Bateman EW 12 2002 mar2003 1uF electrolytic or film.pdf

(619 KB) Pobierz
Updated & expanded March 2003
Original version Pub Electronics World December 2002.
Many capacitors do introduce distortions onto a pure sinewave test signal. In some instances distortion results from the
unfavourable loading which the capacitor imposes onto its valve or semiconductor driver. More often, the capacitor generates
the distortion within itself.
Capacitor generated distortions, for too long the subject of much speculation and opinion, can now be measured. Capacitors are
not categorised for distortion in manufacture, so a distorting capacitor would not be accepted as reject by its maker. Using my
easily replicated test method, audio enthusiasts can select capacitors when upgrading their equipment and designers can select
capacitors for each circuit requirement.
For 100 nF capacitance we find the lowest distortions are generated by choosing either C0G multilayer ceramic, metallised film
Polyphenylene Sulphide (PPS) or double metallised film electrodes with Polypropylene (PP) film.
Ref.1
At 1
µF,
C0G ceramic types are not generally available, reducing our low distortion choice to the above two film types or a
selected metallised Polyethylene Terephthalate (PET). To guarantee low distortion we found that metallised PET types should
be distortion tested and used with no bias or with modest DC bias voltages. The PPS and PP capacitor types produce
exceptionally low distortions but are larger and more expensive.
see
Fig. 1
Fig. 1)
Top row 0.1µF,
the 50v and 100v SMR
capacitors second and
fourth, the B32652 fifth
from left. Far left is the
BC Components type
470 met PET, third from
left is the 100 nF COG
multilayer ceramic.
Bottom row left 1µF, the
best electrolytic, the Bi-
polar, was outperformed
by the 470 type 63v
metallised PET capacitor.
The SMR capacitor is
fourth and the B32653
fifth from left. Finally we
have a type 338 MKP
class ‘X2’ capacitor and
378 MKP both stocked
by most distributors.
To minimise costs at 1
µF
and above, many designers elect to use low cost polar aluminium electrolytic capacitors. We now
explore this option.
Electrolytic capacitors.
At room temperature and 1 kHz, a typical 1
µF
63 volt polar electrolytic capacitor can sustain some 30 mA AC ripple current.
By measuring its distortion using our two test signals at 1 kHz 100 Hz, we obtain a direct comparison of polar electrolytic
distortions with the film capacitors of my last article. see
Fig. 1
Aluminium Electrolytic capacitor myths.
As with other capacitor types, much has been written about the sound distortions they cause. However of all capacitor types,
electrolytics are the most complex and the least well understood. Many false myths, specific to electrolytics have emerged,
based more on speculation than on fact:-
a) Aluminium electrolytic capacitor dielectric has extremely high ‘k’.
b) Electrolytic capacitor distortion is mostly third harmonic.
c) For minimum distortion, electrolytic capacitors should be biased to half rated voltage.
d) Back to back polarised capacitors, biased by the supply rail, minimise distortion.
e) High ESR Electrolytics degrade sound quality, low ESR is always best.
f) Electrolytics are highly inductive at audio frequencies.
g) High voltage electrolytics sound the best.
1
Capacitor Sounds 5 - 1
µ
F choice - Electrolytic or Film ?
As we shall see, a working knowledge of electrolytic capacitor construction combined with careful distortion measurements,
leads to somewhat different conclusions.
Polar Aluminium electrolytic construction.
To begin to understand an aluminium electrolytic capacitor we must explore how it differs from other capacitor types including
Tantalum. Every aluminium electrolytic capacitor comprises two polar capacitors in series, connected back to back.
Ref.2
The dielectric for the wanted capacitance is a thin aluminium oxide coating which intimately covers the ‘Anode’ foil. The metal
core of this anode foil, acts as one capacitor electrode. The second electrode is provided by a conductive electrolyte which
permeates and surrounds the anode foil.
A ‘Cathode’ foil is used to make electrical contact between this electrolyte and the lead-out wire. This cathode foil is also
intimately covered by a much thinner, naturally occurring aluminium oxide, the dielectric for our second capacitor. Electrically
similar to oxide produced using a 1 to 1.5 volt ‘forming’ voltage, capacitance of this cathode is many times that of the anode.
