Den 2007-12-17 02:14:54 skrev Apptech <apptech@paradise.net.nz>:

>> I'm working on a product that would use 6 series Ni-MH
>> cells and would like a
>> simple way to detect when the pack is at 6.0 and/or 5.4
>> volts.  This is a
>> second generation version of my Battery Top Power Supply,
>> see:
>> http://www.prc68.com/I/BTPS.shtml
>> so a minimalist solution is the only type that will work.
>
>
> Super minimalist.
> NPN transistor.
> Ra supply to base.
> Rb base to ground.
> Transistor is on when divider supplies above ABOUT 0.6V.
>
> Use that to do whatever.
>
> So threshold is R/Rb = (Vsupply-0.6)/0.6
>
> This is approximate due to the soft Vbe knee and what "on"
> means to you. Also watch temperature variability.
> BUT it's probably good for 10% accuracy (maybe better,
> depending on ... ). Assuming it's for battery protection
> even 20% would probably save your batteries well enough, as
> cell voltage goes down very rapidly below about 1V under
> normal loads. So a cell that is meant to cutout at say 1V
> and actually does so at 0.9V may not bother you too much
> (again "depending on ... ".)

If you only have acpuple cells in seres that is OK.
But as cells are not identical, especially after half life, sone will be below 1V while other at 1.2V, and it is nice to cut of before causing unecessary stress oin the less god cells.

> At the risk of being labelled as a "NZ designer" I'll admit
> to currently considering a similar scheme for a rough low
> battery cutout. It has limitations but is also good enough
> for some purposes.
>
> If you want precision then a cheap op amp (eg LM358) and a
> reference diode would do. Or even a 2 transistor long tailed
> pair and a reference diode - small and under $US0.20 parts
> cost all up (under $US0.10 in good volume). A zener MAY do
> well enough as a reference depending on what precision you
> want. As the cutout point is well defined the zener will
> have a well defined feed voltage at cutout (infinite
> regress) so it's better than in many zener applications.
>
>
>         Russell

I believe it is less total cost to simply change that NPN for a three pin uC reset IC with open drain output.
For example, Currently I use STM1061 to pull a FET gate to GND when one voltage is low.

They usually need very little supply and reference current, which is a plus in your case.
And use to be pretty accurate.  Some also have hysteresis and/or delay.
They aslo come with inverted logic variants. If not inverting, you can add hysteresis witha feedback R, and with series diode to that R make it latching.
Beware most such reset cirquits do not like more than 5V on the output.

If you want it to output positive current when batt is OK, consider a similar solution in positive side using LM385
(If i remember correctly...)


-- 
Morgan Olsson

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist