PDA

View Full Version : High power a123 balancing charger



Ramboman
01-11-2009, 03:42 PM
Hello world !
With two operational projects for home decoration, I am very new to propeller programming, but it's so easy...
My main hobby in electric model aircraft.·
In that context, I have build a 1/4 scale Piper Cub powered with a large brushless motor and a pack of 12s2p a123.
The key advantage of thoce cells is that they can be reloaded in the plane in less than 15 min.
Presently, I reload the pack at the field with a four 38Ah lead batteries in parallel, a 1200W 12V-220V converter, a CC-CV power unit and a diode to avoid reverse current in case of failure.
The procedure is very simple: you set current to 20A (2p) and voltage to 43.2 (12s).·
In phase 1, in constant current mode, you charge at 20A;
in phase 2, in constant voltage mode, you charge at 43.2V and current decreases to next to zero;
then you have finished...
Converter and power unit are pretty expensive and heavy.
In a first approach I had a very simple procedure for a 6s1p pack:
Connect the pack to a 24V battery thru one unique resistance (0.5 ohms, 50W); it works!
Time to come to my project:
1. Connect my four 12V batteries in a string to have 48V;
2. Connect the 12s2p a123 pack to 48V thru a ESC (RC electronic speed controller);
3. Drive the ESC (like a servo) with a Propeller gathering current and voltage to the a123 pack.
I would appreciate any recommandation in gathering current and voltage...
Next, a123 cells should be balanced (same voltage).
Propeller can check voltage and shunt (with FET or relay) any cell thru a resistance when voltage cecomes too high.
I would appreciate any recommandation is checking cell voltage...
There is no 15min field charger for a123 on the market.
This would be the first one, for less than $100.
Thanks for your support.
·

Leon
01-11-2009, 06:51 PM
Wouldn't it be better to use a much cheaper, smaller MCU with a built-in ADC, like a PIC or AVR? They only cost about $2.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Post Edited (Leon) : 1/11/2009 12:07:35 PM GMT

Beau Schwabe (Parallax)
01-11-2009, 09:51 PM
Ramboman,
·
Use extreme caution when charging LiFePO4 batteries.· They can be damaged if they are overcharged or balanced improperly.
I would stay as close to the manufacturer's specs as possible through the three phases of charging...

CC - Current Charge
VC - Voltage Charge
TC - Trickle Charge
·
In the CC stage, to maintain constant current, you need to do one of two things... Either adjust the voltage
as the battery charges to maintain a constant current or adjust the value of the current limiting resistor(s) to maintain a constant current.· Using a single fixed resistor value with a single fixed supply voltage won't cut it for constant current charging because the current demand will change as the battery charges.
·
In the VC stage the voltage on the battery should constantly be monitored and maintained for a pre determined amount of time indicated by the manufacturer.· This is usually a percentage of the time required during the CC stage.
·
In the TC stage the Supply voltage is fixed at the recommended charge voltage per cell, and the battery is allowed to trickle or float at that level indefinitely.


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe (mailto:bschwabe@parallax.com)

IC Layout Engineer
Parallax, Inc.

Ramboman
01-11-2009, 10:23 PM
Thank you, Beau,

with my CC-CV power unit, I follow strictly manufacturer's instructions, I have no problem.

Resistance approach is very rough, not an operational procedure.

The key issue in my project is not to drive the ESC, it's like a servo, but to get Current and Voltage.

Having that information, Propeller will drive the ESC to emulate the CC-CV power unit.

Do I really have to go to another processor as suggested by Leon, or is it easy to do it with a Propeller.

simonl
01-11-2009, 10:40 PM
@Ramboman: Leon's suggestion is one for reducing the cost. The Propeller could do all you want, but would likely need an external A/D chip (MCP3803? MCP3208). Leon's just pointing-out that a $2 PIC, with an integral A/D, would also be capable of achieving your task. (Of course, he's assuming you have suitable PIC programming knowledge; compiler; and PIC programmer...)

