPT (Maximum Power Point Tracking). Despite sounding like a Microsoft presentation add on, MPPT uses a DC to DC converter to present a maximum load to the solar cell while providing the desired current and voltage to the load. MPPT is what [Abid Jamal] implemented to manage his solar charger.

In addition to the solar panel and DC to DC converter, [Abid’s] project uses an Arduino, an LCD, some indicator LEDs, and some discrete components. He even included an ESP8266 to provide wireless data logging. The finished project resides on perf board and lives in an acrylic case.

There’s a similar MPPT project in the Hackaday prize competition (in fact, the video below is from the creator of that project), and we’ve seen other MPPT builds, too. It might be interesting to contrast the different designs.

21 thoughts on “Maximizing A Solar Panel

  1. If your installation in very simple, you can get away with modifying a LM2596 board (the same board as Al used in this project) into a fixed power point converter. You set it once on a shiny day and forget it.

    I build one that boosts the solar voltage: http://atoomnet.net/solar-energy-fixed-power-point-boost-converter/
    It uses the feedback pin of the step up chip to balance around the fixed power point.

    And I build one that steps down the voltage. I do not have a description of that, but it is as simple as connecting a zener diode from the solar panel to the enable pin of the LM2596. I use it to charge a LiFePo4 cell from a smallish 12v solar panel.

  2. I would suspect that motorized tracking would also significantly increase output, and failing that some sort of calculation based on your specific location and nearby barriers to find the optimal fixed mounting angle.

    The number of panels that I see locally on random things which are nowhere near optimally aimed is frustrating.

    For example there are some solar powered park lights that the city has added in various areas, but the solar panel itself is a fixed part of the entire light assembly. So based on where you want the light aimed, that dictates where the panel faces. Because of this several of them point in the opposite direction as necessary, then only provide a few hours of light at night and people complain. Seriously, why would they manufacture them that way?

    1. Motorized trackers deliver less than a 5% increase for the cost. It’s better to spend the tracker money on more solar panels and get a greater than 5% increase in capacity.

      This is why most installs are not motorized but fixed. now the very far north where the sun rises in the north north East and sets in the north north west, then you use trackers because they extend the solar gathering day from 9 hours to 18 hours. but anything below Ohio it’s a waste of money to put in trackers.

    1. It seems that most MPPT implementations can operate in a much more “dumb” fashion than you would imagine and still be effective. IE, change how the source is being loaded, and if the charging power decreased, then change the load in the other direction.

      The algorithm effectively corrects itself and seeks the highest output……. It doesn’t actually need to predict what that output is.

  3. I’ve implemented something similar using Attiny861A and AVR Studio 4 for programming, but on far smaller scale than this. 11.55Vmp, 4.8W solar panel charging 4.5Ah 6V SLA battery. In certain situations InCond algorithm managed to squeeze almost 6W out of the solar panel, while maintaining conversion efficiency around 93%. When MPPT was not needed PID regulator kept battery voltage within 1% of desired battery voltage. Too bad I didn’t build on top of it some remote weather/sensor station, since night time current draw from battery was around 200uA.

  4. MPPT is falling out of favour on solar systems that don’t have strict limits on collector area. The relatively small gain that MPPT yields can be had by adding panels or buying larger panels. This is usually cheaper. Another issue is that the DC-DC converter in an MPPT system is expensive and short-lived compared to the panels.

    MPPT does have the nice side-effect of drastically increasing the input voltage range.

    The drastic fall in price of solar panels (thank-you China) has made many systems for maximizing solar panel performance (eg. MPPT and tracking platforms) uneconomic. But for applications where you have modest limits on collector area, these modifications can still make sense.

    1. “an MPPT system is expensive and short-lived compared to the panels.” …………
      There are numerous MPPT systems in production right now with longer lifetimes expected than the panels which they operate. This is usually achieved by the avoidance of large electrolytic capacitors in the control board which is allowed due to increased switching frequencies, interleaving of multiple parallel ‘circuits’ and other techniques…..

      MPPT of course has a value that is related to how and where the panels are installed. Sometimes MPPT is a very cost effective way of meeting a system’s needs (satellites etc), sometimes it’s not.

