LT3652 Solar Cell MPPT

Maximum power point tracking (MPPT) is a technique that solar battery chargers and similar devices use to get the maximum possible power from one or more solar panels (several solar cells connected in series and parallel). Solar cells have a complex relationship between solar irradiation, temperature and total resistance that produces a non-linear output efficiency known as the I-V curve. It is the purpose of the MPPT system to sample the output of the cells and apply the proper load to obtain maximum power for any given environmental conditions.

The following solar battery charger circuit is based on a LT3652 from Linear Technology. It’s simple design just based on the data sheet and application notes. Another advantage apart from simplicity is the fact the board using the LT3652 can be stacked in parallel and therefore provide more output power if the solar panel can handle it.

My circuit is designed to charge a LiPo battery in 3S configuration with a current of max 2A out of a 36 series cells solar panel providing a nominal voltage of ~18V. The output provides a constant current or constant voltage to the battery depending on the charging state.

A plot showing the input and output current as well as input power under ideal conditions:

PCB is double sided with only one side populated while the other one serves as heat sink. A exposed pad on the bottom side (just below the LT3652) can be used to attach the board to a separate heat sink.

Two boards stacked.

Visible are the connections between input and output and in the middle two extra connections for input reference voltage and output voltage feedback. Basically one board provides these two voltages for all others stacked to this one. In other words a master board driving several slave boards.

My design uses a P-channel MOSFET instead of a Schottky diode to protect the solar panel from reverse currents. This trick minimizes the voltage drop on the protection circuit to <10mV compared to a diode with roughly 400-600mV. Efficiency of the circuit increases this way.

Two LEDs indicating the status of the charger board. A green LED indicating the charging status when the board is suppling more than 1/10 of the programmed charging current. The second red LED indicates any faults which can be over temperature, internal charging timer overflow or battery fault.

Useful design documents:

LT3652 Development Board

LTJournal-V20N4-02-df-LT3652-Jay_Celani

LT3652 Datasheet

LTMag-V19N4-04-LT3652-JimDrew

BOM LT3652 MPPT 14.2V
BOM LT3652 MPPT 14.2V
BOM_LT3652_MPPT_14.2V.txt
942.0 B
960 Downloads
Details
BOM LT3652 MPPT 7.2V
BOM LT3652 MPPT 7.2V
BOM_LT3652_MPPT_7.2V.txt
945.0 B
642 Downloads
Details
LT3652 MPPT BOM
LT3652 MPPT BOM
LT3652_MPPT_BOM.txt
1.4 KB
1420 Downloads
Details
LT3652 MPPT Schematic
LT3652 MPPT Schematic
LT3652_MPPT_Schematic.pdf
69.1 KB
4291 Downloads
Details
Schematic LT3652 MPPT Iinput 15V Output 7.2V
Schematic LT3652 MPPT Iinput 15V Output 7.2V
LT3652_MPPT_INPUT_15V_OUTPUT_7.2V.pdf
12.2 KB
1000 Downloads
Details
Schematic LT3652 MPPT Input 18V Output 14.2V
Schematic LT3652 MPPT Input 18V Output 14.2V
LT3652_MPPT_INPUT_18V_OUTPUT_14.2V.pdf
12.3 KB
1376 Downloads
Details

57 thoughts on “LT3652 Solar Cell MPPT

  1. Hi! I’ve been looking into MPPT and solar power for a little while now, and I have a LT3652 in the mail which I’m going to start prototyping with. I was wondering what you meant when you say that you’ve replaced the Schottky diode with a MOSFET. Which is to say how did you hook up the MOSFET in your circuit?
    Near as I can figure you have the source on the solar panel side, the drain towards the rest of the circut, and the gate shorted to the source.
    How much reverse current protection does this provide? And as near as you can figure why would the Linear guys pick a diode instead of a transistor?

    Thanks so much for any help that you can provide!

    PS What are the chances the schematic and layout might be posted for your board?

  2. I am very interested in this device, I would like to buy one to put in my FPV plane with two solar panels.
    I would like to know the price and when you send it.

