Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 48, 1915
This text is also available in PDF
(1 MB) Opens in new window
– 464 –

Art. XLVIII.—Investigation into the Resistance of Earth Connections.

[Read before the Philosophical Institute of Canterbury, 7th July, 1915]

The subject of this investigation is one which is being constantly referred to by the electrical engineer, and called for by innumerable regulations framed for the protection of life and plant; but in practice it is often so vague and uncertain, and so few engineers actually measure their earth-resistances, that there is room for much investigation before we can claim to understand the principles involved or their effect when applied to any particular set of circumstances. The electrical engineer, in practice, early admits the difficulty of the problem, and realizes the absence of any comprehensive treatise on the subject. There is such diversity in the data extant, and so strange a variation between his own experience and the stated results of others, that he is only too glad to escape further attention from official inspectors by obtaining permission to “earth” to a water-pipe. And in this way the matter, for the most part, is quietly shelved.

The present paper aims primarily at the publication of data obtained during an investigation of earthing devices, for use on the Canterbury Plains (New Zealand), in connection with the high-voltage electric-power transmission-line recently erected by the Public Works Department of New Zealand for the supply of Christchurch and district from Lake Coleridge, a distance of sixty-two miles. Some description of the country traversed is essential to an intelligent apprehension either of the data itself or of the difficulty experienced in attaining the objective.

The Lake Coleridge power-house is situated in the Rakaia River valley, sixty-two miles west of Christchurch, and from this point two 66,000-volt three-phase transmission-lines are run to the city on independent pole-lines For the first sixteen miles from the power-house the lines are located about 3 chains apart, and pass along the terraces and low hills of the river-valley, in many places over screes or fans of shingle detritus. At Windwhistle Point they divide. In the next fourteen miles the north line continues over low hills and rolling downs to Glentunnel, while the south stretches almost directly across the thirty-two miles of shingle-beds which constitute the Canterbury Plains. At the city end the lines meet again at Bealey Road, seven miles from the substation, and continue about 1 chain apart For this distance the soil is largely of a sandy to peaty nature, the remains of a marsh formed by the sandhills of a former seashore and in the bed of an ancient river. It is thus evident that, although the major part of the line runs through dry shingle country, there is at the same time extensive variety. In many places across the plains water in ordinary soakage wells is almost unobtainable, not being met till below 60 ft, and in some cases 120 ft to 140 ft. Evidently the permanent water-level may thus be at a very great depth, leaving at the top a considerable thickness of dry shingle. The shingle is, in the mam, a hard, highly crystalline, bluish-grey sandstone, commonly known as greywacke. In size it ranges from sand to boulders 18 m in diameter.

The line conductors consist of 7/-135 bare aluminium wires, and are carried on Thomas No. 4000 pm insulators, mounted on wooden (jarrah)

– 465 –

cross-arms, and carried on ironbark poles standing 36 ft. out of the ground and sunk 6 ft. into the ground. There are 864 poles in the northern line and 873 in the southern line. The standard pole-spacing is 6 chains (396 ft.) —i e, about thirteen poles to the mile. In order to obviate the burning of poles and cross-arms, it was decided to earth the whole of the high-tension insulator-pins; and owing to the large inductance with high-frequency current and considerable capacity of a ground wire, not to mention the greater cost, it was decided to earth each pole with an individual earth-strip. In addition to this, the apparatus in telephone depots along the line required to be thoroughly earthed, while the stations at either end, and the numerous transformers in the distribution system, all required very efficient earths. The station lightning-arresters in particular depend for their effectiveness on a low-resistance connection to ground.

The substance of the British Board of Trade specification in regulations concerning station earths and earths for tramways and general power purposes is that the resistance shall be less than 1 ohm between two plates of copper, cast iron, or galvanized iron, packed in coke, and 60 ft. apart.

Regulations under the Coal-mines Act (Great Britain), 1911, specifies that “All conductors of an earthing system shall have a conductivity at all parts and at all joints at least equal to 50 per cent, of that of the largest conductor used solely to supply the apparatus a part of which it is desired to earth.” In practice this would involve an earth-resistance never greater than 10 to 15 ohms, and usually not more than 2 or 3 ohms. “Sparks” (journal of the Institution of Electrical Engineers), 15th March, 1915, recommended packing-plates placed vertically at 6 ft. or more beneath the surface, with a foot of coke on each side: “Tests made on the earth plates constructed on the lines recommended by the General Regulations, the plates being buried in excavations 4 ft by 2 ft. by 8 ft. deep in clay, under favourable conditions as to moisture, vary from 1–8 to 2–2 ohms, while the resistance of similar earth plates in another district, in excavations 4 ft. by 2 ft by 6 ft deep in well-consolidated marl and clay, the earth plates resting on clay in damp positions, reached 2–6 and 2–7 ohms…. If the conditions are not favourable a much higher resistance to earth will be found, as the resistance to earth is directly affected by the nature and temperature of the surrounding strata and by the amount of moisture.”