      The main reason that it is ‘expensive’ is due to current market anomalies in the developing solar power tech world….
      You can point to many other items out there that are dirt cheap at the consumer level and marvel that they were ever economical to manufacture this cheap….. I’m looking at you- smart phones.

      When you get down to it though, most charging systems whether they are connected to battery banks or the electrical grid will require a relatively sophisticated control system….. You rarely see the old simple zener “on-off” regulators even in small simple devices these days.

      And if conversion of the output/charging voltage is implemented at all in the device, it is usually not very hard or much more expensive to implement MPPT. In fact, it is often said that a solar charge controller which DOESN’T implement MPPT is often pointlessly underengineered.

    2. The argument for extra panels is still not clear cut.

      As mentioned, MPPT generally has higher input voltage and this gives a lot of benefit:
      – Cheaper panels as you can use very common 36V “grid tie” modules
      – Reduced wiring cost though higher voltage / lower current
      – Simple future upgrades by connecting additional panels in series.

      By comparison, some of the savings from a dumb charge controller get eaten up by the increased wire gauge required, and the matching panels are more expensive. There’s not a lot left to buy additional panels anyway, and there’s no opportunity to expand without upgrading the main wiring run too.

  5. While MPPT is one of those things that can be difficult to wrap your mind around, mine anyway. I’m certainly no expert as I understand it the I-V curve the charge controllers see is a curve, that depends on the solar irradiance. So in effect MPPT does take in account the light reaching the photovoltaic components. Then again after I click on [post comment] I may read response that reveal how wrong I am.

    1. I’m going to agree with you on everything you said. Not because you are actually correct, as I know less about all this than you do, but simply because I liked how you said it. Aside from that, it is common knowledge that if the majority of people agree with you, you are “right”. So far there are two of us, so I’d say we have a strong start.

  6. It’s an algorithm that “tracks” the optimal impedance which when presented to the source will yield the greatest power level…….

    Most MPPT implementations work based on the “perturb and observe” principle.
    The idea is that say 10 watts is coming out of a panel, and the controller tries to raise the current being drawn, and the power level drops below 10 watts. The controller now knows that it went the wrong way and tries to decrease the current being drawn thus walking the other way on the I-V curve. If the power lever goes above where it was on the last timeframe, it knows that it took a step in the right direction.

    And so forth.

    Current status -> change status and measure result -> if “good” change, keep going, otherwise reverse change -> start over……..

    This cycle can happen many thousands of times per second in a well tuned and configured system, but there are even charge controllers operating on the order of one cycle per second that are suprisingly effective given the average rate of change of solar irradiance/temperature change etc etc is very slow……

    The PLUS SIDE to all this gobbledygook is that it allows the MPPT unit to be completely agnostic as to what factors are actually effecting the system, or even what it’s outputs are controlling. This is kind of like the PID algorithm in that it can “tune” a solar panel, a wind turbine, or even guide a missile or steer a ship or what have you using the same principle.

  7. Functionally in many charge controller systems for batteries or for the grid, this all amounts to changing the voltage conversion ratio flowing through a buck and/or boost converter usually by varying the duty cycle of the switch(es)……

    This is making me want to write an article about MPPT as is applied in solar charging systems as it seems the topic is poorly understood and usually only very technically explained.

    1. And most importantly, MPPT is really only relevant when you have a load that can accept more or less power – eg a battery during their bulk charge phase, or maybe a heating element.

      If you have a fixed wattage load (such as modern switchmode electronics) then it’s likely you’ll collapse the voltage on the solar panel as it tries to pull an ever increasing current.

        1. By far, most solar panels (by watts of capacity, not by number of panels) are connected to the grid or else connected to a system with batteries. There is your odd flashing road sign or pocket calculator here and there, but for the most part, if you’re going to have a DC-DC converter, then all you have to do to make it MPPT is (usually) control the duty cycle of the switching element based on a feedback signal derived from the power produced.

          Once the rest of the DC-DC converter is in place, the additional trouble to implement the tracking is almost trivial…..

Leave a Reply to timgray1 Cancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.