    I await your reply.

    Best regards.

  3. Hello Michael,
    looks good to me, great job.
    Can you show us please a top-, bottom-layer and parts-layer.pdf or GIF-file or whatever for PCB etching.
    I saw on the picture you made several connections between when you stack the pcb’s.

    Thank you for this MPPT-device, really appreciated it.

    Stephan

  4. In the BOM list, do you have a resitor of 107K, but in the schematic this resistor is not connected. Best regards!

  5. This looks amazing!

    I wonder though, how would this be adapted to charge a 4S (14.8v) lipo pack, instead of 3S(11.1v)? I assume the panel voltage would have to be raised, but what else would change? Can this IC even handle it?

    • R1+R2 needs to be re-calculated for the MPP of the new panel voltage. R7+R8+R9 needs to be re-calculated for new battery voltage.

      Check the design documents above how to perform these calculations.

    • I assume you mean IC2, yes chosen because of low RDSon, therefore voltage drop, compared to a Schottky diode.

  6. Hi Michael,

    I’m currently in the process of laying out a board with the lt3652 aswell. I was wondering what your considerations were in selecting such a high value input cap?

    Cheers!

    Gijs

    • To cover the wide input voltage range of the LT3652, up to 40V max.
      So the circuit can be used for other input/output voltages as well.

      • Hi,

        Thanks for the quick reply! Did you also choose the 470uF value to counteract large voltage changes when connected to a solar panel? Also, one last question if I may: Would replacing R1 in your schematic with say a 600K trim pot, to adjust the voltage regulation, be wise?

        Thanks again!

        • The circuit is based mainly on that shown at page 14 in the above document LTJournal-V20N4-02-df-LT3652-Jay_Celani.

          R1 can be replaced by a series combination of a resistor and poti, let’s say 470k resistor with 220k poti in series. This way the input voltage can be set in a wide range.
          Otherwise you have to calculate the combination of R1/R2 using the formulas given at page 14 of LTMag-V19N4-04-LT3652-JimDrew.

  7. Hi Michael,
    I am working on a solar powered weather station project using a mobile phone to send the data back to a server. I had found you Solar Cell MPPT page after finding the LT3652 and concluding it would work really well to charge the phone battery.

    I was thinking of buying the LT demo board until I was the price. Your project looks really good and a better design than the demo board. Are you able to provide the artwork for the layout or blank boards to populate. I am just not sure I want to go through the layout development for a one off project.

    Thanks
    Hugh

  8. Hi Michael,
    danke für Deine soziale Ader 🙂 Sehr gut, ich habe von Dir den Trick mit dem P-Ch.-MOSFET gelernt. Danke dafür.

    Mit diesem LT-Chip bin ich auch am ringen, mein erster Prototyp funzt auch gut in seinen Grundfunktionen, sowohl für 1S LiPo als auch für eine Bleigel mit 12 Volt. Nur wird der LT sehr schnell sehr warm und regelt sich dann intern ab, bis er nicht wärmer als 92 °C wird. Diese Temp. hat er in meinem Layout nach ca. 60 bis 70 Sek. erreicht, das grob geschätzt nicht weniger Kupferfläche zur Kühlung bietet als Deines. Wie sind denn Deine Beobachtungen bzw. Messungen zum Temp.-Verlauf Deiner Schaltung?
    Leider habe ich auf beiden Seiten Bauteile verteilt, so daß ich mir die Chance auf Montage eines Kühlkörpers verbaut habe. Mein Proto2, der gerade in der Mache ist, bekommt noch viel mehr Kupferfläche und der LT ist alleine auf einer Seite, so daß ich ihm auch zur Not was auf den Kopf kleben könnte …

    VG, Jan.

