How much ripple on DC mains?

[emeritus]

25Hz was used by one of the railways south of the Thames ( I forget which one) for their overhead wire electrification before WWI. When the railways were grouped in the 1920's, the Southern Railway decided to standardize on third rail 750V DC electrification and the old 25Hz AC system was abolished. A book I once had did refer to part of the old 25Hz system lingering on for many years afterwards for lighting, where the lights some of the London termini continued to flicker at the old rate. IAFAIR, the larger thermal mass of the straight filaments of older vacuum filament lamps meant that they were less susceptible to flicker than the later coiled and coiled coil filaments.

[Peter.N.]

From memory it certainly looked about 25hz but by the time I started work in '55 the flicker had gone.

There were some odd electrical arrangements though even then, I worked in Cheapside for a year, not far from Cannon Street and we had some sort of split phase system that gave 110-0-110 volts, they were in the process of changing it and had fitted a 'modern' consumer unit with solid neutrals leaving one side unfused - guess who managed to short the 'neutral' side to earth and put the lights out in the insurance company next door as well as ours.

Peter

[hannahs radios]

I think the DC output from power stations would have been fairly smooth because I think most stations used big battery banks across the supply (how would you wire them on the negative mains side) so the batteries would soak up the ripple. For those who don't know the generators were shut down at night and the customers were supplied just off the batteries.

[Synchrodyne]

This short excerpt from Electrical Engineer's Reference Book, 9th, 1958, provides an overview of the smoothing of rectifier outputs to avoid telecomms interference. I'd guess that such smoothing would have been used generally on rectifier-fed public DC reticulation systems given the high probability that somewhere in the chain, they would be contiguous with telecomms systems.


Cheers,

[broadgage]

SOME DC supplies did indeed use a battery, either at night or even all the time, but this practice was in decline by about 1920 due to the rapidly growing load and the cost and bulk of the battery.

The variation in voltage of a battery was much too great to be connected directly to a public supply.
The usual approach was to connect the battery to the DC network via a reversible booster or a battery booster. These machines consisted of two electrically independent rotating machines coupled together or built on the same shaft.
With say a 500 volt battery, the "motor" side of the machine was wound for about 600 volts and connected across the battery, the "dynamo" side of the machine was wound for say 100 volts and was connected in series between the battery and the DC mains serving customers.
To charge the battery from the 500 volt DC mains would need about 600 volts, this was achieved by adjusting the field currents such that the dynamo produced about 100 volts IN ADDITION to the 500 volts from the mains and thereby charged the battery.

If the DC supply from the generators failed, then the battery via the reversible booster supplied the load at a constant voltage.
At the beginning of the discharge, the battery voltage might be say 540 volts and the machinery would generate 40 volts in opposition to the battery and reduce the 540 volts from the battery to the 500 volts needed for distribution.
As the discharge progressed the battery voltage would fall, and the machinery would produce a lesser voltage in opposition to the battery volts.
When the battery voltage fell below that of the mains, the "dynamo" side of the machine would start generating a voltage in addition to the battery voltage, thereby keeping 500 volts on the mains as the battery voltage dropped.
Note that the change from subtracting volts to adding volts was seamless and that the physical direction of rotation was not altered.
It was claimed that the voltage was steady within 1%.

Such machinery was reasonably efficient, note that it only had to be sized for the difference between mains voltage and battery voltage.
Take a system with a load of 1000 amps at 500 volts or 500KW, the battery booster only had to be sized at about 100 volts at 1000 amps or 100KW.

IIRC the largest battery erected in the UK was at Dickinson street in Manchester. Relying on my imperfect memory I think it had a normal charging current of several thousand amps at hundreds of volts.

Growing loads made such schemes non viable.

Some modern power stations still have large batteries, but these are to operate lighting and essential auxiliary plant in case of major breakdown, not to maintain supplies to customers.
The present day trend is to much reduce such batteries and rely instead on relatively small diesel generators of a few MW capacity. These start from modest size batteries, as small as 24 volt 400AH sometimes, or from air bottles.

