Page
1
of
16
BATTERY NEGATIVE RIPPLE CURRENTS IN BATTERY CHARGING
CIRCUITS
Colin F Bonham
oxbow consulting cc
Box 8963 Cinda Park 1463 RSA
Abstract:

Negative ripple currents in battery charging circuits are reputed to cause damage
to batteries and considerably short
en the lifetime. This paper does not set out to
analyse or explain the chemistry or the damage to batteries, but to indicate when
and where negative ripple currents are likely to occur, how to measure them and
how to reduce their magnitude. Some empirical
data is given on the various
types of chargers with varying levels of negative ripple current and the lifetime
performance of their associated batteries.
Introduction
DC supplies obtained from the ac mains by means of controlled or uncontrolled rectifier
s
(thyristor and diode rectifiers) have traditionally contained fairly large ripple voltages and
when applied to battery charging, lead to fairly large ripple currents into the battery. It has
generally been accepted that provided these ripple currents hav
e been limited to the order of
14% of the Ah
10
rating of the battery i.e. 14A rms for each 100Ah of battery capacity, the
effect on battery life is negligible. Because of changes to the structure of cells for economic
reasons this effect may no longer be n
egligible. In particular, when a battery is subjected to a
ripple current which is effectively charging the battery on the positive part of the ripple cycle
and discharging on the negative part, this effect may be particularly severe.
What are negative r
ipple currents?
Negative ripple currents occur where a ripple current exists at typical mains frequency
harmonics, 100,150 and 300Hz, and the magnitude of the peak of the negative part of the
ripple is greater than the dc current flowing into the battery
which is recharging or more likely,
float charging the battery. The battery sees this as repetitive charge

discharge many times
per second. Since batteries spend most of their lives in float there is plenty of time for this
negative ripple to do its damage
. If the ripple current is for example 50 amps peak positive
and negative as shown in Figure 1, then when it is superimposed on a large dc charge current
of 70A flowing into the battery, figure 2, there Is no negative region, it is all positive. As soon
as
the dc current falls below 50 amps a negative region starts to appear. Figure 3 shows a
large negative region with a dc current of 20A. It is very likely that a 50A ripple may be
present with a float current of below 1A!.
How are these large ripple curr
ents generated?
When a 50Hz sine wave is rectified it becomes two half sine waves side by side, which in
itself is a dc level with a distorted 100Hz ac waveform superimposed. The more phases that
are used in the rectification process, the smoother becomes
the dc or the smaller becomes
the ripple voltage appearing on the dc and the higher the ripple frequency. This is shown in
figure 4 for uncontrolled rectification (diodes). The waveforms are more complicated and the
ripple greater for controlled rectifica
tion (thyristors) but the principle is the same.
Page
2
of
16
I
Page
3
of
16
If this ripple voltage were to be applied directly to a battery, extremely large ripple currents
would flow because there is nothing to limit their flow into the effectively large capacitance
of
the battery. This would result in the rapid destruction of the system power components. A dc
choke is inserted between the battery and the bridge rectifier to limit the ripple current flow.
This is shown in Fig. 5a. Typically this choke is sized to allo
w a ripple current equivalent to
20% of the charger dc current which allows for an economical design. On a single phase
thyristor controlled system the choke is almost as large as the main transformer, because of
the low frequency and large ripple voltage.
On three phase systems, it is only about 10% of
the size of the main transformer. This is then likely to result in a maximum possible ripple
current of 5

