Designing a MPPT solar charger using the LT3652 - one question ...

[evalon]

Hello

I'm in the process of designing a solar charger based on the LT3652 (see attachment). However, I'm not sure how to polarize C3.

Linear in their datasheet (see attachment) writes on p. 9 - boost pin - that the operating range of the boost pin relative to the switch pin is 0 to 8.5 volts. I'm just not sure if it is always 0 to + 8.5 volts or it's an AC value. The LT3652's block diagram is on p. 9.

I'd appreciate your feedback.

Best regards,

Jesper

 

[erikl]

... I'm not sure how to polarize C3.

... I'm just not sure if it is always 0 to + 8.5 volts or it's an AC value.

The designated polarization of C3 is correct: during operation, the BOOST voltage is always above the SW voltage (which is necessary for low SW losses). The 0V state exists only at start time.

Even if the max. value is guaranteed to be 8.5V , for safety reason I'd use a 16V rating elco, because it has to stand some switching AC voltage. May be a 0.1..1µF ceramic cap in parallel.

 

[evalon]

Hi erikl,

Thanks for considering and answering. I now consider my question well answered ;-)

Greetings,

Jesper

 

[evalon]

Hi again,

I am now testing the circuitry based on the LT3652.

However, it performs "below par" for some reason (I have attached the schematic to this thread).

First, it doesn't start up until the voltage reaches ~15 volts even if the boost supply is designed as I can see it should be. This is less critical though, as the input supplies will from time to time reach 15 volts.

However, more critical is that the zener diode between the BAT and BOOST pins gets very hot, even if it is a ½ watt version. The 1 uF capacitor is a film capacitor (Wima MKS or MKP).

Also, when the input voltage drops below app. 12.6 volts the circuitry cease to charge which is different from what I understand should be the case from reading the datasheet (pin 2 - Vin reg - at this point is well above 2.7 volts). This is described on p. 13 of the attached datasheet, second paragraph.

For your information I have attached a picture of the actual circuitry so that you can see the layout (not the most aesthetic I have made, though) - please note that the TO-220 to the left is a LM7505 used for something else.

I cannot figure out what can be the issue so hope one of you have an idea.

Help is much appreciated ...

Greetings,

Jesper

P.S: The zener is a 3V9 and not a 3V3...
P.P.S: The diode connected to the zener is an EM513 diode - see datasheet attachment.

 

[frj]

hi
i am also planing to assemble a mppt charger, i collect some circuits, BUT the results are alarming what i find here ,,,

any update?

 

[Einar M]

Vin must exceed Vout for IC to work. Hot zener probably caused by defective part.

 

[Kozzito]

Hi there!

Just finished my LT3652 design (started at 2011.04., test started 2011.06, nice summer time, corrected major issue and get full power 2011.12.), and at first, i was scared to hell too. I used Alu-elco conds for Cin and Cout (low ESR, 3-3 paralelled), and SK54 Schottky diodes. For Vin solar protection too. When i tried to charge a battery, i got almost the same issues. Panel (4 layer with power plates) got very hot, not just the IC itself, but the Cin Cout array too. My Icharge never got above 500mA, no matter what i did. And, the charging stopped if i put some load on it, even if i got 500mA from solar, and loaded only with 300mA. Technically, always had an enormous leakage. After some months break, i just deciced to change something. I was very lucky, because my Alu-elco footprints made myself, so i was be able to solder Tantalum condensators, instead of Alu-elco. Magic happened. My Solar panel (23V Vo, 17V MMP, 0.7A Imax) did its max, at december sunlight!!. The circuit managed to float my solar panel at 17V, and charged my akku with its programmed max 1.0A! Approximately. Almost all the heat disappeared, now just the IC itself gets warm (lot colder than before), and all capacitors and diodes are cold.

So, i hope anybody reading this who needs some tips about this really marvelous IC in this low-power category. Use tantalums, they rulez

A.