Monitoring multi cell LiPo voltages

[DeepThought]

For a RC plane telemetry module, I would like to monitor individual lipo cell voltages.

The module is based on a 5V Arduino. The ADC converters are easily capable of reading in the cell voltages at 4.2V max. But obviously that is only true for the first cell referenced from ground.

How can I read three or more cells?
I was thinking of using analog opto couplers? Are they any good?

Also, I guess you could use a ladder of op-amps and use the cell - tab s a references. I would have to think hard about this one. I haven't really done any op-amp circuits on my own.
And it would be quite messy to wire in a space constrained environment like a glider fuselage.

I guess the easiest option would have to be a single chip solution. This way I may not even have to make a board.

What would you think? Is there a popular go-to solution? I would be thankful for any pointers here.

Thanks
 Tilman

[Paul Price]

Just use the magic of math, use subtraction. Have a measuring point on each battery's plus terminal and subtract the difference between it's + voltage and the + voltage of the batteries below it. You can use optoisolators to connect the battery voltage point to the each voltage divider resistor momentarily so the voltage divider resistors don't slowly discharge your batteries when you RC device is switched off.

[amyk]

There are ICs that will do this for you automatically and have an SPI/1-wire interface. TI makes a bunch.

[digsys]

Linear optos have their place, but too complicated for this application. Use differential op-amps eg - LT1990 LT1991 LT1995 LT1996
Some go to +- 250V input. Extremely easy to use and very accurate. I use these on 200VDC packs+

[Paul Price]

Digsys,

Perhaps I should clarify.  A bjt output optocoupler is just a LED turning on a bjt transistor, so it is a cheap and easy way to switch test points, that is, a way to connect each of the battery terminals to a voltage divider to a common A2D input, Each of the optp's are momentarily strobed on to make a quick measurement of a desired battery's + voltage in a series battery connection circuit. Then voltage subtraction does the rest.

I was not recommending using a linear in to linear output optoisolator IC to do this simple switching job.

[digsys]

I was not recommending using a linear in to linear output optoisolator IC to do this simple switching job. 

Don't sweat it PP :-) I was referring to the OPs reference in using a linear opto, not your reference !! Simple misunderstanding.
Now on my 8th version of BMs for solar/EV cars, I've tried ALL configurations, including versions that switch off voltage dividers.
What I didn't like about the voltage divider method - every cell with unique divider ratio / so either sacrifice resolution, "source"
impedance matching, temp drift etc. Pus, when switched off, the "switch" has to carry full pack voltage (maybe not a big deal with
just a few cells, but serious problems with 200-300V packs etc etc) Then I found the LT1990, and I fell in love. pix to follow :-)

[CodiJack]

First of all: what accuracy do you need?
Is 0.1V enough to view via telemetry the pack unbalance?
Are you going to have only info, that one of the cell is dropping below 3V?
Some rough calculations I did (no drift, tolerance, AD parameters, … considered):
Assuming 5 cells (4.2V max) gives 21V. By simple divider 4:1 max on analog input is 4,2V.
Dividing it by 8bits, it gives <20mV resolution.
Having divider for each cell and simple calculations you might get accuracy <0.1V.
Is such precision of pack unbalance enough to be displayed via telemetry?
P.S. I also fly RC and better precision for LiPo I require only from charger 

[DeepThought]

Thanks for all your answers.

Accuracy will not be critical . Basically, what I need is a low voltage alarm for the individual cells.

ti has some cool IC's indeed. Just needed the right search word. 
But I think I will really just go with a stack of op-amps. The more I think the more I realise how simple that is going to be. I just need a quad package if I can find one and I have my single chip solution. I don't care if I have to use three or four Arduino ADC.

The datasheet of the TI bq77PL157A4225 gave me the idea. In the internal block diagram you see the op-amp cascade.
I just need a bunch of differencial voltages at 1:1 amplification. Even I should be able to figure that out.

Will see if I can get this figured out in spice...

[salvix]

Hey DeepThought, did you get anywhere with this project?

I also fly R/C but I'm tired of the cheap cell meters that are uncalibrated (when compared to my Powerlab 8 charger), so I'm looking for a DIY solution too (not for telemetry, though).

Please let me know how far you got with your plan!

Thanks!