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Cheers,
Simon
-------------------------------

www.norfolkhelicopterclub.com (http://www.norfolkhelicopterclub.com)

You'll always have as many take-offs as landings, the trick is to be sure you can take-off again http://forums.parallax.com/images/smilies/wink.gif
BTW: I type as I'm thinking, so please don't take any offence at my writing style http://forums.parallax.com/images/smilies/smile.gif

Post Edited (simonl) : 1/12/2009 1:27:51 PM GMT

Beau Schwabe (Parallax)
01-11-2009, 10:41 PM
Ramboman (http://forums.parallax.com/member.php?u=52868),

As far as an ESC, I would probably just break it down to a single mosfet with a logic level input.· There are a few mosfets designed for automotive applications from International Rectifier that would do nicely.

for the processor, it depends on what you want to do in the long run, and what you are most comfortable with.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe (mailto:bschwabe@parallax.com)

IC Layout Engineer
Parallax, Inc.

Leon
01-11-2009, 10:49 PM
If it's a one-off and you haven't got much experience with other MCUs, the Propeller would be fine. No-one would use one in a charger that was going into production, though.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Post Edited (Leon) : 1/11/2009 8:43:20 PM GMT

Ramboman
01-12-2009, 12:16 AM
You are all experts guys, I am not...

I understand a123, ESC, propeller (basic), but have no experience with PICs, mosfets...

I am ready to learn about ADC converters.

So keep it simple:

how do I measure up to 20A current (probably on a shunt)?

how do I measure up to 50V voltage?

Having that, I make it work, for my own pleasure ;-)

Thanks

Drone
01-12-2009, 03:03 AM
Hi, I'm just glancing through this thread; need to go to sleep. So sorry if this isn't so relevant...


Ramboman said...
I would appreciate any recommendation in gathering current and voltage...


Have a look at this part: LTC4151 - High Voltage I2C Current and Voltage Monitor. Perhaps it has some worth...

Data sheet available via www.linear.com (http://www.linear.com). I'm not affiliated with Linear Technologies in any pecuniary way.

Best Regards, David

Ken Peterson
01-12-2009, 03:10 AM
I rarely recommend alternatives to the Propeller in this forum, but in this case a Propeller may be overkill. If you don't have experience with a PIC, you can also try a PICAXE controller, which is much easier to program. They are available in different configurations at www.sparkfun.com (http://www.sparkfun.com).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·"I have always wished that my computer would be as easy to use as my telephone.· My wish has come true.· I no longer know how to use my telephone."

- Bjarne Stroustrup

simonl
01-12-2009, 05:25 AM
Hi again. I just saw this Voltage and Current Sense Breakout (http://www.sparkfun.com/commerce/product_info.php?products_id=9028) on the Sparkfun site - looks like it fits your requirements... (still might need a proper A/D chip though).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Cheers,
Simon
-------------------------------

www.norfolkhelicopterclub.com (http://www.norfolkhelicopterclub.com)

You'll always have as many take-offs as landings, the trick is to be sure you can take-off again http://forums.parallax.com/images/smilies/wink.gif
BTW: I type as I'm thinking, so please don't take any offence at my writing style http://forums.parallax.com/images/smilies/smile.gif

Ramboman
01-12-2009, 02:41 PM
simonl said...
Hi again. I just saw this Voltage and Current Sense Breakout (http://www.sparkfun.com/commerce/product_info.php?products_id=9028) on the Sparkfun site - looks like it fits your requirements... (still might need a proper A/D chip though).


http://forums.parallax.com/images/smilies/roll.gif
Second piece to my puzzle!
Next to fond: most appropriate ADC to nterface to Propeller
http://forums.parallax.com/images/smilies/confused.gif

simonl
01-12-2009, 08:24 PM
If you can't use the Prop with an R/C circuit, there are a couple of objects in Obex for the MCP3208 (which Microchip will give you a free sample of...)

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Cheers,
Simon
-------------------------------

www.norfolkhelicopterclub.com (http://www.norfolkhelicopterclub.com)

You'll always have as many take-offs as landings, the trick is to be sure you can take-off again http://forums.parallax.com/images/smilies/wink.gif
BTW: I type as I'm thinking, so please don't take any offence at my writing style http://forums.parallax.com/images/smilies/smile.gif

Leon
01-12-2009, 08:44 PM
Ramboman:

Where did you get your A123 batteries from? They might be useful for amateur radio applications. 12V gel cells work fine but are rather heavy and take a long time to charge.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Fred Hawkins
01-12-2009, 08:51 PM
TI's bq2954 might be just what you need. One of their application notes shows a circuit using the bq2954 that fast charges 6 cells in thirty minutes with a 5A current. The input voltage doing the charging is 30v.

http://focus.ti.com/docs/prod/folders/print/bq2954.html

The application note is called 'using the bq2954', linked·from the bottom of that page.