It is now fairly common knowledge that the Board of Trade test is seldom obtained, and that values such as those quoted above are decidedly the exception rather than the rule.

Some tests were made locally on plate earths of the Christchurch City Council. These plates were ⅛ in. copper, 2 ft. square, and were placed horizontally in a bed of coke 1 ft. thick above and below the plate, and at a depth of 6 ft. to 8 ft. These were tested against the city high-pressure water-mam, and (assuming the water-main as of zero resistance to earth) these gave the following results. Beckenham transformer-house plate earth = 51 ohms (loam and shingle); Sydenham transformer-house plate earth = 17.2 ohms (clay); Montreal Street transformer-house plate earth = 13 ohms (clay).

The Christchurch Tramway Board also had some considerable difficulty with earth connections. Finally two large copper plates 8 ft. square were placed some 10 ft below the surface (60 ft. apart), and well packed with coke. Tests on these plates failed to reach that required by the Board of Trade. To avoid further trouble and expense the ground wire was earthed

– 466 –

to the city water-main. In July, 1915, tests of these earths by the authors gave the following results: Plate 1 to plate 2 = 6 ohms; plate 1 to water-main = 4 ohms; plate 2 to water-main = 5 ohms.

During the course of the present investigation plates were tried on the transmission-line. Three copper plates 12 in square were buried at a depth of 5 ft. 6 in. to 6 ft, and bedded in clay. In one case the country was loam to shingle, while the others were clay to sand. The plates were placed at intervals of 6 chains, and tests of pairs in series gave the following values: Resistance of plates immediately after placing—No 760 plate to No 762 plate = 260 ohms; No 759 plate to No. 760 plate = 200 ohms, No. 759 plate to No. 760 plate = 180 after 4 gallons water had been poured over each; No 760 plate to No. 762 plate = 130 after 3 gallons of brine had been poured over No 762 plate. Values by A.C bridge test.

The contrast of these values with those quoted is obvious, and these are from by no means in the worst districts. This chiefly indicates the unreliability of the earth connection by means of small copper plates. At the same time, since each earth plate in place costs from £1 to £1 10s., the cost for such inadequate result precludes the adoption of the plate earth on any extended scale.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

At the erection of the transmission-line poles provision was made for earthing by carrying a strip of galvanized hoop iron (1¼in. by 1/12 in.) down the pole and carried round the butt three times below ground-level, terminating underneath the bottom of the pole. The connection was thus buried at a depth of at least 6 ft. No precautions in the way of selecting earth to make the contact were taken when filing in, and at bends and in special cases the poles were concreted. Subsequently tests of these earths were made with an instrument consisting of an ordinary Wheatstone bridge and 4-volt dry cell.

To get the resistance of each separate earth connection, the strips were tested in groups of three throughout the 1,737 poles, and the values computed from the three simultaneous equations obtained. (Notes concerning both the test and the method of computation of individual values will be found later).

The values obtained ranged from 10 ohms to 5,000 ohms for individual connections. In some districts the resistances were consistently low; in others consistently high; while still others gave resistances indiscriminately high and low. Of the 864 earths on the north line, more than 30 per cent, were over 1,000 ohms, and 83 per cent, over 100 ohms, while on the south line, out of 873 poles, 18 per cent. were over 1,000 ohms, and 90 per cent, over 100 ohms. As earths of this order are quite useless for practical purposes, attention was turned to pipe earths.

By far the most useful data on the subject available consists of a paper by E. E F. Creighton in the “General Electric Review,” vol. 15, page 66, February, 1912. The most important and pertinent conclusions are here quoted:—

“1. Resistance of a pipe earth varies inversely as the depth of the pipe after the pipe has reached a uniformly conducting stratum.

“2 Practically all the resistance is in the earth in the immediate vicinity of the pipe. This resistance depends on the specific resistance of the material. The specific resistance depends on acids, salts, or alkalies in solution about the plate. To get the lowest possible resistance strong salt water should be poured around the pipe. The chemical action of salt on the iron of the pipe under average conditions is negligible.

– 467 –

“3. If an iron pipe 1 in. in diameter is driven into normally moist earth to a depth of 6 ft. to 8 ft. it will usually have a resistance of about 15 ohms—8 ohms may be considered unusually low—while in dry soils it may give a resistance of 50 ohms and upwards. When it is desired to lower the resistance to earth below that of a single-pipe earth, others should be driven at distances of not less than 6 ft. apart. Then the total conductances will be only slightly less than the sum of the conductances of the individual pipes.

“4. Half the total resistance lies within 6 in. of the pipe. At a distance apart of 6 ft the resistance reaches nearly a constant value.

“5 Since the resistance of a pipe earth lies mostly in the immediate vicinity of the pipe, the greatest potential drop when the current flows will also be concentrated there. Heating and drying out of the soil will tend to magnify this effect. The more salt water placed round a pipe earth the less the potential gradient.

