Letter Re: Low-Budget Photovotaically Charged Drill Batteries

Permalink | Print

Dear Mr Rawles,

This letter is in response to the piece on using photovoltaics to charge batteries by C.K.. I feel I should spend some time discussing some potential problems with charging Lithium Ion ("Li-ion") batteries the way C.K. proposes. However first a few general notes.

- I am all for scavenging if you know what you are doing like some of us. But 'newbies' are better off not trying to disassemble anything more complicated than a desktop power supply for safety reasons as JWR pointed
out. I should add that the process that causes the 'bounce' after discharge also keeps some capacitors charged much longer than you would imagine possible.

Secondly if you only need some diodes or resistors, you can buy a batch of 100 for a few dollars on-line which will give you a lifetime supply. Just look for vendors with a 99%+ positive rating and 10,000+ feedback. If you don't like eBay et al., try Aussie firm Futurlec. Your spare parts will come in handy someday when we realize we cannot afford to throw everything away.

- I have no problems with C.K.'s instructions and wouldn't have bothered to write in if he had used NiCd or NiMH batteries because they can handle
some degree of overcharging, which is bound to happen in his setup if you are shooting for full capacity. For NiCd and NiMH the rule of thumb is that if they start to feel warmer than ambient temperature, they are either full or you are charging them at too high a rate. So even if you have no voltmeter handy, you can have a pretty good idea what is going on. And if you charge them in a solar charger, make sure the batteries themselves (and their housing) are shaded because heat from any source degrades battery life expectancy rather quickly.

- Li-ion cells, on the other hand, cannot cope with overcharging without some form of problems popping up. A 3.7V cell should not be charged to a voltage higher than 4.2V. Prolonged charging above 4.3V destabilizes the cell and causes CO2 to be formed inside it. The charging current will be automatically cut off if the cell's internal pressure rises to 200 psi. However this doesn't immediately stop the chemical reactions and if pressure rises to 500 psi the cell starts venting the CO2. Depending on the exact circumstances, thermal runaway may occur and the cell can burst into flames.

- Charging current isn't a big issue with Li-ion as it can absorb large initial currents and when the cell reaches capacity it will throttle that current anyway. Of course for charging with solar panels this is a problem because it will cause their output voltage to rise which is exactly what we don't need.

- Li-ion cells cannot be trickle charged; they must be disconnected once full because constant charging causes metallic lithium plating which can compromise the safety of the cell.

- 3.7V Li-ion cells need to be charged for 3 hours at 4.2V to reach maximum capacity. Shorter charge times and/or lower voltages lead to reduced charges (= shorter runtime). Consumer products chargers usually are programmed for maximum runtime, but if you can live with shorter runtime between charges, its better to charge at a lower voltage which will give you a longer useful battery life.

- The 3 hour time frame is predicated on the fact that your charger can deliver 0.8C to 1C of current for the first stage of the charging process (i.e. until the cell reaches 4.1V). If your cell is rated at 2000mAh (=2Ah), then 1C = 2 Amps. For large battery packs that means a lot of amps. Lower maximum current is actually beneficial for the cell but requires a longer charge time.

- Taking all of the above together, I would say that charging Li-ion without a good deal of checking voltages can be rather tricky. This is exacerbated by the large variance in the output of a solar cell related to its angle to the sun, cloud cover, etc. Please keep in mind that just because charging a cell works fine once or twice doesn't mean its safe. A lot of the damage done by overcharging Li-ion batteries is cumulative because the chemical processes involved are irreversible. That is, your battery may kill itself (or worse) after ten trips to your solar charger.

- If you are a new or wannabe tinkerer, I would say make small panels that can deliver around 10V-12V (open circuit) and 1 Amp. If you connect that to a 7.4V battery pack (= multiple cells), its unlikely you will seriously overcharge the pack unless you leave it out in the full sun for several days. If you want to use large panels: do yourself a favor and buy a commercial 12V Li-ion charger.