The effective surface area of the anode and cathode foils is much enlarged, by mechanical brushing and electro-chemical
etching. Low voltage capacitor foil areas may be increased perhaps one hundred times larger than the foils superficial or visible
area. In this process a myriad of minute tunnels are created in the aluminium foils, which become sponge like and porous.
Ref.2
An extremely thin layer of dielectric, aluminium oxide AL
2
O
3
with a ‘k’ of eight,
Ref.3
is electro-chemically ‘formed’ or grown
on the surface of the anode foil using a non-aggressive electrolyte. Depending on the desired end use, a general purpose
capacitor anode foil may be formed at 1.25 times, a long life capacitor anode foil to double its rated voltage.
In many ways this is similar to the more familiar ‘anodising’ process, long used to provide a decorative and protective finish on
aluminium. The main difference being the anodising oxide is formed using an aggressive electrolyte, which by simultaneously
dissolving away some of the freshly grown oxide, produces a porous oxide layer. This porous layer accepts colouring dyes
which can be sealed in situ, by boiling in water to hydrate and seal the outer oxide layers.
The thickness of our capacitor dielectric oxide is self limiting, being controlled by the voltage used in the forming process. As
thickness approaches 14 Angstrom for each forming volt applied, oxide growth slows down and almost ceases.
Ref.2
This electro-chemically ‘formed’ hard, non-porous, aluminium oxide produces an excellent, almost perfect insulator, which can
be formed for use at least to 600 volts DC. It has a dielectric strength approaching the theoretical strength as predicted by the
ionic theory of crystals.
Because aluminium oxide takes up more space than the aluminium which is converted in the ‘forming’ process, different
etching methods are used according to the intended forming voltage. For the lowest voltage capacitors, the most minute tunnels
are etched into both foils.
Formed to 50 volts, oxide growth would completely fill these minute tunnels. To avoid this the etching process is adapted to
produce somewhat larger tunnels, which can be formed, perhaps to 100 volts. For higher voltages, progressively larger tunnels
must be etched.
Ref.2.
Becromal, one supplier of capacitor foils, lists some fourteen different grades of etched anode and an
even bigger selection of cathode foils.
As capacitor rated voltage increases, less conductive electrolytes and thicker, denser, separator tissues must be used. To reduce
element size and cost, thinner, lower gain cathode foils will usually be chosen. These changes combine to produce a near
optimum quality, low tanδ, low distorting capacitor when rated for 40 to 63 volt working. With notably poorer audio qualities
above 100 volt and at the lowest voltage ratings.
Assembly.
The required length of anode and a slightly longer length of cathode foil are wound together, cathode foil out, onto a small
rotating spindle. To minimise mechanical damage to the extremely thin, dielectric oxide coating, the foils are interwound
together with soft insulating separators. Thin ‘Kraft’ or ‘Rag’ tissue paper the most common.
Aluminium has an electro-chemical potential of +1.66v. To avoid corrosion, no metal other than aluminium may be used inside
the capacitor case. The external lead wires, copper at -0.337v or steel at +0.44v, must be excluded from all contact with
electrolyte, to avoid corrosion of these metals.
Prior to winding the element, thin aluminium connecting ‘tabs’ are mechanically and electrically connected to both foils. Many
years ago, these tabs were attached near the outer end of the winding. In 1968 I introduced into UK manufacture the use of
‘Central’ foil tabbing, which dramatically reduces the aluminium foil resistance, enhancing ripple current ratings and almost
totally eliminates self inductance from the wound element. The most common tab attachment method is called ‘eyeletting’,
when a shaped needle pierces both the connecting tab and its foil. Small ‘ears’ of tab material are burst through the foils, turned
over and well flattened down effectively riveting both parts together. see
Fig. 2
2
Fig 2)
The ‘eyeletting’ type
connections most often used
to connect aluminium lead
out tabs to the centres of both
cathode and anode foils.
In this case because the
winding was central tabbed,
for clarity the outermost,
almost half the wound turns,
of both anode and cathode
foils, have been removed.