“6 The resistance of a pipe earth does not decrease in direct proportion to the increase in diameter of the pipe. A pipe 2 in in diameter has a resistance only about 6 per cent. to 12 per cent, less than a pipe 1 in. in diameter.”

Mr. H. P. Liversidge, before the International Association of Municipal Electricians, 19th August, 1913, gives data recording resistances with pipe earths (2 in. in diameter and 6 ft. to .12 ft. deep) ranging from 0.08 ohms to 138 ohms. He states that the general average of some other twenty-five different readings, representing fairly average conditions in reference to character of soil, contained moisture, &c, give the following contact resistances: Clay, 13–60 ohms; gravel, 6–01 ohms; top soil, 1–80 ohms. There is no specific mention of the method of test employed, but apparently a voltmeter-ammeter method, with about 220 volts D.C., was used. All these results were obtained after a salt solution had been poured round the pipe, and probably these values are about one-half of the initial resistance. Mr. Liversidge suggests as desirable values for an earth for station work 1 to 2 ohms, and for other purposes 2 to 6 ohms.

As it was apparently necessary to investigate in the actual country, some preliminary pipe earths were then tried on the transmission-line, at Bealey Corner, where the line passes into the heavy shingle of the old Wai-makariri river-bed. Galvanized-iron pipes 1 ½in. diameter were driven 4 ft to 6 ft deep beside line-poles, 6 chains apart. In sand and clay the average individual resistance of seven pipes 5 ft. deep was 126 ohms by the D.C. bridge method. In heavy shingle the average individual resistance of four pipes 4 ft deep was 1,800 ohms (D.C. bridge method). These pipes were driven into the ground with a 14 lb. hammer, and no brine or other electrolyte was added.

Further pipes were then driven in the neighbourhood of three other poles. The ground about the three poles selected was loam for a foot at the top, to clay and sand or clay and light shingle below. The uncertainty of the substratum was demonstrated by the fact that, though one pipe was easily driven in sand and clay, another, 6 ft. away, when driven a little more than 2 ft refused to go farther on account of heavy shingle.

The pipes were drilled with two ⅜in. holes at intervals of 1 ft. along the length of the pipe, and after starting with a crowbar were driven 5 ft. to 8 ft deep with a hammer. They were then tested as driven, and results are tabulated below. Pipes A and B were then “salted” by filling with salt, pouring water in, and filling again with salt. They were then tested again, and at intervals subsequently. Pipes C were salted by pouring in

– 468 –

at intervals hot saturated solutions of brine up to 20 gallons, testing between successive charges, and also subsequently. In pipes D cold saturated solutions (up to 9 gallons) of copper sulphate were used.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Table I—test pipes, bealey corner.
(Average Individual Pipe-resistances to Earth.)
Particulars of Pipes, 6 Chains apart. Pipes as driven Immediately after salting Seven Days after salting Twenty-one Days after salting Three Months after salting Seven Months after salting
Ohms Ohms Ohms Ohms Ohms Ohms
A. Three 2 in. pipes, 7 ft. deep 115 95 84 48 38 45
B. Three 1 ¼ in pipes, 9 ft. deep . 138 105 92 56 51 70
C. Two 1 ¼ in. pipes, 9 ft. deep . 175 85 63 . 36 46
D. Two 1 ½ in pipes, 7 ft 6 in. deep(salted with copper sulphate) 78 68 94

These are all individual pipe-resistances obtained with a low-voltage D C bridge method.

Some discussion is here necessary concerning the various methods of test that have been used. The low-voltage D C. Wheatstone bridge method has been most commonly used, owing to its ease of application and the portable nature of the apparatus. This test cannot give the true resistance under working-conditions, when thousands of volts A.C are concerned, but gives a proportional result, as indicated by the following observations: Some earths at Addington Substation were tested by three different methods —(a) By D C. bridge, (b) by voltmeter-ammeter D C 110 volts; (c) by Wheatstone bridge, using surging A C. voltage and telephone-receiver.

In reference to (a) and (b), in all tests two values were obtained with reversed polarity, to eliminate as far as possible electrolytic effect, and the tabulated values are the mean of the two. Test (c) was effected by placing an interrupter (a small buzzer) in the primary of a small transformer, with a 4-volt cell as source of e.m.f, and connecting the secondary to the “battery” terminals of the bridge. A telephone-receiver replaces the galvanometer for indicating a balance of the arms. The results are tabulated below for three earths marked A, D, and P.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Table II—Comparison of Methods of Testing
(a) (b) (c)
Earth-resistances (Two in Series) D C Bridge D C Ammeter A C Budge.
Ohms Ohms Ohms
A and P 31 3 23 6 22 8
D and P 37 3 31 2 28 8
A and D 15 3 13 0 12 8

Although the tests are not numerous, they are sufficient to indicate a large but fairly constant error in the ordinary low-pressure D C bridge test, and a fairly close approximation of the A C bridge test to the positive voltammeter determination.

