AVR Freaks

ALD Energy Harvesting Module & IC Products

ALD EH301A would be marginal for the load (the PCB); half the load would be a better fit.

P.S.

Lead-acid (in earlier comments) is an inexpensive fit; nickel electro-chemistry is a fit though more expensive and reduced weight.

EH301A Advanced Linear Devices | Mouser

TLI Rechargable (Tadiran Batteries)

As someone who has built and sold solar controllers, there is a lot of mis-information and marketing hype on the interwebs around the term MPPT.   Usually that means some form of buck converter, but not always.

For simple 12v lead acid charging, PWM work very well and is only about 12-15% less efficient then using a buck converter WHEN the battery is fully discharged, and that advantage decreases as the battery charges.  Note: MPPT makes sense when your trying to maximize the ROI for your whole house, grid tie system, and that is the source of much of the confusion.  

In the OP case, he has a 12v solar panel, 21v max open circuit in full Sun, and wants to charge a 4.2v li-poly battery, do I have that right? 

The link showing the panel, also shows a charge controller, the question has come up whether that is a PWM or MPPT controller, but the OP has not said if they are using it. (sounds like they are not).

For low voltage charging, the OP needs a series diode between the panel and the battery to prevent the battery from discharging into the panel at night, most 12v panels do not have blocking diodes built in, although it may have a bypass diode, which is not the same as a blocking diode.  This component is critical.  

So if I have read the OP correctly, they have the following:   12v Solar panel = 24v/5v buck converter = 5v/4.2v linear charger = 4.2v li-poly battery.

The buck converter needs a minimum voltage/current in order to operate, and the panel alone will not guarantee that due to changing Sun shine and day/night cycles.  Also OP has indicated they only have two days of capacity in the current battery.   So I would recommend a couple of changes, OP needs more capacity, use the solar controller that came with the panel (it will also have the needed blocking diode) to a  12v SLA battery to extend capacity to 7 days or more, the rest of the circuit, buck + linear charger should be ok.   Other configurations are possible, including using a solar li-poly charge controller.

Let us know what you decide to do, glad to answer solar questions.

Jim

I would be worried about the Li-Poly charger losing its state if its input voltage goes away (bird fly over the solar panel.) Li-Po battery is not beginner-friendly, and not providing reliable power to the charger is probably a critical mistake. Having the solar panel charge a 12V lead-acid could be a workaround for giving reliable power to the Li-Po charger. Charging a lead-acid does not require knowing its history (it is more or less stateless); the charger can lose input power and then start again without degradation.

Charging Information For Lead Acid Batteries – Battery University

The Li types are very popular now, they pack a lot of power into small spaces, but have one disadvantage with solar apps in that they prefer not to be float charged, rather they are charged fully and then discharged in cycles. 

So with solar, the Sun comes up, the battery is fully charged and the charging stops (when using a proper Li charger), so now the load discharges the battery even though the Sun power is available. There are ways around that, but not using your simple power modules.

With float charging, the Sun comes up, charges the battery fully, and continues to float fully charged and run the load while the Sun is shining, only when it gets dark does the load draw from the battery.

Float charging a Li battery will shorten its life and may cause it to over charge and spontaneously give off fire works! And Li batteries are very hard to extinguish once they catch fire.  Note: some Li batteries have a built in charge/discharge controller that prevents over charging, over discharging to prevent the above disasters. But to rely on that to prevent over charging is still abusing the battery and not a good practice.

So bottom line, what you are doing may work ok for a one off, short term project, but is not best practice for a commercial project or a long term project. 

No, that was very good!  I would only add that with fixed (i.e. non-tracking) panels, you only get about 4 hrs of good charging a day, a couple of hours either side of local noon time, out of that window of time the power output from the panel drops off quickly.

LTC4070 Li-Ion/Polymer Shunt Battery Charger System | Analog Devices

...

• 1% Float Voltage Accuracy Over Full Temperature and Shunt Current Range
• 50mA Maximum Internal Shunt Current (500mA with External PFET)
• Pin Selectable Float Voltage Options: 4.0V, 4.1V, 4.2V

...

LTC4071 is 50mA max.

These lack a 0C condition.

on a cloudy day there is not much charge ...Is there a better panel to use for such weather? 

If you create one, you probably won't be on the freaks anymore--retire to your own island, large yacht, or penhouse suites. 

Do solar panels light up if they get powered backwards? Serious question...