10% of Ah
10
. If however, the battery is for a short standby time, this percentage
may approach 25% or
even more. With ripple currents of this magnitude regardless of
whether negative ripple exists, they are damaging to a battery. With a single choke filter only,
all ripple currents generated are taken up by the battery.
A capacitor bank is often us
ed in the smoothing of a charger, as shown in Fig. 5b, partly
because the charger must be able to deliver a relatively smooth dc output without a battery
connected and partly because some of the ripple that would otherwise go into the battery can
Page
4
of
16
absorbed
by the capacitors. The percentage of the ripple current absorbed by the capacitor
bank varies between 50 and 80% depending on the impedance presented by the connecting
cables of the battery.
This arrangement has an obvious advantage and is clearly mor
e expensive, but the problem
still exists that a charger that is providing current to a large standing load, say 70% of its full
load rating and producing its maximum ripple voltage and hence its maximum ripple current,
with a fully charged battery taking
a float current of less than 1 amp will still push a
proportionally far greater current ripple current into the battery and cause the negative ripple
problem.
A further problem is that the smoothing capacitors fitted in battery chargers are usually
electr
olytic capacitors and these have a finite lifetime depending on the operating voltage,
ripple current and operating ambient temperature. As the capacitors age, their electrolyte
dries out and their ability to absorb ripple current diminishes, forcing more
ripple current into
the battery until eventually the capacitors dry out altogether and all the ripple flows into the
battery.
The capacitor banks on some chargers and particularly the Siemens power station type are
fitted with series fuses. This results i
n a series inductance and resistance not only of the fuse
but also the extra wiring to reach the fuse mounted on a plate. This reduces the effectiveness
of the capacitor banks and the removal of the fuses and shortening of the wiring should be
seriously co
nsidered. Where the capacitors have good voltage, current ripple and
temperature margins this is an acceptable safety risk.
How can these ripple currents be measured?
Equipment
The usual methods of measuring currents may be employed, i.e. current shunt
s, current
transformers and Hall Effect probes, but care must be exercised making these measurements
as the currents may be small and not easily measured amongst the electrical noise pollution
that may exist due to thyristors or transistors switching. A Ha
ll effect probe with a modern
digital storage oscilloscope/multimeter is probably the best means of observing and
measuring the interaction of the dc and ac currents into the battery and the resultant negative
ripple. The waveform can be seen together with
the offset caused by the dc charge current
and will clearly show the negative ripple region below the zero line. The rms value of the
ripple current and the dc current into the battery can be measured on the digital multimeter. If
the charger is associate
d with a modern UPS a lot of electrical noise and interference can be
expected from the high frequency switching of the inverter. If Hall Effect equipment is not
available, a good digital multimeter or oscilloscope can be used across the battery shunt, but
a lot of noise can be expected in the readings because of the very low output voltage from the
shunt.
The measurement of battery and capacitor ripple currents may also be carried out
using standard ac current tongs and an associated ammeter or voltmeter d
epending on the
type of tong.
Page
5
of
16
In my experience the simplest and best equipment for ripple measurement is an ordinary tong
C.T. used with a high quality R.M.S. reading digital multimeter. The tong and meter must of
course be in proper calibration and the d
c currents must be very low at the time of
measurement, but otherwise are easy to use, robust, insensitive to noise and interference and
give very consistent readings, regardless of temperature, position of cable in tong aperture or
any other influencing f
actors. In particular, an excellent comparative figure of ripple current
into the capacitors versus the ripple current into the battery can be obtained.
It is worthwhile to measure the ripple current into the smoothing capacitor if fitted, to ascertain
t
he proportion of ripple current going to the capacitor. This may be done by the same means
as used for the battery current. The ripple voltage across the capacitor bank should also be
measured
with a true rms reading digital voltmeter,
as this with the rip
ple current can be used
to calculate the effective capacity of the capacitors and used to determine whether the
capacitors should be replaced or not.
An oscilloscope should also be used to examine the actual ripple waveform, particularly the
balancing. Th
e modern handheld Scopemeter can also make the full range of measurements
including the rms voltage and current measurements.
The size of the current tong aperture is critical when measuring the battery current ripple on
the large chargers for example. If
for example the copper size is 60 x 10mm and the battery
cable is 630mm
² in cross sectional area which requires an aperture of greater than 60mm
diameter.
The tongs must also be able to accept a small dc current without the accuracy being affected.
This
dc current is the battery float charging current that will always be present even when the
battery is fully charged. The dc currents associated with the capacitor ripple current
measurements are negligible. Once a set of instruments has been decided upon,
it is
advisable to continue using these instruments to make all the necessary measurements
within the region or power station. They should also be made available for comparative tests
with equipment from other regions and power stations to assist in buildi
ng standard
procedures and data bases that can be directly compared to provide information on any
peculiarities of particular sites.
The capacitor bank dc capacitance may also be measured using a suitable power supply
300V @ >100mA with voltage regulation
and current limiting and a timer. If current limiting is
not available then an accurate 1K0 series power resistor must be used instead.
Method
It is very important to first check that the charger system is operating properly, otherwise the
results obtai
ned may be meaningless. In particular the ripple balance must be checked and
adjusted, if necessary according to the operating manual instructions. A final check may be
made by connecting the oscilloscope across the capacitor bank and adjusting the balance
to
give the smoothest 300Hz ripple with virtually no trace of 100Hz ripple superimposed on the
300Hz.
The integrity of the capacitor circuit must also be checked. If there is a series fuse it must be
intact and the wiring must have good connections. The
slightest resistance in the circuit will
have a dramatic effect on the balance of the ripple current between the capacitor and the
battery.
The charger/battery/load system to be measured must be in a quiescent state. That means
the charger is in Float, th
e load current is normal and as constant and as high as possible, the
battery current is as close to zero as is possible and the mains supply is balanced and stable.
This will ensure that the dc currents into the capacitor bank and the battery remain const
ant
and at a minimum, ensuring low magnetisation of the current tong core.
Page
6
of
16
The current into the battery is measured simply by placing the current tong around the battery
cable at any convenient point between the battery isolator and the battery terminals
on the
negative or positive pole.
The current into the capacitor bank is measured simply by placing the current tong around the
conductor feeding the capacitor at any convenient point on the negative or positive pole. It is
important not to include any co
nductors carrying dc currents.
The ripple voltage is measured directly across the capacitor terminals.
The ripple voltage must also be measured on the main busbars and this will give an idea of
the effectiveness or efficiency of the capacitor circuit.
T
he dc capacitance is measured by removing the capacitor fuse or otherwise disconnecting
the capacitor bank. The power supply is adjusted to the rated voltage of the capacitor bank or
in the case of the capacitors on high voltage chargers as high as possibl
e and the current limit
set to 100mA. The capacitor is initially fully discharged. The time to charge the capacitor from
zero to the voltage applied to the capacitor from the power supply, is measured.
If the resistor must be used, then the time from zero
to 63,2% of the voltage applied to the
capacitor from the power supply.
Results and Calculations
To determine the existence of a negative or discharging ripple current:

DC battery current
–
rms battery ripple current x 1,414
If this results in a nega
tive value then negative ripple exists and additional
smoothing is required.
To determine the efficiency of the capacitor bank:

Capacitor bank ripple / Busbar ripple voltage x 100%
To determine the proportion of total ripple current absorbed by the cap
acitor bank:

Capacitor bank ripple x 100%/(Capacitor bank ripple + Battery ripple)
To determine the proportion of total ripple current absorbed by the battery:

Battery ripple x 100%/(Capacitor bank ripple + Battery ripple)
To determine the effective
capacity of the capacitor bank:

X
c
= Capacitor ripple voltage/capacitor ripple current
C
eff
= 1/(2 x pi x f x X
c
) in
µF
C
eff
1/(2 x 3,1416 x 300 x X
c
) in
µF
C
rated
= Number of capacitors in bank x capacitance
If C
eff
/C
rated
= <0,8 then capacitors sho
uld be replaced
To determine the dc capacity of the capacitor bank:

C
dc
= (I
lim
x t x 10
6
)/V in
µF
Where I
lim
is the current limit setting of the power supply in amps
V is the voltage limit setting of the power supply in volts
t is the time measure
d in seconds
or
C
dc
= (R x t x 10
6
) in
µF
Where t is the time measured in seconds
R is the series resistance in ohms
The measured values can be inserted into the attached Excel spreadsheet and the
calculations will be made automatically. The charger
type, reference no., dc voltages and
currents should also be recorded for future reference.
Page
7
of
16
Table 1 Ripple measurements on Siemens Battery Chargers supplied to Eskom
Site
Charger Type
Charger DC Current
Battery
Capacitor
Battery
Capacito
r
Measured
Rated
Comments
Float Current
AC Current
AC Current
AC Voltage
Capacity
Capacity
Stikland 110V Main 1
110V/35A (BA17)
6,1A = 17,4% of FL
0,5A
4,2A = 69%
1,9A = 31%
0,22V
10 128uF
9 400uF
Caps OK
Stikland 110V Main 2
110V/35A (BA17)
6,8A =
19,4%of FL
0,3A
4,7A = 69%
2,1A = 31%
0,25V
9 972uF
9 400uF
Caps OK
Stikland 50V Main 1
50V/100A (BA17)
12,1A = 12,1% of FL
0,5A
10,4A = 98%
0,2A = 2%
0,30V
18 391uF
20 000uF
Caps low
Muldersvlei 110V Main 2
220V/35A (BA17)
4,3A = 12,3% of FL
0,2A
3,7A
N
ot meas.
Not meas.

4 400uF
Muldersvlei 110V Main 2
220V/35A (BA17)
10,1A = 28,9% of FL
0,2A
5,7A
Not meas.
Not meas.

4 400uF
Brenner 220V Main 1
220V/75A (BA17)
5,1A = 6,8% of FL
0,7A
5,0A = 59%
3,3A = 41%
0,54V
4 912uF
7 200uF
Caps low
Brenner 22
0V Main 2
220V/75A (BA17)
5,1A = 6,8% of FL
0,7A
6,8A = 67%
3,4A = 33%
0,58V
6 241uF
7 200uF
Caps low
Brenner 50V Main 1
50V/120A (BA9