In the same connection Mr. Creighton may be quoted—“that this method (A.C bridge) gives values within 5 per cent. of those by D C of A C voltammeter”.

– 469 –

Some measurements of earth-resistances by D.C. and A.C. voltmeter-ammeter method were made by Mr. Parry, Chief Electrical Engineer, in Wellington. These were made on hemispheres 3 in. and 6 in. in diameter, placed in soil without addition of any salt. Results are tabulated below, with percentage of the D.C. value greater than the A.C.:—

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Table III— Comparison of d.c. and a.c. Voltammeter Methods
(Two Earths in Series.)
Number of Test Pair. D.C. A.C. Ratio.
Ohms Ohms.
1 213 177 1.21
2 158 138 1.14
3 161 138 1.16
4 164 139 1.19
5 115 101 1.14
6 158 135 1.18
7 112.5 100 1.12
Average 1.16

This gives an average of D.C. values greater than the corresponding A.C. values by 16 per cent. The A.C. ammeter used in these tests had a very close scale, and results are not entirely satisfactory.

Some later comparative values were obtained at Addington Substation and on the transmission-line, with the following results:—

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Table IV—Comparison of d.c. and a.c. Bridge Tests.
(Two Pipes in Series.)
Number of Pairs Distance apart By D C Bridge. Ratio
Pipes unsalted— Ohms Ohms.
4 18 in. 26.3 16–9 1.55
3 3 ft. 29.7 20.2 1.47
2 4 ft. 6 m. 33 2 23.5 1.41
1 6 ft. 35 5 25.0 1.42
1 30 ft. 40.5 28.0 1.45
2 6 ch. 1,770 1,390 1.27
3 6 ch . 287 238 1.21
1 100 ft. 41.8 32 0 1.31
Pipes salted—
2 6 ch. 89 79 1.13
1 6 ch. 92 70 1.31
1 6 ch. 187 143 1.31
3 6 ch. 383 328 1.17
3 6 ch 142 111 1 28
Average 1.33

The D.C. bridge thus gives a resistance about 33 per cent. higher than the A.C. bridge. There is a considerable range in the percentage difference, showing that the D.C. test probably gives a rather erratic result. Although results are not sufficiently numerous to warrant a definite statement, there is an indication that the difference unsalted is in the region of 30 to 40 per cent, while salted it is from 20 to 30 per cent.

– 470 –

Thus resistances by D C. measurement are all higher than those by A.C. As the resistance is—at least, very largely—the resistance of an electrolyte, this is to be expected by reason of the polarization effect with D.C. Obviously, values by voltmeter-ammeter methods must be accepted as correct, and whether D C. or A C. determinations should be considered in any specific case will depend on working-conditions.

The A C. bridge was found both quicker and more convenient in operation than the D C bridge. A galvanometer, however well packed, will not stand continuous transport by motor bicycle without losing in sensitiveness, while a telephone-receiver is easily packed, and with reasonable care does not suffer from carriage.

A very compact form for the transformer and interrupter can be obtained by rewinding the electro-magnet of an ordinary 4-ohm buzzer, and using one leg as primary and for working the interrupter, and winding a secondary on the other leg. Eight layers of No. 36 copper on the primary, and eleven layers of same gauge on secondary, gave very good results.

Some comment is necessary relative to the practice of solving simultaneous equations to obtain the resistances of individual earths. The solutions obtained will be correct as long as the unknowns are independent quantities—i.e, as long as one earth does not interfere with another. This depends mainly on the distance between the earths being greater than some critical value, which, again, is dependent on the class of soil. Where individual resistances are tabulated herein, unless otherwise stated, the earths are sufficiently far apart not to interfere with each other in any way.

It became evident at once that in the class of ground encountered earth-resistances below 100 ohms were going to be very difficult and expensive in practice.

From the tests made it appears that once a pipe reaches a depth of 6 ft. to 7 ft., increase in depth has little effect on the final value. Any improvement produced by increased depth is so small as to be lost in the effect of other factors. It is common experience that in considering ground connections the character of the soil is the main factor concerned. However, it is not the solids constituting the soil so much as the moisture and the salts in solution which command attention. Messrs McCollum and Logan (Proc. A.I E.E., June, 1913) publish some material bearing on the conductivity of soils with varying moisture-content (diagram No 1). The curve shown was obtained from a sample of red-clay soil which had been dried out at 105° F, and water afterwards added. It will be noticed that the resistance is practically constant for a content greater than 20 percent. In the same paper are recorded ninety-two values of specific resistances of “a wide variety of different kinds of soil.” In 50 per cent of the determinations the moisture-content, which is also recorded, lies between 19 and 30 per cent.—i e, they were moist soils—and at these values the resistance-variation with moisture-content is almost negligible. Yet the recorded specific resistances range from 41,490 to 470 ohms. In some cases—few rather than many—the quantity of moisture is distinctly responsible for the variation in resistance, but in general the main factor affecting the conductivity of the pipe earth is the electrolytic quality of this moisture.