- For dyed-in-the-wool tinkerers there is yet another solution: you can build yourself a small charge controller that drives the charge current to near zero as it approaches a preset voltage. Its parts list contains 6 items and its fits on a square inch if you are really pressed for real estate. What you need is: - 3 metal film resistors (1% tolerance - 2x 10K and 1x 3.9K) - 1 zener diode (6.2V - any wattage is fine - other voltages work too but require different resistor values) - 1 op-amp (rail to rail switching - I use a CA3140) - 1 solid state switch (power MOSFET or transistor with high amp rating - I prefer IRL7833)

The circuit works very simple: the zener diode creates a reference voltage for the op-amp. The op-amp compares the battery voltage to this reference voltage. If the battery voltage is lower it closes the switch and if the battery voltage is higher it opens the switch. One of the 10K resistors limits the current through the zener diode and the other 2 resistors form a voltage divider that maps the battery voltage to the reference voltage range. To calculate the proper resistor sizes for the voltage divider use the following formula (this formula only works if your base resistor is tied to ground): resistor size = target voltage / reference voltage * base resistor size - base resistor size

A 6.2V zener diode gives a 6.1V reference voltage when fed through a 10K resistor.
Targeting 8.4V (2x4.2V) while using a 10K base resistor gives us: resistor size = 8.4/6.1*10K-10K = 3.77K.
I would use a 3.9K resistor here because wires and solder joints have small resistances too so the voltage measured at the battery tends to be .1V - .2V below the charge controller's calculated target voltage and you quickly lose a lot of capacity if you charge Li-ion at voltages below 4.1V. The narrow band of target voltages (4.1V-4.2V) is also the reason to use metal film resistors. Carbon type resistors can have tolerances between 5% and 20%. Putting those numbers in the above formula quickly points out its a waste of time building the charge controller with those.

Connections:
- The circuit uses a common ground for batteries, solar panel and other components; so all ground references must be tied together with the negative leads of the solar cells and batteries.
- IRL7833 pin 1 (left most pin if front facing) connects to op-amp pin 6
- IRL7833 pin 2 connects to solar panel positive lead
- IRL7833 pin 3 connects to battery positive lead
- op-amp pin 1 not connected
- op-amp pin 2 connects to voltage divider center position (between
resistors)
- op-amp pin 3 connects to positive side zener diode (where the band is)
- op-amp pin 4 connects to ground
- op-amp pin 5 not connected
- op-amp pin 6 connects to IRL7833 pin 1
- op-amp pin 7 connects to IRL7833 pin 2 / solar panel positive lead
- op-amp pin 8 not connected
- zener diode positive side (band) and op-amp pin 2 connect to battery
positive lead through a 10K resistor
- zener diode negative side connects to ground
- voltage divider = battery positive lead -> 3.9K resistor -> 10K resistor -> ground

Heat sinks:
For very low currents (< .5A) your solid state switch doesn't need a heat sink. For currents up to 2 amps a small heat sink will do (think soup can lid). Beyond that you should look into using an aluminum heat sink. If you really want to go overboard (the IRL7833 handles 250A): seal your circuit in a peanut butter jar full of vegetable oil and submerge it in a brook - you now have a near infinite heat sink.

This controller's output is not an ideal match for Li-ion batteries but comes close enough to the requirements that you can leave it out in the sun all day without endangering your batteries. Though in sunny weather I would think 4-5 hours charging time is plenty if your solar panel is adequately sized. Most likely you will notice the batteries charging somewhat slower during stage one and converging close to the ideal curve during the saturation stage of the process.

With a volt meter it may look like this controller acts as a variable resistor but it doesn't. Connecting it to an oscilloscope shows it to be a real pulse charger (your batteries will thank you for this!) with variable duty cycle and operating frequency. A 12V version of the controller connected to an old motorcycle battery ran at around 300 kHz while topping up the battery. Its duty cycle was mostly determined by the amount of power absorbed by the battery at any given time.

- For advanced tinkerers: you can replace the op-amp with a micro-controller, omit the zener diode and add a circuit to deliver the proper voltage for the micro-controller. Read the battery voltage through the voltage divider. Again the use of metal film resistors is crucial here. The charging algorithm for Li-ion is very simple and straightforward to program but you may have already realized that from reading the points above.

And finally if you want lots of info on all kinds of batteries: spend some time at BatteryUniversity.com. Regards, - D.P.

All Content on This Web Site Copyright 2005-2013 All Rights Reserved - James Wesley, Rawles - SurvivalBlog.com

About this Entry

This page contains a single entry by Jim Rawles published on July 17, 2013 12:46 AM.

Letter Re: Pro-Gun or No-Gun America, Which Saves More Lives? was the previous entry in this blog.

Note from JWR: is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.

Monthly Archives

Visitor Map

Map

Statistics

counter customisable
Unique visits since July 2005. More than 300,000 unique visits per week.