G)
23A = 19,2% of FL
1,0A

11A = 100%



No caps
Brenner 50V Main 2
50V/120A (BA9

G)
17A = 14,2% of FL
1,0A

8A =
100%



No caps
Tutuka Station 1
220V/280A
55A = 19,6% of FL
<1,0A
37,8A = 76%
11,9A = 24%
1,38V
14 535uF
16 200uF
Caps low
Tutuka Station 2
220V/280A
55A = 19,6% of FL
<1,0A
38,1A = 75%
12,6A = 25%
1,37V
14 737uF
16 200uF
Caps low
Bacchus 220V Main
1
220V/75A (BA12)
6,4A = 8,5% of FL
0,19A
0A = 0%
6,0A = 100%


14 400uF
Caps N/C
Bacchus 220V Main 2
220V/75A (BA12)
5,4A = 7,2% of FL
0,16A
2,6A = 48%
2,8A = 52%
0,385
7 165uF
14 400uF
Caps v/low
Bacchus 50V Main 1
50V/240A (BA9

G)
24,0A = 10,0% of F
L
0,7A

23,9A = 100%



No caps
Bacchus 50V Main 2
50V/240A (BA9

G)
19,5A = 9,1% of FL
1,4A

19,0A = 100%



No caps
Palmiet Station 1
220V/200A
34A = 17% of FL
0,5A
27,5A = 71,2%
11,1A = 28,8%
1,23V
4 787uF
10800uF
Caps v/low
Palmiet Station 2
220V/200A
15A = 7,5% of FL
0,3A
16A = 67%
8,0A = 33%
0,74V
5 735uF
10800uF
Caps v/low
Palmiet 50V Main 1
50V/60A (BA9

G)
6A = 10% of FL
0,7A

6,5A = 100%



No caps
Palmiet 50V Main 2
50V/60A (BA9

G)
8,5A = 14,2% of FL
0,2A

7,0A = 100%



No cap
s
Palmiet 50V Telephone
50V/60A (BA9

S)
9,5A = 15,3% of FL
0,1A
5,5A = 71,4%
2,2A = 28,6%
0,142V
61 644uF
60 000uF
Caps OK
Palmiet UPS Main 1
Fuji 220V/136A
40A = 29,4% of FL
0,86A

