21-01-2018, 06:47 PM
CAPACITANCE AND RESISTANCE BRIDGE
article by W.T.Cocking, M.I.E.E.
EXTRACT FOM THE WIRELESS SERVICING MANUAL,
TENTH EDITION, ILIFFE BOOKS Ltd., 1963
THERE is no doubt that the best method of measuring capacitance is by means of a bridge. This is a comparison method by which the unknown capacitor is compared with another of known value, and the accuracy depends upon how closely the relative values of the two can be determined and upon how accurately the capacitance of the standard is known. The basic circuit is shown in Fig. A 3.1, and it is easy to see that if the reactances of Cl and C2 are the same and R1 equals R2 there will be no current through the telephones. If R1 does not equal R2 however, there will be a difference of potential between the two sides of the phones and, consequently, an audible note in them. As the slider of the potentiometer is moved, therefore, the strength of the sound in the phones will vary and at the exact balance point when R1 equals R2 will disappear.
[attachment=17108]
It is not, of course, necessary for the two capacitances to be equal, but if they are not equal the balance point will be different. The balance point is connected with the relative values of the
capacitors by a very simple relation, Rl I/wC2 :: R2 : I/wCl. Consequently RI/R2 = C1/C2. If one of the capacitors, say C2,
has an accurately known value, and the resistances can be accurately measured each time a balance has been obtained, the unknown capacitance is readily calculated as C1 = R1C2/R2. This is naturally inconvenient for general use, but it is easily possible to provide a scale giving a direct calibration in capacitance, and the operating of measuring capacitance then becomes no more than connecting the capacitor to the terminals, adjusting the control knob to give the balance, and reading off the capacitance directly from the scale.
In theory it is possible to measure all capacitances from zero to infinity in a single range, but this is hardly possible in practice with ordinary potentiometers, for the scale would become extremely cramped at the ends. With only two ranges, however, it is easy to measure any capacitance between 10 pF and 10 µF.
The same bridge may be used also for the measurement of resistance if the standard capacitor is replaced by a standard resistance, and if the values are suitably chosen the same scale and calibration hold for both resistance and capacitance measurement. This has been done in the bridge described here, and two ranges are provided. On one range resistances of 10ohm to 100,000 ohm can be measured, and capacitances of 10 pF to 100,000 pF (0.1 µF), while on the other scale the reading is multiplied by 100 and so gives a range of 1000ohm to 10,000,000ohm (10M) or 1,000pF (0.001µF) to 10,000,000pF (10µF).
[attachment=17109]
The complete circuit diagram of the bridge is shown in Fig. A 3.2, and it will be seen that the control consists of a 20,000ohm potentiometer R4. The switches S2 and S3 change over the connections for capacitance or resistance measurements, and also change the range. There are two resistance standards, R5 and R6, of 1,000ohm and 100,000ohm respectively, and two capacitance standards, C6 and C5, of 0.001µF and 0.1µF respectively. The values of these components must be accurately known, and if the bridge is to have any pretensions to accuracy and reliability good quality components having values within 1 per cent of their ratings must be used.
Compared with ordinary components accurate and stable resistors and capacitors are expensive. Only two of each are needed, however. Firms which specialize in laboratory apparatus can supply suitable parts and the cost rises rapidly with the accuracy required. For this bridge 1 per cent components are desirable and anything better would be rather wasted. This accuracy comes at about borderline between ordinary and laboratory parts and represents about the lowest degree of accuracy available in such special parts. It is, too, a degree of accuracy at which many ordinary parts are obtainable to special order.
Of course, if one has access to accurate measuring equipment and a fair selection of components one can often pick them out for oneself, or make up the required value from several. Resistors should, preferably, be wire-wound; if not, they must be highstability types.
The total value of the balancing potentiometer R4 has no effect on the accuracy, and it is only necessary that the ratio of the two halves be the same in different potentiometers for all positions of the slider in order to permit a printed scale to be used. This requirement is met quite well when a uniformly wire-wound potentiometer is used, and although the accuracy may suffer slightly through the absence of individual calibration, this is the inevitable price which must be paid for avoiding the trouble of calibration.
