| |
G4WYW > TECH 09.04.08 13:00l 154 Lines 5817 Bytes #999 (0) @ WW
BID : 550283G4WYW
Read: GUEST
Subj: Charging Nicads part 3 of 4
Path: IZ3LSV<IW2OHX<IQ0LT<IK2XDE<DB0RES<ON0AR<GB7FCR
Sent: 080409/1014Z @:GB7FCR.#16.GBR.EU #:1553 [Blackpool] FBB-7.03a $:550283G4W
From: G4WYW@GB7FCR.#16.GBR.EU
To : TECH@WW
Part 3 of 4
THE REFLEX (R) CHARGE METHOD.
The ICS1700 Rapid Charge Controller employs a unique method of charging
batteries that solves some of the problems addressed previously. This
method,
patented as the Reflex (R) method and used under licence by ICS, consists
of a
positive current charging pulse followed by a high current, short duration
discharge pulse. The amplitude of the positive charging pulse is
determined by
the current capability of the power supply, the desired charge rate and
the
cell capacity. The ICS1700 is capable of controlling four user-selectable
rapid charge rates of 4C, 2C, 1C and half C, where is the rated battery
capacity in ampere-hours. These charge rates roughly translate into 20
minute,
45 minute, 1 and a half hour and 3 hour charge times.
The discharge pulse has an amplitude set at -2 and a half times the
charging
current. Thus the negative charge energy also varies as a function of cell
capacity but remains a fixed ratio to the positive charge rate.
"BURPING" THE BATTERY.
The primary purpose of the discharge pulse is to prevent the accumulation
of
gas bubbles on the cell plates. These gas bubbles are created under charge
by
the normal recombination of nickel hydroxide into nickel oxyhydroxide and
cadmium hydroxide into cadmium.
The generation of oxygen gas accumulates as bubbles across the cell
plates,
reducing the effective surface area and raising the internal cell
impedance.
Since the plate area is diminished, the time needed to completely charge
the
batteries increases.
The discharge pulse acts to "burp" a battery by stripping the bubbles away
from the plates and assisting in the recombination of oxygen at the
negative
plates. This depolarising process helps reduce the cell's internal
pressure,
temperature and impedance, turning the majority of the applied charge into
stored energy instead of gas and heat.
The charge/discharge pulse cycling helps to restore the crystal structure
of
the cell plates by breaking down crystal formations, eliminating "memory"
problems. The process also helps restore the crystal structure of the
cadmium annodes. Reducing the effect of crystal size problems enhances the
charging efficiency, allowing quicker charges at higher currents.
TROUGH VOLTAGE SENSING.
A 10 millisecond delay immediately follows the discharge pulse. No charge
is
applied during this delay, allowing the cell chemistry to recover. The
ICS1700
reads the no-load battery voltage during this "quiet" window. The no-load
voltage that it measures here contains less noise and is a more accurate
representation of a battery's true charge state. Since no charging current
is
flowing the measured cell voltage is not obscured by any internal or
external
IR drops or distortions caused by excess plate surface charge.
MAINTENANCE MODE.
Once the charger has determined that the cells under charge have reached a
full
charge state, the charger drops into a maintenance mode. The purpose of
this
mode is to keep the nickel-cadmium batteries primed at peak charge for
later
use. Since the charge on a nickel-cadmium battery naturally decays over
time at
a varying rate of C/30 to C/50, the maintenance mode provides a C/30
charge for
as long as the cells are connected to the charger.
The applied C/30 charge consists of the same charge/discharge pulse used
in
Reflex (R) Rapid Charging, only the duty factor of the pulse sequence has
been
extended. For example, the same charge/discharge pulse that happens every
second at 2C charge rate would now occur every 60 seconds for a
maintenance
rate of C/30. The charge/discharge pulses in the maintenance mode prevent
dentrite formations from propogating across the plates and separators. The
pulsing also helps maintain the crystal structure of the cell plates.
SLOPE TERMINATION.
By far the most distinctive point to look for is the peak in the cell
voltage
curve, indicating the transition from charge to overcharge. The voltage
peak
is characterised by a steep positive slope that flattens out, then turns
sharply negative. By taking the first derivative (dv/dt) of the cell
voltage,
a second curve can be drawn showing the change in voltage with respect to
time.
The first derivative will change its slope sharply in response to small
perturbations in the cell voltage curve. This sensitivity to the cell
voltage
helps monitor changes in the voltages of individual cells. The first
derivative
will peak just before the actual peak in the cell voltage. The rapid
charge
controller can then accurately terminate charge before the battery begins
overcharging.
To enhance the shape of the first derivative, a linear regression
algorithm can
find the best fit curve for any set of data points.
CHIP OPERATION.
The core of the ICS1700 is essentially a specialised reduced instruction
set
(RISC) microprocessor, optimised for efficient numerical calculations
since the
mathematics needed to derive the linear regression slope and determine the
correct termination point are fairly complex.
The controller uses a 10 bit successive approximation analogue to digital
converter (ADC) to change the measured analogue voltages to digital
voltage
numbers. These voltage numbers are then averaged with successive numbers
to
create an average voltage number. The averaging process limits the effect
of
voltage jumps caused by battery and ADC noise. The number of successive
voltage
numbers used in the average depends on the charge rate.
An infinate impulse response (IIR) filter weights the averaged voltage
number
to filter out any large aberrations in the cell voltage curve. The
filtered
average is stored in a twelve sample first-in first-out (FIFO) queue. The
twelve point FIFO holds the voltage numbers used to generate the slope.
73 - Mel, G4WYW
Read previous mail | Read next mail
| |
