How much ripple on DC mains?

[Panrock]

Discussion elsewhere about AC/DC televisions and 50/60Hz mains set me wondering...

When there were DC mains supplies in some districts, how 'clean' were they? Was there any ripple produced by the generators and if so, how much and what frequency?

To answer this, you'll need to know more than I do about large DC generators. That should include many folk here.

Were these like motors in reverse, with commutators?

Steve

[TrevorG3VLF]

I used to use a big DC motor to drive a brake dynamometer, supplied by a Ward Leonard set.

The third stage of the set looked just like a motor and ran without any sparking until the operator turned the torque too high when the commutator would become a ring of fire. I was told that there should be less than 10V between commutator segments so for 240V there should be 12 pairs giving 12 phase rectified output.

The frequency will depend on the rotating speed, in the case of the W/L set, this was 3000rpm (50rps) so the frequency ripple would be 12*50 = 600.
Of course, the rotating speed of the generator will depend on the driver, steam engines will be quite slow, steam turbines quite fast. The faster the speed, the smaller and cheaper the machine.

Car dynamos are DC generators which run at a wide range of speeds.

[GMB]

... but are lucky in that the car battery acts to smooth it all out, otherwise it would not have been nice.

In addition to the commutator switching there is also the noise created by the non-electronic field control that tries to stabilise the voltage.

So I guess the sets would still be needing their smoothing capacitors!

[AC/HL]

A lot of DC mains were associated with electric tramlines I've read, I would think that they were quite dirty.

[TrevorG3VLF]

Trams were very dirty RF wise due to the intermittent pick-up connection and were connected to massive aerials.
Amateurs in Derby welcomed the demise of the trams but now have to live in the dirty atmosphere from the diesel busses.

In a recent thread about a DC radio there was mentioned it had a 'flying' choke in the input circuit before anything else, an indicator to how rough DC mains was.

[emeritus]

I do wonder if the capacitor that is connected directly across the mains input on AC/DC radios such as the Bush DAC90 was provided to deal with hash such as brush ripple found on DC mains. Removing it seems to have no effect on operation on AC mains. My understanding is that, for mobile military radios, DC supply was specified to be via twisted pair to mitigate the effect of radiated ripple when the vehicle's engine was running and its alternator or dynamo was in use.

I do remember one of the instructors on a workshop course I attended reminiscing about his days as a radio repair man. On being called to a customer whose new radio wasn't working, he found that its lamps and valves were glowing dimly, with no sound. He recalled that that part of Ilford was still on DC, and the radio was an AC-only model. On investigating, he found that the local dynamo/rotary converter that had been used to supply the remaining pockets of DC mains had failed. As DC was on the way out, as a stop-gap, DC was being provided from the AC supply via a mercury arc rectifier arrangement. One of the arms had failed, and there was enough ripple to allow the transformer in the set to produce a reduced output.

[Panrock]

Originally Posted by TrevorG3VLF

The frequency will depend on the rotating speed, in the case of the W/L set, this was 3000rpm (50rps) so the frequency ripple would be 12*50 = 600.

So there'd be a 'tone' heard under the sound on your radio rather than hum. Easier for the smoothing to suppress but then less tone would be clearly audible to the ear than hum.

If you take a three-phase AC generator and rectify and combine the phases, do you end up with pretty well perfectly flat DC?

Steve

[TrevorG3VLF]

If you full wave rectify three phase, you get six pulses each cycle. It will be nowhere near perfectly flat DC but the ripple will be small.

When alternators (three phase) were first used in cars a variable frequency whine could be heard in the radio. HF suppresion is better these days.

[Ed_Dinning]

Hi Gents, according to the maths, fullwave 3 phase rectification (6 phase) give a ripple of approx. 4%.
This was often used for plating supplies without any cap smoothing and worked well enough on vehicle power supplies that no noise was apparent when it was floating a small battery.

Ed

[Lucien Nunes]

I'm too young to ever have 'listened' to DC mains but now I think about it, there might have been a lot to be learned about the particular supplies in use from the noise signature.