Considering the exceedingly high resistances of both approximately dry soil and approximately pure water, we must conclude that the conductivity —at least, in the immediate neighbourhood of the earth connection—is almost solely due to the salts, alkalies, or acids in solution. The exceed-

– 471 –

ingly large range of resistances is then explicable by electrolytes of variable specific conductivity, determined by the solutes or the degrees of concentration. That such is a feasible explanation is evident from the following specific conductivities of electrolytes (“Physical and Chemical Constants,” by Kaye and Laby), which admit of combinations to provide almost infinite range of values (diagram No. 2).

Picture icon

Fig. 1.

Picture icon

Fig. 2.

In the course of this investigation a few comparatively very low-resistance “earths” were obtained, ranging from 4 to 40 ohms—e.g., at Wind-whistle Point single pipes gave 20 ohms; at Acheron River single pipes gave 35 ohms; and see values in Table VIII: all by A.C. bridge method

– 472 –

These invariably occurred either in marshy places or where there had been extensive accumulation of vegetable matter and a good depth of soil. It is probable that in these localities there is a liberal proportion of those decomposition products which may be included under “humus acids.” Low-percentage solutions of acids have many times the conductivity of equivalent solutions of neutral salts or alkalies, and the existence of these in particular localities probably accounts for the comparatively low resistances encountered there.

Further valuable data were obtained from tests of artesian wells about Christchurch. Since these are from 70 ft. to several hundred feet deep, they might be expected to give resistances approaching zero.

Four wells consisting of 2 in pipes, 70 ft. to 80 ft deep, at Addington, gave, on the average, a resistance of 6–2 ohms (A.C. bridge and voltammeter determinations). A 2 m. well, 80 ft deep, at the city waterworks, Beckenham, had a resistance of 10 ohms (A C. bridge). The average of three 8 in wells, 80 ft deep, at the same place gave resistance of 8 ohms (A.C. bridge).

Thus these well pipes have resistance of the same order as earth pipes 6 ft deep in the same districts. It would therefore appear that depth and area of pipe in contact affect resistance only to a limited extent. The pure artesian water in the vicinity of the pipe is probably responsible for this effect.

A trial was made in sand and shingle ground at Sockburn to ascertain the effect of treating earth pipes with a higher-conductivity electrolyte—viz., sulphuric acid. Three 1 in. pipes 7 ft. long were placed (well tamped) about 1 to 2½ chains apart in much the same class of soil. These were then tested against an independent pipe 1 to 1½ chains distant. These gave individual resistances as follows (A.C bridge test): The earth-resistance of the first pipe, on treatment with 5-per-cent. sulphuric acid, fell from 460 to 170 ohms—i e, by 63 per cent; that of the second, on treatment with 15-per-cent. sulphuric acid, fell from 910 to 350 ohms—i.e., by 61–5 per cent; that of the third, on treatment with cold concentrated sodium chloride, fell from 360 to 170 ohms—i.e., by 52–8 per cent.

The higher-conductivity solution gives a bigger reduction of the initial resistance, but soil, though mechanically uniform, is apparently so variable chemically that a great many tests would be necessary to give a definite result. The degree of penetration into the surrounding soil and the composition of the soil are both very uncertain. There was, of course, considerable action of the acid on the pipe, and there is no suggestion whatever to use such a method commercially.

It has been previously stated that the resistance lies almost wholly within a radius of a few feet of the pipe. This is shown clearly in diagrams Nos. 3 and 3A of the resistances obtained in different localities between pipes at varying distances apart. The effect of saturating the ground round these pipes with an electrolyte (a strong solution of sodium chloride) is also shown by these curves. The Wellington pipes were 3 in. pipes, 6 ft long, and were placed in reclaimed ground with a good admixture of surface soil. The Addington pipes were 1 in in diameter and 10 ft deep, driven in peaty to sandy soil. The pipes at Bealey Corner were 1¼ in. and 1½ in. pipes, 6 ft. deep. At Sockburn 1 in. pipes were placed in shingle and sand over 7 ft deep. The pipes from which curve 3A was obtained were placed in line in the same class of soil. The pipes were all salted with 4 gallons of cold concentrated brine poured into each. When salting, the most dis-

– 473 –

tant pipe was salted and tested, then the next nearer salted and tested, and so on, to avoid any interference-effect due to salted soil between. The curve of retests after two or three days, when the soil between was more or less impregnated with salt, is also shown. The introduction of the salt has substantially reduced the resistance of the pipe earth, and has at the same time very considerably flattened the curve of potential drop.

Picture icon

Fig. 3.

Picture icon

Fig. 3A.

Messrs Liversidge and Creighton both record that salting had the effect of reducing the resistance to about 50 per cent. of the original value. Apparently the tests on which this statement is based were of earths of fairly low initial resistance—i.e, below 150 ohms.