Constructionally, the components are all secured to a panel which may be of wood or plastic, and which forms the lid of a shallow box. A metal panel should not be used for the increased stray capacitance results in errors at low capacitance and a poor null point on the resistance ranges.
The only non-standard part employed is the pointer, which is best cut from a strip of brass and screwed to the control knob by two 8 BA screws. Holes must be drilled and tapped in the knob to receive these screws, and some care is needed in carrying out this operation, for the Bakelite is apt to crumble somewhat.
A source of alternating current is necessary to energize the bridge, a.c. mains are not suitable, and the simplest arrangement is to use a buzzer, connecting R4 across the coil of the buzzer. The writer must confess, however, to a dislike of buzzers on account of their unreliability and their mechanical noise. Cheap buzzers are very troublesome, being difficult to start buzzing properly and liable at any moment to cease work. A good reliable buzzer usually costs as much as a valve or transistor oscillator and suffers from the disadvantage that the direct noise often renders it difficult to distinguish the exact balance point in the phones.
In the writer's experience a valve or transistor oscillator is the better, since it is both reliable and silent, and it is not unduly expensive; and of the two the transistor is the more convenient. The circuit of a suitable oscillator is given in the upper part of Fig. A 3.z. It can built separately from the bridge if required or the whole can form a single unit. A single transistor is used and the power is supplied by a 4.5V battery. The transformer is wound on a core comprising a 0.4 in. stack of No. 74. Magnetic and Electric Alloys laminations. All windings are of No. 38 enamelled wire; A has 1,000 turns, B 200 turns and C 50 turns.
A pair of high-resistance phones must be used or low-resistance with a transformer, and with nothing connected to the " X " terminals the oscillator note should be quite loud. The capacitor or resistor to be measured should be connected to the " X " terminals and the switches S2,S3 set appropriately. The balance point will be found to be quite definite, although complete extinction of the note may not always be obtained. The minimum, however, is so definite that it cannot be mistaken; it is not obscured by oscillator harmonics.
The reason why a null-point is not always obtained is that no provision is made for balancing out the effect of any resistance in the capacitor being tested nor for balancing out the effect of capacitance in the resistor being measured. Normally, these effects are small with good components, and the extra complication to the bridge and its operation is not worth while. If it is found, therefore, that with a particular component a bad balance is secured, it is possible that, if a capacitor, it has a bad power factor, or if a resistor, it has a high self-capacitance.
It must be remembered that the reading of capacitance obtained includes the self-capacitance of the bridge and connecting leads. This can readily be determined by balancing the bridge for capacitance with nothing connected to the " X " terminals, and was about 30 pF for the original instrument. In order to obtain the true capacitance of a capacitor, therefore, this figure must be deducted from the reading obtained. This correction is unimportant for capacitances higher than 1,000 pF.
[attachment=17110]
If at all possible the bridge should be calibrated. This can be done on the " Low R range " with the aid of decade resistance boxes. In case this is not possible a scale is included in Fig. A 3.3, but with this one can hardly rely on the absolute accuracy to better than 5 per cent. When it is employed care must be taken to set the pointer correctly. This may, of course, be done by the end marks which indicate the limits of travel of the pointer, but for the greatest accuracy it would be wise to obtain a 1,000ohm resistance within ± 1 per cent (such resistances are not expensive) and to balance the bridge with this resistance connected to the " X " terminals. Leaving the bridge balanced, set the pointer to read exactly 1,000.
It may be remarked that no earth connection must be used on any point of the bridge or oscillator. The capacitor or resistor under test must be disconnected from other gear, otherwise appreciable errors may be introduced.
Any good quality components may be used, but the scale of Fig. A 3.3 will be reasonably accurate only if the potentiometer is of the kind used originally. This is the Reliance Type TW linear wire-wound potentiometer.
article by W.T.Cocking, M.I.E.E.