    • Ich habe ohne zusätzliche Kühlung im Dauertest mit 12.6V/2A ca. 60° am Chip gemessen. Du solltest unbedingt die Empfehlungen im Datenblatt bezüglich Layout und Wärmeverteilung beachten. Ich habe unter dem chip 3 Durchkontaktierungen mit 1mm Durchmesser, über diese kann ich die Bodenfläche des Chips verlöten. Falls du Platzprobleme hast eignet sich ein ausreichend langer, gebogener 1,5-2mm² Kupferdraht als Kühlkörper, der Rückseitig auf dem Layout direkt unter dem Chip verlötet ist. So kannst du Wärme punktgenau abführen ohne einen sperrigen Kühler zu benutzen. Quasi eine Heatpipe wie in aktuellen PCs.

  9. may i know, it is true if we use lipo battery such 11.1v, we need to balance it after charging??? Im trying to use 3 unit 7.4V 1000mah in parallel to get 7.4v 3000mah.

    • Depends on the situation but usually balancing is recommended. For 7.4V you need to calculate new values for resistors R7, R8 and R9. Also depending on your solar cells R1 and R2. See LT3652 datasheet for details and the two application notes provided.

  10. Hi Michael,

    Very interesting article! Just what I was looking for.

    I saw you don’t sell the circuit. But how and where do you get the PCB? Is there any chance to get one from your side?

    I have 2 projects in mind -> 3S and also 2S with nominal 7.2v solar panels. For 2S I wonder if the input voltage would be good enough for the LT3652 (range 11.5 to 32v)? I can use the panels in series to reach 14.4v, but with half amps 🙁

    What do you think?

    Thank you
    Patrick

    PS. The other articles are also awesome

    • PCBs are not available.

      Running panels in series with a higher voltage than required is way more efficient than at minimum voltage.

  11. Hi Michael

    Thank you for your reply and advise. In this case where did you make your PCB.
    I think you go through a manufacturer. Do you have an address?

    Thank you again
    Patrick

  12. Hi! I read the datasheet of LT3652, it says the Vin should be at least 3.3 V higher than the Vbat. If my battery is a 2s Lipo battery, 7.4V, does it means my solar panal should at least provide 10.7V?
    Thank you very much

    • That’s correct.
      See the plot above. The LT3652 in my circuit was set to 12.6V output. In th plot you see that the input voltage needs to be just above 16V to start the output current. 12.6V + 3.3V = 15.9V

      • Thank you very much. I have one more question, what is the D5 diode (BZV 55C 4.7V) used for?

          • Hi Michael, one more question, what do the R10 and R11 do? how do you define their resistance? why do you connect them with the CHRG pin?

          • R10 and R11 are limiting current through the CHRG pin. Values are not critical and defined to keep the CHRG pin current <10mA max.
            CHRG will pull the gate of transistor IC2 low, and therefore driving IC2 with low impedance, as long as the charging current is >1/10 of the programmed maximum current.
            See also page 13 of the document “LTJournal-V20N4-02-df-LT3652-Jay_Celani” linked above.

    • I never released my board data to public, so that’s not my design, hence you need to ask the designer about JP1 and JP2. There are no jumpers in my schematic.

    • Hi Michael,

      Thank you very much for your work!

      3 boards itched by OSH park (min order) were only $12;
      2 sets of electronic components from DigiKey was worth of $48;
      Butan gas soldering heat gun (a big jet lighter) was $7 ebay;
      Soldering paste with tin-lead-flux was only $5;
      Total for two solar chargers: $72

      40 solar cells + tabbing + flux pen = $24
      $100 for 2 controllers and some extra cells in case if they break during soldering/landings – NOT BAD !!!

      I will solder all together soon and I hope I’ll fly to 5km clouds, thanx to your solar controller! :))

      I’m thinking of using this resistors/capacitors assortment for repairs:
      http://www.ebay.com/sch/i.html?_odkw=SMD+Value+Resistor+Capacitor&LH_BIN=1&_osacat=0&_from=R40&_trksid=p2045573.m570.l1313.TR0.TRC0.H0.XSMD+Value+Resistor+Capacitor+kit&_nkw=SMD+Value+Resistor+Capacitor+kit&_sacat=0

      It doesn’t contain some of the resistors in your BOM lists (549K, 825K, 14.3k, 442K, 390K, 71.5K) and most of capacitors, but do you think your schematics might use only resistors/capacitors from this assortment? Just to spend $9 on ebay instead of $40 on digikey for res/caps 🙂 I didn’t check LEDs/inductor/diodes but I’m sure ebay sells them cheaper than Digikey :))

      Thanx once again for sharing!

