MPPT buck converter ICs

Flip the cell over and do the buck/inverter trick. ;)

Cheers

Phil Hobbs

An MPPT buck converter?

'Duh' hacks?:

- Coupled windings (flyback, close enough)

- Buck-boost (who needs common ground?)

- Add on an external bootstrap gate driver (and current sense if applicable)

- Level-shift any control signals and fool it into "boost"ing its ground instead

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website:

Oh, SEPIC too, of course.

Tim

You can use a regular buck converter IC that has an enable pin and do the MPPT with say an ATTiny PWMing that pin a few hundred Hz. AFAIK Li-ions don't care if they're charged with pure DC or PWM. You might even be able to get away with high-side current sensing with no external components other than sense resistor for P+O on the 6V panels cuz the ATTiny series can run up to 5.5 and the differential ADC channel pins will work up to around 0.4 volts above the supply rail.

Yes, excellent suggestion, and it should work well. Experimenting with six MPPT boost-converter IC eval kits, I learned that most have less than ideal MPPT algorithms, so programming one's own would be good.

But hey, I'm charging a 4V cell from a 6-volt solar panel, and am eager to zoom past this aspect of my project, using a standard charge-controller IC. But most are linear-regulating with 7-volt max ratings, and I worry somebody might plug in a 12V panel, etc. To handle this, and maybe get a bit more efficiency, I'm adding a simple buck converter with a say 4.5V output setting. The question becomes, how much more work is warranted to get an extra 20% or so? My 2W panel has a 4x safety margin to handle cloudy days, work hard to get 5x? OTOH, if some blokes then do use 12V panels, the improvement could be much more.

BTW, my project is a bee-counter and hive-monitoring system, which doesn't need to work in the winter.

Looks like I showed up at just the right time. Concerning the MPPT algorit hm -> I know there are many ways to skin this cat. Why wouldn't looking at just the current into battery be enough to control PWM? If you change pwm and the current into battery goes up then you are good. No?

Regards, Bob

I just realized this is wrong. The charge controller has a constant charge current, which is programmable by resistor-value, but it won't take advantage of the maximum output of the solar panel at any given time.

this cat. Why wouldn't looking at just the current into

Yes good, increase converter PWM, get higher current. Now increase it some more, panel voltage sags dramatically, get lower current. Oops, that's not an ordinary control loop.

What are you talking about? There are tons of ic's out there, and they all use buck-boost and MPPT to accommodate solar PV power sources. Some of them have a fairly huge input voltage range.

I think one oft-used method is to add dither so the loop is forever hunting around the optimum point.

piglet

MPPT controllers always 'hunt', you can look at the original patent. I think its was a Exxon patent.

Cheers

Some panel users charge a very large cap. When the voltage exceeds some limit, turn on the buck or boost converter until the voltage drops to another limit. Repeat as needed.

That's what I've decided to do, it's very simple. The cap doesn't have to be too big, for a fast converter with a high clock rate. I found a buck converter with a precision enable input with hysteresis, TPS54302. Add just two resistors to prevent pulling the panel too far down at low light levels. At high light levels I'll be throwing away extra juice, if a small battery is fitted. The Li-ion charging manager is an MCP73831, in a Feather ARM Cortex M0+ controller board.

There are dozens of algorithms out there, the later ones are based on clever observations of the of the IV behavior.

How about just using a constant-voltage-output synchronous switcher?

Yes, you're right, and I have dozens of candidate buck-boost MPPT datasheets stored in my computer. But they're mostly rather complicated and expensive. For example, your LT8490 part costs $12 and comes in a 64-lead QFN package. Sheesh!

For my inexpensive project with a dinky 2W panel and a 500mAh cell, I'd like to see under a buck in a sot-23 package. :-) I imagined there'd be some simple low-cost buck-only MPPTs out there. E.g., for a 6V solar panel with a single Li-ion cell.

Yes, indeed. And bitrex suggestion to use a micro- controller is right on target to roll-your-own MPPT. My beehive project uses an ARM Cortex M0+ controller, cost 40-cents qty 100 (!), but I hesitate to add to its complexity by diving into one more can of worms. Spring/summer is fast approaching. Bee time!!

Interesting info at