EXTRACT FOM THE WIRELESS SERVICING MANUAL,
TENTH EDITION, ILIFFE BOOKS Ltd., 1963
THERE is no doubt that the best method of measuring capacitance is by means of a bridge. This is a comparison method by which the unknown capacitor is compared with another of known value, and the accuracy depends upon how closely the relative values of the two can be determined and upon how accurately the capacitance of the standard is known. The basic circuit is shown in Fig. A 3.1, and it is easy to see that if the reactances of Cl and C2 are the same and R1 equals R2 there will be no current through the telephones. If R1 does not equal R2 however, there will be a difference of potential between the two sides of the phones and, consequently, an audible note in them. As the slider of the potentiometer is moved, therefore, the strength of the sound in the phones will vary and at the exact balance point when R1 equals R2 will disappear.
[attachment=17108]
It is not, of course, necessary for the two capacitances to be equal, but if they are not equal the balance point will be different. The balance point is connected with the relative values of the
capacitors by a very simple relation, Rl I/wC2 :: R2 : I/wCl. Consequently RI/R2 = C1/C2. If one of the capacitors, say C2,
has an accurately known value, and the resistances can be accurately measured each time a balance has been obtained, the unknown capacitance is readily calculated as C1 = R1C2/R2. This is naturally inconvenient for general use, but it is easily possible to provide a scale giving a direct calibration in capacitance, and the operating of measuring capacitance then becomes no more than connecting the capacitor to the terminals, adjusting the control knob to give the balance, and reading off the capacitance directly from the scale.
In theory it is possible to measure all capacitances from zero to infinity in a single range, but this is hardly possible in practice with ordinary potentiometers, for the scale would become extremely cramped at the ends. With only two ranges, however, it is easy to measure any capacitance between 10 pF and 10 µF.
The same bridge may be used also for the measurement of resistance if the standard capacitor is replaced by a standard resistance, and if the values are suitably chosen the same scale and calibration hold for both resistance and capacitance measurement. This has been done in the bridge described here, and two ranges are provided. On one range resistances of 10ohm to 100,000 ohm can be measured, and capacitances of 10 pF to 100,000 pF (0.1 µF), while on the other scale the reading is multiplied by 100 and so gives a range of 1000ohm to 10,000,000ohm (10M) or 1,000pF (0.001µF) to 10,000,000pF (10µF).
[attachment=17109]
The complete circuit diagram of the bridge is shown in Fig. A 3.2, and it will be seen that the control consists of a 20,000ohm potentiometer R4. The switches S2 and S3 change over the connections for capacitance or resistance measurements, and also change the range. There are two resistance standards, R5 and R6, of 1,000ohm and 100,000ohm respectively, and two capacitance standards, C6 and C5, of 0.001µF and 0.1µF respectively. The values of these components must be accurately known, and if the bridge is to have any pretensions to accuracy and reliability good quality components having values within 1 per cent of their ratings must be used.
Compared with ordinary components accurate and stable resistors and capacitors are expensive. Only two of each are needed, however. Firms which specialize in laboratory apparatus can supply suitable parts and the cost rises rapidly with the accuracy required. For this bridge 1 per cent components are desirable and anything better would be rather wasted. This accuracy comes at about borderline between ordinary and laboratory parts and represents about the lowest degree of accuracy available in such special parts. It is, too, a degree of accuracy at which many ordinary parts are obtainable to special order.
Of course, if one has access to accurate measuring equipment and a fair selection of components one can often pick them out for oneself, or make up the required value from several. Resistors should, preferably, be wire-wound; if not, they must be highstability types.
The total value of the balancing potentiometer R4 has no effect on the accuracy, and it is only necessary that the ratio of the two halves be the same in different potentiometers for all positions of the slider in order to permit a printed scale to be used. This requirement is met quite well when a uniformly wire-wound potentiometer is used, and although the accuracy may suffer slightly through the absence of individual calibration, this is the inevitable price which must be paid for avoiding the trouble of calibration.