      Pavel

  13. Hi Michael!

    Are you sure that the values for R7 and R9 in the 7.2V schematic are correct? shouldn’t they be calculated for a BATT voltage of 8.4?

    Nice project by the way! 😉

    Andre

  14. Hey Michael – nice work. Have you considered incorporating a balancer circuit into this design? I’m considering either rolling my own based on something like an attiny or incorporating this design: http://www.flyelectric.ukgateway.net/lithium-balancer.htm

    I think the CHRG pin can handle the current needed to activate either of those circuits (DS says it’s good to source 10mA).

  15. Hallo Michael ,

    ich habe eine Frage zur berechnung des Dreier Widerstandsnetzerk ( Seite15 Datasheet)
    Ich möchte 4,2V Vbat(Flt) haben.
    Folgendes konnte ich ermitteln , aber die Formel für Rfb3 verstehe ich nicht , können Sie mir da helfen evt. auch kurz erklären , vielen Dank schon mal.
    Rfb2= 330K
    Rfb1= 90K
    Rfb3= ??

    Gruß Andi

    • Die Berechnung geht wie folgt:

      Rfb2 = 330K
      Vbattfloat = 4,15V (etwas weniger um Bauteiltoleranzen zu kompensieren, minimiert Risiko von Überladung)
      Rfb2/Rfb1 = 3,3/(Vbattfloat-3,3) = 3,3/(4,15-3,3) = 3,88
      Rfb1 = Rfb2/ (Rfb2/Rfb1) = 330K / 3,88 = 85K
      Rfb3 = 250-(1/((1/Rfb1)+(1/Rfb2))) = 250-(1/((1/85K)+(1/330K))) = 182,4K

      Wenn die Ladung der Zelle als Backup für irgendwas dient, und immer geladen wird, dann solltest du die maximale Ladespannung auf 70-80% reduzieren, das fördert signifikant die Lebensdauer des LiPos.

  16. Hi,

    I see in you schematic that you left out the reverse current blocking Schottky diode. That means that with a charging current 200mA. The body diode in the FET you use drops more voltage than a Schottky diode, so if you want things to be efficient for low charging currents also, you may want to put the Schottky back in (in parallel with the FET).

    Regards,

    Oscar.

  17. Michael,

    I’m looking to use the LT3652, and I have a 4S configuration of 4.2V Li-ion batteries, which in all are higher than the 14.4V max of the IC. You mention at the beginning that these are easy to parallel – was wondering if there is any particular additional circuitry required to allow this?

    I’d like to take the solar cell input and feed two LT3652s, each hooked up to a 2S battery configuration. I need the batteries in 4S (instead of 2S2P) to feed an inverter (easier to bump 16.8V to 110V than 8.4V). Is it as simple as putting two of these boards in parallel (with adjustments for battery specs), or is some form of regulator needed to make sure the incoming power is evenly distributed between the ICs? Apologies for the simple question, I am pretty new to circuit design.

    • For your idea there will be a problem. You need to feed two LT3652 boards, each by its own solar cells, to prevent a common ground. Keep in mind when you will charge a 4S total in separate 2S then for one charger board the “ground” will be the others V+ from the first 2S pack. So if you short the ground of both charger boards, either at input or output, you will short one 2S pack! Bad idea…

  18. Thank you for the information! So in this case, assuming I only have the one solar cell to supply power, my only real solution would be to configure the batteries in a 2S2P configuration – putting my charge voltage at 8.4V, which is within the specs of the LT3652?

    • Yes thats correct and works.
      But personally I would go with the LT8490. That IC is much more flexible in charger configuration and way more efficient than the LT3652.

