1Charger

Description

A wattmeter and step-down charger for LiFePO4 batteries with CC &CV

* Vin = 4-30V ( peak 36V )
* CV / Vout 1.7-25V
* CC = 0.3 - 5A ( peak 10A )

The charging technique is simple CV and CC, and the LiFePO4 batteries are treated as if they were capacitors. The power source may be TEG (like the powerpot), solar panels, or a dynamo. It can also be charged quickly directly from a car battery. The supply current for the wattmeter is with :
NOKIA 5110 : 7.5 mA.
JLX12864G-086 : 15mA

The device is not (yet) a MPPT charger, even though the wattmeter circuit provides a good basis for taking that step in the future.

Extra feature, “or-ing” diodes:
By connecting several diodes (as many as you wish) in parallel to the primary blocking diode, it is possible to charge multiple batteries at once. However, the most discharged battery should be placed directly after the primary diode.

Details

Background:

I wanted a charger for portable power that could be used with solar panels, TEG elements and dynamo. It was also important that the charger could take all available power. I quickly realized that LiFePO4 are the best batteries available. LiFePO4 batteries have a long shelf-life, can tolerate > 2000 charge cycles, and can accept high charge/discharge currents. They have very little internal resistance, not far from that of a super-capacitor. One of the toughest brands is the A123 Nanophosphate, with 7000 cycles at 1C before it is reduced to 80% of capacity (it can be used for spot welding or to start a car). I also wanted a wattmeter that could work down to 2.8 V (discharged LiFePO4). Long life and durability had highest priority.

Usage:

Adjust CV (Vout) on the step-down to battery float voltage
Adjust CC (current limit) for smaller batteries (or use 5 A to protect the charger)
Estimate min. Vin needed for: step-down = battery float voltage (Vout) + 1 V
Connect power source in serial or parallel to achieve sufficient Vin
Suggested open voltage: TEG = 2xVin (MPPT)
Suggested open voltage: solar panels = 1.25xVin (MPPT)
Protect your TEG and solar panels with reverse blocking diodes
When charging cells in series, use balancing circuit

Example:

Single cell LiFePO4 (8 Ah)
2x5 W solar panels (5.5 V)in parallel, or 2xTEG (12 V) in parallel (PPXspecial edition)
CV = 3.55 V ( most of the energy is between 3–3.4 V )
CC = 5 A (to protect circuit when charging from car battery)
Reverse blocking diodes in cables (solar panels and TEG)

Charger design:

The heart of the charger is the step-down converter from prodctodc. I have tried at least 10 different models and what I like about this one is it does not heat up much, has a high efficiency, and can take a lot of beating. The modifications I have made to this circuit are simple; I have just inserted a reverse blocking Schottky diode and then moved the voltage sense resistor to the new +Vout. I also replaced the 470 µF Sanyo capacitors with equivalents from Rubycon. This step is not necessary; it is only done as a precaution to ensure a long life. The original Sanyo capacitor has low ESR and seems to work fine. Finally, I inserted a 0.02 Ω current sense resistor before −Vin. This resistor is used by the wattmeter. By connecting several diodes (as many as you wish) in parallel to the primary blocking diode, it is possible to charge multiple batteries at once. However, the most discharged battery should be placed after the primary diode (with CV sense). This is very useful when you leave your basecamp for a while and want to charge several cells. This is also a poor man's balancing circuit; that is, the batteries will soon have equal charge. Personally, I prefer four single cell LiFePO4 that can be combined into one 12 V battery, instead of one hard–wired 12 V battery.

Charging a battery is similar to a short-circuit, where the CC is the current limit. When the power supply is too weak, Vin will break down to approximately 1 V above Vout (battery voltage). To achieve a kind of poor man's MPPT, one should select an appropriate open voltage for the power supply (solar panel = 1.25xVin). A better solution would be MPPT (a future improvement). However, it is still possible to catch 90% of available energy from a 5.5 V panel using this simple approach.

Wattmeter design:

The heart of the wattmeter is a Attiny861 which has differential gain x1, x8, x20, x32. This means there is no need for an op-amp. Attiny861 also has four different Voltage reference sections ( 1.1V, 2.56V, Vcc (3.3V),Aref ). This gives a lot of options. I prefer to use the internal 1.1V and 2.56V Vref ( since this makes the circuit less dependent on the 3.3V supply). The 3.3V (buck-boost) voltage supply to the wattmeter is a gutted ebay voltmeter. It also has a voltage divider. The wattmeter is supplied from the battery ( Vout ). The buck-boost circuit ensure 3.3V when battery voltage varies between 1.7-25V. And finally a low power...