Constructionally, the components are all secured to a panel which may be of wood or plastic, and which forms the lid of a shallow box. A metal panel should not be used for the increased stray capacitance results in errors at low capacitance and a poor null point on the resistance ranges.
The only non-standard part employed is the pointer, which is best cut from a strip of brass and screwed to the control knob by two 8 BA screws. Holes must be drilled and tapped in the knob to receive these screws, and some care is needed in carrying out this operation, for the Bakelite is apt to crumble somewhat.
A source of alternating current is necessary to energize the bridge, a.c. mains are not suitable, and the simplest arrangement is to use a buzzer, connecting R4 across the coil of the buzzer. The writer must confess, however, to a dislike of buzzers on account of their unreliability and their mechanical noise. Cheap buzzers are very troublesome, being difficult to start buzzing properly and liable at any moment to cease work. A good reliable buzzer usually costs as much as a valve or transistor oscillator and suffers from the disadvantage that the direct noise often renders it difficult to distinguish the exact balance point in the phones.
In the writer's experience a valve or transistor oscillator is the better, since it is both reliable and silent, and it is not unduly expensive; and of the two the transistor is the more convenient. The circuit of a suitable oscillator is given in the upper part of Fig. A 3.z. It can built separately from the bridge if required or the whole can form a single unit. A single transistor is used and the power is supplied by a 4.5V battery. The transformer is wound on a core comprising a 0.4 in. stack of No. 74. Magnetic and Electric Alloys laminations. All windings are of No. 38 enamelled wire; A has 1,000 turns, B 200 turns and C 50 turns.
A pair of high-resistance phones must be used or low-resistance with a transformer, and with nothing connected to the " X " terminals the oscillator note should be quite loud. The capacitor or resistor to be measured should be connected to the " X " terminals and the switches S2,S3 set appropriately. The balance point will be found to be quite definite, although complete extinction of the note may not always be obtained. The minimum, however, is so definite that it cannot be mistaken; it is not obscured by oscillator harmonics.
The reason why a null-point is not always obtained is that no provision is made for balancing out the effect of any resistance in the capacitor being tested nor for balancing out the effect of capacitance in the resistor being measured. Normally, these effects are small with good components, and the extra complication to the bridge and its operation is not worth while. If it is found, therefore, that with a particular component a bad balance is secured, it is possible that, if a capacitor, it has a bad power factor, or if a resistor, it has a high self-capacitance.
It must be remembered that the reading of capacitance obtained includes the self-capacitance of the bridge and connecting leads. This can readily be determined by balancing the bridge for capacitance with nothing connected to the " X " terminals, and was about 30 pF for the original instrument. In order to obtain the true capacitance of a capacitor, therefore, this figure must be deducted from the reading obtained. This correction is unimportant for capacitances higher than 1,000 pF.
[attachment=17110]
If at all possible the bridge should be calibrated. This can be done on the " Low R range " with the aid of decade resistance boxes. In case this is not possible a scale is included in Fig. A 3.3, but with this one can hardly rely on the absolute accuracy to better than 5 per cent. When it is employed care must be taken to set the pointer correctly. This may, of course, be done by the end marks which indicate the limits of travel of the pointer, but for the greatest accuracy it would be wise to obtain a 1,000ohm resistance within ± 1 per cent (such resistances are not expensive) and to balance the bridge with this resistance connected to the " X " terminals. Leaving the bridge balanced, set the pointer to read exactly 1,000.
It may be remarked that no earth connection must be used on any point of the bridge or oscillator. The capacitor or resistor under test must be disconnected from other gear, otherwise appreciable errors may be introduced.
Any good quality components may be used, but the scale of Fig. A 3.3 will be reasonably accurate only if the potentiometer is of the kind used originally. This is the Reliance Type TW linear wire-wound potentiometer.