[cmjones01]

Very interesting to hear about the batteries used on DC supplies. What goes around comes around - see the recent thread on electric vehicles and energy storage on the grid. Of course, today's batteries are smaller, lighter, longer-lasting, need no maintenance and are much safer than those of a century ago, but the principle is the same. The rotating machinery is replaced with electronics, which is quieter and more efficient but much less fun...

CHris

[broadgage]

To add to the points already made, very small DC plants almost always used a battery.
Such plants were more akin to country house lighting plants, than to DC mains, some larger systems were installed to serve "the big house" and then extended for street lighting, staff cottages, or local businesses.
So they just about count as DC mains or public supplies.

Smaller systems were often 24 volt or 32 volt on these the voltage was either un regulated as in vehicle practice, or was regulated by inserting resistance between battery and lamps.

Larger systems were often 110 volt. Some of these used a reversible booster to keep a constant voltage on the lamps, but this was rather wasteful since the booster set might consume hundreds of watts in loses when the load was a single lamp.
A more common technique was "end cell switching" A nominal 110 volt system might have a 120 volt battery of 60 cells.
The battery would be charged in the day and utilised at night. At the beginning of the discharge 52 cells might produce 110 volts and that number of cells would be in circuit.
When the voltage dropped to 109 volts, another cell would be inserted to raise the voltage to 111 volts, this sequence being repeated until all 60 cells were in circuit.
The automatic switching was rather complex, remembering that it had to avoid extinguishing the lamps when changing from say 55 cells to 56 cells, but also not short circuit one cell whilst making the change.

In the morning the engine would charge the battery, starting with the whole battery and cutting out the end cells one by one as they become charged.

Various refinements were possible including sensing the voltage at the centre of the load rather than at the switchboard.
Alternatively it might be set up to maintain 111 volts on minimum load, 114 volts at half load and 117 volts at full load, thereby compensating for voltage drop between plant and load.
A slightly higher battery voltage of 128 volts, 64 cells was usual in such cases.

The prime mover was usually a steam traction engine, a huge investment even for a wealthy landowner but useful for other purposes.
Such plants were often modernised by selling the steamer and fitting a small petrol engine.
Some continued in use into at least the 1970s !

[Peter.N.]

That's very interesting, I never knew that early electrical supplies were so complicated or went back as far as traction engines. I know some of them had fairly hefty generators as they were used to run fairgrounds but it hadn't occurred to me that they were also in use for domestic purposes.

Peter

[G6Tanuki]

A relative in Kent had a fairly-large farm that until the early-1960s generated its own power - 110VDC - using a 2-cylinder 'hot bulb' oil-engine mounted on a big 4-wheeled trolley. There were two huge lead-acid batteries - allegedly ex-Submarine - that provided power for some days to cover for when the engine was hauled out to power threshing machines and suchlike elsewhere on the farm.

Several farm-cottages and the stables [still used for farm-horses: the "Field Marshall" tractor was seen as very much a novelty] were also supplied with power from this setup.

[broadgage]

A traction engine was a hugely costly item, but quite a few large country estates had one.
The dynamo for such an installation was normally fixed in the battery room, and the engine backed into place when needed and a drive belt used to work the dynamo.

Large country estates often produced most of their own food, and affluent persons travelled on horseback or in horse drawn carriages.
The main external "input" to such an estate was often coal, and a traction engine was ideal to haul this from the nearest rail yard or canal dock.

To return to strictly electrical matters, such lighting plants often had a "power" circuit as well as the lighting supply.
The lighting circuit was regulated by end cell switching as previously described.
The power circuit was not regulated but was permanently at the full voltage of the dynamo or battery. This explains why 15 amp 2 pin outlets in old grand houses are sometimes marked "140 volts" this being an estimate or approximation of the maximum likely voltage.

Our dear Lucien has this on his site... http://www.electrokinetica.org/d1/1/3.php

These days with most stuff not really bothered about volts or AC/DC a 120V (battery nominal, 150V fully charged) could easily run a house. Assuming gas heating/cooking.