17 Jan - 880W/m2
mr.jb • 01/17/2016 at 18:54 • 0 comments

888w/m2 - it's all about the angle ;-)
http://solarelectricityhandbook.com/solar-angle-calculator.html
0.85A - panel area 3.2 dm2
Parabolic mirror 8cm
RT8272 => Tiny-1charger
mr.jb • 07/24/2015 at 08:01 • 2 comments

The RT8272 looks very promising ( especially with a 22uF cap at Vin ), so far it has no problem charging with my smallest 0.2A to my largest 1.5 A solar panel ( with clouds or clear blue sky ).
http://tw.richtek.com/assets/product_file/RT8272/DS8272-02.pdf
Veery cheap
http://www.ebay.com/itm/10PCS-Mini-DC-DC-Buck-Converter-Step-Down-Module-Power-Supply-For-aeromodelling-/201234570492?pt=LH_DefaultDomain_0&hash=item2eda83dcfc
With a big cap at Vin I think it is regulating around under-voltage lockout ( 4.2V ) when it's cloudy
http://www.fieldlines.com/index.php/topic,144628.msg977808.html#msg977808
This might be taken one step further with the EN pin.....( if it's worth the trouble ). The under voltage lockout is close to the Mppt point for smaller "5V" panels.
But I think it's use should be limited to one panel ( <= 1.5 A )...
Putting several RT8272 in parallel is probably not a good idea since Vref "0.92v" (FB) will differ from circuit to circuit, meaning one circuit will probably get more heated than the others.

4 display types supported

mr.jb • 07/22/2015 at 09:36 • 0 comments

You can choose different display types by activating #define of your choice

( by using CE you could have multi-screen ... ;-)

#define LX12864G_086_SPI   (3)
//#define SSD1306_SPI      (1)
//#define SSD1306_IIC      (2)
//#define NOKIA_5110_SPI   (0)


// Selected display type
#define DISP_TYPE          LX12864G_086_SPI
#define DISP_WIDTH         128
#define DISP_HEIGHT        64
#define DISP_NCOL_SMALL    21  // 128/6 small font

Nokia 5110 replacement - JLX12864G-086-P
mr.jb • 07/21/2015 at 11:21 • 0 comments
The JLX12864G-086 is in my opinion a good replacement for the Nokia 5110.
Similar price 5$, better connection but slightly higher power consumption ( like 10 mA ).
The one I work with has a feature of sometimes forgetting "screen orientation", my workaround is repeating this command before writing any text. My guess it's a light sensitive part at the bottom of the display that triggers the effect. Once I've covered it with tape..problem seems to be gone.
```
spi_write_byte(0xa6, 0 ); /*  General: Positive Display */

/*  bottom 0XC8: reverse scanning order: from bottom to top*/
spi_write_byte(0xc8, 0 );

/* 0xA0: General: column address from left to right */
spi_write_byte(0xa0, 0 );
```
http://www.electrodragon.com/product/12864min-graphic-lcd/
http://www.mikrocontroller.net/attachment/252515/JLX12864G-086.pdf
The alternative is trying to get a good connector for Nokia
http://www.starlino.com/nokia_lcd_connect.html

poor man's pyranometer

mr.jb • 07/04/2015 at 06:03 • 0 comments

Most lux sensors has a limited range ( < 100 000 lux ) and I plan a different approach.

Current source current is directly proportional to the solar radiation

Measuring Solar Radiation, Chuck Wright

In a solar cell, the parameter most affected by an increase in temperature is the open-circuit voltage -2.2 mV/°C

The short-circuit current, Isc, increases slightly with temperature, since the band gap energy, EG, decreases and more photons have enough energy to create e-h pairs. However, this is a small effect and the temperature dependence of the short-circuit current from a silicon solar cell is

A change form 25 to 50 °C will increase Isc current 1.5%.