A box of ‘nine squares’, tool
marks indicative of the cold
pressure welds used to
reliably connect these tabs to
the tag rivets, can be clearly
seen.
Cold pressure welds, as seen in this photo connecting the aluminium ‘tabs’ to the outer tag rivets, provide a most reliable, low
and linear resistance, connection of aluminium to aluminium. By applying pressure over small areas, metal is forced to flow
between the two items which become intimately bonded and permanently welded together. This method is often also used to
replace ‘eyeletting’ of tabs to foils in the best constructed capacitors.
The completed winding is vacuum impregnated with the electrolyte which becomes absorbed into both foils and separator
papers. Producing a low resistance connection between the anode and cathode foil capacitances.
Bi-polar Aluminium electrolytic capacitor construction.
A Bi-polar electrolytic is made in exactly the same way as a polar capacitor, with one significant difference. In place of the
cathode foil, we use a second formed anode foil.
We still have two polar capacitances in series, back to back. Both now the same value and working voltage. This Bi-polar
capacitor will measure as half the capacitance of either anode foil. To make the required capacitance value, two anode foils,
each double the desired capacitance are used.
Aluminium electrolytic capacitor designers are accustomed to mixing and matching their available materials, to suit the
capacitor’s end application. So it should not surprise that some designs are semi Bi-polar, i.e. they are made using a lower
voltage deliberately ‘formed’ anode foil as cathode.
Equivalent circuit.
Using this constructional background, we deduce an equivalent circuit for a polar aluminium electrolytic capacitor. see
Fig. 3
Fig 3)
This simplified
equivalent schematic
illustrates how a polar
electrolytic capacitor
behaves. For clarity,
components needed to
account for dielectric
absorption, have been
omitted.
3
Fig 3A)
Sectional view of
anode and cathode foils
showing their dielectric
oxide layers and how the
‘electrolyte/paper’ function
acts to provide a good
electrical connection
between the aluminium
oxide dielectric capacitances
of both foils.
Box
Capacitance of an electrolytic.
The high capacitances available in an electrolytic are the result of the effective surface area of the etched and ‘formed’ anode
foil combined with its exceptionally thin dielectric. This effective area is many times larger than the apparent or visible surface
area. The extremely thin, electro-chemically ‘formed’ dielectric oxide film, has a modest ‘k’ value of eight.
Ref.3
Capacitance = Electrode area
×
‘k’
×
0.0885 / Dielectric thickness.
in pF/cm.
Ref.6
This increase in area or ‘gain’, is greatest for very low voltage rated capacitors, reducing with increasing voltage.
This ‘k’ of eight, compared to the ‘k’ of 3.3 for PET, more than doubles capacitance, but far more significant is the extremely
thin dielectric thickness used in aluminium electrolytics and the much increased effective area resulting from the etching
process. As a result, assuming a 50 volt rated capacitor, the aluminium electrolytic’s oxide film produces some 1000 times
more capacitance per unit of apparent electrode area. This gain increases significantly to some 5,000 times for an electrolytic
capacitor rated for 6 volt working.
The cathode foil is covered by a naturally occurring, transparent oxide film, which coats all aluminium surfaces once exposed
to air. Some 20 Angstroms thick, it is equivalent to a 1.5 volt electro-chemically formed oxide. Much thinner than that ‘formed’
on the anode foil even for the lowest voltage capacitors. This cathode foil oxide creates our second capacitor.
For example to make a 100
µF
6.3 volt rated capacitor we might use anode foil formed to 8 volts. This would have a dielectric
thickness of some 110 Angstroms, almost 6 times thicker than the cathode foil’s natural oxide film. We use an anode
capacitance around 118
µF
in series with a cathode capacitance around 660
µF
to obtain our 100
µF
capacitor. The oxide on
the cathode foil, which creates our second capacitor, has a small usable voltage and much larger capacitance than the anode
foil.
Ref.2
see
Fig.3
This naturally occurring, extremely thin, low quality cathode foil oxide, has a larger voltage coefficient than has the anode foil.
It is this cathode capacitor which allows a ‘polar’ aluminium electrolytic to operate on small AC voltages, without polarisation.