Rectification for DC supply was usually at least 6-pulse, IIRC there were 12-pulse installations in service. With the aid of a very large choke in the cathode and anode reactors to achieve tidy commutation between anodes of each group (mainly to avoid overlap, where the transformer secondary inductance causes two phases to conduct at once) the ripple was not as deep as one might expect. I don't know what constituted acceptable depth but might be able to look it up.

Generators were a different kettle of fish. Slot ripple amplitude is minimal on a large machine because so many coils are in circuit at once. What there is, however, is often fairly high in the audio band and very noticeable to the ear. The frequency is much more variable than that of rectified AC mains from the grid, and there might be multiple generators running at similar but different speeds, possibly beating with each other. What might have been much more of a problem than the ripple fundamental itself, would be the hash such as brush arcs, modulated at ripple frequency.

A kind of ELF ripple also occurred with house lighting plants with large low-speed single-cylinder engines. Unless fitted with massive flywheels, these could impose a noticeable cyclic variation of both voltage and slot ripple frequency, in time with the power strokes of the engine. For example, my big Crossley/GEC plant runs at 290rpm, being a 4-stroke the speed variation is 2.4Hz, although the 7-foot flywheel reduces the amplitude to an acceptably low level - you cannot see the lamps pulsating. I expect that anything audibly reproducing the slot ripple, would be made all the more objectionable with this modulating 'wow'.

Re. generator field controls and regulation, I don't think there would have been any audible contribution from this source due to the very long time constants of large generator field circuits, often many seconds due to the enormous inductance of the field coils.

[broadgage]

I suspect that DC mains supply was reasonably smooth in practice.

In the earlier days of DC supply, it was produced from dynamos worked by steam engines. The many segments of the commutator give a reasonably smooth supply, and also all but the very smallest stations would have a number of dynamos sharing the load, and also thereby smoothing the supply.
Later rotary converters were employed and produced a very similar result.

Towards the end of the DC era most supplies were obtained by rectification of AC. It was common practice to start with 6 phase AC feeding a 12 pulse rectifier.

Although tramways used DC this was invariably a separate system to the DC lighting mains. Often the same dynamo could be used to serve EITHER the 550 volt 2 wire negative earth tramway system, OR the 3 wire 500 volt centre earthed lighting system. The machine would be switched onto whichever system it was needed on, but there was no connection between the two.
Some districts had a separate DC "power" system for factories often at 550 volts or so, this WAS supplied from the tramway system and would have been very poor quality. Not suitable for lighting, apart from the high voltage, which could be overcome by running lamps in series, the voltage would drop each time a tram accelerated.
Factories with DC power generally also had a separate lighting supply, or retained gas lighting.

[Panrock]

Some fascinating answers. Thank you.

As an aside... with the tram wires acting as giant radiating aerials, maybe the past wasn't such a golden age for AM radio reception after all, at least in urban areas.

Steve

[Synchrodyne]

Regarding interference from tramway and trolleybus overheads, I don’t think it was too bad in the days when stepped series resistance/field weakening/motor grouping were the norm in control systems, but there was a step-change for the worse when thyristor chopper thyristor control systems arrived in the 1970s. In Wellington in the mid-1970s there was not that much interference from the trolleybus system, and at one stage I lived adjacent to a trolleybus route, actually looking the down on the wires from an upstairs window, but wideband AM reception of local stations was unimpaired apart from an occasional random click. In the 1980s, when a new thyristor chopper-equipped bus fleet entered service, the interference heard on the car radio was certainly worse, although one had to be in the vicinity of an accelerating bus to experience it.

I also recall driving around Europe in 1979 and often listening to BBC R4 on LW. In some of the German cities that had thyristor-equipped trams, the interference whine was quite pronounced and one did not have to be that close to a tram to hear it. But then Germany was probably well outside of the primary groundwave service area of the Droitwich transmitter. On the other hand, I recall very little interference from the overhead of the Arnhem trolleybus system, which at that time had only conventionally-controlled vehicles. In any city that had trams or trolleybuses, it was easy to tell whether there were any thyristor chopper-controlled vehicles in the fleet.