      • I didn’t see your response until just now – and I am currently doing that exact thing! Not sure how I missed the 8490 in my previous passes through Linear’s catalogue…it looks like an amazing IC.

  19. Michael,

    Can you verify R7/8/9 from your original 3S work.

    The FB pin wants an equivalent resistance of 250k. In your other designs the FBr3 is adjusted correctly, but in your 3S design, that 28.7k resistor is far short of making up the 280k/100k divider. Unless my math is wrong, the 280k/100k is 73.6k equiv and would need a 176k resistor for FBr3 to make up the 250k equivalent needed?

    Nice work, I went back through all the documentation and verified all the ones that use 3 resistors there and in all cases they were effectively 250k, but you 3S isn’t?

    Can you explain?
    Thanks
    Alan

    • I checked components of my two remaining 3S boards. Both are identical with 280k/100k/28.7k and output is ~12.4V floating.
      I know it’s against the math and might be not equal the required 250k but this is what’s working for me.

    • I replaced the 28.7k resistor with 178k. The results is the same. The max battery voltage settles at 12.5V. That means Rfb3 has not much influence on the charging voltage.

      So to please the design requirements I will change the schematic and BOM for Rfb3 = 178k (easier to get than 176k).

      Thanks anyway for pointing this out.

  20. Hello Michael ,

    Thank you so much for all information sharing !
    I have a solar charger project for bicycle trip. With 5 W solar panel (Vmpp = 15.4 V), I would like to charge 2 Li-ion batteries (but not in the same time) : a camcorder battery ( 3.7 V) and a battery to the digital camera ( 8.4 V – 700 mAh). The range of temperature will be from -10°C to +40 °C. I would make my own solar charger but I am learning in electronics. I read the datasheet of the LT3652 and LT8490 and several articles from linear technology (LT journal). I understand that LT8490 can do lot of things (MPPT, temperature compensation for panel and battery…)and it is very efficient, but it is more complicated.

    Can you tell me the differences between these 2 devices ? According to you, which is the best device for me : LT3652 or LT8490 ?

    Thanks in advance.

    Best regards
    Pascal

  21. Hello.
    Thanks for sharing all of this information.

    In your schematic, you have an output for a battery and for a load. Both will powered at the voltage you specify in Bat pin, right?

    Can you comment if LT3652 works this way:
    Given a constant solar insulation level, say the panel is providing 1A, can you power the load through the panel and if in case the current drops, will the battery kick in and power the load?

    Regards

    • The load must be connected directly to the battery like in the schematic. The solar panel provides power through the charger to the battery and supplies the load and charges battery when current is sufficient. In case load is higher than what the charger provides (2A) then it shares load with the battery, hence the battery will be drained. As long as the load current is less than what the charger supplies your load will be powered from the solar panel only and the battery will be charged too.

      • This sounds similar to what I’m looking for – I made an ebike using 10 x 6S 5Ah lipos, connected as 5P x 2 to give ~48V and want to be able to keep it float charged in use using 2 x 120W panels at around 4.13 V per cell with balancing. It would probably need one for each 5P pack.
        The batteries 10C so with 5 in parallel can supply 250A, though the motor only takes about 30A max, and can regenerate at 50A.
        The batteries are rated at 2C max charge, so charge current could be anything below 50A. I guess charging between 10A – 25A would be good for longer life.

        Alternatively, it might be easier to use MPPT to just supply whatever the panels can produce without any charge controlling, and use any DC between 10-28V to power the charger (Turnigy Reaktor) which already has the electronics for DC-DC boost and balancing, though it doesn’t float charge as far as I can tell. Once it reaches the desired level then it stops if a time limit wasn’t exceeded.

        There are painfully few reasonable options on the net for solar charging decent sized lipos, so it looks like we have to make our own until the market catches up, and I’m sure it is going to be big.

        Thanks for sharing!

Leave a Reply

Your email address will not be published. Required fields are marked *

Comments will be moderated! Spam deleted immediately!
Before you submit form:
Human test by Not Captcha