Re: [jallib] LiFePo or Lead-Acid battery management. Failure predicament.

[Rob Hamerling]

Hi,
This group is specifically about Jallib (developments) and I'm afraid
your message is 'off topic'. You'll be better off in PicList or maybe
Jallist.
Regards, Rob,



On 08/07/12 04:56 pm, Bogdan Mihai Octavian wrote:
> Hello everybody! I'm not new into microcontrollers, actually all my
> projects, personal or those built for the company I'm working for, were
> built around PIC family. Now I'm working at a project that involves
> battery management and battery failure predicament. It is a industrial
> process automation powered from a Solar panel (UPS with MPPT charger and
> all the other stuff involved). Because it has to work without the Sun,
> sometimes, from the internal UPS powered from one or two LiFePO cells, I
> have not decide yet, I have to know the battery condition. The
> automation it is remotely located and it is visited by the maintenance
> stuff once a week or in case that something goes wrong (radio
> communication also implemented). I have founded some hardware solutions,
> battery gas gauge, provided by Linear Technology or Texas Instruments or
> Maxim, but those are little too complicated. I was thinking do do that
> myself and this is where I will need some help.
> So, does anybody have an idea how to do that? Measure the battery
> capacity and predict battery failure.....
> Thanks for your time! (Allocated to read this)


--
R. Hamerling, Netherlands --- http://www.robh.nl

[vasile surducan]

Hi Bogdan
In my opinion, failure predicament for Lead Acid accumulators is not too complicated, but I do know nothing about LiFePO cells so I can't pronounce... I've designed and manufactured as prototype, about five years ago (if I don't wrong) a charger-discharger for NiMH  with one channel using a simple PIC12F675 and a PC for analyzing data.. I think you have to implement two things: a charging algorithm and a discharging algorithm on a known load. Both are quite simple. You can also measure dV on dt with load, but this might need more than 10 or 12bit AD resolution.
The biggest problem I've faced over the time since  I'm playing with PIC micro is not the algorithm itself, but the problems of long time running. For such application you'll have to consider deeply how a watchdog should help you against bottlenecks.

best wishes,
Vasile

On Tue, Aug 7, 2012 at 5:56 PM, Bogdan Mihai Octavian <bogdan.mih...@gmail.com> wrote:

       Hello everybody! I'm not new into microcontrollers, actually all my projects, personal or those built for the company I'm working for, were built around PIC family. Now I'm working at a project that involves battery management and battery failure predicament. It is a industrial process automation powered from a Solar panel (UPS with MPPT charger and all the other stuff involved). Because it has to work without the Sun, sometimes, from the internal UPS powered from one or two LiFePO cells, I have not decide yet, I have to know the battery condition. The automation it is remotely located and it is visited by the maintenance stuff once a week or  in case that something goes wrong (radio communication also implemented). I have founded some hardware solutions, battery gas gauge, provided by Linear Technology or Texas Instruments or Maxim, but those are little too complicated. I was thinking do do that myself  and this is where I will need some help.

       So, does anybody have an idea how to do that? Measure the battery capacity and predict battery failure.....

                                                                                                                          Thanks for your time! (Allocated to read this) 

--
You received this message because you are subscribed to the Google Groups "jallib" group.
To view this discussion on the web visit https://groups.google.com/d/msg/jallib/-/vxYiirFkNs4J.
To post to this group, send email to jal...@googlegroups.com.
To unsubscribe from this group, send email to jallib+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/jallib?hl=en.

[Joep Suijs]

Hi, 

Lipo charge/discharge is not complicated, just keep the current within the bounadies until certain voltage levels are reached. But these cells are quite sensitive to overcharging ( resulting in fire risk) and deep discharge (destroying the cell). When puting cells in series, balancing and/monitoring of each cell is required.

So proper testing and reliable circuit (including the watchdog Vasile mentioned) are manatory. 

I'm note sure about failure prediction, but i guess capacity measurement during a significant charge part of a charge cycle os a good indication for any battery. For nimh, reduced discharge curreny is a failure mode too.

Joep

Op donderdag 9 augustus 2012 schreef vasile surducan (vsur...@gmail.com) het volgende:

[Bogdan Mihai Octavian]

       Well, thank you for your answer! I guess that I'm responding back to Vasile Surducan? Cause your book, "Microcontrolere pentru toti" was the best book in the University and the book that help me started with PIC programming. I am still writing code with the help of JAL....

       I wish to thank again for the good job writing that book!

Back to my problem now! After a few revises I have dropped Lead_acid battery and switched to LiFePo, just one cell (http://www.all-battery.com/3.2V12.5AhLiFePO4PrismaticEnergyCell-30251.aspx). Because Linear Technology as well as Texas Instruments had developed already LiFePO gas gauges (http://www.ti.com/lit/ds/symlink/bq27501.pdf, for example) my job just got easier. Because I will need to power my 5vdc electronics just from a single cell LiFePo the boost converter efficiency is not so bad especially If I use the converters from Murata (http://www.farnell.com/datasheets/369793.pdf). Because I will need to power the 12VDC/1Watt coil of an servo-valve I will use isolated DC-DC converters from XP POWER. The valve should work for 5 to 15 minutes every 12 to 36 hours...not so bad. The other sensors (4-20mA pressure sensors and the other two optical level sensors) will be powered from a different isolated DC-DC converter that will be on for 2 seconds every 5 to 10 minutes, not so bad also.