[Ed_Dinning]

Hi Gents, in the Arcs & Sparks gallery at Newcastle Discovery museum we have the power system from a large country estate near Thirsk in N Yorks. It is a 110v DC system and comprises a large dynamo that was powered by either a waterwheel or gas engine and charged batteries. This fed both the main house and the workshops. A very comprehensive motor driven AVR is fitted as well as comprehensive switching between the various power sources.

Also in N Yorks, at Reeth was a power station built by a local hayrake manufacturer which supplied the town until about 1950 when it finally went on grid. It was in operation from the early 1900's, drawing water from the river swale. The leats are still visible and the turbine hall is now holiday lets.
It was updated to a gas/ oil engine quite early on as the Swale flow was variable, with generation at 110v. The same gent who installed this system had also installed the power station in the mill at Aysgarth to the south. Details are available at the Swaledale museum in Reeth.

Ed

[Hartley118]

A few weeks ago, I came across these interesting remnants of a Country House switchboard at the National Trust Killerton House in Devon. The local manager was helpful, but had little information. Apparently, in the room behind the switchboard were batteries, whilst the dynamo set was well away from the house in the stable yard.

With those clues, I've made some guesses on the functioning of the switchboard:

The ammeter on the top left (marked AMPERES) would normally show the total current being drawn from the battery by the house lighting. The zero on the scale is actually in the middle because the needle can swing both ways to indicate battery discharge current (probably to the right) or charge current (probably to the left). It would have been normal to charge the batteries from the dynamo during the day and supply the lamps from the batteries during the evening. The rotary switch with the black handle and brass sectors below and between the two meters probably switched the batteries between charging and discharging.

The voltmeter (top right) would be all-important for the house electrician to keep an eye on: I guess that the aim was probably to keep the pointer at 120 volts, around the middle of the scale. As the batteries discharged during the evening, their voltage would drop and the lamps would grow dim. . That would be the signal for the electrician to grasp a big rotating black knob (missing) turning a spring contact (missing) over those brass studs below the ammeter (labelled 39 to 56 and then DIRECT). This big rotary switch selected the number of battery cells in circuit (from 39 to 56). Each cell would give a nominal 2 volts. As the battery discharged, so more cells would be brought into circuit to boost the voltage and keep the lamps bright. If an evening's entertainment was continuing into the small hours, that may well have proved too much for the battery capacity and a mechanic would then have been sent to the engine house to start up the engine to drive the dynamo . The electrician would then select the 'DIRECT' position of the switch to run the lamps directly from the dynamo.

The other (unlabelled) remains of a rotary stud switch to the right of the cell selector probably controlled the dynamo output to charge the battery at its recommended rate.

To left and right we see main fuses and at the bottom 'knife' switches to switch supplies to various parts of the house.

Judging from the ammeter scale, the system would probably have been capable of between 6 and 8 kilowatts. Each carbon-filament lamp was probably rated at a rather dim 16 candlepower and would draw around 50 watts. So, assuming that the electric supply was only used for lighting, as was usual in those days, that would indicate there being around 100 to 150 lamps in the house.

Locating the generator away from the house in the estate yard would seem sensible because it was probably driven by a rather noisy and smelly oil engine, or indeed a traction engine as suggested earlier.

i'd be very interested in other views on how the installation might have worked back in the day.

Martin

[emeritus]

No personal experience, but the GEC catalogue of 1893 illustrates equipment for this type of installation. The only description of operation is on Page 79, which seems to be unusual in that it only designed for three states: charge battery with load isolated; supply load from with dynamo disconnected; supply load from dynamo with battery disconnected. My copy of the catalogue is an early issue and a note says that illustrations and descriptions of switchboards will appear in the final issue, which I I do not have.

[Hartley118]

Originally Posted by emeritus

No personal experience, but the GEC catalogue of 1893 illustrates equipment for this type of installation and has a brief explanantion of how it can be used.