On the majority of the pipe earths tested during the course of our investigations, with initial resistances ranging from 400 to 4,000 ohms, we found much larger percentage reduction of the initial resistance as the result of

– 474 –

salting. This is very clearly shown in diagram No. 4. This is based on the result of tests on over 100 pipes, each plotted point being the value from a group of three to eight 1 in. pipes, each 6 ft. to 7 ft. long. No two pipes were less than 6 ft. apart or more than 12 chains, and they were placed in groups of three to eight over the tract of country described, and therefore in all classes of soil. Salting was effected by pouring into each 4 gallons of saturated-brine solution, and then filling the pipe with salt. Tests were repeated within an hour or so of salting. Initial resistances up

Picture icon

Fig. 4.

to 250 ohms are in loam, sand, or clay, and over 700 are in heavy shingle; others are in a mixture of earth and shingle. As the degree of penetration of the salt into the surrounding soil is a matter of considerable uncertainty, the percentage decreases in resistance are much more uniform than might be expected. The curve drawn through the plotted points suggests the probable average reduction. Thus it appears that the reduction of resistance by salting is largely dependent on the magnitude of the initial resistance, and the salting is highly efficacious in reducing high resistances.

A further interesting feature of salting is the improvement of the “earth” with time. Apparently the salt continues to work into the surrounding soil for a considerable time after the initial salting. In time the salt must be dissolved by successive rains, and finally the “earth” must deteriorate. Curves are given demonstrating this in diagram No 5. Apparently the brine continues to percolate into the surrounding soil, partly by soakage and partly by capillarity. Eventually the solution is weakened by the addition of rain and subsoil water until the resistance again increases.

From a cursory inspection of these values it is evident that—at least, in many cases—single-pipe earths, however well salted, will not give the value of resistance desired. The only loophole remaining is by way of placing a number of pipes in a group. The question immediately arises as to how close such pipes may be placed and still remain effective. A consideration of curves of the decrease in resistance with increase of distance indicates

– 475 –

that the minimum useful distance between two pipes in good ground is about 6 ft. when the pipes are unsalted. After pipes have been salted the minimum distance must be increased.

In general, when pipes are salted, they must be placed at least 12 ft. apart to retain 90 per cent. of the efficiency of each pipe. The actual distance

Picture icon

Fig. 5.

necessary in any particular case depends on the conductivity of the soil in which the pipes are placed, and the higher the conductivity the farther apart must the pipes be to retain their efficiency.

Groups of pipes were placed on the transmission-line for the purpose of earthing depot telephones and apparatus. The following table gives particulars and results obtained with several pipes in parallel. No two pipes were placed less than 12 ft apart.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Table VI—Pipes in Parallel.
(All A.C. Bridge.)
Average Resistance of each Pipe Resistance of all Pipes in Parallel.
Locality Number of Pipes Class of Soil. Unsalted Salted Immediately after salting Five Months after salting.
Ohms Ohms Ohms Ohms
Greendale 5 1,900 244 47 32 Very dry clay and shingle.
Charing Cross 4 840 141 44 57 Heavy shingle cemented with clay.
Aylesbury 7 2,000 440 80 78
Heavy shingle. West Melton 4 910 305 55 80 Sand and clay to heavy shingle.
Bealey 4 534 150 34 61 Loam to heavy shingle.
Glenroy 4 180 94 26 Loam and clay, heavy boulder, somewhat marshy.
Highfield. 4 556 126 31 Loam and sand for 4 ft. to shingle.
Sandy knolls 5 2,334 417 90 Heavy running shingle.
– 476 –

It remains now to cite some actual methods employed, with the results attained.

Pipe earths, as well as the plate earths before mentioned, were used by the Christchurch City Council (diagram No. 6). These were placed in

Picture icon

Fig. 6.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

excavations 6 ft to 7 ft deep. The bottom of the pipe was packed with a mixture of rock salt (refuse salt from tanneries) and coke, while round the top was placed a drainpipe filled with rock salt and coke, and protected by a cover. Water was then poured in. The lead was 7/16 S.W.G braided

– 477 –

copper. These tested against independent earths (either water or gas mains) gave the following results (A.C. bridge method):—

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Table VII.—Christchurch City Council Earths
Locality. Resistance of Earth. Computed Value of Gas or Water Main.
Corner Madras and Armagh Streets (pipe) 22 Gas = 0.
N E. corner Latimer Square " 19
S E. corner "" 21 Gas = minus 2.* S.W. corner "" 20
Corner Cashel and Madras Streets (N.W) " 23 Gas = plus 3.
" "(S.E.)" 18
Corner Peterborough and Madras Streets " 76.5 Gas = plus 1.5.
Corner Salisbury and Madras Streets " 21.5
Corner Chester and Madras Streets " 44
Corner Victoria and Montreal Streets " 11.5 Gas = plus 0.5, assuming water-main = 1 ohm.
Beckenham transformer " 21
" (plate) 50
Sydenham transformer (pipe) 15
" (plate) 16
Montreal Street (pipe) 12
" (plate) 12
Madras Street, over Bealey Avenue (pipe) 13.8
" (plate) 10.7
Armagh Street - Hanmer Street (pipe) 4
" (plate) 6

At the transformers, the pipe earths and plate earths were placed within a few feet of each other, and tests show in favour of the pipe earth. As might be expected, in good conducting soil there is very little difference between plate and pipe earths, but in high-resistance soil the salted pipe usually has a considerable advantage.