They show another·sample circuit in another pdf. Look for the "Designed to Go: 20W Miniature Dual-Stage Fast Lithium-Ion Charger Using bq2954"·in the same set of application note links.

·@Leon, I like www.batteryspace.com (http://www.batteryspace.com)
I got some RCR123A's plus a charger there, but they have all sorts of batteries...

Post Edited (Fred Hawkins) : 1/12/2009 1:57:29 PM GMT

Leon
01-12-2009, 09:11 PM
Thanks, Fred. Two of the $75.50 12.8V 3000 mAH batteries in parallel would do the job.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Post Edited (Leon) : 1/12/2009 2:16:25 PM GMT

mcstar
01-13-2009, 04:40 AM
I've got some LiFePo4's for sale on ebay as well. They're 10AH industrial quality cylindrical cells. You could use 4 of these in series to give you about 14V with a peak current capacity of 200amps with a weight under 4lbs. You can charge them at up to 20amps. I'll even give a 10% discount to fellow propellerheads! http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&Item=130271429385&Category=40975

Ramboman... I'm currently working on a propeller driven PWM battery charge and discharge circuit. Perhaps the propeller is a bit of overkill, but in my application I'm able use viewport and fuzzy logic on the prop to control a mosfet. This lets me control the load based on various inputs from the ADC.

My first prototype is able to charge/discharge at up to about 3.0 amps and log the voltage and current throughout the process. A shunt works fine, but in order to minimize the I2R losses across it, you want it to have a very small resistance. In fact the one I'm using now is so small, I cannot accurately measure its resistance with my meter! This also means the voltage drop across it is really too small to accurately measure directly, so I've got a op-amp to amplify the signal. It also acts as an integrator circuit to help smooth out the pulses since I'm using PWM.

Ramboman
01-13-2009, 02:35 PM
Hi Leon,
part of my a123 are coming from a123,
part from MEC.
I use also MEC solderless tubes to assemble them, with a small modification to have balancing taps.

Ramboman
01-13-2009, 02:43 PM
Hi Mcstar,
funny, I try to do the same, could we share information ?
Hope you can go up to 20A, 43.2V...
I bought a AD7705 to start with.

ianw
01-21-2009, 08:03 AM
Hi all,

I'm glad I found this thread. I too am working on a battery management system (as a first microcontroller project). One difference from the posts I've seen so far is that I would like to handle more than just lithium chemistries - also nickel an lead chemistries.

I was envisioning the system as a set of modules - specifically, separating the constant voltage/current source, an the discharge resistors, since you may want a larger/smaller power source if dealing with different sized packs, or may not always need discharge capabilities. My own interest is to just charge any old battery around the house, an as a datalogger for my electrochemistry tinkering. So, I actually don't have high power requirements (rarely over 3A), but it would be nice to have the option of greater power e.g. if I ever get that electric bike project in the back of my brain going. To support my experiments, I also planned to add a module for a freescale pressure sensor an temperature monitoring.

Although the Propeller may be overkill, I thought this would be a good first project to learn the Propeller. I was actually considering another micro for this first project, but will go with the propeller since it seems several others here are headed along a similar vector.

mcstar wrote:
>I'm currently working on a propeller driven PWM battery charge and discharge circuit
I'd love to see what people have come up with. I plan to develop my board in Eagle an post my files - this will be my first Eagle project too.

mcstar wrote:
>I've got a op-amp to amplify the signal
your setup sounds similar to what I have on my spaghetti prototype, although the recent sparkfun board mentioned earlier in this thread provide the same thing - so, I will probably move to this since they are fairly cheap. I looked at your blog an saw several LiFePO4 BMS options listed, including this one: http://www.endless-sphere.com/forums/viewtopic.php?t=5416 What made you decide to create your own?