Thanks for this information Emeritus - its comprehensive description and installation instructions fill in some gaps in my guesswork. In particular, it confirms that the lamps could be run directly from the dynamo when required. In those pre-radio days, I don't suppose that there would have been concern about ripple on the dynamo output compared with the smooth DC from the batteries, as long as the lamps didn't obviously flicker.

I guess, however, that there would have been a period in the 1920s-30s when radios might well have been run from this kind of Country House supply, presenting a few challenges.

The introduction of electricity in those early 20th century years must have placed new technical demands on the 'below stairs' staff. Operating such a switchboard safely and consistently must have been pretty challenging. The technician had to avoid supply interruptions and 'brown-outs', accidental over-volting of the lamps etc .

There were also the accumulators and engine to look after: very new technology at the time. Any failure in that technical role would have been very obvious to the Master and Mistress, particularly if the lights were allowed to fail during a prestigious event!

Martin

[kalee20]

Wow that must have been a splendid board! Polished slate - maybe I'll specify that for my next front panel in the day job...

Presumably the voltage adjust switch was a break-before-make action so as to avoid shorting the extra cells during switching, and the operator had to switch it quickly so as to have just a brief flicker to lamps.

Switching in extra cells to compensate for falling voltage would do the job, but if the load changed then more or less would be required (as the internal resistance of the cells rises as they discharge). I'm envisaging a situation where the main cells are nearing discharge, extra ones are switched in to bring the on-load voltage back up, then someone somewhere switches a significant load OFF. Voltage immediately rises, overvolting the remaining lamps in circuit, as each one goes pop, the voltage gets a bit higher, they die faster and faster before the technician can readjust his switch...

I'd want a suitable overvoltage cut-out, maybe an electromagnet pulling a plunger with a critically-adjusted opposing spring, and when the electromagnet pulls hard enough to overcome the spring the plunger operates a latch which disconnects the supply from the load. Embarrassing if it operates, but less so than having to replace a load of expensive bulbs!

[broadgage]

Most end cell voltage regulators operated in one direction only, to insert more cells in circuit as the discharge progressed.
Some types used a spring that had to be wound up each morning.
A modest drop in load was of little consequence, the voltage would rise slightly and the next switching step would be postponed for a bit.
Any substantial drop in load might cause over voltage on the lamps, but this was seldom a problem in practice.

Consider an example of a nominal 110 volt system with 60 cells.
Part way through the discharge 57 cells might be in circuit and producing an on load voltage of 110 volts.
Any sudden drop in load would cause the voltage to increase, but not by much. I would not expect a cell already significantly discharged to produce more than 2.0 volts on load, even a reduced load
So that would be 114 volts on the lamps, acceptable in practice and probably not continued for long. The 114 volts would be steadily reducing, until it dropped to 109 volts at which point the regulator would bring in the 58th cell.

[Lucien Nunes]

That looks like a charge-discharge lighting plant board, an early example with only one ammeter and no automatic cutout. One pole of each of the dynamo and lights fuses have been removed, the tapping selectors are missing and a couple of additional switches have been provided e.g. for direct working.

In response to an enquiry about the method of operation of the standard kind of house-lighting board on another forum I wrote the following description:

This is a very standard type of board that was made by numerous firms for many decades. It is designed for manually-controlled charge-discharge working of a house battery.

LAYOUT
Equipment on the left side is for controlling charge. An isolating switch, fuses and ammeter are provided for the dynamo, and a rotary end-cell selector switch (see below) that determines which battery tap the dynamo current is fed into. Equipment on the right side is for controlling discharge. The switch, fuses and ammeter are for the load circuit (often marked lights), and the end cell selector determines from which battery tap the load current is drawn. In the centre of the board is the voltmeter, usually with a selector switch to read the voltage on the charge (dynamo) or discharge (lights) side or switch it off to avoid it discharging the battery when unattended. Below this are the field rheostat for adjusting the dynamo voltage, and the cutout relay. Any special-function switches, such as for sets with electric start or remote stop, are usually found here too.

