Measuring battery voltage while charging it

[sreenath]

Hi all,

I have a tricky problem on hand, but one that I am sure has been solved
by several of you out there.

I am building a battery charger cum load controller that uses a solar
panel for charging the battery. I use the PWM output of a PIC
microcontroller for charging the battery. However, the problem is
generic for any other microcontroller types.

If I measure the battery voltage while it is getting charge current
from the solar panel, the measured voltage tends to be more than the
actual battery voltage. If I give a small delay after switching off the
charging current(turning off a MOSFET controller by a microcontroller
pin), the voltage seems to be closer to the true battery voltage.

What is the method used commonly for measuring battery voltage while it
is getting charged? Is a delay as indicated above a good way of doing
it?
If so what should be the ideal delay between turning off the charging
current and measuring the battery voltage?

I wonder how the analog PWM controllers measure the voltage?

thanks for any suggestions,
Sreenath

 

[Charles Schuler]

Hi all,

I have a tricky problem on hand, but one that I am sure has been solved
by several of you out there.

I am building a battery charger cum load controller that uses a solar
panel for charging the battery. I use the PWM output of a PIC
microcontroller for charging the battery. However, the problem is
generic for any other microcontroller types.

If I measure the battery voltage while it is getting charge current
from the solar panel, the measured voltage tends to be more than the
actual battery voltage. If I give a small delay after switching off the
charging current(turning off a MOSFET controller by a microcontroller
pin), the voltage seems to be closer to the true battery voltage.

What is the method used commonly for measuring battery voltage while it
is getting charged? Is a delay as indicated above a good way of doing
it?
If so what should be the ideal delay between turning off the charging
current and measuring the battery voltage?

Are you making this too difficult? There is a correlation between the
charging voltage and the true battery voltage. Maybe the charging voltage
is all you need. Or, if you really want to get more useful data, you could
program a discharge cycle every 10 minutes or so and measure the battery
voltage as it supplies current.

 

[sreenath]

Thanks for the response.

It would be really great if I can get the battery voltage from charging
voltage.

From the charging voltage, how can I arrive at the battery voltage? Is

there a fixed voltage to be subtracted to arrive at this?

In the current project, introducing discharge cycles (to measure
battery voltage is not possible) since there are strict rules as to
when the load is to be switched on.

thanks,
Sreenath

 

[CWatters]

Lead Acid, NiCad and NiMH battery chargers typically measure the voltage on
charge...but it can be a problem if the charger doesn't supply a constant
current or is noisy (eg from PWM) and you are looking for the small delta V
peak on an NIMH cell.

 

[Charles Schuler]

In the current project, introducing discharge cycles (to measure
battery voltage is not possible) since there are strict rules as to
when the load is to be switched on.

How about a dummy load?

 

[H. Dziardziel]

Hi all,

I have a tricky problem on hand, but one that I am sure has been solved
by several of you out there.

I am building a battery charger cum load controller that uses a solar
panel for charging the battery. I use the PWM output of a PIC
microcontroller for charging the battery. However, the problem is
generic for any other microcontroller types.

If I measure the battery voltage while it is getting charge current
from the solar panel, the measured voltage tends to be more than the
actual battery voltage. If I give a small delay after switching off the
charging current(turning off a MOSFET controller by a microcontroller
pin), the voltage seems to be closer to the true battery voltage.

What is the method used commonly for measuring battery voltage while it
is getting charged? Is a delay as indicated above a good way of doing
it?
If so what should be the ideal delay between turning off the charging
current and measuring the battery voltage?

I wonder how the analog PWM controllers measure the voltage?

thanks for any suggestions,
Sreenath

Voltage is measured while charging but voltage alone is a very
unreliable charge state indicator. Manufacture data sheets give
the charging voltage-current-temperature-time curves for each cell
chemistry and size and charging methods. Keep in mind the curves
are for new cells but cell impedance increases and charge
acceptance and retention changes as they age.

 

[sreenath]

Hi,
thanks for all the useful information.

I think I will stick to measuring the battery voltage after connecting
the solar panel to it, and see how it goes.
While it is true that measuring the battery voltage while discharging
to a dummy load/actual load gives the TRUE indication of battery
condition, it does not seem to be practical in this case.

thanks,
Sreenath

 

[H. Dziardziel]

Hi,
thanks for all the useful information.

snip
While it is true that measuring the battery voltage while discharging
to a dummy load/actual load gives the TRUE indication of battery
condition, it does not seem to be practical in this case.

Please note unless the cell is nearly completely discharged, the
only thing meaningfully measured with a momentary load is cell
impedance. If the initial charge state and ensuing charging
current-time is unknown the state of charge can be measured only
by discharging the cell to a standard cell voltage and integrating
that current -time.curve area.

Smart charge controllers use cell chemistry parameters to
determine when a cell reaches its fully charged state from a given
initially charged state. E.g., voltage slope changes,
temperature, constant voltage with current level termination etc.
but none are as accurate as starting with fully discharged cells
and then charging fully per the manufacture recommended method/s.

Solar panels are constant current sources so the method may
actually be quite simple if say, NiMh are being used and the
charging current is around C/10.. Please do look over the cell
literarure.

 

[sreenath]

This design is for VRLA and VMLA types lead acid batteries, the battery
type IDP switch selectable. I am provided with peak voltage each cell
can be allowed to reach on "bulk charge" and the float level etc.

A couple of years ago I had made a similar charger using another PIC
device. With the 8 bit ADC of the device, it was quite an exercise to
get the voltages correct.

The customer used a digital multimeter to verify our design and the
voltage indicated by the digital multimeter was different from the
values we used in the design.

Of course the customer insisted that we change the code to match
readings from his multimeter.

-Sreenath