30A = 100%



No caps
Palmiet UPS Main 2
Fuji 220V/136A
30A = 22,0%
of FL
0,80A

24A = 100%



No caps
Tutuka Unit 4 A1
24V/1030A
>250A = >25% of FL
<1,0A
59,1A = 80,1%
14,7 = 19,9%
0,235
133430
168000
Caps low
Tutuka Unit 4 B1
24V/1030A
>250A = >25% of FL
<1,0A
62,0A = 75,2%
20,4 = 24,8%
0,270
121818
168000
Caps lo
w
Page
8
of
16
Table 2 Matimba P/S Charger Ripple Measurements
Charger
Type
Chrgr.
No.
Charger
Volt. (V)
Charger
Curr. (A)
Battery
Curr. (A)
Capacitor
Volt. (mV)
Busbar
Volt. (mV)
Smooth.
Effic. (%)
Capacit.
Curr. (A)
Batttery
Curr. (A)
Capacit.
Curr. (%)
B
attery
Curr. (%)
Measured
Cap. (uF)
Rated
Cap. (uF)
Cap.
(%)
D220/170
1
241
30
0,0
1580
1510
104,6
22,4
17,2
56,6
43,4
7521
7200
104,5
D220/170
2
241
48
0,0
1600
1530
104,6
23,0
16,7
57,9
42,1
7626
7200
105,9
D24/1740P
1A1
29
480
2,0
47
141
33,3
27,9
4,
0
87,5
12,5
314924
280000
112,5
D241740P
1B1
28,8
340
1,0
66
155
42,6
27,3
11,9
69,6
30,4
219441
280000
78,4
D24125N
1A2
26,8
11,9
0,1
118
217
54,4
4,0
4,5
47,1
52,9
17984
28000
64,2
D24125N
1B2
26,8
9
0,1
109
206
52,9
4,5
4,8
48,4
51,6
21902
28000
78,2
D220/485
1C
246
5
1,0
280
238
117,6
11,2
4,4
71,8
28,2
21221
21600
98,2
D24/1740P
4A1
29,1
416
15,0
53,8
112
48,0
26,5
11,3
70,1
29,9
261314
280000
93,3
D241740P
4B1
29,1
373
10,0
55,4
115
48,2
27,9
11,6
70,6
29,4
267173
280000
95,4
D24125N
4A2
26,7
1
2
1,0
98
160
61,3
5,7
4,8
54,3
45,7
30857
28000
110,2
D24125N
4B2
26,8
4,6
0,0
87
120
72,5
5,1
6,4
44,3
55,7
31099
28000
111,1
D220/485
4C
232
10
5,0
350
298
117,4
16,6
4,8
77,6
22,4
25162
21600
116,5
D24/1740P
6A1
28,9
412
3,0
45,2
126
35,9
25,1
12,0
6
7,7
32,3
294601
280000
105,2
D241740P
6B1
28,9
346
3,0
49,5
133
37,2
25,4
12,6
66,8
33,2
272225
280000
97,2
D24125N
6A2
26,7
12
0,5
89,3
153
58,4
4,1
4,5
47,7
52,3
24357
28000
87,0
D24125N
6B2
26,7
8
0,5
85,5
135
63,3
3,8
4,2
47,5
52,5
23579
28000
84,2
D220/485
6C
233
7
2,0
301
296
101,7
12,3
3,0
80,4
19,6
21679
21600
100,4
D24/400P
01
28,7
100
0,0
122
130
93,8
17,9
6,4
73,7
26,3
77838
84000
92,7
D24/400P
02
28,8
100
0,0
132
168
78,6
19,1
7,6
71,5
28,5
76764
84000
91,4
Page
9
of
16
Table 3a Kendal P/S Charger
Ripple Measurements
Charger
Type
Chrgr.
No.
Charger
Volt. (V)
Charger
Curr. (A)
Battery
Curr. (A)
Capacitor
Volt. (mV)
Busbar
Volt. (mV)
Smooth.
Effic. (%)
Cap.
Curr. (A)
Battery
Curr. (A)
Cap.
Curr. (%)
Battery
Curr. (%)
Measured
Cap. (uF)
Rated
Cap. (uF
)
Cap.
(%)
D220/180
243
17
1,0
1040
980
106,1
12,0
4,8
71,4
28,6
6121
7200
85,0
D220/180
243
44
1,5
1140
1070
106,5
13,3
5,8
69,6
30,4
6189
7200
86,0
D24/1740P
1A1
29,1
485
5,0
65
169
38,5
26,7
3,3
89,0
11,0
217920
280000
77,8
D241740
P
1B1
29,1
465
5,0
84
144
58,3
22,3
2,6
89,6
10,4
140839
280000
50,3
D24130N
1A2
29,1
15,6
0,5
88
184
47,8
2,8
3,4
45,2
54,8
16880
28000
60,3
D24130N
1B2
29,1
15,6
0,5
100
202
49,5
2,6
2,8
48,1
51,9
13793
28000
49,3
D220/525
1C
246
11
3,0
333
242
137,6
12,0
0,8
93,8
6,3
19118
21600
88,5
D24/1740P
2A1
29,2
515
5,0
52
140
37,1
22,3
1,5
93,7
6,3
227510
280000
81,3
D241740P
2B1
29,1
355
2,0
69
149
46,3
23,3
1,2
95,1
4,9
179145
280000
64,0
D24130N
2A2
29
15,2
0,6
106
205
51,7
3,5
1,6
68,6
3
1,4
17517
28000
62,6
D24130N
2B2
29,2
7,7
0,6
94
185
50,8
2,6
1,5
63,4
36,6
14674
28000
52,4
D220/525
2C
244
11
4,0
410
345
118,8
13,5
1,7
88,8
11,2
17468
21600
80,9
D24/1740P
3A1
29,1
426
3,0
61
134
45,5
24,7
1,6
93,9
6,1
214816
280000
76,7
D241740P
3B1
29,1
450
3,0
61
141
43,3
24,2
1,4
94,5
5,5
210467
280000
75,2
D24130N
3A2
29,2
16
0,5
88
288
30,6
3,0
2,5
54,5
45,5
18086
28000
64,6
D24130N
3B2
29,1
6
0,5
81
250
32,4
2,1
2,1
50,0
50,0
13754
28000
49,1
D220/525
3C
244
10
2,0
350
290
120,7
12,0
1,3
90,2
9,8
18189
21600
84,2
Page
10
of
16
Table 3b Kendal P/S Charger Ripple Measurements continued
Charger
Type
Chrgr.
No.
Charger
Volt. (V)
Charger
Curr. (A)
Battery
Curr. (A)
Capacitor
Volt. (mV)
Busbar
Volt. (mV)
Cap.
Curr. (A)
Battery
Curr. (A)
Ca
p.
Curr. (%)
Battery
Curr. (%)
Measured
Cap. (uF)
Rated
Cap. (uF)
Cap.
(%)
D24/1740P
4A1
29,2
322
3,0
59
159
37,1
22,2
1,7
92,9
7,1
199618
280000
71,3
D241740P
4B1
29,2
462
3,0
52
150
34,7
21,4
1,5
93,4
6,6
218328