The practice pursued by the Public Works Department (diagram No. 7) in the general transmission and distribution earths is to drive a 1 in. pipe, drilled with small holes every foot, to a depth of 6 ft. to 8 ft. in ground; then to pour 4 gallons of concentrated brine down the pipe, and fill the pipe with rock salt, placing some rock salt round the top. The top of the pipe is stopped with a wooden plug, which can easily be removed for replenishing the supply of salt inside the pipe. The conductor consists of 1 in galvanized-iron strip, riveted round the pipe and bolted through. In attaching the lead it is essential to make a good connection mechanically. Also any combination of metals which would yield to electrolysis should be avoided, or otherwise the joint must be thoroughly covered with some waterproof material. On this account, as well as for economy in cost, the joint is not sweated.

[Footnote] * Since the assumption is that values are independent quantities, and in a city pipes run in unknown directions, some of the values will be slightly in error, as in this case, owing to the unknowns of the simultaneous equations being in some cases not entirely independent. Hence the above negative values.

– 478 –

Resistances with this device, in soils similar to those of the previous tests (Table VII), are given in Table VIII. Values are of individual pipes, and are obtained by the A.C. bridge method.

Picture icon

Fig. 7.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Table VIII—Public Works Department Earths About Christchurch
Locality Resistance in Ohms Remarks
Substation, Addington 16 1 in pipe, 10 ft, unsalted, peaty to sand.
Christchurch Brick Company 6 1 in pipe, 8 ft, unsalted; marshy loam.
" 4 1 in pipe, 8 ft, salted; marshy loam.
Corner Lincoln Road and Barrington Street 17 1 in pipe, 12 ft, salted; loam to peat.
Lincoln Road 15 1 ¼ in pipe, 12 ft, salted; loam to peat.
Colombo Street 24 1 in pipe, 12 ft, unsalted; loam to clay.
" 34 1 in pipe, 10 ft, unsalted; loam to clay.
Allen's Mill 15 1 in pipe, 10 ft, unsalted; loam to clay.
Sunnyside transformer 14 1 in unsalted, loam to clay.
Lincoln Road 20 1 in pipe, 10 ft, unsalted; loam to clay.
Barrington Street 100 1 in pipe, 7 ft, unsalted; loam to gravel.
" 40 1 in pipe, 7 ft., salted; loam to gravel.
– 479 –

A comparison of these two sets of values (Tables VII and VIII) indicates that the bare pipe salted is more efficient than the pipe packed with coke and costing twice as much to place. The more elaborate earth, however, will probably tend to retain its efficiency longer.

A number of sample values (A.C. bridge method) of the resistances of pipes in various parts of the transmission-line are indicated in curve No. 4. The initial resistance of individual pipes ranges from 12 ohms to 4,000 ohms, and averages several hundred ohms. Immediately after salting, the range is 7 ohms to 560 ohms.

Pipes are now being placed at each pole on the transmission-line. Where there is nothing heavier than ordinary gravel, pipes can easily be started with a crowbar, and then driven to 6 ft. or 7 ft. with a hammer. The lower end of the pipe is flattened and closed, while the top is protected for driving by a loosely fitting cap turned from mild steel. In heavy shingle ground it is sometimes necessary to make an excavation in depth almost equal to the length of the pipe, and this very considerably increases the cost of placing. The lead from the pipe, as before, is 1 in. galvanized-iron strip, riveted round and bolted through by ¼in. bolt to the pipe, and riveted on to the earth-strip running up the pole. Bolting and riveting were resorted to after it was found that the sweated joint broke away while driving. Particulars of some samples of these earths are given below (sum of two resistances in series values equivalent to A.C. bridge test):—

