what about a PC based charging/balancing system?

robots4joey

10 µW
Joined
Feb 18, 2009
Messages
6
I'm trying to figure all this out... I'm wanting to charge a bunch of 18650 li-ion 2200mAh cells. I was thinking- if we had a full size EV battery, computers are so cheap (free for old ones) why can't we base a charging, balancing, LVC, SOC remaining, and everything else through the serial port?

I ask this because I'm one of those crazy's who animate their Christmas lights to music... ( http://www.vimeo.com/2695308 if you care to see) I do all home build electronics with a great group on the web a lot like this one. We can easily turn on and off 120 VAC or DC by computer control through the serial or parallel port. so theoretically we should be able to turn on and off a 3.7v 20 amp charging circuit (for each parallel group of 18650 cells) with a really big mosfet of course!

We'd be able to put a small screen in the dash or something that could display all this inf real time like the 'expensive EV's'

!!!the issue is we need a input ability figured out(through the serial port?) Then we need a computer program to take care of everything

What do you think? I unfortunately sound like I kinda know what I'm doing... but I really have very little knowledge (Mech. Eng. student not EE) So this is an idea I'm hoping someone has done or wants to do or can help me figure out...
 
Well there are a number of microcontroller-based chargers, which isn't a PC obviously but the logic is done in software. But since I too know nothing of EE, I don't know the advantages, if any, of an analog circuit.
 
robots4joey said:
I ask this because I'm one of those crazy's who animate their Christmas lights to music... ( http://www.vimeo.com/2695308 if you care to see)
Wow, fantastic light show ! :eek:

Dude, you could rent some pro audio (for concerts), put that out your lawn, blast the neighborhood and charge money.

Seriously. Good show ! :D

What's that first track ? It's wicked ! It gave me visions of Rocky Balboa in a Santa costume just slugging his opponents with a left-hook then bopping them over the head with his sack of chrismuss presents with his right, over and over. Be cool if you can work such images with your lights, he he !

Excuse my brain fart, i need some sleep. :oops:

J
 
Sooooo.......

Back to computer driven BMS? It just seems to be an easy thing to do- I just don't know where to begin... It could monitor everything and more!

The only issue I see right now is the power requirements for a PC would take power from the batteries when running- but you could use an old laptop or a really old PC that doesn't take much juice...

Anyone?
 
How about a waterproof Pocket-PC phone incl. GPS which docks to the dashboard / handlebars and does all the logging and BMS funcions?

A problem with that would be the theft risk when charging in public.

The device could also have the function of an electronic key / engine immobiliser.

The touch screen could be programmed with BIG buttons so that it can all be operated with leather gloves on.
 
I think ZEVA already use the LABJACK DAQ simple interface to monitor the test discharge for the cells he test....

He use the E3 model..

search for ZEVA...
 
First of all, it is not all that simple:
e.g.:
for state of health you need to measure all cell voltages at once and the current too. This means every cell needs a sample&hold circuit with a bunch of multiplexers. The multiplexer provides data to an AD-converter and finally to a chip for the communication port. ( OK a PIC-Controller has most of the functions included ) But still you need a lot of hardware around it.
Serial communication comes with lots of errors and perhaps you have to send the same data several times. ( So there needs to be some intelligence too.

Once the data is in the PC it is a fingersnip of programming, compared to the neccessary hardware to collect/send it to the PC.

-Olaf
 
olaf-lampe said:
First of all, it is not all that simple:
e.g.:
for state of health you need to measure all cell voltages at once and the current too. This means every cell needs a sample&hold circuit with a bunch of multiplexers. The multiplexer provides data to an AD-converter and finally to a chip for the communication port. ( OK a PIC-Controller has most of the functions included ) But still you need a lot of hardware around it.
Serial communication comes with lots of errors and perhaps you have to send the same data several times. ( So there needs to be some intelligence too.

Once the data is in the PC it is a fingersnip of programming, compared to the neccessary hardware to collect/send it to the PC.

-Olaf

I believe that you do not need a sample and hold for each cell. But you do need a current sensor.

With the M-BMS that I built for my Vectux it takes more than a second to switch between cells for voltage measurements (until the display on the auto-switching DMM has stabilised). It would take about two minutes to measure each of the 102 cells and it is virtually impossible to do this at the same current draw on a real road.

But I nevertheless never miss a reverse charging event, because I know each cells individual characteristics, and I "measure where the money is", so to say.

Similarly, an intelligent (learning) BMS would use a reference value and compare the cell voltages periodically to the reference values. Because it has a memory of the measurements, it can quite easily detect discrepancies between the usual behavior of each cell and the behavior measured in the last measurement.

If the Intelligent BMS (I-BMS) finds that all cells are showing more voltage drop than usual at a given amp draw, then the SOC is lower than expected, maybe due to long storage and self discharge, or due to cold temperature or charger error, or because the IR is creeping up with age. If this change is persistent, or develops gradually, then the reference value gets adjusted and battery is recognized as having aged somewhat.

If a single cell (or a few cells) behave unusual, then they are out of balance or worse, have aged more or been damaged more than the others.

An I-BMS would know which cells are most likely going to be empty first, and monitor these cells frequently during low SOC and high current draw situations. Once the lowest cell hits the low voltage cutoff point (or rather the "current reduction point"), it can monitor that cell continuously and reduce the current draw to the maximum that this cell can produce without falling below it's safe voltage. That is then the ideal emergency crawl-home-mode tailored to protect the weakest cell.

During recharging the job of the I-BMS is much easier, because it has lengthy periods of constant current during which it can compare cells "at leisure". That makes it much easier to pick the highest and lowest voltage, and the highest and lowest temperature cells and monitor them more and more frequently as they approach critical values.

Again, the measurement results can be constantly compared to the average value = reference value for each cell under those conditions; if the measurement result is too different from the average, then the I-BMS could review the standard deviation and the minimum and maximum values for the individual cell under those circumstances (and repeat the measurement to rule out errors). That will almost instantaneously show if the cell is behaving just as it usually does, or not. Depending on how far off the (repeated = verified) measurement results are from the reference value, the cell gets flagged for more frequent or less frequent measurements.
 
Back
Top