The one component I am uncertain of is how to construct the constant voltage/current source. What would others recommend?
-I've seen buck regulators like the LM2576 used for constant voltage/current up to 3A. an they seem relatively efficient compared to e.g. LM317's.
-I've used H-bridges (as I understand it, 4 FET's) for motor control before an thought 1 half of a bridge might be usable for 1 pack an the other for a separate pack - by sending a pwm I could vary the voltage level (constant voltage), an if monitoring with a voltage/current sense board I could alternatively modulate voltage level as a function of current consumption (constant current). But I was not sure if this scheme was wise or efficient(?).
-I've also found a few random FET designs via Google, but did not find anything I had much confidence in - given my limited electronics background, particularly with high power applications involving FET's.

Anyway, neat to see others are also interested in this - hopefully we can collaborate. Also hoping someone has some recommendations regarding an efficient constant voltage/current supply.

ianw
01-28-2009, 06:47 AM
I thought about posting this in a new thread, since I'm not only interested in A123 cells, but any chemistry - despite the thread title - wasn't sure if this was acceptable forum protocol though.

I was not familiar with the rcgroups forums prior to my last post, but after reading through their archives, I discovered there is a lot of information about chargers in the forums. I assume the Ramboman listed here is the same Ramboman in the rcgroups forum - I saw your postings in the one off circuit being used for A123 cells. A quick review of the rcgroups forums indicates people constantly pursuing various one off circuits over the years for charging the latest battery chemistry, a microcontroller soultion seems to be a more logical solution over the long term (?).

After further reading (mostly wikipedia, and rcgroups), I decided to update this thread with my findings, and answer some of my questions.
Some good sources when considering the power portion of a charger:
http://en.wikipedia.org/wiki/Smps
http://en.wikipedia.org/wiki/DC-DC_converter
http://en.wikipedia.org/wiki/Buck-boost_converter
http://www.rcgroups.com/diy-electronics-199/
http://www.rcgroups.com/batteries-and-chargers-129/

The rcgroups forums are so dense, I'm still wading through all of the information.

To answer my own questions (realizing I should have spent some time with Google before posting):

>I've seen buck regulators like the LM2576 used for constant voltage/current up to 3A. an they seem relatively efficient compared to e.g. LM317's.
The linear regulators are very inefficient, on average 30-40%. A boost buck topology is >80% efficient and with synchronous rectification ~95%. More information on synchronous rectification here: http://www.maxim-ic.com/appnotes.cfm/an_pk/652

This is a list of various National Semiconductor buck and boost IC's suitable up to 5A (including the 2576) with design software and Spice models - but I could not find any IC's that integrate both buck and boosts: http://www.national.com/analog/power/switchers_made_simple
These would be fine for a smaller capacity charger, but I will use logic level IRF FET's in my design rather than an IC.

>I've used H-bridges for motor control before ...was not sure if this scheme was efficient
H-bridges are typically ~70% efficient and are designed for inductive loads, so although an H-bridge would work, it does not seem to be an optimal design. I'm not familiar with the many RC ESC designs though, and some seem to be more efficient than others e.g. sychronous rectification, etc. - so, perhaps an ESC would make sense under certain conditions.

>With a single mosfet, what would be the power limits of the circuit?
This is dependent on the specs of the MOSFET, but by running several MOSFET's in parallel, higher currents can be managed.

My current approach for the CC/CV module is to use a buck boost FET topology with synchronous rectification. The http://en.wikipedia.org/wiki/Split-pi design seems interesting too, but I've not read enough about the design to know if it is appropriate for a battery charger.

>Anyway, neat to see others are also interested in this - hopefully we can collaborate.
Any comments/criticisms are very much appreciated. It seems interest has waned in a propeller charger, so I'm currently considering other microcontrollers that have a larger user base and may be better suited for a first uC project.

Ramboman
02-01-2009, 11:16 PM
Keep It Simple and Stupid...
Source (48V) is higher than destination (43.2V).
PWM is OK for an ESC (that I understand) or a MOSFET able to switch 20amps (which one, how to connect to propeller?).
Driving PWM is function of current and voltage of destination.
I bought a "hall" current sensor (ACS750SCA-075) that will be coupled to my AD7705.
The rest is pure logic:
- set current to 20amp until voltage reaches 43.2V (CC)
- reduce current to maintain 43.2V (CV)
- after "n" minutes, switch to trickle...
An oldtimer should do that with a rheostat and an eye on the wattmeter!