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Pole Numbers Resistance Original Earth-strip Initial Resistance of Pipe Resistance immediately after salting Ten to Fourteen Days after salting Forty-four Days after salting Five Months after salting Character of Soil.
Ohms Ohms Ohms Ohms Ohms Ohms
759 and 758 305 170 98 84 65 104 Loam to sand.
757 " 756 241 170 126 99 91 129 "
755 " 754 209 216 118 93 71 100 Loam to sand and clay.
753 " 752 309 330 160 135 131 189 Sand to shingle.
751 " 750 890 1,600 493 387 290 420 Heavy shingle.
749 " 748 700 1,667 440 393 333 420 "
747 " 746 667 1,200 360 327 267 374 Heavy shingle and sand.
745 " 744 680 1,093 467 360 267 327 "
743 " 742 660 1,067 393 363 260 360 Heavy shingle.
742 " 741 593 1,467 433 380 213 297 Heavy shingle to sand.
603 " 604 7,000 1.060 800 Heavy shingle.
605 " 606 7,000 1,000 800 "
607 " 608 7,000 667 800 "
609 " 610 4,667 847 780 Heavy shingle and loam.
611 " 612 2,667 380 400 "
482 " 483 500 360 Clay to sand and shingle.
483 " 484 560 253 277 Sand, loam, and shingle.
485 " 486 280 143 143 Sand, loam, and clay (scrub).
487 " 488 173 83 110 "
489 " 490 313 123 154 Shingle and sand to loam.
491 " 492 833 187 190 Clay to shingle.
493 " 494 713 145 176 "
495 " 496 1,800 293 230 "

For salting these pipes, ordinary coarse salt is best for the initial charge, as it readily dissolves, while rock salt is better for filling and leaving about the pipe, on account of its taking a considerable time to dissolve. At present refuse rock salt from tanneries is being used, and found equal to ordinary salt for the purpose and much cheaper.

– 480 –

It still appears that in some districts single pipes will not provide a reasonably low-resistance earth connection. In such districts it will be necessary either to drive several pipes round each hole or to run a ground wire along that portion of the line and connect it with the single pipe at each pole.

This discussion would not be complete without some reference to the efficiency of the earth connection when called on to dissipate current.

Concerning the behaviour of pipe earths under working - conditions Mr. Creighton notes, “With an applied potential of 1,000 volts the current was so great that the earth round the pipe was quickly dried out, and 90 per cent of the drop of potential took place within 1 ft of the pipe (unsalted). The pipe had lost its effectiveness as a ground. With 900 volts drop in the immediate vicinity, it was a dangerous condition.”

The effect of salting on the potential drop has been illustrated by previous curves (Nos. 3 and 3A) as the potential curve is of the same order as the curve of resistance. Mr. Creighton publishes curves of potential distribution which are very instructive, and further insists on the value of thorough salting. With a potential difference applied of 120 volts D C a drop of 70 volts took place in the first 6 in. unsalted, while the same drop was extended over 3 ft 6 in. after the pipe was salted.

The capacity of an earth to discharge current over an extended period is also treated by Mr. Creighton: “The quantity of electricity that can be passed through a pipe earth without materially changing its resistance increases directly with the wetness of the earth in contact with the iron, and the area of the iron surface exposed to the passage of current, and decreases as the resistance of the earth in contact with the pipe increases Certain critical values of the current may be carried continuously by a pipe earth without varying the resistance. The higher the current above this critical value, the more rapid the drying-out. To increase the ampere-hour capacity, keep the pipe earth wet with salt water”.

Curves are also published showing the advantage of the pipe earth over a solid od in both states, and particularly its value when salted. A pipe earth thoroughly salted carried a current of 56 amperes for a period of forty hours without appreciable variation, while the same pipe earth initially carried only 30 amperes, and was baked out to 4 amperes in four hours and a half.

Mr. Liversidge also made some tests, passing a current of 25 amperes D C.: “In most of the tests it was found that the pipe connection or other grounding devices were able to carry a current of 25 amperes continuously without any marked increase in contact resistance. The results which were obtained indicated in most cases that the limiting feature of the current carrying-capacity of a ground connection is the ampere-discharge per unit area of contact surface. If the current-density was high enough to drive off the electrolytic moisture as a consequence of excessive heating of the earth immediately surrounding the grounding device, then the contact resistance would gradually increase”.

With alternating current it appears that the ampere-hour capacity of a given earth is, in general, considerably higher than with D C. This is consequent on the excessive formation of films in the interspaces, consequent on electrolysis, these films forming to a very much lesser degree, and almost negligible, when the current is alternating. With a high frequency the inductance of the earth and lead is a proportionally large factor for consideration, and this must be reduced as far as possible by using shortest

– 481 –

possible leads. This question will be the subject of further investigation as opportunity offers.

To conclude, the observations here set out demonstrate the uniformly high resistance of earth in the country investigated. Comparison with published data accentuates the contrast with the average values found elsewhere. It is not suggested that this is the only place of the kind, but it is evident that some circumstance is responsible for the abnormal high resistances.

A further item which points to a high resistance of soil in the district is the fact that very little electrolysis of waterpipes, &c., has been noted.

Although the dryness of the shingle plain may be responsible for some of the high resistances, it cannot account for their persistent occurrence. Christchurch and environs is extensively supplied with artesian water which itself is almost pure, containing the very lowest percentage of salts or impurities of any kind. Since pure water has such a high resistance, this may account for the high “earth” resistances encountered.