Just one cell LiFePO (LiIon are not doing it so well at lower temperature, as -20 Celsius degrees) is easier to monitor and manage. The best solution for MPPT charger from Solar panel comes from Linear Technology, LT 4000-1. So,  I am pretty much covered in this corner of the problem. 

      Now, I have the battery gas gauge with I2C communication, 12 bits A/D converter for the pressure sensor and COG LCD on the same I2C bus. That wouldn't be so difficult to control.  

      Any idea about what I2C COG LCD will work better with JAL COG library? I have selected a few from Farnell, but it's my first time that I have to send data to a COG LCD ( I have to display a few graphical parameters of the process)

    P.S.: I quite new with forums and threading so it would take a while to accommodate myself with this. Should I move my thread to JALLIST forum, as Rob mentioned?

[Bogdan Mihai Octavian]

          Thank you Joep! I'm pretty familiar with LiPO because I'm also an RC passionate. I have carbon fiber racing 1/10 touring cars, CF helicopters and racing RC boats. I have also LiPO packs and a specialized Lipo balancer/charger/discharger that I had to modify because it was not measuring right the upper voltage threshold, instead of 4.2 volts it was displaying 4.1, so the charge process was not terminated at 4.2 but at 4.3 which is not so good. 

Anyway, I think I'm gonna go with specialized IC for battery capacity measuring and MPPT solar charger, as I just told Vasile. Because it is an industrial process automation and it has to work at least 3 years without maintenance, I can't take to much risks! 

         Still have some questions left. I will start with the most needy. Can you advise me about the best choice for an LCD that works with JAL routines for graphic lcd without problems. I have selected a few from Farnell but it's my first time  that I have to display data onto a COG display. 

        Thank you again for your time!   

[Bogdan Mihai Octavian]

marți, 7 august 2012, 17:56:28 UTC+3, Bogdan Mihai Octavian a scris:

       Hello everybody! I'm not new into microcontrollers, actually all my projects, personal or those built for the company I'm working for, were built around PIC family. Now I'm working at a project that involves battery management and battery failure predicament. It is a industrial process automation powered from a Solar panel (UPS with MPPT charger and all the other stuff involved). Because it has to work without the Sun, sometimes, from the internal UPS powered from one or two LiFePO cells, I have not decide yet, I have to know the battery condition. The automation it is remotely located and it is visited by the maintenance stuff once a week or  in case that something goes wrong (radio communication also implemented). I have founded some hardware solutions, battery gas gauge, provided by Linear Technology or Texas Instruments or Maxim, but those are little too complicated. I was thinking do do that myself  and this is where I will need some help.

       So, does anybody have an idea how to do that? Measure the battery capacity and predict battery failure.....

                                                                                                                          Thanks for your time! (Allocated to read this) 

marți, 7 august 2012, 17:56:28 UTC+3, Bogdan Mihai Octavian a scris:

       Hello everybody! I'm not new into microcontrollers, actually all my projects, personal or those built for the company I'm working for, were built around PIC family. Now I'm working at a project that involves battery management and battery failure predicament. It is a industrial process automation powered from a Solar panel (UPS with MPPT charger and all the other stuff involved). Because it has to work without the Sun, sometimes, from the internal UPS powered from one or two LiFePO cells, I have not decide yet, I have to know the battery condition. The automation it is remotely located and it is visited by the maintenance stuff once a week or  in case that something goes wrong (radio communication also implemented). I have founded some hardware solutions, battery gas gauge, provided by Linear Technology or Texas Instruments or Maxim, but those are little too complicated. I was thinking do do that myself  and this is where I will need some help.

       So, does anybody have an idea how to do that? Measure the battery capacity and predict battery failure.....

                                                                                                                          Thanks for your time! (Allocated to read this) 

[vasile surducan]

Probably the best place to post questions for many (and quite different)  opinions is piclist: http://www.piclist.com/techref/piclist/index.htm or jallist: http://tech.groups.yahoo.com/group/jallist/
So far I didn't saw major problems in discussing pic related things or electronics at jallib, but here the number of answers will be limited. The topic of jallib is related to jal libraries and standards, Rob is the informal chief in charge here.

I can say only how I would do it:
- I will choose a PIC running from 2V to 4V, there are many, also the other IC's should run at 3V +/- 1V (LCD, communication etc.)  If this will not be entirely possible, then I'll choose your DC-DC converter (but my supply option will be 3.3V, PIC resources running at lower voltage are better, see only the hardware RTCC for example).
- I will keep the LIPO voltage range between 2V and 4V (3.6V if mandatory) using a simple crowbar-like  protection on the accumulator end and a high constant current generator driven from PIC. A TL431 (variable zener) and a high current MOS-FET can do the protection. Once the LIPO voltage reached 4V, the crowbar will source the entire charging current, until the PIC will stop the charging process. You always need two levels of protection: one hardware which should work even when the software (the PIC) has a malfunction (and will have until firmware versions will be greater than 10...20).
- I will choose a power discharge resistor and a high current MOS-FET switch for the failure predicament. This will simplify the measurement (you will not need current measurement only voltage measurement). Also the version in which you'll use the available load for failure predicament and measure the voltage and the current between two moments of time is feasible.

To view this discussion on the web visit https://groups.google.com/d/msg/jallib/-/ERMbALr12soJ.