CUTOUT
The cutout serves the same purpose as on a vehicle dynamo; preventing the battery discharging into the dynamo and running it as a motor, if the output voltage is less than the battery voltage (e.g. if the engine runs out of fuel). As the dynamo voltage builds up, a shunt winding closes the contacts, which is then assisted by a series winding when current flows out of the dynamo. If that current reverses, the series winding opposes the shunt and the relay releases to prevent further discharge. Cutout relays often used open mercury cups for the contacts. These offered the advantage of requiring very little operating force from the sensitive relay, and could handle heavy current surges without any danger of welding shut or developing high resistance. The mercury was normally covered with a little paraffin to reduce outgassing.

END-CELL SWITCHING
Whenever the generating plant is not running, the house load is taken by the battery. As it runs down, its voltage falls significantly, which on incandescent lighting causes a very noticeable drop in brightness. To mitigate this, a manual method of regulating the voltage was provided in the form of end-cell switching. The inter-cell connections between the last few cells of the battery are wired to the contact studs of the end-cell selector switches, often termed 'regulating' switches. With all the end-cells switched out, the voltage of the main section of the battery would be equal to the nominal system voltage when fully charged. Once the main section voltage has fallen by 2V through discharge, one end-cell can be switched-in by rotating the discharge (lights) end-cell selector, to restore the circuit voltage to normal. This can be repeated after each 2V drop until all the end-cells are in circuit, at which time the voltage on the main battery is at its fully-discharged level.

As soon as the set is started to recharge the battery, the voltage rises substantially so the discharge end-cell selector can be returned to the all-out position. The charging end-cell selector can be started at the all-in position so that the end cells receive a charge. However, they will not be so heavily depleted as the main battery section so they can be bypassed sequentially as charge progesses, until only the main section of the battery is on-charge. At this time, the load will be receiving the full dynamo voltage, which (in order for the battery to charge at a reasonable speed) will be significantly higher than nominal. It may be necessary to temporarily switch off the load to prevent lamps burning out, during the latter stages of charging.

The end-cell selectors must not short-circuit adjacent battery taps, so their contact studs are widely spaced to prevent the wiper bridging between them (unlike those of a rheostat). To prevent an annoying interruption of supply whenever the tap is changed, an auxiliary contact is often fitted to the discharge selector only. This is linked to the main one via a low-value resistance coil that can briefly withstand 2V across it. As the handle is moved, the incoming tap connects to one side of the resistance before the other side disconnects, thus the load circuit is never broken. Boards supplied with sets capable of electric starting from the house battery may have a special contact on the charge side end-cell selector, that bypasses the cutout and energises the dynamo series-field directly from the battery.

Regarding the 'runaway' condition suggested above, where a reduction in load current could cause a progressive blowing of light bulbs, I do not think this occurred in practice because of the low dynamic internal resistance of the battery. Voltage variations were principally due to state of charge, not drop incurred by any reasonable amount of load current.

I have a large automatic board of the type that Broadgage describes above, with motorised end-cell switching controlled by a voltage sensing relay. The set delivers 230V DC nominal with an output of 18kW, installed in Firle Place in 1923 to supply the house, stables, and a few important buildings in the village. To date I have not had an opportunity to erect the plant (which requires massive foundations - the dynamo alone weighs 3 tons) and find out just how it behaves.

[emeritus]

Here's some more pages from the 1893 GEC catalogue showing further examples of switchboards and some of the individual components, such as the automatic cut-outs, used in switchboards.

Also estimates of the cost of plants of various capacities powered by steam engines, turbines, oil engines, and gas engines. The price of export lamps was cheaper than the UK price because Swan's patent was then stlill in force in the UK, meaning that royalties had to be paid to Swan on lamps sold for use in the UK. The catalogue also offered turbines for use where a suitable water supply was available, said to have an efficiency of 75%, for which cost estimates would be provided on request.

[dseymo1]

Page 10 of the first publication above illustrates the sorts of appliance which could be run from the board. What is meant by 'mallet' in this context?