The impact of increasing temperature is shown in the figure below

Area [dm²]	Isc [mA]	Lux	0.000079 x lux [Watt/dm²]
0.71 ( 1W panel )	185.5	120000	9.48
	152	93500	7.39
	129	76400	6.03
	20.9	14300	1.13
3.1 ( 5.15w panel )	1010	119500	9.44
	850	95400	7.54
	730	76500	6.04
	610	58800	4.65
5.7 ( 7.5watt panel )	1560	113200	8.94
	1390	96000	7.58
	1170	76200	6.02
	1040	61100	4,83

http://www.pveducation.org/pvcdrom/solar-cell-operation/short-circuit-current

http://www.pveducation.org/pvcdrom/solar-cell-operation/effect-of-temperature

Pyranometers and reference cells
Pyranometers and Reference Cells, What’s the Difference ?

April 30 - 2015 more than 100 000 lux
mr.jb • 04/30/2015 at 10:17 • 0 comments

Often more than 100 000 lux ( resulting in unreliable lux measurements )
A 8h LiFePO4 battery is fully charged in 3.5h compared to 2.5h from an wall outlet.
Sun with a clear blue sky, peak 55000 lux and -1 °C
mr.jb • 01/05/2015 at 13:09 • 0 comments

I have compared the 6.4 dm² panels with my 11.2 dm² a sunny winter day with a clear blue sky. The light intensity peaked at 55000 lux.
The theoretically available power P at 55000 lux and a 11.2 dm² panel with 0.16% efficiency is:
P = 55000*0.000079*11.2*0.16 = 7.8 Watt ( I measured close to 6.6 Watt on the battery side )
Plotly 5 jan, comparison 6.4 dm² vs 11.2 dm²
Plotly 4 jan,clear blue sky 11.2 dm² panel
Very cloudy winter day, 4300 lux and 0 °C
mr.jb • 12/30/2014 at 21:35 • 0 comments

I have compared the 6.4 dm² panels with my 11.2 dm² under harsh conditions. The light intensity peaked at 4300 lux and it was impossible to distinguish the contour of the sun. I managed to charge close to 0.5Ah. I think at least 22 dm² ( 30W panel ) is necessary to do something meaningful in this kind of light. ( When it's snowing I've measured 170 lux )
The theoretically available power P at 4300 lux and a 11.2 dm² panel with 0.16% efficiency is:
P = 4300*0.000079*11.2*0.16 = 0.6 Watt ( I measured close to 0.4 Watt on the battery side )
Plotly from 30 december , comparison 6.4 dm² vs 11.2 dm²
Plotly from 31 december sun and clouds, 11.2 dm²
A cloudy winter day, 12000 lux and -8 °C
mr.jb • 12/29/2014 at 15:35 • 0 comments

A cloudy winter day, but I could see the contour of the sun behind the clouds. I measured around 12000 lux peak. For this test a geared up and tested two of my "new" 7.5W panlels. You can read more about them at http://cottonpickersplace.com/ . I think these panels are very good and they come in many sizes. I prefer panels that are waterproof and has exposed soldering points. I'm evaluating if the slightly smaller panels (1.6x2dm) or this size (2.8x2 dm) is the best.
26 December a sunny winter day, -5 °C
mr.jb • 12/26/2014 at 12:20 • 0 comments

With just 6.4 dm² panel area in december (winter in Sweden) I could charge a 2Ah LiFePO4 in less than three hours. The light sensor is just a rough estimate ( It's still uncalibrated ). It is however worth to note that with a drop in temperature the efficiency of the panel increases. As an example a cell that produces 0.6A at 25 °C will produce 0.68A at 0 °C.
( December 22 was the shortest day 2014 )
Plotly from 28 december - sun and clouds

View all 14 project logs

Build Instructions

Collapse

Step 1

Essential parts ( less is more ;-)

Step 2

Initial modifications to step-down circuit

* Replace Sanyo capacitors with Rubycon 470µF ( optional )

* Solder reverse blocking diode to Vout+

* Desolder CV sense potentiometer and solder a wire ( orange )

* Solder 0.02 Ω current sense resistor to Vout-

Step 3

buck-boost (3.2)-3.3V from a gutted ebay voltmeter

( my guess is it's something similar to a Dickson Doubler with a 3.3V regulator )

* Desolder CPU

* Desolder LED ( 7-segment )

* Solder wire to regulated voltage 3.3 V ( purple )

* Solder wire to GND ( gray )

* Solder wire to voltage divider ( yellow )

* Apply voltage and verify you get 3.3V ( grey-purple) when you vary supply voltage 1.7-25V ( black-red ). Voltage sense( yellow) should vary accordingly if soldered correctly

Discussions

Nathaniel VerLee wrote 07/23/2014 at 15:30

You minght consider checking these out: http://www.all-battery.com/Tenergy3.2V10Ah38120SLiFePO4EnergyCellRechargeableBattery-30074.aspx

They are similar to the A123s, cant get quite as high in terms of charge and discharge but they are a lot cheaper and easy to configure in packs with simple screws, washers and plates along with battery holders.