280000
78,0
D24130N
4A2
29
10
0,5
127
2
12
59,9
3,8
1,7
69,1
30,9
15874
28000
56,7
D24130N
4B2
29,2
12
0,5
124
214
57,9
3,9
1,6
70,9
29,1
16686
28000
59,6
D220/525
4C
244
9
1,0
329
248
132,7
11,6
1,0
92,1
7,9
18705
21600
86,6
D24/1740P
5A1
29,2
349
5,0
70
167
41,9
19,7
2,0
90,
8
9,2
149302
280000
53,3
D241740P
5B1
29,2
426
5,0
125
192
65,1
17,0
1,6
91,4
8,6
72150
280000
25,8
D24130N
5A2
29,2
14
0,5
140
265
52,8
4,7
2,4
66,2
33,8
17810
28000
63,6
D24130N
5B2
29,1
8
0,5
109
231
47,2
3,5
2,0
63,6
36,4
17035
28000
60,8
D220/525
5C
243
?
?
251
195
128,7
10,6
0,9
92,2
7,8
22404
21600
104
D24/1740P
6A1
29,2
425
2,0
57
130
43,8
19,1
1,4
93,2
6,8
177770
280000
63,5
D241740P
6B1
29,2
369
2,0
91
142
64,1
20,0
1,3
93,9
6,1
116597
280000
41,6
D24130N
6A2
29,2
10
0,5
113
219
51,6
3,5
1,7
67,3
32,7
16432
28000
58,7
D24130N
6B2
29,2
11
0,5
119
222
53,6
3,8
1,9
66,7
33,3
16941
28000
60,5
D220/525
6C
246
8
2,0
260
205
126,8
10,5
1,0
91,3
8,7
21425
21600
99,2
Page
11
of
16
What results have been measured on various installations aroun
d South Africa?
Measurements have been carried on many Eskom sites throughout South Africa on various
radio sites, distribution stations and in particular Power stations. In Tables 1, 2 & 3, the
pattern is very clear in every case: where no capacitors are
fitted all the considerable ripple
current is taken by the battery, where capacitors are fitted generally less than half of this
ripple is taken by the battery but these currents are still magnitudes higher than the battery dc
float current and consequent
ly always produce negative ripple.
On the power stations, measurements on a relatively large number of chargers were made,
with the chargers and their batteries of various ages depending on the point in time when
each unit was commissioned. At Matimba for
example, on Unit 1, the oldest chargers, the
capacitors are beginning to age.
Another interesting comparison is between the virtually identical D24/1740P chargers at
Matimba and Kendall. At Matimba, the connection to the battery is very short indeed to
m
inimize dc volt drop, by means of busbars about 4 metres long through the separating wall
to the battery. This gives on average a battery ripple current of 30% of the total. At Kendal,
the station architecture necessitated long cables, the two poles being
separately routed
between the charger and battery. This gives a battery ripple current of 10% or less of the total
which is quite remarkable and has led to much longer battery lifetimes for the positive 24V
systems at Kendal than at Matimba.
How can the
magnitude of these ripple currents be reduced?
There is one group of thyristor battery chargers that do not have any measurable ripple
currents. These are the chargers specially designed for telecommunications use and
particularly for telephone use. To pr
event audible noise on the telephone circuits a double
choke capacitor filter is used to reduce the ripple voltage to below 100mV rms with or without
a battery. The ripple currents produced by this low voltage are negligible.
Obviously this equipment
is a lot more expensive with its double filter, but all that is required
to improve the ripple current on a charger with one LC filter is one more small choke. This is
connected in series with the battery anywhere between the existing filter capacitor and
the
battery. Its effect is to increase the impedance through the battery circuit so that the ripple
flows virtually entirely through the filter capacitor. Obviously the filter capacitor must be
already fitted in the charger and must be in good condition.
If not one must be fitted taking
into consideration the ripple current from the choke, operating temperature and voltage.
Page
12
of
16
The following is a list of Siemens chargers built to a wide variety of specifications and having
a variety of ripple characteris
tics and have correlated very closely with battery lifetime
problems. Not all types of Siemens chargers exhibit the ripple problem and these have had
long lived batteries. They can be grouped as follows:

1.
50V chargers
–
contract supervised by John Wessels
beginning 1979.
Single Special 50V/15A, 30A, 60A & 120A
–
Psophometric smoothing
–
no ripple
problem, batteries have lasted well.
Single General 50V/15A, 30A, 60A & 120A
–
Choke only, all ripple into battery
–
big
problem, batteries have lasted for half
normal lifetime.
Dual General 50V/60A, 120A & 240A
–
Choke only, all ripple into battery
–
big problem,
batteries have lasted for half normal lifetime.
2.
110/220V chargers

contract supervised by Ron Coney beginning 1983.
Single 110V and 220V/20A & 40A
–
Psophometric smoothing
–
no ripple problem,
batteries have lasted well.
Dual 110V and 220V/20A & 40A
–
Psophometric smoothing
–
no ripple problem, batteries
have lasted well.
Dual 220V/75A & 100A Special
–
Choke & capacitor, <50% ripple into battery
–
problem, battery lifetime is shortened.
3.
50V & 110/220V chargers
–
contract supervised by Leon Drotsche beginning 1987.
Single 50V/25 to 50A
–
Psophometric smoothing
–
no ripple problem, batteries have
lasted well.
Dual 50V/25 to 400A
–
Psophometric sm
oothing
–
no ripple problem, batteries have
lasted well.
Single 110V/25 to 50A
–
Choke & capacitor, <50% ripple into battery
–
problem, battery
lifetime is shortened.
Dual 110V/25 to 400A
–
Choke & capacitor, <50% ripple into battery
–
problem, batte
ry
lifetime is shortened.
Single 220V/25 to 100A
–
Choke & capacitor, <50% ripple into battery
–
problem,
battery lifetime is shortened.
Dual 220V/25 to 300A
–
Choke & capacitor, <50% ripple into battery
–
problem, battery
lifetime is shortened.
Page
13
of
16
4.
220
V & 24V power station chargers
–
contracts supervised by various Eskom staff.
Single 220V/180 to 660A
–
Choke & capacitor, <50% ripple into battery
–
problem,
battery lifetime is shortened.
Single 24V/125 to 1740A
–
Choke & capacitor, <50% ripple into
battery
–
problem,
battery lifetime is shortened.
5.
Bearer 50V chargers
–
contract supervised by Derrick Dekker beginning 1993.
Dual 50V/100, 200 & 300A
–
Psophometric smoothing
–
no ripple problem, batteries have
lasted well.
What improvements have be
en measured where the additional choke has been fitted?
A few chargers have been fitted with an additional choke but unfortunately not all of them
have been properly measured again to assess the improvement. For example, Bacchus,
Muldersvlei and Stikland.
Everest had chokes fitted just after manufacture, but the site has
not been visited since installation to measure the effects of the chokes.
One of the chargers on a D50/240 dual at Beta substation was modified to improve the
smoothing of the load circui
ts to allow the telephone PABX to be connected to it. This was
done by adding a capacitor bank and including a choke in series with the battery. However,
the final measurements with the loads connected, showed a much higher ripple current into
the battery
than had been expected. This appears due to ripple currents generated by the
loads and should be the subject of further investigation.
The best example of the effectiveness of the chokes is at Palmiet Pumped Storage Scheme,
where four Siemens chargers and
two Fuji chargers (in UPS’s) were modified. The two
Station chargers had the capacitor banks replaced with new capacitors and chokes fitted to
reduce the battery ripple from 22,5% to a calculated 1% of charger output current, i.e. 2A. The
two 50V comms ch
argers had capacitors and chokes fitted to reduce the battery ripple from
20% to a calculated 1% of charger output current i.e. 0,6A. The two chargers in the Fuji
UPS’s also had capacitor banks and chokes fitted to reduce the battery ripple to 1% of
charge
r output i.e. 1,4A.
Generally this was a very successful exercise, resulting in achieving and exceeding the ripple
levels calculated, so that the ripple was divided between the capacitor and the battery in the
ratio of 24:1 or better. A problem occurred o
n the UPS chargers because the calculations
were made on the basis of a 600Hz ripple from the charger which was correct, but
cogniscence was not taken of the ripple generated by the inverter in the UPS which is
generated at 100Hz and is not affected very m
uch by the inclusion of the choke. The
improvement was a current division of 2:1 between capacitor and battery, but still a reduction
of the ripple into the battery by a factor of 3. New chokes have been designed and will be
supplied and fitted at a later
date.
Again this problem outlines the need for investigation into the types of noise, interference and
ripple generated by the loads themselves and feeding currents into the battery which could
otherwise be absorbed by a capacitor bank.
All the batteries
for the Palmiet chargers will be replaced in the next few years and it will be
interesting to see how the battery lifetimes improve.
Page
15
of
16
Table 4 Charger Ripple Measurements Where Additional Chokes Have Been Fitted
Charger
Site
Charger
Type
Chrger
Volt.(V)
Chrger
Curr.
(A)
Battery
Curr. (A)
Capacitor
Volt. (mV)
Busbar
Volt. (mV)
Smooth.
Effic. (%)
Capacit.
Curr. (A)
Batttery
Curr. (A)
Capacit.
Curr. (%)
Battery
Curr. (%)
Measured
Cap. (uF)
Rated
Cap. (uF)
Cap.
(%)
Stikland
D110/35
119,6
480
2,0
47
141
33,3
27,9
4,0
87,5
12,5
314924
280000
112,5
Bacchus
D220/75
239,2
30
0,0
1580
1510
104,6
22,4
17,2
56,6
43,4
7521
7200
104,5
Muldersvlei
D220/35
239,2
48
0,0
1600
1530
104,6
23,0
16,7
57,9
42,1
7626
7200
105,9
Beta (before)
D50/240
55,2
25
<1,0