Neil

Yes, I knew about LEDs working backwards - I've used them as cheap'n'cheerful light sensors for timing pulses. That was why I wondered about PV panels. I'd expect the reverse bias to produce photons of the same frequency as those absorbed - at least with a semiconductor bandgap doing the work - but I have an idea that it may not be the same mechanism for PV generation.

Neil

The following has a cold start that's an order of magnitude less power :

AEM10941 Solar Energy Harvesting IC - e-peas | Mouser

20mA at 1.8V to MCU; AVR DA, DB, and DD have extended SOA (1.8V, 24MHz) and approx 7mA typ for AVR DB.

What are the AVR28DA128 AVR32DA128 AVR48DA128 AVR64DA128 ?? | AVR Freaks

Photovoltaic Energy Harvesting | e-peas

edit :

EVK10941M Mini Evaluation Board - e-peas | Mouser

ETA tomorrow :

DEMPV-BLE Batteryless Photovoltaic IoT Demo Kit - e-peas | Mouser

would that better for efficiency

Maybe, maybe not...depends on your converter, probably a switcher.   Remember higher current===> higher loss, due to I2R.  Double the current, and even the loss in the wires & connectors goes up by 4x 

Your example, such losses go up by around 10X to 13X !!!  

I am hoping to plug the output of the solar pannel via diode directly into Li-Poly charger

Note that is not MPPT, but at a few watts, probably nobody cares.  If you are dealing with getting 100W or 1000W , then you care about % delivery from the cell

MPPT is based along the lines of an optimally matched load.

Imagine that a cell voltage varies as the load increases or decrease, maybe in a nonlinear way.

Imagine an extremely heavy current drawn  (short) and thus a near zero voltage from the cell    V=0, W=0

Imagine a rather heavy current drawn and thus a rather low voltage from the cell   W=V*I

Imagine a rather light current drawn and thus a rather high voltage from the cell    W=V*I

Imagine an extremely light current drawn (open) and thus the highest voltage from the cell      I=0, W=0

In each case the power drawn//delivered is volts*amps

As you can see from the above, at both extremes the delivered power is zero.  

The other conditions, of course, have higher power.

These slopes must connect by a curve of some sort.  Somewhere along this curve, at least one maximum power delivery point will occur. 

MPPT tres to dynamically find this point at all times.   

By adjusting the duty cycle, the conversion ratio is changed such that the input resides at this optimal setting to deliver the required load voltage.

Normally a fixed input supply requires an exact duty cycle to give a desired output.   Here, the input varies so the duty cycle will be "mushy" ... more than one duty cycle will give the desired output & at least one of those be the highest cell efficiency.

However, if you aren't trying to draw the max output possible from the cell, it really wouldn't matter how efficient the cell is acting (perhaps other than cell heating considerations).

If you connect them all in parallel, don't you need blocking diodes?

Hook 'em all in series for lots of volts, and a step down regulator can be pretty efficient... amps, volts, what you're collecting is electrons. Joules, if you insist.

Neil

The solar panel acts as a voltage source (somewhat) until the load pulls the current caused by photons generating carriers (in the depletion region of a P-N junction). If the load wants more current, the voltage will drop, but not much more current is provided; it is then acting more like a current source.

I have tinkered with the LT3652, did a fair number of board spins before I could figure out what my problems were. I had the MPPT input voltage set at about 16V at room temp; it would drop to a lower voltage as the solar panel heated up (datasheet is a good read). It charged a 6V lead-acid and was in absorption mode most of the time (e.g., not MPPT).

It was a good educator, but in the end, I went with a 12V battery and solar panel. The solar panel was sized, so the short circuit current was a fair match for the charging current.

After all the games, I ended up with a solar panel directly charging a battery (sized to match each other). Also, a PMOS is used to disable charging when the battery is full. An MP2467 based converter steps down the battery voltage to regulated 5V. It minimized the number of converters I needed, and matching the battery and solar also maximized the storage efficiency. I need to do this with Li-ion at some point; I probably need to start learning that stuff soon since it's growth is past the exponential nee.

interesting use of the solar panel as a very poor "light source" for their inspection

The DTU group’s new technique involves electroluminescence — the glow produced by photovoltaics in response to an applied voltage. This phenomenon can enable detection of cracks, broken interconnects, shunts, and other problems.
“Electroluminescence makes it possible to find a defect before it’s severe,” said Peter Poulsen, senior scientific officer at DTU and a co-author of the paper.

https://www.photonics.com/Articl...

you can even buy a tester from China

https://www.made-in-china.com/sh...