Are you sure? yes | no

mr.jb wrote 07/23/2014 at 15:49

Thanx !! I'm very interested good LiFePO4. I'll check them out.
I prefer prismatic ... with aluminum case. But they rarely have A123 performance.
In the end, durability is my No1 concern.

Are you sure? yes | no

davedarko wrote 07/18/2014 at 07:18

Can you post a schematic / logic-outline of your parts/boards used? Doesn't have to be very detailed but more of a setup sketch would be really cool!

Are you sure? yes | no

mr.jb wrote 07/25/2014 at 21:21

http://tinyurl.com/muticharger-schem

Are you sure? yes | no

davedarko wrote 07/29/2014 at 19:52

thx, hope you will keep it here for a while, because I will need some time figuring out all the other stuff first ;)

Are you sure? yes | no

davedarko wrote 07/16/2014 at 21:55

Hey, do you have any reading material on lifepo4 batteries and how to charge them? I'm planning a bicycle computer project and figured those batteries would be great to use at german wheather conditions (can get goo hot or cold for lipos). Anyways great project!

Are you sure? yes | no

mr.jb wrote 07/17/2014 at 10:59

Check out http://batteryuniversity.com/ ( for the true story ;-)

My approach is to avoid over/under voltage and stay between 2.9 - 3.55V
Most of the energy in a LiFePO4 is 3-3.4V. Charging with < 1-2C is probably more efficient than fast charging ( check the spec for your battery ). In short I treat the LiFePO4 battery like a capacitor. In my case I don't have a power source strong enough to stress it (except car battery )...my charger will break before. My reason for CC is to limit the stress on the charger,. Another reason is when I use it to charge other batteries like Li-ion ( 2s1p camera batteries ).

A123 Nanophosphate is probably one of the best brands on the market
http://www.batteryspace.com/prod-specs/6610_1.pdf
http://insideevs.com/a123-updates-next-gen-nanophosphate-ext-batteries-solves-lithium-battery-heat-issues/

Maximum Continuous Discharging :70A, 28.0C rate
Maximum Impulse Discharging (< 10 sec): 120A, 48.0C rate

10C>=1000 cycles
1C >=7000 cycles

I love this demonstration
https://www.youtube.com/watch?feature=player_detailpage&v=ECzEaTzqFsY&t=53

/JB

Note...when you charge a LiFePO4 battery with a PWM ( step-down circuit ) the load is similar to a short circuit ( due to very low inner resistance ).

Are you sure? yes | no

davedarko wrote 07/18/2014 at 06:37

wow, quite a read. Thanks, definitely going to read that because I know nothing or less about batteries and have to get in to that.

Are you sure? yes | no

1Charger

Description

Details

Components

Project Logs

Collapse

17 Jan - 880W/m2

RT8272 => Tiny-1charger

4 display types supported

Nokia 5110 replacement - JLX12864G-086-P

poor man's pyranometer

April 30 - 2015 more than 100 000 lux

Sun with a clear blue sky, peak 55000 lux and -1 °C

Very cloudy winter day, 4300 lux and 0 °C

A cloudy winter day, 12000 lux and -8 °C

26 December a sunny winter day, -5 °C

Build Instructions

Collapse

Discussions

Similar Projects

30W (=~240W halogen) LED balcony solar lanterns

ARDUINO MPPT SOLAR CHARGE CONTROLLER

555 MPP solar charger

Testing cheap linear Li-ion chargers for solar

1Charger

Become a Hackaday.io member

Just one more thing

Description

Details

Components

Project Logs Collapse

Build Instructions Collapse

Enjoy this project?

Discussions

Become a Hackaday.io Member

Similar Projects

Does this project spark your interest?

Report project as inappropriate

Send message

Remove Member

Project Logs

Collapse

Build Instructions

Collapse