47

100



Beta (after)
D50/240
55,2
25
<1,0
290
275
94,8
58
2
96,7
3,3
159155
120000
132,6
Everest
D220/300
246










26400

Palmiet (before)
D220/200
232
34
<1,0
720
675
93,8
16,2
8,1
66,7
33,3
11936
10800
110,5
Palmiet (before)
D220/20
0
232
15
<1,0
1010
1100
108,9
24,5
11,1
68,8
31,2
12869
10800
119,1
Palmiet (before)
E50/60
56,0
9
<1,0

321


7,7

100



Palmiet (before)
E50/60
56,0
10
<1,0

287


6,6

100



Palmiet (before)
D220/136
259
40
<1,0

1320


24,5

100



Palmiet (before)
D220/136
259
30
<1,0

1160


22,2

100



Palmiet (after)
D220/200
234
35
<1,0
650
650
100
17,6
0,7
96,2
3,8
14365
13200
108,8
Palmiet (after)
D220/200
234
17
<1,0
900
870
103,4
22,5
0,83
96,4
3,6
13263
13200
100,5
Palmiet
(after)
E50/60
56,2
30
<0,5
333
299
111,4
9,3
0,39
96,0
4,0
44449
40000
111,1
Palmiet (after)
E50/60
56,2
15
<0,5
215
211
101,9
6,2
0,21
96,7
3,3
45896
40000
114,7
Palmiet (after)
D220/136
255
40
<1,0
1230
1230
100,0
16
8
66,7
33,3
20703
13200
156,8
Pa
lmiet (after)
D220/136
255
30
<1,0
1100
1100
100,0
13
7
65,0
35,0
18809
13200
142,5
Page
16
of
16
What effects do these high negative ripple currents have on batteries?
Batteries are damaged by three main factors, namely over voltage, over temperature and
ripple cu
rrent. The effects of ripple current and its corresponding ripple voltage cause
overcharging and undercharging due to the peak ripple voltage, temperature rise due to the
heating effects of the ac current flowing in the battery and where the ripple current
is larger
than the dc charging current causes shallow but rapid cycling of the battery.
The lead acid battery responds rapidly to charging and discharging up to frequencies of
approximately 1000Hz, but beyond that the battery does not recognize these cha
nges in
current flow direction.
A new battery has the ability to provide a number of ampere hours down to a defined depth of
discharge and repeat this for a certain number of cycles, i.e. 1000Ah to 100% depth of
discharge for 500 cycles, This means that i
n theory the battery will be able to provide a total
discharge of 500,000 Ah over its useful lifetime.
The negative ripple current constitutes a real discharge of the battery during the period of the
negative ripple. This can be calculated from the area o
f the negative ripple waveform below
the zero line and multiplying by the number of cycles. More simply and readily the rms ripple
current may be measured, subtract the measured dc float current and divide by two, giving
the discharging amperes. The ripple
current is usually more than a factor of ten greater than
the float current making the dc offset insignificant.
If the above 1000Ah battery is fed by a 250A charger that can produce a maximum ripple of
50A (limited to 20% of dc output), the negative part
of the ripple is in the worst case (50

0,5)/2A = 25A approx. If this current flows from the battery for 20 000 hours the total ampere
hours will have been used up. 20 000 hours is only 2,3 years.
A case like the one above is unlikely in a well maintained
system, but it does illustrate the
rapid effects of discharging ripple and in practice the available ampere hours may be used up
in 5

8 years.
It is also possible that the situation could be considerably worse if a charger has not been
properly maintaine
d and is no longer properly balanced. This may be due to drift of the
settings or even worse thyristor or electronic failure resulting in lower frequency ripple
voltages, the resulting currents of which pass almost unhindered through the smoothing
circuits
.
Conclusions
If it seems that the lifetime of batteries in a particular installation is being shortened, the
possibility that it is cause by negative ripple effects is very possible. The presence of negative
ripple currents can easily be measured and f
or a cost of as little as 5% of the cost of the new
charger, the lifetime of the battery may be substantially extended. On new thyristor equipment
it may be advisable to fit the additional choke during manufacture to reduce costs even further
and not expos
e the battery to ripple currents at all during its working life.
Comments 0
Log in to post a comment