Correctly polarised the ‘formed’ aluminium oxide dielectric on the anode foil is an excellent insulator. When reverse polarised
it becomes a low resistance as though a diode has been connected in parallel with a good capacitor.
In similar fashion, the naturally occurring cathode oxide film behaves like a capacitor in parallel with a diode. This diode’s
polarity is in opposition with that of the anode. Because the cathode oxide is thinner, it produces a more leaky diode.
Because a ‘Bi-polar’ electrolytic is made using two anode foils connected back to back in opposition, it can be used on
relatively large AC voltages without polarisation voltage provided the resulting through current does not exceed the rated ripple
current for that frequency and temperature. The Bi-polar electrolytic capacitor can also be used polarised in either direction.
The ‘polar’ capacitor should never be reverse polarised. Any DC polarisation voltage must be correctly applied with the
end of Box.
positive voltage to the capacitor’s anode terminal.
4
Dielectric Oxide.
Aluminium oxide has a ‘k’ of eight,
Ref.3
similar to that of C0G ceramic or impregnated paper capacitors. It is rather higher
than PET, which at 3.3, has the highest ‘k’ of commonly used films. A low value compared to the ‘k’ of several thousand,
found in BX, X7R and Z5U ceramics.
Ref.4
While the impregnant used in paper capacitors is an insulator and acts as the dielectric, the electrolyte impregnant used in
electrolytic capacitors is a good conductor so cannot be a dielectric. This electrolyte is needed to provide a low resistance
connection between the two capacitors.
More significant than ‘k’ value is dielectric thickness. Large capacitance values are possible because the dielectric of a 50 volt
aluminium electrolytic capacitor is some 100 times thinner than that used in a film capacitor.
Ref.2
As a result, electrolytic
capacitors are sensitive to dielectric absorption effects.
The dielectric oxide films have a measurable voltage coefficient of capacitance. When DC biased, the measured capacitance of
a 1
µF
63 volt capacitor increased 0.15% at -0.5 volt. Initially decreasing 0.05% at +0.5 volt, capacitance then increased to
+0.16% at +10 volt.
Voltage effects.
I explored these voltage effects by measuring the distortion produced by a 1µF 63 volt polar electrolytic capacitor, subjected to
different AC test voltages. Commencing with 0.1 volt, capacitor distortion was measured at 0.1 volt increments to 1 volt then
with a test at 2 volts. Initially I test with no bias, then with various DC bias voltages. Remember these voltages are those
actually measured across the capacitor terminals and not the generator set voltage.
Small test voltages reduce measurement dynamic range. To compensate for this, distortion from the test capacitor will be
compared with those produced by a near perfect film capacitor, tested exactly the same as reference. All tests for this article use
my DC bias buffer and two frequencies, 100 Hz/1 kHz to observe intermodulation effects.
Electrolytic capacitor behaviour varies with small changes in temperature. To minimise the affect of temperature changes, all
reported tests were performed at constant room temperature. Unless otherwise stated, all voltages are RMS as measured using a
DMM.
Without DC bias.
Notably larger distortions were produced by this electrolytic than the film capacitor, even with a test signal as small as 0.1 volt,
across the capacitor.
Tested with a 0.3 volt signal and no bias, distortion of this typical 1
µF
63 volt polar electrolytic capacitor, clearly dominated
by second harmonic, measured 0.00115%. Almost three times greater than for the reference capacitor.
see
Fig. 4.
Fig 4)
Distortions
measured on our
1
µF
63 volt polar
capacitor, using a
0.3 volt test signal
without DC bias.
Note how the large
second harmonic
component dominates
all others.
When the peak of the AC voltage applied across this unbiased polar capacitor exceeds some 0.5 volt, the cathode foil’s voltage
dependency has more noticeable effect.
Tested at 0.4 volt RMS, both harmonics increase relative to the small change in test signal. Second harmonic voltage has almost
doubled compared to the 0.3 volt test. Distortion is now four times greater than our reference capacitor.
see Fig.
5
5

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika:

Zgłoś jeśli naruszono regulamin