I have made a solar panel charger, it consists of a 1.2 watt panel connected to a buck converter, then to the phone.

Now, when I connect it, it charges the phone and in the battery settings it shows it is charging as "plugged ac" and slow charging.

I wanted to test if it can slow down the discharge rate, so I came up with a test procedure where I:

  • Downloaded "battery drainer" app (which essentially turns on cpu stress, gps, vibrator, full brightness etc);
  • Measure the time it take for the phone to get to 90% from 100% with solar panel attached and without solar panel attached.

But the time it takes to discharge is almost the same. I am wondering what am I missing. Wouldn't the solar charging increase the discharge time?

When I measured the solar panel output it was 5 V and around 110 mA.

Thank you, any help will be appreciated.


closed as unclear what you're asking by PeterJ, Daniel Grillo, Adam Davis, Matt Young, Adam Haun Feb 23 '16 at 21:46

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 2
    \$\begingroup\$ Charging a LiPo battery is really inefficient in general; having a 100 mA 5V supply should allow the phone to sip a bit of power to either run off of or charge, but this largely depends on the implementation of the battery charging circuitry. You said "almost the same" time -- but how close? How is the battery charger's fuel gauge determining the capacity left? Under high load, is it ignoring the 100 mA supply altogether and running straight from battery? Most of these questions probably can't be answered without the schematics for your smartphone. \$\endgroup\$ – Jay Carlson Feb 23 '16 at 5:09
  • \$\begingroup\$ For example, it took 20.12 minutes to discharge 10% with solar panel attached and 20.25 minutes without solar panel. What do you mean by schematics of smartphone? \$\endgroup\$ – Tanvir Feb 23 '16 at 5:14
  • \$\begingroup\$ You're really asking about why the phone's internal charging and power usage isn't performing as you expected, which isn't an appropriate question for this site. If you can remove the phone from your design altogether and present the question as charging a regular battery or load, then we may be able to help you. As it is, though, this is a question about a consumer device. \$\endgroup\$ – Adam Davis Feb 23 '16 at 13:06

Your Experimental Setup:

  1. You listed your chargind data as 0.11A@5V, or 0.55W
  2. Your stated test protocol uses many/all power consuming parts in your phone at high output (in many phones this can reach well over 10W total load)
  3. You then timed fall time from '100% charged' to '90% charged,' as displayed on your phone's screen to gather test results.

Anticipated Problems for your Experimemt:

  1. Charging at ~0.55W while draining at >10W is a very small 'drop in the bucket' charge rate to be testing, so your error % will be raised and your accuracy & precision of timing & charge state measurement become incredibly demanding.
  2. Using the on-screen charge-status indication adds 2 major sources of accuracy/precision loss to your results:
    • The OSD charge status doesn't update with a very high refresh rate, due to the normal timing of charge/discharge the refresh rate of this indicator could be several seconds long
    • The OSD charge status indicator on the vast majority of phones relies on a 'fuel guage' circuit to guess about the battery's state-of-charge. This is usually accomplished by measuring total input/output current, then compared against past 'learning' measurements to 'fill in' data which is not directly measurable in the small increments displayed. Due to this, an error factor lf at least +/-5% should be considered highly possible for the OSD state-of-charge.

Recommended improvements:

  1. Find a lower-drain (but still as steady/predictable as possible) method for your test discharge rate to increase your SNR and allow easjer gathering of meaningful results (maybe open a GPS nav app, with location enabled, power save disabled, and screen brightness as 25%)
  2. Run your test for long enough that the 'gas guage' ic can give more meaningful output by passing a few of its 'threshold voltages' which it can actually measure, rather than only estimating (I'd recommend staring testing as soon as 'your phone is done charging' is registered on-screen, then run until <=25% charge remaining)
  3. To be a scientific test, the results must be repeatable, so running this test 3-4 times (both with and without the PV charger attached) can help "weed out" inaccurate results that can be caused by variables (phone ringing, GPS signal fluctuations, clouds, etc.) you can't easily control.

Make sure your testing conditions are the same, if you measured the panel outside in direct sunlight, make sure you test it in direct sunlight.

If you really want to answer this question, it will take some experimentation on your part. Get a power supply with a current limiting option. Hook it up to your phone. Then set it to 5V and set it to limit at 100mA then observe the rate of charging. Then set it to 150mA and observe the rate of charging and so on. If this isn't consistent with what your getting out of your charger then you might have a problem.

A DC to DC is not the best way to load a solar panel, it could potentially load the solar panel that it wouldn't produce much power at all. If your test conditions were a resistor on the DC to DC, this is going to be much different from your phone which has a variable load. You need a maximum peak power tracker to ensure your getting the optimal power from your solar panel. The MPPT will find the optimal point for getting the most power from the panel. See the picture from solarhome.ru.

Another issue is the inefficiencies of the system. A DC to DC could be 70% to 95% efficient. And then the phone is going to take its cut, and that could be 70% to 90% efficient. You measured 5V and 0.1A that is half of a watt. So you start out with 0.5W and multiply that by 70% as a worst case efficiency of your phone, that's 0.35W and isn't much to work with. Even at 0.5W you are still at 10% of what a normal phone takes which is 1A and 5W.

check out the picture from solarhome.ru

  • \$\begingroup\$ Do you think it will help if I connect two 10F 2.5V super capacitor and a diode to the panel so it will give a somewhat consistent input voltage and current? \$\endgroup\$ – Tanvir Feb 23 '16 at 6:19
  • \$\begingroup\$ No, that would not help. The panel is a variable voltage source, it has the highest efficiency when you are drawing just the right amount of current out of it. They have IC's (and sparkfun has the modules on a dev board) that do this. You should at minimum look at the wiki page for maximum peak power tracker. Another option is to come up with the curve on your own and see for yourself by measuring the voltage and current simultaneously and using a variable resistor to change the load, then you could calculate the efficency of the DC to DC converter. \$\endgroup\$ – Voltage Spike Feb 23 '16 at 6:28
  • \$\begingroup\$ Thanks for the reply. I actually bought a 1-4 V boost converter from sparkfun but will you be able to provide a link of the dev board from sparkfun for this setup that you talked about? I did study about mppt but dont know how to include that or where to get one to attach it to the panel or load side since my panels are only 1.2 watt small ones \$\endgroup\$ – Tanvir Feb 23 '16 at 6:35
  • \$\begingroup\$ I have looked at the sparkfun mppt solar charger but its primarily to charge a lithium ion battery. I know there is a load option there. But i have adafruits solar charger which is the same and I connected a usb to the load side of that battery charger, but the test result is same. I have this mppt solar charger from adafruit : adafruit.com/product/390 \$\endgroup\$ – Tanvir Feb 23 '16 at 6:45
  • \$\begingroup\$ Then you need more power... \$\endgroup\$ – Voltage Spike Feb 23 '16 at 6:46

Not the answer you're looking for? Browse other questions tagged or ask your own question.