Re: My Battery Monitor

[Lee Hart <leeahart@earthlink.net>]

Steve Richardson stevenr@flash.net wrote:
> it seems to me we first have to define what we want to accomplish.

Excellent point. Different circuits are required for different
functions, or degrees of accuracy.

> To measure battery voltages only, I think can be done using the 555
> timer.

It does have a lot to recommend it. It's a simple, inexpensive circuit,
and it works. However, recognize that it is likely to have low accuracy
and high drift. Here are a few reasons.

Your power supply voltage regulator is also your reference voltage. You
didn't say what part you used, but let's assume it is an LM78L05. Its
specs are: 

 - initial output voltage at 25 deg.C, 40ma    5v +/-0.2v   (4%)
 - line regulation (Vin = 7v to 20v)           10mv typical (0.2%)
 - load regulation (Iout = 1ma to 40ma)         5mv typical (0.1%)
 - temperature stability (0 to 70 deg.C)       40mv typical (0.8%)
 - long term stability (drift per 1000 hours)  12mv typical (0.25%)

You can trim out the initial error, but the remaining error sources
remain. This regulator alone will thus typically be responsible for a
1-2% error.

Do the same thing for each part in your circuit. For an LM555 timer:

 - initial timing accuracy (25 deg.C, 5v supply)  2.25% typical
 - temperature stability (0 to 70 deg.C)          3.0% typical
 - drift with supply voltage (for a 1v change)    0.3% typical

Accuracy is also affected by the resistor and capacitor you use in the
timing circuit. For ordinary carbon film resistors:

 - initial resistance accuracy                    5%
 - temperature stability (-55 to +155 deg.C)      4%
 - long-term stability (per 1000 hours)           0.25%

For the 4.7uF capacitor, I hope you didn't use an electrolytic! They are
typically 20% initial accuracy, and change 100% over time and
temperature. Tantalums are 10% initial accuracy, and 10% over
temperature. A capacitor good enough so it is not the dominant source of
error will cost more and be larger than everything else in the circuit
put together! For example, a 4.7uF polypropylene film capacitor would be
2" long and 0.6" diameter, and cost about $5.

 - initial capacitance accuracy                    2%
 - temperature stability (0-70 deg.C)              1%
 - long-term stability (per 1000 hours)            0.2%

The final error tolerance is the sum of the parts. You can play some
statistical games (not all part errors will be in the same direction, or
at their maximum at the same time. But I only used *typical* errors
here; worst-case errors are several times larger. The manufacturer only
promises the part will meet worst-case specs, so the odds are that at
least some parts will be at their worst-case limits.

You can see that if you just "slap the circuit together", the initial
errors are huge. 4% + 2.25% + 5% + 2% = 13.25%. A 13.25% error in
measuring battery voltage makes the number useless; there is less than a
10% difference between 0% and 100% SOC for a lead-acid battery.

You can calibrate out the initial errors, but are still left with the
temperature, line, load, and other long-term stability errors. They
still total over 10% for this circuit. You can use better parts ($5
capacitors), hand-pick the parts, tweak each circuit, measure
temperature and compute out its errors, etc. but this gets to be so much
time and money that the circuit loses its "cheap and simple" appeal.
That's why I don't think it is a good choice for measuring absolute
voltage.

If you used just *one* of these circuits, with relays etc. to switch it
between batteries, you *could* make accurate relative measurements of
the voltage difference between batteries. If the circuit makes a 1v
error reading one battery, it makes the same error for all of them and
the difference cancels out. This is basically why I chose to use one A/D
converter in my Balancer.
-- 
Lee A. Hart                Ring the bells that still can ring
814 8th Ave. N.            Forget your perfect offering
Sartell, MN 56377 USA      There is a crack in everything
leeahart_at_earthlink.net  That's how the light gets in - Leonard Cohen