Cheers,

[grampy2]

I've got a feeling I've said this before, lived in Sheffield 60 years ago, alongside a tram route, could hear trams up to a mile away through a pair of 2000 ohm phones connected to a couple of Earth stakes about 20 feet apart, obviously due to earth leakage currents.

[Synchrodyne]

Here is some information (from the early 1950s) about the radio interference suppression equipment that was fitted to trolleybuses back then. Presumably something similar would have been used on tramcars.



Returning to DC mains, I imagine that when fed by rectifiers, these would have required to have adequate suppression of rectifier harmonics in order to avoid interfering with any nearby telecoms circuits, the presence of which was quite likely in respect of street distribution. That suggests that they were quite “clean” when leaving the substation, with contamination coming more from connected machinery and equipment than from the power source. In that case, were equipment input filters such as that mentioned for the Bush DAC90 installed primarily to stop interference getting in, or to stop it getting out?


Cheers,

[Peter.N.]

I remember the lights on Cannon Street station flickering at a steady rate, never seen that before, I was quite fascinated. Must have been in the late '40's early '50's. I assume that was DC.

Peter

[Synchrodyne]

Might it have been 25 Hz AC rather than DC? I understand that some of the earlier Southern Electric network substations were originally supplied with 3-phase 25 Hz AC (from dedicated generating stations), because the rotary converters of the era (say before the 1930s) worked better at lower frequencies. And the LBSCR had originally electrified at 6.6 kV 25 Hz single-phase, which would have required a 25 Hz supply network, probably retained after 3rd rail DC conversion. Given that background, the use of 25 Hz for ancillary functions would not have been too surprising, the more so in the early days when standardized domestic/industrial supplies were not always readily available. (And why buy in electricity when you are generating your own for traction purposes.) Lamps operating on 25 Hz would, I imagine, suffer from quite visible flicker.

Cheers,

[G6Tanuki]

Originally Posted by broadgage

Some districts had a separate DC "power" system for factories often at 550 volts or so, this WAS supplied from the tramway system and would have been very poor quality. Not suitable for lighting, apart from the high voltage, which could be overcome by running lamps in series, the voltage would drop each time a tram accelerated.
Factories with DC power generally also had a separate lighting supply, or retained gas lighting.

Some relatives lived in a pit-village in the East Midlands and up until about 1965 their electric supply was 220V DC supplied by the power-house at the pit (they got electricity for free from the NCB just as the miners got a weekly domestic coal-allowance).

Periodically, the lights would pulsate in brightness, the frequency of the pulsations starting at around 0.5Hz and rising in frequency as the amplitude of the pulsations also decreased until after a few seconds the effect disappeared.

This was apparently due to the pit starting-up the big coal-conveyor to load the day's output onto rail wagons so it could be hauled to one of the big Trent power-stations. The locals called the effect "Cogging" and even though it sometimes caused frame-roll on TV sets they didn't complain too much about it because, well, free electricity!

The move to replace the DC supply with AC-from-the-grid was unwelcomed by many in the village since for the first time they got an electricity-meter and had to pay for what they used.

[When the relatives in question wanted to buy a new radiogram they had been told that the DC mains was due to be replaced by AC sometime soon - so my late father built them an inverter using a bunch of big top-cap-anode valves in order that they could buy an AC radiogram but run it on DC in the interim]

[broadgage]

Low frequency supplies were used on some railways, and may still be used, for signalling circuits.
DC electrified railways tended to use 50 cycles for signalling.
AC electrified railways tended to use either DC signalling or AC at a different frequency to the traction current, so as to avoid traction current in the running rails interfering with track circuits.

If this low frequency supply was generated in bulk it would make sense to use it also for some of the station lighting, with the rest on AC mains, so as to ensure some lighting if either supply failed.